THE

CAMBRIDGE NATURAL HISTORY

EDITED BY

S. F. HARMER, Sc.D., F.R.S., Fellow of King's College, Cambridge; Superintendent of the University Museum of Zoology

AND

A. E. SHIPLEY, M.A., Fellow of Christ's College, Cambridge; University Lecturer on the Morphology of Invertebrates

VOLUME VI

INSECTS

PART II. Hymenoptera continued (Tubulifera and Aculeata), Coleoptera, Strepsiptera, Lepidoptera, Diptera, Aphaniptera, Thysanoptera, Hemiptera, Anoplura.

By David Sharp, M.A. (Cantab.), M.B. (Edinb.), F.R.S.

London
MACMILLAN AND CO., Limited
NEW YORK: THE MACMILLAN COMPANY
1899

All rights reserved

"Men are poor things; I don't know why the world thinks so
much of them."—Mrs. Bee, by L. & M. Wintle.

CONTENTS

SCHEME OF THE CLASSIFICATION ADOPTED IN THIS BOOK

Order.		Sub-order, Division, or Series.		Family.		Sub-Family or Tribe.		Group.
HYMENOPTERA (continued from Vol. V)		Petiolata. (continued from Vol. V).
		Tubulifera (p. [1])		Chrysididae (p. [1]).
		Aculeata (p. [4])		Anthophila (p. [10]) Apidae (p. [10])		Archiapides (p. [21]). Obtusilingues (p. [22]). Andrenides (p. [23]). Denudatae (p. [29]). Scopulipedes (p. [32]). Dasygastres (p. [35]). Sociales (p. [53]).
				Diploptera (p. [71])
				Eumenidae (p. [72]). Vespidae (p. [78]). Masaridae (p. [88]).
				Fossores (p. [90]) Scoliidae (p. [94])		Mutillides (p. [94]). Thynnides (p. [96]). Scoliides (p. [97]). Sapygides (p. [99]). Rhopalosomides (p. [100]).
				Pompilidae (p. [101]).
				Sphegidae (p. [107])		Sphegides (p. [107]). Ampulicides (p. [114]). Larrides (p. [116]). Trypoxylonides (p. [118]). Astatides (p. [119]). Bembecides (p. [119]). Nyssonides (p. [123]). Philanthides (p. [124]). Mimesides (p. [127]). Crabronides (p. [128]).
				Heterogyna (p. [131]) Formicidae (p. [131])		Camponotides (p. [144]). Dolichoderides (p. [157]).
						Myrmicides (p. [158])		Myrmicini (p. [159]). Attini (p. [165]). Pseudomyrmini (p. [168]). Cryptocerini (p. [169]).
						Ponerides (p. [170]).
						Dorylides (p. [174])		Ecitonini (p. [175]). Dorylini (p. [177]).
						Amblyoponides (p. [180]).

Order.		Sub-order, Division, or Series.		Family.		Sub-Family or Tribe.
COLEOPTERA (p. [184])		Lamellicornia (p. [190])		Passalidae (p. [192]). Lucanidae (p. [193]).
				Scarabaeidae (p. [194])		Coprides (p. [195]). Melolonthides (p. [198]). Rutelides (p. [198]). Dynastides (p. [199]). Cetoniides (p. [199]).
		Adephaga or Caraboidea (p. [200])		Cicindelidae (p. [201]).
				Carabidae (p. [204])		Carabides (p. [206]). Harpalides (p. [206]). Pseudomorphides (p. [206]). Mormolycides (p. [206]).
				Amphizoidae (p. [207]). Pelobiidae (p. [207]). Haliplidae (p. [209]). Dytiscidae (p. [210]).
		Polymorpha (p. [213])		Paussidae (p. [213]). Gyrinidae (p. [215]). Hydrophilidae (p. [216]). Platypsyllidae (p. [219]). Leptinidae (p. [220]). Silphidae (p. [221]). Scydmaenidae (p. [223]). Gnostidae (p. [223]). Pselaphidae (p. [223]). Staphylinidae (p. [224]). Sphaeriidae (p. [227]). Trichopterygidae (p. [227]). Hydroscaphidae (p. [228]). Corylophidae (p. [228]). Scaphidiidae (p. [229]). Synteliidae (p. [229]). Histeridae (p. [230]). Phalacridae (p. [231]). Nitidulidae (p. [231]). Trogositidae (p. [232]). Colydiidae (p. [233]). Rhysodidae (p. [234]). Cucujidae (p. [234]). Cryptophagidae (p. [235]). Helotidae (p. [235]). Thorictidae (p. [236]). Erotylidae (p. [236]). Mycetophagidae (p. [237]). Coccinellidae (p. [237]). Endomychidae (p. [239]). Mycetaeidae (p. [239]). Latridiidae (p. [240]). Adimeridae (p. [240]). Dermestidae (p. [241]). Byrrhidae (p. [242]). Cyathoceridae (p. [243]). Georyssidae (p. [243]). Heteroceridae (p. [243]). Parnidae (p. [243]). Derodontidae (p. [244]). Cioidae (p. [245]). Sphindidae (p. [245]). Bostrichidae (p. [246]).
				Ptinidae (p. [246])		Ptinides (p. [246]). Anobiides (p. [246]).
				Malacodermidae (p. [248])		Lycides (p. [248]). Drilides (p. [248]). Lampyrides (p. [248]). Telephorides (p. [248]).
				Melyridae (p. [252]). Cleridae (p. [253]). Lymexylonidae (p. [254]). Dascillidae (p. [255]). Rhipiceridae (p. [256]).
				Elateridae (p. [256])		Throscides (p. [260]). Eucnemides (p. [260]). Elaterides (p. [260]). Cebrionides (p. [260]). Perothopides (p. [260]). Cerophytides (p. [260]).
				Buprestidae (p. [261]).
		Heteromera (p. [262])		Tenebrionidae (p. [263]). Cistelidae (p. [264]). Lagriidae (p. [264]). Othniidae (p. [265]). Aegialitidae (p. [265]). Monommidae (p. [265]). Nilionidae (p. [265]). Melandryidae (p. [265]). Pythidae (p. [265]). Pyrochroidae (p. [266]). Anthicidae (p. [266]). Oedemeridae (p. [266]). Mordellidae (p. [267]). Cantharidae (p. [269]). Trictenotomidae (p. [275]).
		Phytophaga (p. [276])		Bruchidae (p. [276])
				Chrysomelidae (p. [278])		Eupoda (p. [280]). Camptosomes (p. [281]). Cyclica (p. [282]). Cryptostomes (p. [282]).
				Cerambycidae (p. [285])		Prionides (p. [287]). Cerambycides (p. [287]). Lamiides (p. [287]).
		Rhynchophora (p. [288])		Anthribidae (p. [290]). Curculionidae (p. [290]). Scolytidae (p. [294]). Brenthidae (p. [295]).
				Aglycyderidae (p. [297]). Protorhinidae (p. [298]).
		Strepsiptera (p. [298])		Stylopidae (p. [298]).

Order.		Sub-order, Division, or Series.		Family.		Sub-Family or Tribe.
LEPIDOPTERA (p. [304])		Rhopalocera (p. [341])		Nymphalidae (p. [343])		Danaides (p. [344]). Ithomiides (p. [346]). Satyrides (p. [347]). Morphides (p. [348]). Brassolides (p. [349]). Acraeides (p. [350]). Heliconiides (p. [351]). Nymphalides (p. [352]).
				Erycinidae (p. [354])		Erycinides (p. [355]). Libytheides (p. [355]).
				Lycaenidae (p. [356]). Pieridae (p. [357]). Papilionidae (p. [359]). Hesperiidae (p. [363])
		Heterocera (p. [366])		Castniidae (p. [371]). Neocastniidae (p. [372]). Saturniidae (p. [372]). Brahmaeidae (p. [374]). Ceratocampidae (p. [375]). Bombycidae (p. [375]). Eupterotidae (p. [376]). Perophoridae (p. [377]). Sphingidae (p. [380]). Cocytiidae (p. [382]). Notodontidae (p. [383]). Cymatophoridae (p. [386]). Sesiidae (p. [386]). Tinaegeriidae (p. [387]). Syntomidae (p. [388]). Zygaenidae (p. [390]). Himantopteridae (p. [392]). Heterogynidae (p. [392]). Psychidae (p. [392]). Cossidae (p. [395]). Arbelidae (p. [396]). Chrysopolomidae (p. [396]). Hepialidae (p. [396]). Callidulidae (p. [400]). Drepanidae (p. [400]). Limacodidae (p. [401]). Megalopyogidae (p. [404]). Thyrididae (p. [404]). Lasiocampidae (p. [405]). Endromidae (p. [406]). Pterothysanidae (p. [406]). Lymantriidae (p. [406]). Hypsidae (p. [408]). Arctiidae (p. [408]). Agaristidae (p. [410]). Geometridae (p. [411]). Noctuidae (p. [414]). Epicopeiidae (p. [418]). Uraniidae (p. [419]). Epiplemidae (p. [420]). Pyralidae (p. [420]). Pterophoridae (p. [426]). Alucitidae (p. [426]). Tortricidae (p. [427]). Tineidae (p. [428]). Eriocephalidae (p. [433]). Micropterygidae (p. [435]).

Order.		Sub-order, Division, or Series.		Family.		Sub-Family or Tribe.
DIPTERA (p. [438])		Orthorrhapha Nemocera (p. [455])		Cecidomyiidae (p. [458]). Mycetophilidae (p. [462]). Blepharoceridae (p. [464]). Culicidae (p. [466]). Chironomidae (p. [468]). Orphnephilidae (p. [470]). Psychodidae (p. [470]). Dixidae (p. [471]).
				Tipulidae (p. [471])		Ptychopterinae (p. [472]). Limnobiinae (p. [473]). Tipulinae (p. [475]).
				Bibionidae (p. [475]). Simuliidae (p. [477]). Rhyphidae (p. [478]).
		Orthorrhapha Brachycera (pp. [455], [478])		Stratiomyidae (p. [478]). Leptidae (p. [479]). Tabanidae (p. [481]). Acanthomeridae (p. [483]). Therevidae (p. [484]). Scenopinidae (p. [484]). Nemestrinidae (p. [484]). Bombyliidae (p. [485]). Acroceridae (p. [489]). Lonchopteridae (p. [490]). Mydaidae (p. [491]). Asilidae (p. [491]). Apioceridae (p. [492]). Empidae (p. [492]). Dolichopidae (p. [493]).
		Cyclorrhapha Asciza (pp. [455], [494])		Phoridae (p. [494]). Platypezidae (p. [496]). Pipunculidae (p. [496]). Conopidae (p. [497]). Syrphidae (p. [498]).
		Cyclorrhapha Schizophora (pp. [456], [503])		Muscidae Acalyptratae (p. [503]). Anthomyiidae (p. [506]). Tachinidae (p. [507]). Dexiidae (p. [510]). Sarcophagidae (p. [510]). Muscidae (p. [511]). Oestridae (p. [514]).
		Pupipara (pp. [456], [517])		Hippoboscidae (p. [518]). Braulidae (p. [520]). Streblidae (p. [521]). Nycteribiidae (p. [521]).
APHANIPTERA(pp. [456], [522])				Pulicidae (p. [522]).
THYSANOPTERA (p. [526])		Terebrantia (p. [531]). Tubulifera (p. [531]).

Order.		Sub-order.		Series.		Family
HEMIPTERA (p. [532])		Heteroptera (pp. [543], [544])		Gymnocerata (p. [544])		Pentatomidae (p. [545]). Coreidae (p. [546]). Berytidae (p. [548]). Lygaeidae (p. [548]). Pyrrhocoridae (p. [549]). Tingidae (p. [549]). Aradidae (p. [550]). Hebridae (p. [551]). Hydrometridae (p. [551]). Henicocephalidae (p. [554]). Phymatidae (p. [554]). Reduviidae (p. [555]). Aëpophilidae (p. [559]). Ceratocombidae (p. [559]). Cimicidae (p. [559]). Anthocoridae (p. [560]). Polyctenidae (p. [560]). Capsidae (p. [561]). Saldidae (p. [562]).
				Cryptocerata (p. [562])		Galgulidae (p. [562]). Nepidae (p. [563]). Naucoridae (p. [565]). Belostomidae (p. [565]). Notonectidae (p. [567]). Corixidae (p. [567]).
		Homoptera (pp. [543], [568])		Trimera (p. [544])		Cicadidae (p. [568]). Fulgoridae (p. [574]). Membracidae (p. [576]). Cercopidae (p. [577]). Jassidae (p. [578]).
				Dimera (p. [544])		Psyllidae (p. [578]). Aphidae (p. [581]). Aleurodidae (p. [591]).
				Monomera (p. [544])		Coccidae (p. [592]).
		Anoplura (p. [599])				Pediculidae (p. [599]).

CHAPTER I

HYMENOPTERA PETIOLATA CONTINUED

SERIES 2. TUBULIFERA OR CHRYSIDIDAE—SERIES 3. ACULEATA—GENERAL—CLASSIFICATION—DIVISION I. ANTHOPHILA OR BEES

The First Series—Parasitica—of the Sub-Order Hymenoptera Petiolata was discussed in the previous volume. We now pass to the Second Series.

Series 2. Hymenoptera Tubulifera.

Trochanters undivided; the hind-body consisting of from three to five visible segments; the female with an ovipositor, usually retracted, transversely segmented, enveloping a fine, pointed style. The larvae usually live in the cells of other Hymenoptera.

The Tubulifera form but a small group in comparison with Parasitica and Aculeata, the other two Series of the Sub-Order. Though of parasitic habits, they do not appear to be closely allied to any of the families of Hymenoptera Parasitica, though M. du Buysson suggests that they have some affinity with Proctotrypidae; their morphology and classification have been, however, but little discussed, and have not been the subject of any profound investigation. At present it is only necessary to recognise one family, viz. Chrysididae or Ruby-wasps.[^[1]] These Insects are usually of glowing, metallic colours, with a very hard, coarsely-sculptured integument. Their antennae are abruptly elbowed, the joints not being numerous, usually about thirteen, and frequently so connected that it is not easy to count them. The abdomen is, in the great majority, of very peculiar construction, and allows the Insect to curl it completely under the anterior parts, so as to roll up into a little ball; the dorsal plates are very strongly arched, and seen from beneath form a free edge, while the ventral plates are of less hard consistence, and are connected with the dorsal plates at some distance from the free edge, so that the abdomen appears concave beneath. In the anomalous genus Cleptes the abdomen is, however, similar in form to that of the Aculeate Hymenoptera, and has four or five visible segments, instead of the three or four that are all that can be seen in the normal Chrysididae. The larvae of the Ruby-flies have the same number of segments as other Hymenoptera Petiolata. The difference in this respect of the perfect Chrysididae from other Petiolata is due to a greater number of the terminal segments being indrawn so as to form the tube, or telescope-like structure from which the series obtains its name. This tube is shown partially extruded in Fig. 1; when fully thrust out it is seen to be segmented, and three or four segments may be distinguished. The ovipositor proper is concealed within this tube; it appears to be of the nature of an imperfect sting; there being a very sharply pointed style, and a pair of enveloping sheaths; the style really consists of a trough-like plate and two fine rods or spiculae. There are no poison glands, except in Cleptes, which form appears to come very near to the Aculeate series. Some of the Chrysididae on occasions use the ovipositor as a sting, though it is only capable of inflicting a very minute and almost innocuous wound.

Fig. 1.—Chrysis ignita, ♀. England.

Although none of the Ruby-flies attain a large size, they are usually very conspicuous on account of their gaudy or brilliant colours. They are amongst the most restless and rapid of Insects; they love the hot sunshine, and are difficult of capture. Though not anywhere numerous in species, they are found in most parts of the world. In Britain we have about twenty species. They usually frequent old wood or masonry, in which the nests of Aculeate Hymenoptera exist, or fly rapidly to and fro about the banks of earth where bees nest. Dr. Chapman has observed the habits of some of our British species.[^[2]] He noticed Chrysis ignita flying about the cell of Odynerus parietum, a solitary wasp that provisions its nest with caterpillars; in this cell the Chrysis deposited an egg, and in less than an hour the wasp had sealed the cell. Two days afterwards this was opened and was found to contain a larva of Chrysis a quarter of an inch long, as well as the Lepidopterous larvae stored up by the wasp, but there was no trace of egg or young of the wasp. Six days after the egg was laid the Chrysis had eaten all the food and was full-grown, having moulted three or four times. Afterwards it formed a cocoon in which to complete its metamorphosis. It is, however, more usual for the species of Chrysis to live on the larva of the wasp and not on the food; indeed, it has recently been positively stated that Chrysis never eats the food in the wasp's cell, but there is no ground whatever for rejecting the evidence of so careful an observer as Dr. Chapman. According to M. du Buysson the larva of Chrysis will not eat the lepidopterous larvae, but will die in their midst if the Odynerus larva does not develop; but this observation probably relates only to such species as habitually live on Odynerus itself. The mother-wasp of Chrysis bidentata searches for a cell of Odynerus spinipes that has not been properly closed, and that contains a full-grown larva of that wasp enclosed in its cocoon. Having succeeded in its search the Chrysis deposits several eggs—from six to ten; for some reason that is not apparent all but one of these eggs fail to produce young; in two or three days this one hatches, the others shrivelling up. The young Chrysis larva seizes with its mouth a fold of the skin of the helpless larva of the Odynerus, and sucks it without inflicting any visible wound. In about eleven days the Chrysis has changed its skin four times, has consumed all the larva and is full-fed; it spins its own cocoon inside that of its victim, and remains therein till the following spring, when it changes to a pupa, and in less than three weeks thereafter emerges a perfect Chrysis of the most brilliant colour, and if it be a female indefatigable in activity. It is remarkable that the larva of Chrysis is so much like that of Odynerus that the two can only be distinguished externally by the colour, the Odynerus being yellow and the Chrysis white; but this is only one of the many cases in which host and parasite are extremely similar to the eye. Chrysis shanghaiensis has been reared from the cocoons of a Lepidopterous Insect—Monema flavescens, family Limacodidae—and it has been presumed that it eats the larva therein contained. All other Chrysids, so far as known, live at the expense of Hymenoptera (usually, as we have seen, actually consuming their bodies), and it is not impossible that C. shanghaiensis really lives on a Hymenopterous parasite in the cocoon of the Lepidopteron.

Parnopes carnea frequents the nests of Bembex rostrata, a solitary wasp that has the unusual habit of bringing from time to time a supply of food to its young larva; for this purpose it has to open the nest in which its young is enclosed, and the Parnopes takes advantage of this habit by entering the cell and depositing there an egg which produces a larva that devours that of the Bembex. The species of the anomalous genus Cleptes live, it is believed, at the expense of Tenthredinidae, and in all probability oviposit in their cocoons which are placed in the earth.

Series 3. Hymenoptera Aculeata.

The females (whether workers or true females) provided with a sting: trochanters usually undivided (monotrochous). Usually the antennae of the males with thirteen, of the females with twelve, joints (exceptions in ants numerous).

These characters only define this series in a very unsatisfactory manner, as no means of distinguishing the "sting" from the homologous structures found in Tubulifera, and in the Proctotrypid division of Hymenoptera Parasitica, have been pointed out. As the structure of the trochanters is subject to numerous exceptions, the classification at present existing is an arbitrary one. It would probably be more satisfactory to separate the Proctotrypidae (or a considerable part thereof) from the Parasitica, and unite them with the Tubulifera and Aculeata in a great series, characterised by the fact that the ovipositor is withdrawn into the body in a direct manner so as to be entirely internal, whereas in the Parasitica it is not withdrawn in this manner, but remains truly an external organ, though in numerous cases concealed by a process of torsion of the terminal segments. If this were done it might be found possible to divide the great group thus formed into two divisions characterised by the fact that the ovipositor in one retains its function, the egg passing through it (Proctotrypidae and Tubulifera), while in the other the organ in question serves as a weapon of offence and defence, and does not act as a true ovipositor, the egg escaping at its base. It would, however, be premature to adopt so revolutionary a course until the comparative anatomy of the organs concerned shall have received a much greater share of attention; a detailed scrutiny of Prototrypidae being particularly desired.

Fig. 2.—Diagram of upper surface of Priocnemis affinis ♀, Pompilidae. o, ocelli; B¹, pronotum; B², mesonotum; B³, scutellum of mesonotum; B⁴, post-scutellum or middle part of metanotum; B⁵, propodeum or median segment (see vol. v. p. 491); B⁶, combing hairs, pecten, of front foot: C¹, first segment of abdomen, here not forming a pedicel or stalk: D¹, coxa; D², trochanter; D³, femur; D⁶, calcaria or spurs of hind leg: 1 to 15, nervures of wings, viz. 1, costal; 2, post-costal; 3, median; 4, posterior; 5, stigma; 6, marginal; 7, upper basal; 8, lower basal; 9, 9, cubital; 10, the three sub-marginal; 11, first recurrent; 12, second recurrent; 13, anterior of hind wing; 14, median; 15, posterior: I to XI, the cells, viz. I, upper basal; II, lower basal; III, marginal; IV, V, VI, first, second and third sub-marginal; VII, first discoidal; VIII, third discoidal; IX, second discoidal; X, first apical; XI, second apical.

We have dealt with the external anatomy of Hymenoptera in Vol. V.; so that here it is only necessary to give a diagram to explain the terms used in the descriptions of the families and sub-families of Aculeata, and to discuss briefly their characteristic structures.

Fig. 3—Sting of bee. A, One of the needles separated; a, the barbed point; b, piston; c, arm. B, Transverse section of the sting: dd, the two needles; e, bead for guiding the needles; f, director; g, channel of poison. (After Carlet.)

The Sting of the bee has been described in detail by Kraepelin, Sollmann, Carlet[^[3]] and others. It is an extremely perfect mechanical arrangement. The sting itself—independent of the sheaths and adjuncts—consists of three elongate pieces, one of them a gouge-like director, the other two pointed and barbed needles; the director is provided with a bead for each of the needles to run on, these latter having a corresponding groove; the entrance to the groove is narrower than its subsequent diameter, so that the needles play up and down on the director with facility, but cannot be dragged away from it; each needle is provided with an arm at the base to which are attached the muscles for its movement. This simple manner of describing the mechanical arrangement is, however, incomplete, inasmuch as it includes no account of the means by which the poison is conveyed. This is done by a very complex set of modifications of all the parts; firstly, the director is enlarged at the anterior part to form a chamber, through which the needles play; the needles are each provided with a projecting piece, which, as the needle moves, plays in the chamber of the director, and forces downwards any liquid that may be therein; the poison-glands open into the chamber, and the projections on the needles, acting after the manner of a piston, carry the poison before them. The needles are so arranged on the director that they enclose between themselves and it a space that forms the channel along which the poison flows, as it is carried forwards by the movement of the pistons attached to the needles. If the needles be thrust into an object quite as far as, or beyond, the point of the director much poison may be introduced into a wound, as the barbs are provided with small orifices placed one above the other, while if this be not the case much of the liquid will flow on the outside of the object.

According to Carlet the poison of the bee is formed by the mixture of the secretions of two glands, one of which is acid and the other alkaline; it is very deadly in its effects on other Insects. We shall see, however, that the Fossorial Hymenoptera, which catch and sting living prey for their young, frequently do not kill but only stupefy it, and Carlet states that in this group the alkaline gland is absent or atrophied, so that the poison consists only of the acid; it is thus, he thinks, deprived of its lethal power. Moreover, in the Fossoria the needles are destitute of barbs, so that the sting does not remain in the wound. Bordas, however, states[^[4]] that in all the numerous Hymenoptera he has examined, both acid and alkaline glands exist, but exhibit considerable differences of form in the various groups. He gives no explanation of the variety of effects of the poison of different Aculeata.

The larvae (for figure of larva of Bombus, see Vol. V. p. 488) are, without known exception, legless grubs, of soft consistence, living entirely under cover, being protected either in cells, or, in the case of social Hymenoptera, in the abodes of the parents. The larvae of Ants and fossorial Hymenoptera have the anterior parts of the body long and narrow and abruptly flexed, so that their heads hang down in a helpless manner. All the larvae of Aculeates, so far as known, are remarkable from the fact that the posterior part of the alimentary canal does not connect with the stomach till the larval instar is more or less advanced; hence the food amongst which they live cannot be sullied by faecal matter. The pupa is invariably soft, and assumes gradually the colour of the perfect Insect. Almost nothing is known as to the intimate details of the metamorphosis, and very little as to the changes of external form. According to Packard a period intervenes between the stadium of the full-grown larva and that of the pupa, in which a series of changes he speaks of as semi-pupal are passed through; these, however, have not been followed out in the case of any individual, and it is not possible to form any final idea about them, but it seems probable that they are largely changes of external shape, in conformity with the great changes going on in the internal organs. Owing to the fragmentary nature of observations, much obscurity and difference of opinion have existed as to the metamorphosis of Aculeate Hymenoptera. Sir S. Saunders gives the following statement as to the larva of a wasp of the genus Psiliglossa,[^[5]] just before it assumes the pupal form: "The respective segments, which are very distinctly indicated, may be defined as follows:—The five anterior, including the head, are compactly welded together, and incapable of separate action in the pseudo-pupa state; the third, fourth, and fifth bearing a spiracle on either side. The thoracical region terminating here, the two anterior segments are assignable to the development of the imago head, as pointed out by Ratzeburg." This inference is not, however, correct. We have seen that in the perfect Insect of Petiolate Hymenoptera the first abdominal segment is fixed to the thorax, and Saunders' statement is interesting as showing that this assignment of parts already exists in the larva, but it in no way proves that the head of the imago is formed from the thorax of the larva. It has been stated that the larvae of the Aculeata have a different number of segments according to the sex, but this also is incorrect. The difference that exists in the perfect Insects in this respect is due to the withdrawal of the terminal three segments to the interior in the female, and of two only in the male. The larva consists of fourteen segments, and we find this number distributed in the female perfect Insect as follows: one constitutes the head, four segments the thorax and propodeum, followed by six external segments of the restricted abdomen, and three for the internal structures of the abdomen. This agrees with Forel's statement that in the ants the sting is placed in a chamber formed by three segments.

The development of the sting of the common bee has been studied by Dewitz.[^[6]] It takes place in the last larval stage. Although nothing of the organ is visible externally in the adult larva, yet if such a larva be placed in spirit, there can be seen within the skin certain small appendages on the ventral surface of the penultimate and antepenultimate abdominal segments (Fig. 4, A) placed two on the one, four on the other; these are the rudiments of the sting. In the course of development the terminal three segments are taken into the body, and the external pair of the appendages of the twelfth body segment (the ninth abdominal) become the sheaths of the sting, and the middle pair become the director; the pair of appendages on the eleventh segment give rise to the needles or spiculae. The sting-rudiments at an earlier stage (Fig. 4, C) are masses of hypodermis connected with tracheae; there is then but one pair on the twelfth segment, and this pair coalesce to form a single mass; the rudiments of the pair that form the director are differentiated secondarily from the primary pair of these masses of hypodermis. A good deal of discussion has taken place as to whether the component parts of the sting—gonapophyses—are to be considered as modifications of abdominal extremities (i.e. abdominal legs such as exist in Myriapods). Heymons is of opinion that this is not the case, but that the leg-rudiments and gonapophysal rudiments are quite distinct.[^[7]] The origin of the sting of Hymenoptera (and of the ovipositor of parasitic Hymenoptera) is very similar to that of the ovipositor of Locusta (Vol. V. p. 315 of this work), but there is much difference in the history of the development of the rudiments.

Fig. 4—Development of sting of the bee: A and C, ventral; B, side view. A, End of abdomen of adult larva: a, b, c, d, the last four segments, c being the eleventh body segment, 11; b bearing two pairs, and c one pair, of rudiments. B, Tip of abdomen of adult bee: 9, the ninth, d, the tenth body segment. C, Rudiments in the early condition as seen within the body: c, first pair; b, the second pair not yet divided into two pairs; b″, c′, commencement of external growths from the internal projections. (After Dewitz.)

Dewitz has also traced the development of the thoracic appendages in Hymenoptera.[^[8]] Although no legs are visible in the adult larva, they really arise very early in the larval life from masses of hypodermis, and grow in the interior of the body, so that when the larva is adult the legs exist in a segmented though rudimentary condition in the interior of the body. Dewitz's study of the wing-development is less complete.

Four primary divisions of Aculeates are generally recognised, viz. Anthophila (Bees), Diploptera (Wasps), Fossores (Solitary Wasps), Heterogyna (Ants). Though apparently they are natural, it is impossible to define them by characters that are without some exceptions, especially in the case of the males. Ashmead has recently proposed[^[9]] to divide the Fossores; thus making five divisions as follows:—

Body with more or less of the hairs on it plumose .......... 1. Anthophila.

Hairs of body not plumose.

Pronotum not reaching back to tegulae .......... 2. Entomophila [= Fossores part].

Pronotum reaching back to tegulae.

Petiole (articulating segment of abdomen) simple without scales or nodes.

Front wings in repose with a fold making them narrow .......... 3. Diploptera.

Front wings not folded .......... 4. Fossores [part].

Petiole with a scale or node (an irregular elevation on the upper side) .......... 5. Heterogyna.

We shall here follow the usual method of treating all the fossorial wasps as forming a single group, uniting Ashmead's Entomophila and Fossores, as we think their separation is only valid for the purposes of a table; the Pompilidae placed by the American savant in Fossores being as much allied to Entomophila as they are to the other Fossores with which Ashmead associates them.

Division I. Anthophila or Apidae—Bees.

Some of the hairs of the body plumose; parts of the mouth elongated, sometimes to a great extent, so as to form a protrusible apparatus, usually tubular with a very flexible tip. Basal joint of hind foot elongate. No wingless adult forms; in some cases societies are formed, and then barren females called workers exist in great numbers, and carry on the industrial operations of the community. Food always derived from the vegetable kingdom, or from other Bees.

There are about 150 genera and 1500 species of bees at present known. Some call the division Mellifera instead of Anthophila. The term Apidae is used by some authorities to denote all the bees, while others limit this term to one of the families or sub-divisions. The bees are, as a rule, distinguished from other Hymenoptera by the hairs, by the great development of the mouth parts to form a proboscis (usually, but not correctly, called tongue), and by the modification of the hind-legs; but these distinctive characters are in some of the species exhibited in so minor a degree of perfection that it is not easy to recognise these primitive forms as Anthophila. A few general remarks on the three points mentioned will enable the student to better appreciate the importance of certain points we shall subsequently deal with.

Fig. 5—Hairs of Bees: A, simple hair from abdomen of Osmia; B, spiral hair from abdomen of Megachile; C, plumose hair from thorax of Megachile; D, from thorax of Andrena dorsata; E, from thorax of Prosopis.

The bees are, as a rule, much more covered with hair than any other of the Hymenoptera. Saunders[^[10]] states that he has examined the structure of the hairs in all the genera of British Aculeata, and that in none but the Anthophila do branched and plumose hairs occur. The function of this kind of hairs is unknown; Saunders suggests[^[10]] that they may be instrumental in the gathering of pollen, but they occur in the parasitic bees as well as in the males, neither of which gather pollen. The variety of the positions they occupy on the body seems to offer but little support to the suggestion. Not all the hairs of the bee's body are plumose, some are simple, as shown in Fig. 5, A, and this is specially the case with the hairs that are placed at the edges of the dilated plates for carrying pollen. In some forms there is an extensive system of simple hairs all over the body, and the "feathers" are distributed between these; and we do not see any reason for assuming that the feathered are superior to the simple hairs for gathering and carrying pollen. Some bees, e.g. Prosopis, Ceratina, have very little hair on the body, but nevertheless some plumose hairs are always present even though they be very short.

Fig. 6—A, Worker of the honey-bee (Apis mellifica), with pollen plates laden; B, basal portions of a middle-leg (trochanter with part of coxa and of femur) with plumose hairs and grains of pollen; C, one hair bearing pollen-grains.

The hind-legs of bees are very largely used in the industrial occupations of these indefatigable creatures; one of their chief functions in the female being to act as receptacles for carrying pollen to the nest: they exhibit, however, considerable diversity. The parts most modified are the tibia and the first joint of the hind-foot. Pollen is carried by other parts of the body in many bees, and even the hind-leg itself is used in different ways for the purpose: sometimes the outer face of the tibia is highly polished and its margins surrounded by hair, in which case pollen plates are said to exist (Fig. 6, A); sometimes the first joint of the tarsus is analogous to the tibia both in structure and function; in other cases the hind-legs are thick and densely covered with hair that retains the pollen between the separate hairs. In this case the pollen is carried home in a dry state, while, in the species with pollen plates, the pollen is made into a mass of a clay-like consistence.[^[11]] The legs also assist in arranging the pollen on the other parts of the body. The males do not carry pollen, and though their hind-legs are also highly modified, yet the modifications do not agree with those of the female, and their functions are in all probability sexual. The parasitic bees also do not carry pollen, and exhibit another series of structures. The most interesting case in this series of modifications is that found in the genus Apis, where the hind-leg of male, female, and worker are all different (Fig. 25); the limb in the worker being highly modified for industrial purposes. This case has been frequently referred to, in consequence of the difficulty that exists in connection with its heredity, for the structure exists in neither of the parents. It is, in fact, a case of a very special adaptation appearing in the majority of the individuals of each generation, though nothing of the sort occurs in either parent.

The proboscis of the bee[^[12]] is a very complex organ, and in its extremely developed forms exhibits a complication of details and a delicacy of structure that elicit the admiration of all who study it. In the lower bees, however, especially in Prosopis, it exists in a comparatively simple form (Fig. 9, B, C), that differs but little from what is seen in some Vespidae or Fossores. The upper lip and the mandibles do not take any part in the formation of the bee's proboscis, which is consequently entirely made up from the lower lip and the maxillae, the former of these two organs exhibiting the greatest modifications. The proboscis is situate on the lower part of the head, and in repose is not visible; a portion, and that by no means an inconsiderable one, of its modifications being for the purpose of its withdrawal and protection when not in use. For this object the under side of the head is provided with a very deep groove, in which the whole organ is, in bees with a short proboscis, withdrawn; in the Apidae with a long proboscis this groove also exists, and the basal part of the proboscis is buried in it during repose, while the other parts of the elongate organ are doubled on the basal part, so that they extend backwards under the body, and the front end or tip of the tongue is, when in repose, its most posterior part.

For the extrusion of the proboscis there exists a special apparatus that comes into play after the mandibles are unlocked and the labrum lifted. This extensive apparatus cannot be satisfactorily illustrated by a drawing, as the parts composing it are placed in different planes; but it may be described by saying that the cardo, or basal hinge of the maxilla, changes from an oblique to a vertical position, and thrusts the base of the proboscis out of the groove. The maxillae form the outer sheath of the proboscis, the lower lip its medial part (see Figs. 7 and 9); the base of the lower lip is attached to the submentum, which rises with the cardo so that labium and maxillae are lifted together; the co-operation of these two parts is effected by an angular piece called the lorum, in which the base of the submentum rests; the submentum is articulated with the mentum in such a manner that the two can either be placed in planes at a right angle to one another, or can be brought into one continuous plane, and by this change of plane the basal part of the tongue can also be thrust forwards.

Fig. 7.—Side view of basal portions of proboscis of Bombus. a, Epipharyngeal sclerites; b, arrow indicating the position of the entrance to pharynx, which is concealed by the epipharynx, c; d, hypopharyngeal sclerites; e, vacant space between the scales of the maxillae through which the nectar comes: f, lobe; f′, stipes; g, cardo of maxilla: h, encephalic pillar on which the cardo swings; i, angle of junction of lores and submentum lorum; k, mentum; l, base of labial palp; m, maxillary palp.

There is considerable variety in the lengths of these parts in different genera, and the lorum varies in shape in accordance with the length of the submentum. The lorum is a peculiar piece, and its mechanical adaptations are very remarkable; usually the base of the submentum rests in the angle formed by the junction of the two sides of the lorum, but in Xylocopa, where the submentum is unusually short, this part reposes in a groove on the back of the lorum, this latter having a very broad truncated apex instead of an angular one; in the condition of repose the apex of the lorum rests in a notch on the middle of the back of the oral groove, and in some of the forms with elongate submentum, this depression is transformed into a deep hole, or even a sort of tunnel, so as to permit the complete stowing away of the base of the tongue, which would otherwise be prevented by the long submentum; another function of the lorum appears to be that, as it extends, its arms have an outward thrust, and so separate the maxillae from the labium. In addition to these parts there are also four elongate, slender sclerites that are only brought into view on dissection, and that no doubt assist in correlating the movements of the parts of the mouth and hypopharynx; one pair of these strap-like pieces extends backwards from the two sides of the base of the epipharynx; Huxley called them sclerites of the oesophagus; a better name would be epipharyngeal sclerites (Fig. 7, a): the other pair pass from the terminations of the epipharyngeal sclerites, along the front face of the hypopharynx, down to the mentum, their lower parts being concealed by the stipites of the maxillae; these are the hypopharyngeal sclerites, and we believe it will prove that they play a highly important part in deglutition. When the labrum of a bee is raised and the proboscis depressed, the epipharynx is seen hanging like a curtain from the roof of the head; this structure plays an important part in the act of deglutition. The entrance to the pharynx, or commencement of the alimentary canal, is placed below the base of the epipharynx. As we are not aware of any good delineations of the basal parts of the proboscis we give a figure thereof (Fig. 7). The maxillae in the higher bees are extremely modified so as to form a sheath, and their palpi are minute; in the lower bees the palpi have the structure usual in mandibulate Insects.

Returning to the consideration of the lower lip, we find that there is attached to the mentum a pair of elongate organs that extend forwards and form a tube or sheath, enclosed by the maxillary sheath we have previously mentioned; these are the greatly modified labial palpi, their distal parts still retaining the palpar form; and in the lower bees the labial palpi are, like the maxillary, of the form usual in mandibulate Insects. Between the labial palps and the central organ of the lip there is attached a pair of delicate organs, the paraglossae.

There remains for consideration the most remarkable part of the proboscis, the long, delicate, hairy organ which the bee thrusts out from the tip of the shining tube formed by the labial palps and the maxillae, described above, and which looks like a prolongation of the mentum. This organ is variously called ligula, lingua, or tongue.[^[13]] We prefer the first of these names.

According to Breithaupt and Cheshire the structure of the ligula is highly remarkable; it is a tube (filled with fluid from the body cavity), and with a groove underneath caused by a large part of the circumference of the tube being invaginated; the invaginated part can be thrust out by increase of the pressure of the fluid in the tube. A portion of the wall of the invaginate part is thickened so as to form a chitinous rod.

This description will suffice for present purposes, as the other parts of the mouth will be readily recognised by the aid of figure 9, A, B, C. In the exquisitely endowed South American genus Euglossa (Fig. 18), the proboscis is somewhat longer than the whole of the body, so that its tip in repose projects behind the body like a sting.

Fig. 8.—Transverse section of ligula of honey-bee, diagramatic. A, With the long sac invaginate. B, evaginate: a, chitinous envelope with the bases of the hairs; b, rod; c, groove of rod; d, lumen due in A to invagination of the rod, in B to its evagination; n, nerve; tr, trachea.

The correct nomenclature of the parts connected with the lower lip is not definitely settled, authorities not being agreed on several points. The whole of the proboscis is usually called the tongue; this, however, is admittedly an erroneous application of this term. The terminal delicate, elongate, flexible organ is by some called the tongue; but this again is wrong: the lingua in Insects is the hypopharynx; this part is developed in a peculiar manner in bees, but as it is not tongue-like in shape, the term lingua is not suitable for it, and should be dismissed altogether from the nomenclature of the bee's trophi; it is used at present in two different senses, both of which are erroneous. We see no objection to describing the flexible apical portion of the proboscis as the ligula. The lorum is probably a special part peculiar to the higher bees; according to Saunders it is not present as a specialised part in some of the primitive forms.[^[14]] The application of the terms mentum, submentum and hypoglottis is open to the same doubts that exist with regard to them in so many other Insects, and we have omitted the term hypoglottis altogether, though some may think the mentum entitled to that name.

Fig. 9.—A, Proboscis of a "long-tongued" bee, Anthophora pilipes; B, lower, C, upper view of proboscis of an "obtuse-tongued" bee, Prosopis pubescens. a, Labrum; b, stipes; c, palpiger; d, scale: f, lobe; g, palpus; h, cardo, of maxilla: i, lorum; k, submentum; l, mentum; m, labial palp; n, paraglossa; o, ligula; p, tip of ligula (with "spoon" at tip and some of the hairs more magnified); q, hypopharyngeal sclerites.

The way in which the proboscis of the bee acts has been very largely discussed, with special reference to the question as to whether it is a sucking or a licking action. It is impossible to consider either of these terms as applicable. The foundation of the action is capillary attraction, by which, and by slight movements of increase and contraction of the capacity of various parts, the fluid travels to the cavity in front of the hypopharynx: here the scales of the maxillae leave a vacant space, (Fig. 7, e) so that a cup or cavity is formed, the fluid in which is within reach of the tip of the dependent epipharynx (c), which hangs down over the front of the hypopharynx (and is so shaped that its tip covers the cup); it is between these two parts that the fluid passes to reach the pharynx. It is no doubt to slight movements of the membranous parts of the hypopharynx and of the epipharynx that the further progress of the nectar is due, aided by contraction and expansion of the pharynx, induced by muscles attached to it. It should be recollected that in addition to the movements of the head itself, the hypopharynx is constantly changing its dimensions slightly by the impulses of the fluid of the general body cavity; also that the head changes its position, and that the proboscis is directed downwards as well as forwards. Those who wish to pursue this subject should refer to the works of Breithaupt[^[15]] and Cheshire.

The other external characters of the Bees call for little remark. The pronotum is never very large or much prolonged in front, and its hind angles never repose on the tegulae as they do in the wasps,[^[16]] but extend backwards below the tegulae. The hind body is never narrowed at the base into an elongate pedicel, as it so frequently is in the Wasps and in the Fossors; and the propodeum (the posterior part of the thorax) is more perpendicular and rarely so largely developed as it is in the Fossors; this last character will as a rule permit a bee to be recognised at a glance from the fossorial Hymenoptera.

Bees, as every one knows, frequent flowers, and it is usually incorrectly said that they extract honey. They really gather nectar, swallow it, so that it goes as far as the crop of their alimentary canal, called in English the honey-sac, and is regurgitated as honey. Bertrand states that the nectar when gathered is almost entirely pure saccharose, and that when regurgitated it is found to consist of dextrose and levulose:[^[17]] this change appears to be practically the conversion of cane- into grape-sugar. A small quantity of the products of the salivary glands is added, and this probably causes the change alluded to; so that honey and nectar are by no means synonymous. According to Cheshire the glandular matter is added while the nectar is being sucked, and is passing over the middle parts of the lower lip, so that the nectar may be honey when swallowed by the bee. In addition to gathering nectar the female bees are largely occupied in collecting pollen, which, mixed with honey, is to serve as food for the colony. Many, if not all, bees eat pollen while collecting it. The mode in which they accumulate the pollen, and the mechanism of its conveyance from hair to hair till it reaches the part of the body it must attain in order to be removed for packing in the cells, is not fully understood, but it appears to be accomplished by complex correlative actions of various parts; the head and the front legs scratch up the pollen, the legs move with great rapidity, and the pollen ultimately reaches its destination. The workers of the genus Apis, and of some other social bees, have the basal joint of the hind foot specially adapted to deal with pollen (Fig. 25, 2). We have already mentioned the modifications of the legs used for its conveyance, and need here only add that numerous bees—the Dasygastres—carry the pollen by aid of a special and dense clothing of hairs on the underside of the abdomen.

The buzzing of bees (and other Insects) has been for long a subject of controversy: some having maintained that it is partially or wholly due to the vibration of parts connected with the spiracles, while others have found its cause in the vibrations of the wings. According to the observations of Pérez and Bellesme,[^[18]] two distinct sounds are to be distinguished. One, a deep noise, is due to the vibration of the wings, and is produced whenever a certain rapidity is attained; the other is an acute sound, and is said to be produced by the vibrations of the walls of the thorax, to which muscles are attached; this sound is specially evident in Diptera and Hymenoptera, because the integument is of the right consistence for vibration. Both of these observers agree that the spiracles are not concerned in the matter.

The young of bees are invariably reared in cells. These (except in the case of the parasitical bees) are constructed by the mothers, or by the transformed females called workers. The solitary bees store the cells with food, and close up each cell after having laid an egg in it, so that in these cases each larva consumes a special store previously provided for it. The social bees do not close the cells in which the larvae are placed, and the workers act as foster-mothers, feeding the young larvae after the same fashion as birds feed their nestling young. The food is a mixture of honey and pollen, the mixing being effected in various ways and proportions according to the species; the honey seems to be particularly suitable to the digestive organs of the young larvae, and those bees that make closed cells, place on the outside of the mass of food a layer more thickly saturated with honey, and this layer the young grub consumes before attacking the drier parts of the provisions. The active life of the larva is quite short, but after the larva is full-grown it usually passes a more or less prolonged period in a state of quiescence before assuming the pupal form. The pupa shows the limbs and other parts of the perfect Insect in a very distinct manner, and the development of the imago takes place gradually though quickly. Some larvae spin cocoons, others do not.

A very large number of bees are parasitic in their habits, laying an egg, or sometimes more than one, in the cell of a working bee of some species other than their own; in such cases the resulting larvae eat and grow more quickly than the progeny of the host bee, and so cause it to die of starvation. It has been observed that some of these parasitic larvae, after eating all the store of food, then devour the larva they have robbed. In other cases it is possible that the first care of the parasitic larva, after hatching, is to eat the rival egg.

The taxonomy of bees is in a very unsatisfactory state. The earlier Hymenopterists were divided into two schools, one of which proposed to classify the bees according to their habits, while the other adopted an arrangement depending on the length of the parts of the mouth, the development of the palpi, and the form and positions of the organs for carrying pollen. Neither of these arrangements was at all satisfactory, and some entomologists endeavoured to combine them, the result being a classification founded partly on habits and partly on certain minor structural characters. This course has also proved unsatisfactory; this is especially the case with exotic bees, which have been placed in groups that are defined by habits, although very little observation has actually been made on this point. Efforts have recently been made to establish an improved classification, but as they relate solely to the European bees they are insufficient for general purposes.

The more important of the groups that have been recognised are—(1) the Obtusilingues, short-tongued bees, with the tip of the lingua bifid or broad; (2) Acutilingues, short-tongued bees, with acute tip to the tongue; these two groups being frequently treated of as forming the Andrenidae. Coming to the Apidae, or the bees with long and folded tongues, there have been distinguished (3) Scopulipedes, bees carrying pollen with their feet, and (4) Dasygastres, those that carry it under the abdomen; some of the parasitic and other forms have been separated as (5) Denudatae (or Cuculinae); the Bombi and the more perfectly social bees forming another group, viz. (6) Sociales. A group Andrenoides, or Panurgides, was also proposed for certain bees considered to belong to the Apidae though exhibiting many points of resemblance with the Andrenidae. This arrangement is by no means satisfactory, but as the tropical bees have been but little collected, and are only very imperfectly known, it is clear that we cannot hope for a better classification till collections have been very much increased and improved. The arrangement adopted in Dalla Torre's recent valuable catalogue of bees[^[19]] recognises no less than fourteen primary divisions, but is far from satisfactory.

Fig. 10—Prosopis signata. Cambridge. A, Female; B, front of head of female; C, of male.

The two genera Prosopis and Sphecodes have been recently formed into a special family, Archiapidae, by Friese,[^[20]] who, however, admits that the association is not a natural one. The term should be limited to Prosopis and the genera into which it has been, or shortly will be, divided. The primitive nature of the members of this genus is exhibited in all the external characters that are most distinctive of bees; the proboscis (Fig. 9, B, C), is quite short, its ligula being very short, and instead of being pointed having a concave front margin. The body is almost bare, though there is some very short feathered plumage. The hind legs are destitute of modifications for industrial purposes. Owing to these peculiarities it was for long assumed that the species of Prosopis must be parasites. This is, however, known not to be the case so far as many of the species are concerned. They form cells lined with a silken membrane in the stems of brambles and other plants that are suitable, or in burrows in the earth, or in the mortar of walls; individuals of the same species varying much as to the nidus they select. The food they store in these cells is much more liquid than usual, and has been supposed to be entirely honey, since they have no apparatus for carrying pollen. Mr. R. C. L. Perkins has, however, observed that they swallow both pollen and nectar, brushing the first-named substance to the mouth by aid of the front legs. He has ascertained that a few of the very numerous Hawaiian species of the genus are really parasitic on their congeners: these parasites are destitute of a peculiar arrangement of hairs on the front legs of the female, the possession of which, by some of the non-parasitic forms, enables the bee to sweep the pollen towards its mouth. These observations show that the structural peculiarities of Prosopis are correlative with the habits of forming a peculiar lining to the cell, and of gathering pollen by the mouth and conveying it by the alimentary canal instead of by external parts of the body. Prosopis is a very widely distributed genus, and very numerous in species. We have ten in Britain; several of them occur in the grounds of our Museum at Cambridge.

The species of the genus Colletes are hairy bees of moderate size, with a good development of hair on the middle and posterior femora for carrying pollen. They have a short, bilobed ligula like that of wasps, and therein differ from the Andrenae, which they much resemble. With Prosopis they form the group Obtusilingues of some taxonomists. They have a manner of nesting peculiar to themselves; they dig cylindrical burrows in the earth, line them with a sort of slime, that dries to a substance like gold-beater's skin, and then by partitions arrange the burrow as six to ten separate cells, each of which is filled with food that is more liquid than usual in bees. Except in regard to the ligula and the nature of the cell-lining, Colletes has but little resemblance to Prosopis; but the term Obtusilingues may be applied to Colletes if Prosopis be separated as Archiapidae. We have six species of Colletes in Britain.

Sphecodes is a genus that has been the subject of prolonged difference of opinion. The species are rather small shining bees, with a red, or red and black, abdomen, almost without pollen-collecting apparatus, and with a short but pointed ligula. These characters led to the belief that the Insects are parasitic, or, as they are sometimes called, cuckoo-bees. But evidence could not be obtained of the fact, and as they were seen to make burrows it was decided that we have in Sphecodes examples of industrial bees extremely ill endowed for their work. Recent observations tend, however, to prove that Sphecodes are to a large extent parasitic at the expense of bees of the genera Halictus and Andrena. Breitenbach has taken S. rubicundus out of the brood-cells of Halictus quadricinctus; and on one of the few occasions on which this bee has been found in Britain it was in circumstances that left little doubt as to its being a parasite of Andrena nigroaenea. Marchal[^[21]] has seen S. subquadratus fight with Halictus malachurus, and kill it previous to taking possession of its burrows; and similar observations have been made by Ferton. As the older observations of Smith, Sichel, and Friese leave little doubt that Sphecodes are sometimes industrial bees, it is highly probable that we have in this genus the interesting condition of bees that are sometimes parasitic, at other times not; but so much obscurity still prevails as to the habits of Sphecodes that we should do well to delay accepting the theories that have been already based on this strange state of matters.[^[22]] Friese states that in Sphecodes the first traces of collecting apparatus exist; and, accepting the condition of affairs as being that mentioned above, it is by no means clear whether we have in Sphecodes bees that are abandoning the parasitic habit or commencing it; or, indeed, whether the condition of uncertainty may not be a permanent one. It is difficult to decide as to what forms are species in Sphecodes owing to the great variation. The Hymenopterist Forster considered that 600 specimens submitted to him by Sichel represented no less than 140 species, though Sichel was convinced that nearly the whole of them were one species, S. gibbus. It has recently been found that the male sexual organs afford a satisfactory criterion. The position of Sphecodes in classification is doubtful.

Fig. 11.—Sphecodes gibbus ♀. Britain.

The great majority of the species of short-tongued bees found in Britain belong to the genera Andrena and Halictus, and with some others constitute the Andrenides of many writers. Halictus includes our smallest British bees. Their economy escaped the earlier observers, but has recently been to some extent unravelled by Smith, Fabre, Nicolas, Verhoeff, and others, and proves to be of great interest and variety. Fabre observed H. lineolatus and H. sexcinctus[^[23]] under circumstances that enabled him to give them continuous attention, whenever requisite, throughout a whole year. These bees are to a certain extent social; they are gregarious; each bee works for its own progeny, but there is collaboration between members of a colony, inasmuch as a piece of general work is undertaken from which more families than one derive benefit. This common work is a gallery, that, ramifying in the earth, gives access to various groups of cells, each group the production of a single Halictus; in this way one entrance and one corridor serve for several distinct dwellings. The work of excavation is carried on at night. The cells are oval, and are covered on the interior with a delicate waterproof varnish; Fabre considers this to be a product of the salivary glands, like the membrane we noticed when speaking of Colletes. In the south of France both sexes of these species are produced from the nests in September, and then the males are much more numerous than the females; when the cold weather sets in the males die, but the females continue to live on in the cells underground. In the following spring the females come out and recommence working at the burrows, and also provision the cells for the young; the new generation, consisting entirely of females, appears in July, and from these there proceeds a parthenogenetic generation, which assumes the perfect form in September, and consists, as we have above remarked, in greater part of males. Pérez,[^[24]] however, considers that Fabre's observations as to the parthenogenetic generation were incomplete, and that males might have been found a little earlier, and he consequently rejects altogether the occurrence of parthenogenesis in Halictus. Nicolas confirms Fabre's observations, so far as the interesting point of the work done for common benefit is concerned; and adds that the common corridor being too narrow to permit of two bees passing, there is a dilatation or vestibule near the entrance that facilitates passage, and also that a sentinel is stationed at this point.

Smith's observations on Halictus morio in England lead one to infer that there is but one generation, the appearance of which extends over a very long period. He says, "Early in April the females appeared, and continued in numbers up to the end of June"; then there was an interval, and in the middle of August males began to appear, followed in ten or twelve days by females. Hence it is probable that in different countries the times of appearance and the number of generations of the same species may vary. Verhoeff has described the burrows of Halictus quadricinctus with some detail. The cells, instead of being distributed as usual throughout the length of the burrow one by one, are accumulated into a mass placed in a vault communicating with the shaft. This shaft is continued downwards to a depth of 10 cm., and forms a retreat for the bees when engaged in construction. Several advantages are secured by this method, especially better ventilation, and protection from any water that may enter the shaft. The larvae that are present in the brood-chambers at any one moment differ much in their ages, a fact that throws some doubt on the supposed parthenogenetic generation. No cocoons are formed by these Halictus, the polished interior of the cell being a sufficiently refined resting place for metamorphosis. Verhoeff states that many of the larvae are destroyed by mouldiness; this indeed, he considers to be the most deadly of the enemies of Aculeate Hymenoptera. The nest of Halictus maculatus has also been briefly described by Verhoeff, and is a very poor construction in comparison with that of H. quadricinctus.

Fig. 12—Nesting of Halictus quadricinctus. u, Original burrow, with entrance e thereto; n, retreat or continuation of the burrow; w, the vaults; s, the accumulation of cells. (After Verhoeff, Verh. Ver. Rheinl. xlviii. 1891; scale not mentioned.)

The genus Andrena includes a great number of species, Britain possessing about fifty. They may be described in a general manner as Insects much resembling the honey-bee—for which, indeed, they are frequently mistaken—but usually a little smaller in size. Many of the bees we see in spring, in March or April, are of this genus. They live in burrows in the ground, preferring sandy places, but frequently selecting a gravel path as the locality for their operations; they nearly always live in colonies. Great difficulties attend their study on account of several points in their economy, such as, that the sexes are different, and frequently not found together; also that there may be two generations of a species in one year, these being more or less different from one another. Another considerable difficulty arises from the fact that these bees are subject to the attacks of the parasite Stylops, by which their form is more or less altered. These Insects feed in the body of the bee in such a way as to affect its nutrition without destroying its life; hence they offer a means of making experiments that may throw valuable light on obscure physiological questions. Among the effects they produce in the condition of the imago bee we may mention the enfeeblement of the sexual distinction, so that a stylopised male bee becomes less different than it usually is from the female, and a stylopised female may be ill developed and less different than usual from the male. The colours and hair are sometimes altered, and distortion of portions of the abdominal region of the bee are very common. Further particulars as to these parasites will be found at the end of our account of Coleoptera (p. [298]). We may here remark that these Stylops are not the only parasitic Insects that live in the bodies of Andrenidae without killing their hosts, or even interrupting their metamorphoses. Mr. R. C. L. Perkins recently captured a specimen of Halictus rubicundus, from which he, judging from the appearance of the example, anticipated that a Stylops would emerge; but instead of this a Dipterous Insect of the family Chloropidae appeared. Dufour in 1837 called attention to a remarkable relation existing between Andrena aterrima and a parasitic Dipterous larva. The larva takes up a position in the interior of the bee's body so as to be partly included in one of the great tracheal vesicles at the base of the abdomen; and the bee then maintains the parasite in its position, and at the same time supplies it with air by causing two tracheae to grow into its body. Dufour states that he demonstrated the continuity of the tracheae of the two organisms, but it is by no means clear that the continuity was initially due to the bee's organisation.

Fig. 13—Parasitic Dipterous larva in connection with tracheal system of Andrena aterrima. (After Dufour.)

Fig. 14.—D. hirtipes ♀. Britain.

Dasypoda hirtipes appears to be the most highly endowed of the European Andrenides. The Insects of the genus Dasypoda are very like Andrena, but have only two in place of three submarginal cells (just beneath the stigma) on the front wing. The female of D. hirtipes has a very dense and elongate pubescence on the posterior legs, and carries loads of pollen, each about half its own weight, to its nest. The habits of this insect have been described by Hermann Müller.[^[25]] It forms burrows in the ground after the fashion of Andrena; this task is accomplished by excavating with the mandibles; when it has detached a certain quantity of the earth it brings this to the surface by moving backwards, and then distributes the loose soil over a considerable area. It accomplishes this in a most beautiful manner by means of the combined action of all the legs, each pair of these limbs performing its share of the function in a different manner; the front legs acting with great rapidity—making four movements in a second—push the sand backwards under the body, the bee moving itself at the same time in this direction by means of the middle pair of legs; simultaneously, but with a much slower movement, the hind legs are stretched and moved outwards, in oar-like fashion, from the body, and thus sweep away the earth and distribute it towards each side. This being done the bee returns quickly into the hole, excavates some more earth, brings it up and distributes it. Each operation of excavation takes a minute or two, the distribution on the surface only about fifteen seconds. The burrow extends to the length of one or two feet, so that a considerable amount of earth has to be brought up; and when the Insect has covered one part of the circumference of the mouth of the hole with loose earth, it makes another patch, or walk, by the side of the first. The main burrow being completed, the Insect then commences the formation of brood-chambers in connection with it. Three to six such chambers are formed in connection with a burrow; the lower one is first made and is provisioned by the bee: for this purpose five or six loads of pollen are brought to the cell, each load being, as we have already remarked, about half the weight of the Insect. This material is then formed into a ball and made damp with honey; then another load of pollen is brought, is mixed with honey and added as an outer layer to the ball, which is now remodelled and provided on one side with three short feet, after which an egg is placed on the top of the mass; the bee then sets to work to make a second chamber, and uses the material resulting from the excavation of this to close completely the first chamber. The other chambers are subsequently formed in a similar manner, and then the burrow itself is filled up. While engaged in ascertaining these facts, Müller also made some observations on the way the bee acts when disturbed in its operations, and his observations on this point show a very similar instinct to that displayed by Chalicodoma, referred to on a subsequent page. If interrupted while storing a chamber the Insect will not attempt to make a fresh one, but will carry its stock of provisions to the nest of some other individual. The result of this proceeding is a struggle between the two bees, from which it is satisfactory to learn that the rightful proprietor always comes out victorious. The egg placed on the pollen-ball in the chamber hatches in a few days, giving birth to a delicate white larva of curved form. This creature embraces the pollen-ball so far as its small size will enable it to do so, and eats the food layer by layer so as to preserve its circular form. The larva when hatched has no anal orifice and voids no excrement, so that its food is not polluted; a proper moulting apparently does not take place, for though a new delicate skin may be found beneath the old one this latter is not definitely cast off. When the food, which was at first 100 to 140 times larger than the egg or young larva, is all consumed the creature then for the first time voids its refuse. During its growth the larva becomes red and increases in weight from .0025 grains to .26 or .35 grains, but during the subsequent period of excretion it diminishes to .09 or .15 grains, and in the course of doing so becomes a grub without power of movement, and of a white instead of a red colour. After this the larva reposes motionless for many months—in fact, until the next summer, when it throws off the larval skin and appears as a pupa. The larval skin thus cast off contrasts greatly with the previous delicate condition of the integument, for this last exuvium is thick and rigid. Although it voids no excrement till much later the union of the stomach and hind-intestine is accomplished when the larva is half-grown. A larva, from which Müller took away a portion of its unconsumed food-store, began directly afterwards to emit excrement. The pupa has greater power of movement than the resting larva; when it has completed its metamorphosis and become a perfect Insect, it, if it be a female, commences almost immediately after its emergence to form burrows by the complex and perfect series of actions we have described.

Parasitic Bees (Denudatae).—This group of parasitic bees includes fourteen European genera, of which six are British. They form a group taxonomically most unsatisfactory, the members having little in common except the negative characters of the absence of pollen-carrying apparatus. Although there is a great dearth of information as to the life-histories of parasitic bees, yet some highly interesting facts and generalisations about their relations with their hosts have already been obtained. Verhoeff has recently given the following account of the relations between the parasitic bee Stelis minuta and its host Osmia leucomelana:—The Osmia forms cells in blackberry stems, provisions them in the usual manner, and deposits an egg in each. But the Stelis lays an egg in the store of provisions before the Osmia does, and thus its egg is placed lower down in the mass of food than that of the legitimate owner, which is in fact at the top. The Stelis larva emerges from the egg somewhat earlier than the Osmia larva does. For a considerable time the two larvae so disclosed consume together the stock of provisions, the Osmia at the upper, the Stelis at the lower, end thereof. By the consumption of the provisions the two larvae are brought into proximity, and by this time the Stelis larva, being about twice the size of the Osmia larva, kills and eats it. Verhoeff witnessed the struggle between the two larvae, and states further that the operation of eating the Osmia larva after it has been killed lasts one or two days. He adds that parasitic larvae are less numerous than the host larvae, it being well known that parasitic bees produce fewer offspring than host bees. Verhoeff further states that he has observed similar relations to obtain between the larvae of other parasitic bees and their hosts, but warns us against concluding that the facts are analogous in all cases.

Fig. 15.—Nomada sex-fasciata ♀. Britain.

Fabre has made us acquainted with some points in the history of another species of the same genus, viz. Stelis nasuta, that show a decided departure from the habits of S. minuta. The first-named Insect accomplishes the very difficult task of breaking open the cells of the mason-bee, Chalicodoma muraria, after they have been sealed up, and then, being an Insect of much smaller size than the Chalicodoma, places several eggs in one cell of that bee. Friese informs us that parasitic bees and their hosts, in a great number of cases, not only have in the perfect state the tongue similarly formed, but also frequent the same species of flower; thus Colletes daviesanus and its parasite Epeolus variegatus both specially affect the flowers of Tanacetum vulgare. Some of the parasitic bees have a great resemblance to their hosts; Stelis signata, for instance, is said to be so like Anthidium strigatum that for many years it was considered to be a species of the genus Anthidium. In other cases not the least resemblance exists between the parasites and hosts. Thus the species of Nomada that live at the expense of species of the genus Andrena have no resemblance thereto. Friese further tells us that the Andrena and Nomada are on the most friendly terms. Andrena, as is well known, forms populous colonies in banks, paths, etc., and in these colonies the destroying Nomada flies about unmolested; indeed, according to Friese, it is treated as a welcome guest. He says he has often seen, and in several localities, Nomada lathburiana and Andrena ovina flying peacefully together. The Nomada would enter a burrow, and if it found the Andrena therein, would come out and try another burrow; if when a marauding Nomada was in a burrow, and the rightful owner, returning laden with pollen, found on entering its home that an uninvited guest was therein, the Andrena would go out in order to permit the exit of the Nomada, and then would again enter and add the pollen to the store. Strange as this may seem at first sight, it is really not so, for, as we have before had occasion to observe, there is not the slightest reason for believing that host Insects have any idea whatever that the parasites or inquilines are injurious to their race. Why then should they attack the creatures? Provided the parasites do not interfere in any unmannerly way with the hosts and their work, there is no reason why the latter should resent their presence. The wild bee that seals up its cell when it has laid an egg therein, and then leaves it for ever, has no conception of the form of its progeny; never in the history of the race of the Andrena has a larva seen a perfect insect and survived thereafter, never has a perfect Insect seen a larva. There is no reason whatever for believing that these Insects have the least conception of their own metamorphosis, and how then should they have any idea of the metamorphosis of the parasite? If the Andrena found in the pollen the egg of a parasitic Nomada, it could of course easily remove the egg; but the Andrena has no conception that the presence of the egg ensures the death of its own offspring and though the egg be that of an enemy to its race, why should it resent the fact? Is it not clear that the race has always maintained itself notwithstanding the enemy? Nature has brought about that both host and parasite should successfully co-exist; and each individual of each species lives, not for itself, but for the continuance of the species; that continuance is provided for by the relative fecundities of host and guest. Why then should the Andrena feel alarm? If the species of Nomada attack the species of Andrena too much it brings about the destruction of its own species more certainly than that of the Andrena.

Fig. 16.—Melecta luctuosa ♀. Britain.

Such extremely friendly relations do not, however, exist between all the parasitic bees and their hosts. Friese says that, so far as he has been able to observe, the relations between the two are not in general friendly. He states that marauders of the genera Melecta and Coelioxys seek to get out of the way when they see the pollen-laden host coming home. But he does not appear to have noted any other evidence of mistrust between the two, and it is somewhat doubtful whether this act can properly be interpreted as indicating fear, for bees, as well as other animals, when engaged in work find it annoying to be interfered with; it is the interest of the parasite to avoid annoyance and to be well-mannered in its approaches. Shuckard, however, says that battles ensue between the parasite Melecta and its host Anthophora, when the two bees meet in the burrows of the Anthophora.[^[26]]

We shall have occasion to remark on some of the habits of Dioxys cincta when considering the history of the mason-bee (Chalicodoma), but one very curious point in its economy must here be noticed. The Dioxys, which is a much smaller bee than the Chalicodoma, lays an egg in a cell of the latter, and the resulting larva frequently has more food in the cell than it can consume; there is, however, another bee, Osmia cyanoxantha, that frequently takes advantage of an unoccupied cell in the nest of the Chalicodoma, and establishes its own offspring therein. The Dioxys, it seems, cannot, or at any rate does not, distinguish whether a cell is occupied by Chalicodoma or by Osmia, and sometimes lays its egg in the nest of the Osmia, though this bee is small, and therefore provides very little food for its young. It might be supposed that under these conditions the Dioxys larva would be starved to death; but this is not so; it has the power of accommodating its appetite, or its capacity for metamorphosis, to the quantity of food it finds at its disposal, and the egg laid in the Osmia cell actually produces a tiny specimen of Dioxys, only about half the natural size. Both sexes of these dwarf Dioxys are produced, offering another example of the fact that the quantity of food ingested during the lifetime of the larva does not influence the sex of the resulting imago.

The highly endowed bees that remain to be considered are by some writers united in a group called Apidae, in distinction from Andrenidae. For the purposes of this work we shall adopt three divisions, Scopulipedes, Dasygastres, Sociales.

The group Scopulipedes includes such long-tongued, solitary bees as are not parasitic, and do not belong to the Dasygastres. It is not, however, a natural group, for the carpenter-bees (Xylocopa) are very different from Anthophora. It has recently been merged by Friese with Andrenides into a single group called Podilegidae. Four British genera, Ceratina, Anthophora, Eucera and Saropoda (including, however, only seven species), are referred to the Scopulipedes; in some forms a considerable resemblance to the Bombi is exhibited, indeed the female of one of our species of Anthophora is so very like the worker of Bombus hortorum var. harrisellus, that it would puzzle any one to distinguish them by a superficial inspection, the colour of the hair on the hind legs being the only obvious difference. Anthophora is one of the most extensive and widely distributed of the genera of bees. Some of the species make burrows in cliffs and form large colonies which are continued for many years in the same locality. Friese has published many details of the industry and metamorphoses of some of the species of this genus; the most remarkable point he has discovered being that A. personata at Strasburg takes two years to accomplish the life-cycle of one generation. Some of the European species of the genus have been found to be very subject to the attacks of parasites. An anomalous beetle, Sitaris, has been found in the nests of A. pilipes; and this same Anthophora is also parasitised by another beetle, Meloe, as well as by a bee of the genus Melecta.

The genus Xylocopa[^[27]] contains many of the largest and most powerful of the bees, and is very widely distributed over the earth. In Europe only four or five species have been found, and none of them extend far northwards, X. violacea being the only one that comes so far as Paris. They are usually black or blue-black in colour, of broad, robust build, with shining integuments more or less covered with hair. X. violacea is known as the carpenter-bee from its habit of working in dry wood; it does not touch living timber, but will form its nest in all sorts of dried wood. It makes a cylindrical hole, and this gives access to three or four parallel galleries in which the broad cells are placed; the cells are always isolated by a partition; the bee forms this by cementing together with the products of its salivary glands the fragments of wood it cuts out. Its habits have been described at length by Réaumur, who alludes to it under the name of "abeille perce-bois." This bee hibernates in the imago condition, both sexes reappearing in the spring. Possibly there is more than one generation in the year, as Réaumur states that specimens that were tiny larvae on the 12th of June had by the 2nd of July consumed all their stock of provisions; they then fasted for a few days, and on the 7th or 8th of July became pupae, and in the first days of August were ready to emerge as perfect Insects. Thus the whole cycle of metamorphoses is passed through in about eight weeks. This species, though very clever in drilling holes, does not hesitate to appropriate old burrows should they be at hand. Fabre observed that it was also quite willing to save itself labour by forming its cells in hollow reeds of sufficient calibre. We have figured the larva and pupa of this species in the previous volume (p. 170).

Fig. 17.—Xylocopa (Koptorthosoma), sp. near flavonigrescens, ♂. Sarawak.

Xylocopa chloroptera in E. India selects a hollow bamboo for its nidus; it cements together the pieces obtained in clearing out the bamboo, and uses them as horizontal partitions to separate the tube into cells. The species is much infested with a small Chalcid of the genus Encyrtus: 300 specimens of the parasite have been reared from a single larva of the bee; two-thirds of the larvae of this bee that Horne endeavoured to rear were destroyed by the little Chalcid.

The most beautiful and remarkable of all the bees are the species of Euglossa. This genus is peculiar to Tropical America, and derives its name from the great length of the proboscis, which in some species surpasses that of the body. The colours in Euglossa proper are violet, purple, golden, and metallic green, and two of these are frequently combined in the most harmonious manner; the hind tibia is greatly developed and forms a plate, the outer surface of which is highly polished, while the margins are furnished with rigid hairs. Very little is known as to the habits of these bees; they were formerly supposed to be social; but this is doubtful, Bates having recorded that E. surinamensis forms a "solitary nest." Lucas concluded that E. cordata is social, on the authority of a nest containing "a dozen individuals." No workers are known. The species of Eulema have a shorter tongue than Euglossa, and in form and colour a good deal resemble our species of Bombus and Apathus.

The group Dasygastres includes seven European genera, four being British (Chelostoma being included in Heriades). The ventral surface of the hind body is densely set in the females with regularly arranged hairs, by means of which the pollen is carried. In many of the Dasygastres (Megachile, e.g.) the labrum is very large, and in repose is inflected on to the lower side of the head, and closely applied to the doubled-in tongue, which it serves to protect; the mandibles then lock together outside the labrum, which is thus completely concealed. This group includes some of the most interesting of the solitary bees.

Fig. 18.—Euglossa cordata, ♂. Amazons. A, The Insect with extended proboscis; B, outer face of hind tibia and tarsus.

The genus Chalicodoma is not found in our own country, but in the South of France there exist three or four species. Their habits have given rise to much discussion, having been described by various naturalists, among whom are included Réaumur and Fabre. These Insects are called mason-bees, and construct nests of very solid masonry. C. muraria is in appearance somewhat intermediate between a honey-bee and a Bombus; it is densely hairy, and the sexes are very different in colour. It is solitary in its habits, and usually chooses a large stone as a solid basis for its habitation. On this a cell is formed, the material used being a kind of cement made by the Insect from the mixture of a suitable sort of earth with the material secreted by its own salivary glands; the amount of cement used is reduced by the artifice of building small stones into the walls of the cell; the stones are selected with great care. When a cell about an inch in depth has been formed in this manner, the bee commences to fill it with food, consisting of honey and pollen; a little honey is brought and is discharged into the cell, then some pollen is added. This bee, like other Dasygastres, carries the pollen by means of hairs on the under surface of the body; to place this pollen in the cell the Insect therefore enters backwards, and then with the pair of hind legs brushes and scrapes the under surface of the body so as to make the pollen fall off into the cell; it then starts for a fresh cargo; after a few loads have been placed in the receptacle, the Insect mixes the honey and pollen into a paste with the mandibles, and again continues its foraging until it has about half filled the cell; then an egg is laid, and the apartment is at once closed with cement. This work is all accomplished, if the weather be favourable, in about two days, after which the Insect commences the formation of a second cell, joined to the first, and so on till eight or nine of these receptacles have been constructed; then comes the final operation of adding an additional protection in the shape of a thick layer of mortar placed over the whole; the construction, when thus completed, forms a sort of dome of cement about the size of half an orange. In this receptacle the larvae pass many months, exposed to the extreme heat of summer as well as to the cold of winter. The larvae, however, are exposed to numerous other perils; and we have already related (vol. v. p. 540) how Leucospis gigas succeeds in perforating the masonry and depositing therein an egg, so that a Leucospis is reared in the cell instead of a Chalicodoma.

Fig. 19—Chalicodoma muraria. Greece. A, Male; B, female.

This Insect has been the object of some of J. H. Fabre's most instructive studies on instinct.[^[28]] Although it is impossible for us here to consider in a thorough manner the various points he has discussed, yet some of them are of such interest and importance as to demand something more than a passing allusion.

We have mentioned that the nest of Chalicodoma is roofed with a layer of solid cement in addition to the first covering with which the bee seals up each cell. When the metamorphoses of the imprisoned larva have been passed through, and the moment for its emergence as a perfect Insect has arrived, the prisoner has to make its way through the solid wall by which it is encompassed. Usually it finds no difficulty in accomplishing the task of breaking through the roof, so that the powers of its mandibles must be very great. Réaumur has, however, recorded that a nest of this mason-bee was placed under a glass funnel, the orifice of which was covered with gauze, and that the Insects when they emerged from the nest were unable to make their way through the gauze, and consequently perished under the glass cover; and he concluded that such insects are only able to accomplish the tasks that naturally fall to their lot. By some fresh experiments Fabre, however, has put the facts in a different light. He remarks that when the Insects have, in the ordinary course of emergence, perforated the walls of their dark prison, they find themselves in the daylight, and at liberty to walk away; when they have made their escape from a nest placed under a glass cover, they, having no knowledge of glass, find themselves in daylight and imprisoned by the glass, which, to their inexperience, does not appear to be an obstacle, and they therefore, he thought, might perhaps exhaust themselves in vain efforts to pass through this invisible obstacle. He therefore took some cocoons containing pupae from a nest, placed each one of them in a tube of reed, and stopped the ends of the reeds with various substances, in one case earth, in another pith, in a third brown paper; the reeds were then so arranged that the Insects in them were in a natural position; in due course all the Insects emerged, none of them apparently having found the novel nature of the obstacle a serious impediment. Some complete nests were then taken with their inmates, and to the exterior of one of them a sheet of opaque paper was closely fastened, while to another the same sort of paper was applied in the form of a dome, leaving thus a considerable space between the true cover of the nest and the covering of paper. From the first nest the Insects made their escape in the usual manner, thus again proving that paper can be easily pierced by them. From the second nest they also liberated themselves, but failed to make their way out through the dome of paper, and perished beneath it; thus showing that paper added to the natural wall caused them no difficulty, but that paper separated therefrom by a space was an insuperable obstacle. Professor Pérez has pointed out that this is no doubt due to the large space offered to the bee, which consequently moves about, and does not concentrate its efforts on a single spot, as it of course is compelled to do when confined in its natural cell.

The power of the mason-bee to find its nest again when removed to a distance from it is another point that was tested by Du Hamel and recounted by Réaumur. As regards this Fabre has also made some very valuable observations. He marked some specimens of the bee, and under cover removed them to a distance of four kilometres, and then liberated them; the result proved that the bees easily found their way back again, and indeed were so little discomposed by the removal that they reached their nests laden with pollen as if they had merely been out on an ordinary journey. On one of these occasions he observed that a Chalicodoma, on returning, found that another bee had during her absence taken possession of her partially completed cell, and was unwilling to relinquish it; whereupon a battle between the two took place. The account of this is specially interesting, because it would appear that the two combatants did not seek to injure one another, but were merely engaged in testing, as it were, which was the more serious in its claims to the proprietorship of the cell in dispute. The matter ended by the original constructor regaining and retaining possession. Fabre says that in the case of Chalicodoma it is quite a common thing for an uncompleted cell to be thus appropriated by a stranger during the absence of the rightful owner, and that after a scene of the kind described above, the latter of the two claimants always regains possession, thus leading one to suppose that some sense of rightful ownership exists in these bees; the usurper expressing, as it were, by its actions the idea—Before I resign my claims I must require you to go through the exertions that will prove you to be really the lawful owner.

Another experiment was made with forty specimens of Chalicodoma pyrenaica, which were removed to a distance of four kilometres and then liberated. About twenty of the individuals had been somewhat injured by the processes of capturing, marking, and transferring, and proved unable to make a proper start. The others went off well when released, and in forty minutes the arrivals at the nest had already commenced. The next morning he was able to ascertain that fifteen at least had found their way back, and that it was probable that most of the uninjured bees had reached home; and this although, as Fabre believed, they had never before seen the spot where he liberated them.

These observations on the power of Chalicodoma to regain its nest attracted the attention of Charles Darwin, who wrote to M. Fabre, and suggested that further observations should be made with the view of ascertaining by means of what sense these bees were able to accomplish their return. For it must be borne in mind that this bee is very different from the domestic bee, inasmuch as it enjoys but a brief life in the winged state, and it is therefore to be presumed that an individual has no knowledge of such comparatively distant localities as those to which Fabre transported it. Further observations made by the Frenchman have unfortunately failed to throw any light on this point. Darwin thought it might possibly be some sensitiveness to magnetic conditions that enabled the bees to return home, and suggested that they should be tested as to this. Fabre accordingly made some minute magnets, and fixed one to each bee previous to letting them loose for a return journey. This had the effect of completely deranging the bees; and it was therefore at first thought that the requisite clue was obtained. It occurred to the experimenter, however, to try the plan of affixing small pieces of straw to the bees instead of magnets, and on this being done it was found that the little creatures were just as much deranged by the straws as they were by the magnets: thus it became evident that no good grounds exist for considering that the bees are guided by magnetic influences.

One of the species[^[29]] of Chalicodoma observed by Fabre fixes its nests to the small boulders brought down and left by the Rhone on the waste places of its banks. This habit afforded Fabre an opportunity of removing the nests during the process of construction, and of observing the effect this produced on the architects. While a bee that had a nest partially constructed was absent, he removed the stone and the nest attached to it from one situation to another near at hand and visible from the original site. In a few minutes the bee returned and went straight to the spot where the nest had been; finding its home absent it hovered for a little while around the place, and then alighted on the vacated position, and walked about thereon in search of the nest; being after some time convinced that this was no longer there, it took wing, but speedily returned again to the place and went through the same operations. This series of manoeuvres was several times repeated, the return always being made to the exact spot where the nest had been originally located; and although the bee in the course of its journeys would pass over the nest at a distance of perhaps only a few inches, it did not recognise the object it was in search of. If the nest were placed very near to the spot it had been removed from—say at a distance of about a yard—it might happen that the bee would actually come to the stone to which the nest was fixed, would visit the nest, would even enter into the cell it had left partially completed, would examine circumspectly the boulder, but would always leave it, and again return to the spot where the nest was originally situated, and, on finding that the nest was not there, would take its departure altogether from the locality. The home must be, for the bee, in the proper situation, or it is not recognised as the desired object. Thus we are confronted with the strange fact that the very bee that is able to return to its nest from a distance of four kilometres can no longer recognise it when removed only a yard from the original position. This extraordinary condition of the memory of the Insect is almost inconceivable by us. That the bee should accurately recognise the spot, but that it should not recognise the cell it had itself just formed and half-filled with honey-paste, seems to us almost incredible; nevertheless, the fact is quite consistent with what we shall subsequently relate in the case of the solitary wasp Bembex. A cross experiment was made by taking away the stone with the attached nest of the bee while the latter was absent, and putting in its place the nest of another individual in about the same stage of construction; this nest was at once adopted by the bee, which indeed was apparently in no way deranged by the fact that the edifice was the work of another. A further experiment was made by transposing the positions of two nests that were very near together, so that each bee when returning might be supposed to have a free choice as to which nest it would go to. Unhesitatingly each bee selected the nest that, though not its own, was in the position where its own had been. This series of experiments seems to prove that the Chalcidoma has very little sense as to what is its own property, but, on the other hand, has a most keen appreciation of locality. As, however, it might be supposed that the bees were deceived by the similarity between the substituted nests, Fabre transposed two nests that were extremely different, one consisting of many cells, the other of a single incomplete cell; it was, of course, a necessary condition of this experiment that each of the two nests, however different in other respects, should possess one cell each in similar stages of construction; and when that was the case each bee cheerfully adopted the nest that, though very different to its own, was in the right place. This transposition of nests can be rapidly repeated, and thus the same bee may be made to go on working at two different nests.

Suppose, however, that another sort of change be made. Let a nest, consisting of a cell that is in an early stage of construction, be taken away, and let there be substituted for it a cell built and partially stored with food. It might be supposed that the bee would gladly welcome this change, for the adoption of the substituted cell would save it a great deal of work. Not so, however; the bee in such a case will take to the substituted cell, but will go on building at it although it is already of the full height, and will continue building at it until the cell is made as much as a third more than the regulation height. In fact the bee, being in the building stage of its operations, goes on building, although in so doing it is carrying on a useless, if not an injurious, work. A similar state ensues when the Insect ceases to build and begins to bring provisions to the nest; although a substituted cell may contain a sufficient store of food, the bee goes on adding to this, though it is wasting its labours in so doing. It should be noted that though the bee must go through the appropriate stages of its labours whether the result of so doing be beneficial or injurious, yet it is nevertheless to some extent controlled by the circumstances, for it does not in such cases complete what should have been the full measure of its own individual work; it does not, for instance, raise the cell to twice the natural height, but stops building when the cell is about one-third larger than usual, as if at that stage the absurdity of the situation became manifest to it.

Fabre's experiments with the Chalicodoma are so extremely instructive as regards the nature of instinct in some of the highest Insects, that we must briefly allude to some other of his observations even at the risk of wearying the reader who feels but little interest in the subject of Insect intelligence.

Having discovered that a mason-bee that was engaged in the process of construction would go on building to an useless or even injurious extent, Fabre tried another experiment to ascertain whether a bee that was engaged in the process of provisioning the nest, would do so in conditions that rendered its work futile. Taking away a nest with completely built cell that a bee was storing with food, he substituted for it one in which the cell was only commenced, and therefore incapable of containing food; when the bee with its store of provisions reached this should-be receptacle it appeared to be very perplexed, tested the imperfect cell with its antennae, left the spot and returned again; repeating this several times it finally went to the cell of some stranger to deposit its treasure. In other cases the bee broke open a completed cell, and having done so went on bringing provisions to it, although it was already fully provisioned and an egg laid therein: finally, the little creature having completed the bringing of this superfluous tale of provisions, deposited a second egg, and again sealed up the cell. But in no case does the bee go back from the provisioning stage to the building stage until the cycle for one cell of building, provisioning, and egg-laying is completed: but when this is the case, the building of a fresh cell may be again undertaken. This is a good example of the kind of consecutive necessity that seems to be one of the chief features of the instinct of these industrious little animals. Another equally striking illustration of these peculiarities of instinct is offered by interfering with the act of putting the provisions into the cell. It will be recollected that when the bee brings provisions to add to the stock, it carries both honey and pollen; in order to deliver these it begins by entering head first into the cell and disgorging the honey, then emerging it turns round, enters backwards and scrapes off the pollen from its body. If after the honey has been discharged, the bee be interfered with and gently removed to a slight distance with a straw, it returns to complete its task, but instead of going on with the actions at the point at which the interruption took place, it begins the series over again, going in—at any rate partially—head first, although it has no honey to discharge, and having performed this useless ceremony it then emerges, turns round and adds the pollen. This illustration is in some respects the reverse of what might have been expected, for the Insect here does not continue the act at the interrupted point, but begins the series of actions afresh.

It would be reasonable to suppose that an Insect that takes the pains to provide for the safety of its progeny by constructing a complex edifice of cement, secures thereby the advantage of protection for its young. But this is far from being the case. Notwithstanding the cement and the thick dome of mortar, the Chalicodoma is extremely subject to the attacks of parasites. The work performed by the creature in constructing its mass of masonry is truly astounding; Fabre calculated from measurements he made that for the construction and provisioning of a single cell, the goings and comings of the bee amounted to 15 kilometres, and it makes for each nest sometimes as many as fifteen cells. Notwithstanding all this labour, it would appear that no real safety for the larvae is obtained by the work. Some sixteen—possibly more—other species of Insects get their living off this industrious creature. Another bee, Stelis nasuta, breaks open the cells after they have been completely closed and places its own eggs in them, and then again closes the cells with mortar. The larvae of this Stelis develop more rapidly than do those of the Chalicodoma, so that the result of this shameless proceeding is that the young one of the legitimate proprietor—as we human beings think it—is starved to death, or is possibly eaten up as a dessert by the Stelis larvae, after they have appropriated all the pudding.

Another bee, Dioxys cincta, is even more audacious; it flies about in a careless manner among the Chalicodoma at their work, and they do not seem to object to its presence unless it interferes with them in too unmannerly a fashion, when they brush it aside. The Dioxys, when the proprietor leaves the cell, will enter it and taste the contents; after having taken a few mouthfuls the impudent creature then deposits an egg in the cell, and, it is pretty certain, places it at or near the bottom of the mass of pollen, so that it is not conspicuously evident to the Chalicodoma when the bee again returns to add to or complete the stock of provisions. Afterwards the constructor deposits its own egg in the cell and closes it. The final result is much the same as in the case of the Stelis, that is to say, the Chalicodoma has provided food for an usurper; but it appears probable that the consummation is reached in a somewhat different manner, namely, by the Dioxys larva eating the egg of the Chalicodoma, instead of slaughtering the larva. Two of the Hymenoptera Parasitica are very destructive to the Chalicodoma, viz. Leucospis gigas and Monodontomerus nitidus; the habits of which we have already discussed (vol. v. p. 543) under Chalcididae. Lampert has given a list of the Insects attacking the mason-bee or found in its nests; altogether it would appear that about sixteen species have been recognised, most of which destroy the bee larva, though some possibly destroy the bee's destroyers, and two or three perhaps merely devour dead examples of the bee, or take the food from cells, the inhabitants of which have been destroyed by some untoward event. This author thinks that one half of the bees' progeny are made away with by these destroyers, while Fabre places the destruction in the South of France at a still higher ratio, telling us that in one nest of nine cells, the inhabitants of three were destroyed by the Dipterous Insect, Anthrax trifasciata, of two by Leucospis, of two by Stelis, and of one by the smaller Chalcid; there being thus only a single example of the bee that had not succumbed to one or other of the enemies. He has sometimes examined a large number of nests without finding a single one that had not been attacked by one or other of the parasites, and more often than not several of the marauders had attacked the nest.

It is said by Lampert and others that there is a passage in Pliny relating to one of the mason-bees, that the Roman author had noticed in the act of carrying off stones to build into its nest; being unacquainted with the special habits of the bee, he seems to have supposed that the insect was carrying the stone as ballast to keep itself from being blown away.

Fig. 20—Anthidium manicatum, Carder-bee. A, Male; B, female.

The bees of the genus Anthidium are known to possess the habit of making nests of wool or cotton, that they obtain from plants growing at hand. We have one species of this genus of bees in Britain; it sometimes may be seen at work in the grounds of our Museum at Cambridge: it is referred to by Gilbert White, who says of it, in his History of Selborne: "There is a sort of wild bee frequenting the garden-campion for the sake of its tomentum, which probably it turns to some purpose in the business of nidification. It is very pleasant to see with what address it strips off the pubes, running from the top to the bottom of a branch, and shaving it bare with the dexterity of a hoop-shaver. When it has got a bundle, almost as large as itself, it flies away, holding it secure between its chin and its fore legs." The species of this genus are remarkable as forming a conspicuous exception to the rule that in bees the female is larger than the male. The species of Anthidium do not form burrows for themselves, but either take advantage of suitable cavities formed by other Insects in wood, or take possession of deserted nests of other bees or even empty snail-shells. The workers in cotton, of which our British species A. manicatum is one, line the selected receptacle with a beautiful network of cotton or wool, and inside this place a finer layer of the material, to which is added some sort of cement that prevents the honied mass stored by the bees in this receptacle from passing out of it. A. diadema, one of the species that form nests in hollow stems, has been specially observed by Fabre; it will take the cotton for its work from any suitable plant growing near its nest, and does not confine itself to any particular natural order of plants, or even to those that are indigenous to the South of France. When it has brought a ball of cotton to the nest, the bee spreads out and arranges the material with its front legs and mandibles, and presses it down with its forehead on to the cotton previously deposited; in this way a tube of cotton is constructed inside the reed; when withdrawn, the tube proved to be composed of about ten distinct cells arranged in linear fashion, and connected firmly together by means of the outer layer of cotton; the transverse divisions between the chambers are also formed of cotton, and each chamber is stored with a mixture of honey and pollen. The series of chambers does not extend quite to the end of the reed, and in the unoccupied space the Insect accumulates small stones, little pieces of earth, fragments of wood or other similar small objects, so as to form a sort of barricade in the vestibule, and then closes the tube by a barrier of coarser cotton taken frequently from some other plant, the mullein by preference. This barricade would appear to be an ingenious attempt to keep out parasites, but if so, it is a failure, at any rate as against Leucospis, which insinuates its eggs through the sides, and frequently destroys to the last one the inhabitants of the fortress. Fabre states that these Anthidium, as well as Megachile, will continue to construct cells when they have no eggs to place in them; in such a case it would appear from his remarks that the cells are made in due form and the extremity of the reed closed, but no provisions are stored in the chambers.

The larva of the Anthidium forms a most singular cocoon. We have already noticed the difficulty that arises, in the case of these Hymenopterous larvae shut up in small chambers, as to the disposal of the matters resulting from the incomplete assimilation of the aliment ingested. To allow the once-used food to mingle with that still remaining unconsumed would be not only disagreeable but possibly fatal to the life of the larva. Hence some species retain the whole of the excrement until the food is entirely consumed, it being, according to Adlerz, stored in a special pouch at the end of the stomach; other Hymenoptera, amongst which we may mention the species of Osmia, place the excreta in a vacant space. The Anthidium adopts, however, a most remarkable system: about the middle of its larval life it commences the expulsion of "frass" in the shape of small pellets, which it fastens together with silk, as they are voided, and suspends round the walls of the chamber. This curious arrangement not only results in keeping the embarrassing material from contact with the food and with the larva itself, but serves, when the growth of the latter is accomplished, as the outline or foundations of the cocoon in which the metamorphosis is completed. This cocoon is of a very elaborate character; it has, so says Fabre, a beautiful appearance, and is provided with a very peculiar structure in the form of a small conical protuberance at one extremity pierced by a canal. This canal is formed with great care by the larva, which from time to time places its head in the orifice in process of construction, and stretches the calibre by opening the mandibles. The object of this peculiarity in the fabrication of the elaborate cocoon is not clear, but Fabre inclines to the opinion that it is for respiratory purposes.

Other species of this genus use resin in place of cotton as their working material. Among these are Anthidium septemdentatum and A. bellicosum. The former species chooses an old snail-shell as its nidus, and constructs in it near the top a barrier of resin, so as to shut off the part where the whorl is too small; then beneath the shelter of this barrier it accumulates a store of honey-pollen, deposits an egg, and completes the cell by another transverse barrier of resin; two such cells are usually constructed in one snail-shell, and below them is placed a barricade of small miscellaneous articles, similar to what we have described in speaking of the cotton-working species of the genus. This bee completes its metamorphosis, and is ready to leave the cell in early spring. Its congener, A. bellicosum, has the same habits, with the exception that it works later in the year, and is thus exposed to a great danger, that very frequently proves fatal to it. This bee does not completely occupy the snail-shell with its cells, but leaves the lower and larger portion of the shell vacant. Now, there is another bee, a species of Osmia, that is also fond of snail-shells as a nesting-place, and that affects the same localities as the A. septemdentatum; very often the Osmia selects for its nest the vacant part of a shell, the other part of which is occupied by the Anthidium; the result of this is that when the metamorphoses are completed, the latter bee is unable to effect its escape, and thus perishes in the cell. Fabre further states with regard to these interesting bees, that no structural differences of the feet or mandibles can be detected between the workers in cotton and the workers in resin; and he also says that in the case where two cells are constructed in one snail-shell, a male individual is produced from the cell of the greater capacity, and a female from the other.

Osmia is one of the most important of the genera of bees found in Europe, and is remarkable for the diversity of instinct displayed in the formation of the nests of the various species. As a rule they avail themselves for nidification of hollow places already existing; choosing excavations in wood, in the mortar of walls, and even in sandbanks; in several cases the same species is found to be able to adapt itself to more than one kind of these very different substances. This variety of habit will render it impossible for us to do justice to this interesting genus within the space at our disposal, and we must content ourselves with a consideration of one or two of the more instructive of the traits of Osmia life. O. tridentata forms its nest in the stems of brambles, of which it excavates the pith; its mode of working and some other details of its life have been well depicted by Fabre. The Insect having selected a suitable bramble-stalk with a cut extremity, forms a cylindrical burrow in the pith thereof, extending the tunnel as far as will be required to allow the construction of ten or more cells placed one after the other in the axis of the cylinder; the bee does not at first clear out quite all the pith, but merely forms a tunnel through it, and then commences the construction of the first cell, which is placed at the end of the tunnel that is most remote from the entrance. This cavity is to be of oval form, and the Insect therefore cuts away more of the pith so as to make an oval space, but somewhat truncate, as it were, at each end, the plane of truncation at the proximal extremity being of course an orifice.

Fig. 21.—Osmia tricornis, ♀. Algeria.

The first cell thus made is stored with pollen and honey, and an egg is deposited. Then a barrier has to be constructed to close this chamber; the material used for the barrier is the pith of the stem, and the Insect cuts the material required for the purpose from the walls of the second chamber; the excavation of the second chamber is, in fact, made to furnish the material for closing up the first cell. In this way a chain of cells is constructed, their number being sometimes as many as fifteen. The mode in which the bees, when the transformations of the larvae and pupae have been completed, escape from the chain of cells, has been the subject of much discussion, and errors have arisen from inference being allowed to take the place of observation. Thus Dufour, who noted this same mode of construction and arrangement in another Hymenopteron (Odynerus nidulator), perceived that there was only one orifice of exit, and also that the Insect that was placed at the greatest distance from this was the one that, being the oldest of the series, might be expected to be the first ready to emerge; and as the other cocoons would necessarily be in the way of its getting out, he concluded that the egg that was last laid produced the first Insect ready for emergence. Fabre tested this by some ingenious experiments, and found that this was not the case, but that the Insects became ready to leave their place of imprisonment without any reference to the order in which the eggs were laid, and he further noticed some very curious facts with reference to the mode of emergence of Osmia tridentata. Each Insect, when it desires to leave the bramble stem, tears open the cocoon in which it is enclosed, and also bites through the barrier placed by the mother between it and the Insect that is next it, and that separates it from the orifice of exit. Of course, if it happen to be the outside one of the series it can then escape at once; but if it should be one farther down in the Indian file it will not touch the cocoon beyond, but waits patiently, possibly for days; if it then still find itself confined it endeavours to escape by squeezing past the cocoon that intervenes between it and liberty, and by biting away the material at the sides so as to enlarge the passage; it may succeed in doing this, and so get out, but if it fail to make a side passage it will not touch the cocoons that are in its way. In the ordinary course of events, supposing all to go well with the family, all the cocoons produce their inmates in a state for emergence within a week or two, and so all get out. Frequently, however, the emergence is prevented by something having gone wrong with one of the outer Insects, in which case all beyond it perish unless they are strong enough to bite a hole through the sides of the bramble-stem. Thus it appears that whether a particular Osmia shall be able to emerge or not depends on two things—(1) whether all goes well with all the other Insects between it and the orifice, and (2) whether the Insect can bite a lateral hole or not; this latter point also largely depends on the thickness of the outer part of the stem of the bramble. Fabre's experiments on these points have been repeated, and his results confirmed by Nicolas.

The fact that an Osmia would itself perish rather than attack the cocoon of its brother or sister is certainly very remarkable, and it induced Fabre to make some further experiments. He took some cocoons containing dead specimens of Osmia, and placed them in the road of an Osmia ready for exit, and found that in such case the bee made its way out by demolishing without any scruple the cocoons and dead larvae that intervened between it and liberty. He then took some other reeds, and blocked the way of exit with cocoons containing living larvae, but of another species of Hymenoptera. Solenius vagus and Osmia detrita were the species experimented on in this case, and he found that the Osmia destroyed the cocoon and living larvae of the Solenius, and so made its way out. Thus it appears that Osmia will respect the life of its own species, and die rather than destroy it, but has no similar respect for the life of another species.

Some of Fabre's most instructive chapters are devoted to the habits and instincts of various species of the genus Osmia. It is impossible here to find space even to summarise them, still more impossible to do them justice; but we have selected the history just recounted, because it is rare to find in the insect world instances of such self-sacrifice by an individual for one of the same generation. It would be quite improper to generalise from this case, however, and conclude that such respect for its own species is common even amongst the Osmia. Fabre, indeed, relates a case that offers a sad contrast to the scene of self-sacrifice and respect for the rights of others that we have roughly portrayed. He was able to induce a colony of Osmia tricornis (another species of the genus, be it noted) to establish itself and work in a series of glass tubes that he placed on a table in his laboratory. He marked various individuals, so that he was able to recognise them and note the progress of their industrial works. Quite a large number of specimens thus established themselves and concluded their work before his very eyes. Some individuals, however, when they had completed the formation of a series of cells in a glass tube or in a reed, had still not entirely completed their tale of work. It would be supposed that in such a case the individual would commence the formation of another series of cells in an unoccupied tube. This was not, however, the case. The bee preferred tearing open one or more cells already completed—in some cases, even by itself—scattering the contents, and devouring the egg; then again provisioning the cell, it would deposit a fresh egg, and close the chamber. These brief remarks will perhaps suffice to give some idea of the variety of instinct and habit that prevails in this very interesting genus. Friese observes that the variety of habits in this genus is accompanied as a rule by paucity of individuals of a species, so that in central Europe a collector must be prepared to give some twenty years or so of attention to the genus before he can consider he has obtained all the species of Osmia that inhabit his district.

As a prelude to the remarks we are about to make on the leaf-cutting bees of the genus Megachile it is well to state that the bee, the habits of which were described by Réaumur under the name of "l'abeille tapissière," and that uses portions of the leaves of the scarlet poppy to line its nest, is now assigned to the genus Osmia, although Latreille, in the interval that has elapsed since the publication of Réaumur's work, founded the genus Anthocopa for the bee in question. Megachile is one of the most important of the genera of the Dasygastres, being found in most parts of the world, even in the Sandwich Islands; it consists of bees averaging about the size of the honey-bee (though some are considerably larger, others smaller), and having the labrum largely developed; this organ is capable of complete inflection to the under side of the head, and when in the condition of repose it is thus infolded, it underlaps and protects the larger part of the lower lip; the mandibles close over the infolded labrum, so that, when the Insect is at rest, this appears to be altogether absent. These bees are called leaf-cutters, from their habit of forming the cells for their nest out of pieces of the leaves of plants. We have several species in Britain; they are very like the common honey-bee in general appearance, though rather more robustly formed. These Insects, like the Osmiae, avail themselves of existing hollow places as receptacles in which to place their nests. M. albocincta frequently takes possession of a deserted worm-burrow in the ground. The burrow being longer than necessary the bee commences by cutting off the more distant part by means of a barricade of foliage; this being done, it proceeds to form a series of cells, each shaped like a thimble with a lid at the open end (Fig. 22, A). The body of the thimble is formed of large oval pieces of leaf, the lid of smaller round pieces; the fragments are cut with great skill from the leaves of growing plants by the Insect, which seems to have an idea of the form and size of the piece of foliage necessary for each particular stage of its work.

Fig. 22—Nidification of leaf-cutting bee, Megachile anthracina. A, one cell separated, with lid open; the larva (a) reposing on the food; B, part of a string of the cells. (After Horne.)

Horne has given particulars as to the nest of Megachile anthracina (fasciculata), an East Indian species.[^[30]] The material employed was either the leaves of the Indian pulse or of the rose. Long pieces are cut by the Insect from the leaf, and with these a cell is formed; a circular piece is next cut, and with this a lid is made for the receptacle. The cells are about the size and shape of a common thimble; in one specimen that Horne examined no less than thirty-two pieces of leaf disposed in seven layers were used for one cell, in addition to three pieces for the round top. The cells are carefully prepared, and some kind of matter of a gummy nature is believed to be used to keep in place the pieces forming the interior layers. The cells are placed end to end, as shown in Fig. 22, B; five to seven cells form a series, and four or six series are believed to be constructed by one pair of this bee, the mass being located in a hollow in masonry or some similar position. Each cell when completed is half filled with pollen in the usual manner, and an egg is then laid in it. This bee is much infested by parasites, and is eaten by the Grey Hornbill (Meniceros bicornis).

Megachile lanata is one of the Hymenoptera that in East India enter houses to build their own habitations. According to Horne both sexes take part in the work of construction, and the spots chosen are frequently of a very odd nature. The material used is some kind of clay, and the natural situation may be considered to be the interior of a hollow tube, such as the stem of a bamboo; but the barrel of a gun, and the hollow in the back of a book that has been left lying open, have been occasionally selected by the Insect as suitable. Smith states that the individuals developed in the lower part of a tubular series of this species were females, "which sex takes longer to develop, and thus an exit is not required for them so soon as for the occupants of the upper cells which are males." M. proxima, a species almost exactly similar in appearance to M. lanata, makes its cells of leaf-cuttings, however, and places them in soft soil.

Fabre states that M. albocincta, which commences the formation of its nest in a worm-burrow by means of a barricade, frequently makes the barricade, but no nest; sometimes it will indeed make the barricade more than twice the proper size, and thus completely fill up the worm burrow. Fabre considers that these eccentric proceedings are due to individuals that have already formed proper nests elsewhere, and that after completing these have still some strength remaining, which they use up in this fruitless manner.

The Social bees (Sociales) include, so far as is yet known, only a very small number of genera, and are so diverse, both in habits and structure, that the propriety of associating them in one group is more than doubtful; the genera are Bombus (Fig. 331, vol. v.), with its commensal genus or section, Psithyrus (Fig. 23); Melipona (Fig. 24), in which Trigona and Tetragona may at present be included, and Apis (Fig. 6); this latter genus comprising the various honey-bees that are more or less completely domesticated in different parts of the world.

In the genus Bombus the phenomena connected with the social life are more similar to what we find among wasps than to what they are in the genus Apis. The societies come to an end at the close of the season, a few females live through the winter, and each of these starts a new colony in the following spring. Males, females and workers exist, but the latter are not distinguished by any good characters from the females, and are, in fact, nothing but more or less imperfect forms thereof; whereas in Apis the workers are distinguished by structural characters not found in either of the true sexes.

Hoffer has given a description of the commencement of a society of Bombus lapidarius.[^[31]] A large female, at the end of May, collected together a small mass of moss, then made an expedition and returned laden with pollen; under cover of the moss a cell was formed of wax taken from the hind-body and mixed with the pollen the bee had brought in; this cell was fastened to a piece of wood; when completed it formed a subspherical receptacle, the outer wall of which consisted of wax, and whose interior was lined with honey-saturated pollen; then several eggs were laid in this receptacle, and it was entirely closed. Hoffer took the completed cell away to use it for museum purposes, and the following day the poor bee that had formed it died. From observations made on Bombus agrorum he was able to describe the subsequent operations; these are somewhat as follows:—The first cell being constructed, stored, and closed, the industrious architect, clinging to the cell, takes a few days' rest, and after this interval commences the formation of a second cell; this is placed by the side of the first, to which it is connected by a mixture of wax and pollen; the second cell being completed a third may be formed; but the labours of the constructor about this time are augmented by the hatching of the eggs deposited a few days previously; for the young larvae, having soon disposed of the small quantity of food in the interior of the waxen cell, require feeding. This operation is carried on by forming a small opening in the upper part of the cell, through which the bee conveys food to the interior by ejecting it from her mouth through the hole; whether the food is conveyed directly to the mouths of the larvae or not, Hoffer was unable to observe; it being much more difficult to approach this royal founder without disturbing her than it is the worker-bees that carry on similar occupations at a subsequent period in the history of the society. The larvae in the first cell, as they increase in size, apparently distend the cell in an irregular manner, so that it becomes a knobbed and rugged, truffle-like mass. The same thing happens with the other cells formed by the queen. Each of these larval masses contains, it should be noticed, sister-larvae all of one age; when full grown they pupate in the mass, and it is worthy of remark that although all the eggs in one larval mass were laid at the same time, yet the larvae do not all pupate simultaneously, neither do all the perfect Insects appear at once, even if all are of one sex. The pupation takes place in a cocoon that each larva forms for itself of excessively fine silk. The first broods hatched are formed chiefly, if not entirely, of workers, but small females may be produced before the end of the season. Huber and Schmiedeknecht state that though the queen provides the worker-cells with food before the eggs are placed therein, yet no food is put in the cells in which males and females are produced. The queen, at the time of pupation of the larvae, scrapes away the wax by which the cocoons are covered, thus facilitating the escape of the perfect Insect, and, it may also be, aiding the access of air to the pupa. The colony at first grows very slowly, as the queen can, unaided, feed only a small number of larvae. But after she receives the assistance of the first batch of workers much more rapid progress is made, the queen greatly restricting her labours, and occupying herself with the laying of eggs; a process that now proceeds more and more rapidly, the queen in some cases scarcely ever leaving the nest, and in others even becoming incapable of flight. The females produced during the intermediate period of the colony are smaller than the mother, but supplement her in the process of egg-laying, as also do the workers to a greater or less extent. The conditions that determine the egg-laying powers of these small females and workers are apparently unknown, but it is ascertained that these powers vary greatly in different cases, so that if the true queen die the continuation of the colony is sometimes effectively carried on by these her former subordinates. In other cases, however, the reverse happens, and none of the inhabitants may be capable of producing eggs: in this event two conditions may be present; either larvae may exist in the nest, or they may be absent. In the former case the workers provide them with food, and the colony may thus still be continued; but in the latter case, there being no profitable occupation for the bees to follow, they spend the greater part of the time sitting at home in the nest.

Supposing all to go well with the colony it increases very greatly, but its prosperity is checked in the autumn; at this period large numbers of males are produced as well as new queens, and thereafter the colony comes to an end, only a few fertilised females surviving the winter, each one to commence for herself a new colony in the ensuing spring.

The interior of the nest of a bumble-bee (Bombus) frequently presents a very irregular appearance; this is largely owing to the fact that these bees do not use the cells as cradles twice, but form others as they may be required, on the old remains. The cells, moreover, are of different sizes, those that produce workers being the smallest, those that cradle females being the largest, while those in which males are reared are intermediate in size. Although the old cells are not used a second time for rearing brood they are nevertheless frequently adapted to the purposes of receptacles for pollen and for honey, and for these objects they may be increased in size and altered in form.

It may be gathered from various records that the period required to complete the development of the individual Bombus about midsummer is four weeks from the deposition of the egg to the emergence of the perfect Insect, but exact details and information as to whether this period varies with the sex of the Insect developed are not to be found. The records do not afford any reason for supposing that such distinction will be found to exist: the size of the cells appears the only correlation, suggested by the facts yet known, between the sex of the individual and the circumstances of development.

The colonies of Bombus vary greatly in prosperity, if we take as the test of this the number of individuals produced in a colony. They never, however, attain anything at all approaching to the vast number of individuals that compose a large colony of wasps, or that exist in the crowded societies of the more perfectly social bees. A populous colony of a subterranean Bombus may attain the number of 300 or 400 individuals. Those that dwell on the surface are as a rule much less populous, as they are less protected, so that changes of weather are more prejudicial to them. According to Smith, the average number of a colony of B. muscorum in the autumn in this country is about 120—viz. 25 females, 36 males, 59 workers. No mode of increasing the nests in a systematic manner exists in this genus; they do not place the cells in stories as the wasps do; and this is the case notwithstanding the fact that a cell is not twice used for the rearing of young. When the ground-space available for cell-building is filled the Bombus begins another series of cells on the ruins of the first one. From this reason old nests have a very irregular appearance, and this condition of seeming disorder is greatly increased by the very different sizes of the cells themselves. We have already alluded to some of these cells, more particularly to those of different capacities to suit the sexes of the individuals to be reared in them. In addition to these there are honey-tubs, pollen-tubs, and the cells of the Psithyrus (Fig. 23), the parasitic but friendly inmates of the Bombus-nests. A nest of Bombus, exhibiting the various pots projecting from the remains of empty and partially destroyed cells, presents, as may well be imagined, a very curious appearance. Some of the old cells apparently are partly destroyed for the sake of the material they are composed of. Others are formed into honey-tubs, of a make-shift nature. It must be recollected that, as a colony increases, stores of provisions become absolutely necessary, otherwise in bad weather the larvae could not be fed. In good weather, and when flowers abound, these bees collect and store honey in abundance; in addition to placing it in the empty pupa-cells, they also form for it special receptacles; these are delicate cells made entirely of wax filled with honey, and are always left open for the benefit of the community. The existence of these honey-tubs in bumble-bees' nests has become known to our country urchins, whose love for honey and for the sport of bee-baiting leads to wholesale destruction of the nests. According to Hoffer, special tubs for the storing of pollen are sometimes formed; these are much taller than the other cells. The Psithyrus that live in the nests with the Bombus are generally somewhat larger than the latter, and consequently their cells may be distinguished in the nests by their larger size. A bumble-bees' nest, composed of all these heterogenous chambers rising out of the ruins of former layers of cells, presents a scene of such apparent disorder that many have declared that the bumble-bees do not know how to build.

Although the species of Bombus are not comparable with the hive-bee in respect of the perfection and intelligent nature of their work, yet they are very industrious Insects, and the construction of the dwelling-places of the subterranean species is said to be carried out in some cases with considerable skill, a dome of wax being formed as a sort of roof over the brood cells. Some work even at night. Fea has recorded the capture of a species in Upper Burmah working by moonlight, and the same industry may be observed in this country if there be sufficient heat as well as light. Godart, about 200 years ago, stated that a trumpeter-bee is kept in some nests to rouse the denizens to work in the morning: this has been treated as a fable by subsequent writers, but is confirmed in a circumstantial manner by Hoffer, who observed the performance in a nest of B. ruderatus in his laboratory. On the trumpeter being taken away its office was the following morning filled by another individual The trumpeting was done as early as three or four o'clock in the morning, and it is by no means impossible that the earliness of the hour may have had something to do with the fact that for 200 years no one confirmed the old naturalist's observation.

One of the most curious facts in connection with Bombus is the excessive variation that many of the species display in the colour of the beautiful hair with which they are so abundantly provided. There is not only usually a difference between the sexes in this respect, but also extreme variation within the limits of the same sex, more especially in the case of the males and workers; there is also an astonishing difference in the size of individuals. These variations are carried to such an extent that it is almost impossible to discriminate all the varieties of a species by inspection of the superficial characters. The structures peculiar to the male, as well as the sting of the female, enable the species to be determined with tolerable certainty. Cholodkovsky,[^[32]] on whose authority this statement as to the sting is made, has not examined it in the workers, so that we do not know whether it is as invariable in them as he states it to be in queens of the same species. According to Handlirsch,[^[33]] each species of Bombus has the capacity of variation, and many of the varieties are found in one nest, that is, among the offspring of a single pair of the species, but many of the variations are restricted to certain localities. Some of the forms can be considered as actual ("fertige") species, intermediate forms not being found, and even the characters by which species are recognised being somewhat modified. As examples of this he mentions Bombus silvarum and B. arenicola, B. pratorum and B. scrimshiranus. In other cases, however, the varieties are not so discontinuous, intermediate forms being numerous; this condition is more common than the one we have previously described; B. terrestris, B. hortorum, B. lapidarius and B. pomorum are examples of these variable species. The variation runs to a considerable extent in parallel lines in the different species, there being a dark and a light form of each; also each species that has a white termination to the body appears in a form with a red termination, and vice versa. In the Caucasus many species that have everywhere else yellow bands possess them white; and in Corsica there are species that are entirely black, with a red termination to the body, though in continental Europe the same species exhibit yellow bands and a white termination to the body. With so much variation it will be readily believed that much remains to be done in the study of this fascinating genus. It is rich in species in the Northern hemisphere, but poor in the Southern one, and in both the Ethiopian and Australian regions it is thought to be entirely wanting.

The species of the genus Psithyrus (Apathus of many authors) inhabit the nests of Bombus; although less numerous than the species of the latter genus, they also are widely distributed. They are so like Bombus in appearance that they were not distinguished from them by the earlier entomologists; and what is still more remarkable, each species of Psithyrus resembles the Bombus with which it usually lives. There appear, however, to be occasional exceptions to this rule, Smith having seen one of the yellow-banded Psithyrus in the nest of a red-tailed Bombus. Psithyrus is chiefly distinguished from Bombus by the absence of certain characters that fit the latter Insects for their industrial life; the hind tibiae have no smooth space for the conveyance of pollen, and, so far as is known, there are only two sexes, males and perfect females.

Fig. 23—Psithyrus vestalis, Britain. A, Female, x 3⁄2; B, outer side of hind leg.

The Bombus and Psithyrus live together on the best terms, and it appears probable that the latter do the former no harm beyond appropriating a portion of their food supplies. Schmiedeknecht says they are commensals, not parasites; but it must be admitted that singularly few descriptions of the habits and life-histories of these interesting Insects have been recorded. Hoffer has, however, made a few direct observations which confirm, and at the same time make more definite, the vague ideas that have been generally prevalent among entomologists. He found and took home a nest of Bombus variabilis, which contained also a female of Psithyrus campestris, so that he was able to make observations on the two. The Psithyrus was much less industrious than the Bombus, and only left the nest somewhat before noon, returning home again towards evening; after about a month this specimen became still more inactive, and passed entire days in the nest, occupying itself in consuming the stores of honey of its hosts, of which very large quantities were absorbed, the Psithyrus being much larger than the host-bee. The cells in which the young of the Psithyrus are hatched are very much larger than those of the Bombus, and, it may therefore be presumed, are formed by the Psithyrus itself, for it can scarcely be supposed that the Bombus carries its complaisance so far as to construct a cell specially adapted to the superior stature of its uninvited boarder. When a Psithyrus has been for some time a regular inhabitant of a nest, the Bombus take its return home from time to time as a matter of course, displaying no emotion whatever at its entry. Occasionally Hoffer tried the introduction of a Psithyrus to a nest that had not previously had one as an inmate. The new arrival caused a great hubbub among the Bombus, which rushed to it as if to attack it, but did not do so, and the alarm soon subsided, the Psithyrus taking up the position in the nest usually affected by the individuals of the species. On introducing a female Psithyrus to a nest of Bombus in which a Psithyrus was already present as an established guest, the latter asserted its rights and drove away the new comer. Hoffer also tried the experiment of placing a Psithyrus campestris in the nest of Bombus lapidarius—a species to which it was a stranger; notwithstanding its haste to fly away, it was at once attacked by the Bombus, who pulled it about but did not attempt to sting it.

When Psithyrus is present in a nest of Bombus it apparently affects the inhabitants only by diminishing their stores of food to so great an extent that the colony remains small instead of largely increasing in numbers. Although Bombus variabilis, when left to itself, increases the number of individuals in a colony to 200 or more, Hoffer found in a nest in which Psithyrus was present, that on the 1st of September the assemblage consisted only of a queen Bombus and fifteen workers, together with eighteen specimens of the Psithyrus, eight of these being females.

The nests of Bombus are destroyed by several animals, probably for the sake of the honey contained in the pots; various kinds of small mammals, such as mice, the weasel, and even the fox, are known to destroy them; and quite a fauna of Insects may be found in them; the relations of these to their hosts are very little known, but some undoubtedly destroy the bees' larvae, as in the case of Meloe, Mutilla and Conops. Birds do not as a rule attack these bees, though the bee-eater, Merops apiaster, has been known to feed on them very heavily.

The genera of social bees known as Melipona, Trigona or Tetragona, may, according to recent authorities, be all included in one genus, Melipona. Some of these Insects are amongst the smallest of bees, so that one, or more, species go by the name of "Mosquito-bees." The species appear to be numerous, and occur in most of the tropical parts of the continents of the world, but unfortunately very little is known as to their life-histories or economics; they are said to form communities consisting at times of a countless number of individuals; but it has not been thoroughly ascertained whether these are the produce of a single queen, as in the case of the hive-bee, or whether there may be more than one egg-producer in each community. The late F. Smith thought the former of these alternatives would prove to be correct. These mosquito-bees are frequently spoken of as stingless bees, but this is not quite correct, for although they do not sting, von Ihering[^[34]] says that all the essential elements of the sting are present, the pointed or penetrating part of the apparatus being stunted.

It would serve no useful purpose to attempt to construct the social history of these stingless bees from the numerous brief scattered accounts in entomological literature, for they refer to different species; it is, however, positively stated by Smith on the authority of Peckolt[^[35]] that Trigona mosquito sends off swarms after the manner of the hive-bee in this country, and that after searching six hives only one royal female could be found in each.

Fig. 24.—Melipona sp. ♀. Amazons.

The nests of many of these little bees are rich in honey, and they have a host of enemies from man and monkeys downwards; and as they do not defend themselves by stinging, it might be supposed they would have but a poor time of it. From the accounts that have been published we may, however, gather that they are rich in devices for the protection of their nests, and for the exclusion of intruders. Bates has given some particulars as to Melipona interrupta (fasciculata); it is about one-third shorter than the hive-bee, and its colonies are composed of an immense number of individuals. The workers are usually occupied in gathering pollen; but they also collect clay in a similar manner, and convey it to the nest, where it is used for building a wall to complete the fortification of the nest, which is placed either in a suitable bank, or in a trunk of a tree; in either situation it is completely built in with clay. A nest which Bates saw opened contained about two quarts of pleasantly-tasted liquid honey. Forty-five species of these little bees were found in different parts of the Amazons Valley, the largest kind being half an inch in length, the smallest very minute, not more than one-twelfth of an inch. These little creatures are thus masons as well as workers in wax and resin, and they are also gatherers of nectar, pollen, and resin.

According to Gosse, one of these bees is well known in Jamaica, where they are called "Angelitos," in consequence of their not stinging people. He observed a nest of this bee in a tree, and found it to be much infested by black ants anxious to obtain entrance to it; three bees, however, stood sentinel in the entrance, so as to completely block it and keep out intruders, but the middle bee moved on one side out of the way directly one of its fellows wished to come in or out of the nest. The honey accumulated by this species is kept in clusters of cups about the size of a pigeon's egg, at the bottom of the hive and away from the brood-cells. The queen or mother-bee is lighter in colour than the others, and has the hind body twice the length of theirs.

Hockings[^[36]] has given us some details as to the natural history of two of these bees that inhabit Australia, where they are called "Karbi" and "Kootchar," the first being, it is supposed, Trigona carbonaria, Smith: it is usually about three-sixteenths of an inch in length, the queen, when fully developed, being nearly twice that length. The comb is built in a most peculiar form, being, it is said, in the shape of a spiral staircase, and tapering towards the ends: honey-pots and pollen are constructed for the storage of food. The comb is encased in wax, and outside it a labyrinth of waxen passages is formed. The entrance to the colony is guarded by a line of bees who inspect every one that arrives, and it is surprising to see how soon a stranger is discovered and pounced upon before it has time even to alight; the intruder, when caught, is held by several bees, who put it on the rack by holding and stretching out its limbs to their full extent, retaining it in this position for as long as an hour, by which time the unfortunate prisoner is usually dead. These bees, as well as many other allied species, fight desperately with their mandibles, and are apparently of a very fierce disposition. The other species, called "Kootchar," is said to produce a very large number of drones, and the habits and dispositions of the bees differ considerably from those of the "Karbi": the entrance to their hive is guarded by a pipe of propolis (a sort of resinous wax) about an inch in length, having an exceedingly sticky outer edge, and it is by this pipe alone that access to the interior can be gained. At night the entrance is closed by numerous minute globules of semi-fluid gum placed against it, thus forming a thin wall full of air-holes. The colonies of "Kootchar" can be united by taking away a queen and then packing her brood-nest, bees and all, against that of the colony it is to be joined to. This cannot be done with the "Karbi." The account given by Mr. Hockings contains a great many other interesting details, and there can be no doubt that a full account of the natural history of these Insects would be very instructive.

Fritz Müller has recorded a singular case bearing on the instinct of these social Insects: he says that a nest of a small Trigona was built in a hollow tree, and that as a consequence of the irregularity of the hole the bees were obliged to give a very irregular shape to their combs of honey. These bees were captured and put in a spacious box (presumably together with the irregular comb, but this he unfortunately does not mention): after a year, "when perhaps not a single bee survived of those which had come from the canella tree," they still continued to build irregular combs, though quite regular combs were built by several other communities of the same species that he had kept. These bees, he also tells us, do not use pure wax for the construction of their combs, but mix it with resin or gum that gives it a peculiar odour and appearance. He captured two communities of a common Melipona, one of which had the combs made of dark reddish brown, the other of pale yellowish brown, wax, and in captivity in a distant locality each of the two communities continued to form its comb in the same way, thus showing the continuity that prevails in these cases as long as circumstances permit. Müller thinks this due to imitation, but it seems at least as probable that it is due to perception of the properties of the nest. The nest has a certain colour that the worker-bee matches.

Several species of the Melipona and Trigona were imported from Brazil to France, and kept there for some time in captivity by M. Drory. Girard has published[^[37]] some details as to these colonies, and is of opinion that some of them indicate an intelligence or instinct superior to that of the honey-bee. The queen-bee of M. scutellaris seems to display more intelligence than the corresponding sex of A. mellifica. The mode of feeding the larvae apparently differs from that of A. mellifica, a provision of pollen being first placed in the cell, then some honey; when sufficient food for the whole consumption of a larva is accumulated the queen deposits an egg in the cell, which is at once completely closed by the worker. The interior of the abode of these bees is quite dark, only a very small orifice being left, and in this a sentinel is constantly on the alert. The same writer states that Trigona crassipes has the very peculiar habit of always locating its brood-comb in the nest of a species of Termes.

The honey-bee, Apis mellifica (Fig. 6), is considered the highest form attained by the Anthophilous division of the Hymenoptera. The differentiation of the three forms, male, female, and worker, is here carried to a greater degree of perfection than in the other bees. The drones are the males; the individuals we see gathering honey are always workers, neither the male nor the female in this species taking any part in procuring food for themselves or for the colony. In addition to this the colonies formed may be described as permanent: they do not come to an end at the close of one season, and provision is made for the formation of a new colony while the old one still persists, by means of a peculiar process called swarming. The life-history of Apis mellifica and its anatomy and physiology have been discussed in a whole library of works, and we need only notice the chief features. When a swarm of bees leaves a hive it consists of the queen-bee or female, and a number of workers, these latter being, in fact, the surplus population that has been produced in the hive. The swarm is not a nuptial flight, as is often supposed, but an act of emigration. When this swarm has been housed, the bees commence operations in their new quarters, by secreting wax; they are enabled to do this by having consumed much saccharine food; the wax is produced by means of glands in the hind-body over the inner faces of the ventral plates of the abdominal rings, and it makes its appearance there, after passing from the interior of the body through some peculiar membranes on the ventral segments, in the form of thin projecting plates. These the bee takes off with an apparatus on the hind pair of legs and applies, after working up with the mandibles, to form the cells in which young ones are to be reared and food stored. A large number of bees working in common thus produce the regular and beautiful structure known as the comb; the queen afterwards lays an egg in each cell, and as these soon hatch, great labour is thrown on the workers, which have then to feed the young; this they do by eating honey and pollen, which, being formed into a sort of pap by a portion of their digestive organs, is then regurgitated and given to the young, a quantity of it being placed in the cell, so that the larva is bathed by it, and possibly may absorb the food by the skin as well as the mouth. When the colony is in good progress and young bees emerge, these act as nurses, the older ones cease to prepare food and act as foragers, bringing in honey and pollen which are each stored in separate cells. The larva in the cell increases its size and sheds a very delicate skin several times; when the larva has reached its full size no more food is supplied, but the worker-bees seal up the cell by means of a cover formed of pollen and wax, in such a manner as to be pervious to air: sealed up in the cell the larva spins a cocoon for itself, remains therein for a little time as a larva, then changes to a pupa, and thereafter bites its way out through the cover of the cell, and appears for the first time as a new being in the form of a worker-bee; the whole process of development from the egg-state to the perfect condition of the worker-bee occupies about three weeks.

When the denizens of a hive are about to produce another queen, one or more royal cells are formed; these are much larger than the ordinary worker-cells, and of a quite different form. In this cell is placed an egg, not differing in any respect from the egg that, if placed in an ordinary cell, produces a worker; when the egg has produced a larva this is tended with great care and fed throughout its life with royal jelly. This food appears to be the same as that supplied to an ordinary worker-larva when it is first hatched; but there is this difference, that whereas the worker-larva is weaned, and supplied, after the first period of its existence, with food consisting largely of honey, pollen and water, the queen-larva is supplied with the pap or royal jelly until it is full grown. Some difference of opinion exists as to this royal jelly, some thinking that it is a different substance from what the workers are fed with; and it is by no means improbable that there may be some difference in the secretion of the glands that furnish a part of the material composing the pap. The queen is produced more rapidly than workers are, about sixteen days being occupied in the process of her development. Only one queen is allowed in a hive at a time; so that when several queen-cells are formed, and queen-larvae nurtured in them, the first one that is developed into a perfect queen goes round and stings the royal nymphs to death while they are still in their cells. The production of drones is supposed to depend chiefly on the nature of the egg laid by the queen; it being considered that an unfertilised egg is deposited for this purpose. There is still some doubt on this point, however. Though there is no doubt that drones are produced in great numbers from unfertilised eggs, yet there is not evidence that they cannot also be produced from fertilised eggs.[^[38]] The drone-cells are somewhat larger than the ordinary worker-cells, but this is probably not of much import, and it is said that the larvae intended to produce drones receive a greater proportion of pap than worker-larvae do: about twenty-four days are required to produce a drone from the egg.

From this sketch it will be seen that the production of the worker (or third sex, as it is improperly called, the workers being really females atrophied in some points and specially developed in others) is dependent on the social life, in so far at any rate as the special feeding is concerned. There is good reason for supposing that A. mellifica has been kept in a state of domestication or captivity for an enormous period of time; and this condition has probably led to an increase of its natural peculiarities, or perhaps we should say to a change in them to suit a life of confinement. This is certainly the case in regard to swarming, for this process takes place with comparative irregularity in Apis mellifica in a wild condition. The killing of superfluous queens is also probably a phenomenon of captivity, for it varies even now in accordance with the numbers of the colony. It is interesting to notice that in confinement when a swarm goes from the hive it is the old queen that accompanies it, and this swarm as a rule settles down near the old hive, so that the queen-bee being already fertilised, the new swarm and its subsequent increase are nothing but a division of the old hive, the total products of the two having but a single father and mother. When a second swarm goes off from a hive it is accompanied by a young queen, who frequently, perhaps, in the majority of cases, is unfertilised; this swarm is apt to fly for long distances, so that the probability of cross-fertilisation is greatly increased, as the fertilisation of the young new queen is effected during a solitary flight she makes after the colony has settled down. But in a state of nature the colonies do not send off swarms every year or once a year, but increase to an enormous extent, going for years without swarming, and then when their home is really filled up send off, it may be presumed, a number of swarms in one year. Thus the phenomena of bee-life in a wild condition differ considerably from those we see in artificial confinement. And this difference is probably greatly accentuated by the action of parasites, the proportions of which to their guests are in a state of nature liable to become very great; as we have seen to be the case in Bombus.

Under these circumstances it is not a matter for surprise when we find that the honey-bee has formed distinct races analogous to those that exist in the case of the domesticated vertebrate animals. The knowledge of these races is, however, at present very little advanced, and is complicated by the fact that only imperfect information exists as to the true species of the genus Apis. There is a bee very like our common honey-bee found in southern Europe called A. ligustica; this is certainly a variety of A. mellifica, and the same remark applies to a bee found in Egypt, and called A. fasciata. This gives the honey-bee a very wide distribution, extending possibly over the whole of the palaearctic region: besides this, the species has been introduced into various other parts of the world.

According to Karsch the honey-bee shows in Germany several varieties, all of which belong to the northern form, which may be spoken of as the A. domestica of Ray; the A. ligustica and A. fasciata form as we have said distinct races, and it is a remarkable fact that these races remain distinct even when imported into other climates; though for how long a period of time this remains true there is very little evidence to show. The northern form, A. domestica, is now found in very widely separated parts of the world, in some of which it is wild; Smith mentions it as occurring in the West India islands, throughout the North American continent as far south as Mexico, even in Central and Southern Africa, and in Australia and New Zealand. The var. ligustica has been found also at the Cape of Good Hope. The other species known of the genus Apis all belong to the Old World, so that there is very little doubt that A. mellifica is also a true native of the eastern hemisphere, and its original home may possibly have been not far from the shores of the eastern portion of the Mediterranean sea. Seven or eight other species of Apis are known, all but one of which occur in Asia, extending as far as Timor and Celebes. The exceptional one, A. adansonii, occurs in tropical Africa and in Madagascar. Gerstaecker thought these species might be reduced to four, but Smith's statement that the males and even the workers show good distinctive characters seems to be correct. Very little is known as to the honey-bees of China and Japan.

The queen-bee greatly resembles the worker, but has the hind body more elongated; she can, however, always be distinguished from the worker by the absence of the beautiful transverse, comb-like series of hairs on the inner side of the first joint of the hind foot, the planta, as it is called by the bee-keeper: she has also no wax plates and differs in important anatomical peculiarities. The male bee or drone is very different, being of much broader, more robust build, and with very large eyes that quite meet in the middle of the upper part of the head: he also has the hind leg differently shaped. The form of this limb enables the male of A. mellifica to be distinguished from the corresponding sex of allied species of the genus.

Fig. 25.—Portions of hind-feet, 1, of male, 2, of worker, 3, of queen, of the honey-bee; series on the left, outer faces; on the right, inner faces. a, Tip of tibia: b, first joint; c, second joint of tarsus.

We are indebted to Horne for some particulars as to the habits of A. dorsata, an allied East Indian species. He informs us that these bees greatly disfigure buildings, such as the Taj Mahal at Agra, by attaching their pendent combs to the marble arches, and are so pertinacious that it is almost useless to destroy the nests. This bee is said to be so savage in its disposition that it cannot be domesticated; it attacks the sparingly clad Hindoos with great ferocity when they disturb its nest. Notwithstanding its inclination and power to defend its societies this Insect appears to be destroyed wholesale. Colonel Ramsay failed to establish hives of it, because the Insects were eaten up by lizards. The crested honey-buzzard carries off large portions of the comb, and devours it on a branch of some tree near by, quite regardless of the stings of the bees; while the fondness of bears for the honey of the "Dingar," as this species is called, is well known.

Note to P. [33]: It has just been discovered that a most remarkable symbiosis, with structural modification of the bee, exists between the females of Xylocopa, of the Oriental sub-genus Koptorthosoma, and certain Acarids. A special chamber, with a small orifice for entry, exists in the abdomen of the bee, and in this the Acari are lodged.—See Perkins, Ent. Mag. xxxv. 1899, p. 37.

Note to P. [80]: referring to the habits of social wasps in warm countries. The anticipation we ventured to indulge in is shown to be correct by the recent observations of Von Ihering.[^[39]] He states that social wasps in Brazil may be divided into two great groups by their habits, viz. 1. Summer communities, lasting for one year, and founded annually by fertilised females that have hibernated—example, Polistes; 2. Perennial communities, founded by swarms after the fashion of bee colonies—examples, Polybia, Chartergus.

Note to Vol. V. Pp. 545, 546: The development of Encyrtus fuscicollis has now been studied by Marchal, who has discovered the existence of embryonic dissociation. The chain of embryos and the epithelial tube in which they are placed, are formed as follows: the Encyrtus deposits an egg in the interior of the egg of the Hyponomeuta. This does not kill the egg of the Lepidopteron, but becomes included in the resulting caterpillar. The amnion of the Chalcid egg lengthens, and forms the epithelial tube; while the cells within it become dissociated in such a way as to give rise to a chain of embryos, instead of a single embryo.—C.R. Ac. Paris, cxxvi. 1898, p. 662, and translation in Ann. Nat. Hist. (7), ii. 1898, p. 28.

CHAPTER II

HYMENOPTERA ACULEATA CONTINUED—DIVISION II. DIPLOPTERA OR WASPS—EUMENIDAE, SOLITARY TRUE WASPS—VESPIDAE, SOCIAL WASPS—MASARIDAE

Division II. Diploptera—Wasps.

Anterior wings longitudinally plicate in repose; the pronotum extending back, so as to form on each side an angle reposing on the tegula; the basal segments of the hind body not bearing nodes or scales; the hind tarsi formed for simple walking. The species either solitary or social in their habits; some existing in three forms, males, females, and workers.

Fig. 26—Upper aspect of pronotum and mesonotum of a wasp, Eumenes coarctata. a, Angle of pronotum; b, tegula; c, base of wing; d, mesonotum.

This division of Hymenoptera includes the true wasps, but not the fossorial wasps. The name applied to it has been suggested by the fact that the front wings become doubled in the long direction when at rest, so as to make them appear narrower than in most other Aculeata (Fig. 27). This character is unimportant in function so far as we know,[^[40]] and it is not quite constant in the division, since some of the Masaridae do not exhibit it. The character reappears outside the Diploptera in the genus Leucospis—a member of the Chalcididae in the parasitic series of Hymenoptera—the species of which greatly resemble wasps in coloration. A better character is that furnished by the well-marked angle, formed by the pronotum on the dorsal part (Fig. 26). By a glance at this part a Diplopterous Insect can always be readily distinguished.

Three families are at present distinguished in the Diploptera, viz. Eumenidae, Vespidae and Masaridae. We anticipate that Eumenidae and Vespidae will ultimately be found to constitute but one family.

Fam. 1. Eumenidae—Solitary True Wasps.

Claws of the feet toothed or bifid; middle tibiae with only one spur at tip. Social assemblages are not formed, and there is no worker-caste, the duties of nest-construction, etc., being performed solely by the female.

The Eumenidae, or solitary wasps, are very little noticed by the ordinary observer, but they are nevertheless more numerous than the social Vespidae, about 800 species being known. In Britain we have sixteen species of the solitary, as against seven of the social wasps. The Eumenidae exhibit a considerable diversity in form and structure; some of them have the pedicel at the base of the abdomen very elongate, while in others this is so short as to be imperceptible in the ordinary position of the body. A repetition of similar differences of form occurs in the social wasps, so that notwithstanding the difference in habits there seems to be no satisfactory way of distinguishing the members of the two families except by the structure of the claws and tibial spurs.

Fig. 27.—Eumenes flavopicta ♀. Burma. The wings on the left in the position of repose, to show folding.

Fabre has sketched the habits of a species of Eumenes, probably E. pomiformis. This Eumenes constructs with clay a small vase-like earthenware vessel, in the walls of which small stones are embedded (like Fig. 28, B). This it fills with food for the young. The food consists of caterpillars to the number of fourteen or sixteen for each nest. These caterpillars are believed to be stung by the parent-wasp (as is the case in the fossorial Hymenoptera), but complete evidence of this does not seem to be extant, and if it be so, the stinging does not completely deprive the caterpillars of the capacity of movement, for they possess the power of using their mandibles and of making strokes, or kicking with the posterior part of the body. It is clear that if the delicate egg of the Eumenes or the delicate larva that issues from it were placed in the midst of a mass of this kind, it would probably suffer destruction; therefore, to prevent this, the egg is not placed among the caterpillars, but is suspended from the dome covering the nest by a delicate thread rivalling in fineness the web of the spider, and being above the mass of food it is safe. When the young larva leaves the egg it still makes use of the shell as its habitation, and eats its first meals from the vantage-point of this suspension; although the mass of the food grows less by consumption, the little larva is still enabled to reach it by the fact that the egg-shell splits up to a sort of ribbon, and thus adds to the length of the suspensory thread, of which it is the terminal portion. Finally the heap of caterpillars shrinks so much that it cannot be reached by the larva even with the aid of the augmented length of the suspensory thread; by this time, however, the little creature has so much increased in size and strength that it is able to take its place amongst the food without danger of being crushed by the mass, and it afterwards completes its metamorphosis in the usual manner.

Fig. 28—Nidification of solitary wasps: section through nest, A, of Odynerus reniformis; B, of Eumenes arbustorum. a, The suspended egg of the wasp; b, the stored caterpillars. (After André.)

It is known that other species of Eumenes construct vase-like nests; E. unguiculata, however, according to an imperfect account given by Perris, makes with earth a closed nest of irregular shape, containing three cells in one mass. The saliva of these builders has the power of acting as a cement, and of forming with the clay a very impenetrable material. One species, E. coarctata, L. of this genus occurs in Britain. The clay nests (Fig. 29) of this Insect are often attached to the twigs of shrubs, while those of the two species previously mentioned are usually placed on objects that offer a large surface for fixing the foundations to, such as walls. According to Goureau the larva of this species forms in one corner of its little abode, separated by a partition, a sort of dust-heap in which it accumulates the various débris resulting from the consumption of its stores.

Eumenes conica, according to Horne, constructs in Hindostan clay-nests with very delicate walls. This species provisions its nest with ten or twelve green caterpillars; on one occasion this observer took from one cell eight green caterpillars and one black. It is much attacked by parasites owing, it is thought, to the delicacy of the walls of the cells, which are easily pierced; from one group of five cells two specimens only of the Eumenes were reared.

Fig. 29—Nest of Eumenes coarctata: A, the nest attached to wood; B, detached, showing the larva. a, the larva; b, the partition of the cell. (After André.)

Odynerus, with numerous sub-genera, the names of which are often used as those of distinct genera, includes the larger part of the solitary wasps; it is very widely distributed over the earth, and is represented by many peculiar species even in the isolated Archipelago of Hawaii; in Britain we have about fifteen species of the genus. The Odynerus are less accomplished architects than the species of Eumenes, and usually play the more humble parts of adapters and repairers; they live either in holes in walls, or in posts or other woodwork, or in burrows in the earth, or in stems of plants. Several species of the sub-genus Hoplopus have the remarkable habit of constructing burrows in sandy ground, and forming at their entry a curvate, freely projecting tube placed at right angles to the main burrow, and formed of the grains of sand brought out by the Insect during excavation and cemented together. The habits of one such species were described by Réaumur, of another by Dufour; and recently Fabre has added to the accounts of these naturalists some important information drawn from his own observations on O. reniformis.

Fig. 30.—Odynerus antilope ♀. Britain.

This Insect provisions its cell with small caterpillars to the number of twenty or upwards (Fig. 28, A.) The egg is deposited before the nest is stocked with food; it is suspended in such a manner that the suspensory thread allows the egg to reach well down towards the bottom of the cell. The caterpillars placed as food in the nest are all curled up, each forming a ring approximately adapted to the calibre of the cell. Fabre believes these caterpillars to be partly stupefied by stinging, but the act has not been observed either by himself, Réaumur, or Dufour. The first caterpillar is eaten by the wasp-larva from its point of suspension; after this first meal has been made the larva is supposed to undergo a change of skin; it then abandons the assistance of the suspensory thread, taking up a position in the vacant chamber at the end of the cell and drawing the caterpillars to itself one by one. This arrangement permits the caterpillars to be consumed in the order in which they were placed in the cell, so that the one that is weakest on account of its longer period of starvation is first devoured. Fabre thinks all the above points are essential to the successful development of this wasp-larva, the suspension protecting the egg and the young larva from destruction by pressure or movement of the caterpillars, while the position of the larva when it leaves the thread and takes its place on the floor of the cell ensures its consuming the food in the order of introduction; besides this the caterpillars used are of a proper size and of a species the individuals of which have the habit of rolling themselves up in a ring; while, as the calibre of the tube is but small, they are unable to straighten themselves and move about, so that their consumption in proper order is assured. Some interesting points in the habits of an allied species, O. (Pterocheilus) spinipes have been observed by Verhoeff; the facts as regards the construction and provisioning of the cell are almost the same as in O. reniformis. The species of Odynerus are very subject to the attacks of parasites, and are, it is well known, destroyed to an enormous extent by Chrysididae. Verhoeff says that the wasp in question supplied food much infested by entoparasites; further, that a fly, Argyromoeba sinuata, takes advantage of the habit of the Odynerus of leaving its nest open during the process of provisioning, and deposits also an egg in the nest; the Odynerus seems, however, to have no power of discovering the fact, or more probably has no knowledge of its meaning, and so concludes the work of closing the cell in the usual way; the egg of the Argyromoeba hatches, and the maggot produced feeds on the caterpillars the wasp intended for its own offspring. Verhoeff observed that the egg of the wasp-larva is destroyed, but he does not know whether this was done by the mother Argyromoeba or by the larva hatched from her egg. Fabre's observations on allied species of Diptera render it, however, highly probable that the destruction is effected by the young fly-larva and not by the mother-fly.

Mr. R. C. L. Perkins once observed several individuals of our British O. callosus forming their nests in a clay bank, and provisioning them with larvae, nearly all of which were parasitised, and that to such an extent as to be evident both to the eye and the touch. In a few days after the wasps' eggs were laid, swarms of the minute parasites emerged and left no food for the Odynerus. Curiously, as it would seem, certain of the parasitised and stored-up larvae attempted (as parasitised larvae not infrequently do), to pupate. From which, as Mr. Perkins remarks, we may infer that (owing to distortion) the act of paralysing by the wasp had been ineffectual. Mr. Perkins has also observed that some of the numerous species of Hawaiian Odynerus make a single mud-cell, very like the pot of an Eumenes, but cylindrical instead of spherical. This little vessel is often placed in a leaf that a spider curls up; young molluscs of the genus Achatinella also avail themselves of this shelter, so that a curious colony is formed, consisting of the Odynerus in its pot, of masses of the young spiders, and of the little molluscs.

Horne has recorded that the East Indian O. punctum is fond of availing itself of holes in door-posts where large screws have been; after the hole has been filled with provisions, the orifice is covered over level with the surface of the wood so that it eludes human observation. It is nevertheless discovered by an Ichneumon-fly which pierces the covering with its ovipositor and deposits an egg within.

The genus Abispa is peculiar to Australia and includes some very fine solitary wasps, having somewhat the appearance of very large Odynerus: these Insects construct a beautiful nest with a projecting funnel-shaped entrance, and of so large a size that it might pass for the habitation of a colony of social wasps; it appears, however, that this large nest is really formed by a single female.

The species of the genus Rhygchium are also of insecticide habits, and appear to prefer the stems of pithy plants as the nidus for the development of the generation that is to follow them. Lichtenstein says that a female of the European R. oculatum forms fifteen to twenty cells in such a situation, and destroys 150 to 200 caterpillars, and he suggests that, as it is easy to encourage these wasps to nest in a suitable spot, we should utilise them to free our gardens from caterpillars, as we do cats to clear the mice from our apartments.

The East Indian R. carnaticum seems to have very similar habits to its European congener, adapting for its use the hollow stems of bamboos. Horne has recorded a case in which a female of this species took possession of a stem in which a bee, Megachile lanata, had already constructed two cells; it first formed a partition of mud over the spot occupied by the bee, this partition being similar to that which it makes use of for separating the spaces intended for its own young. This species stores caterpillars for the benefit of its larvae, and this is also the case with another Eastern species, R. nitidulum. This latter Insect, however, does not nidificate in the stems of plants, but constructs clay cells similar to those of Eumenes, and fixes them firmly to wood. Rhygchium brunneum is said by Sir Richard Owen to obliterate hieroglyphic inscriptions in Egypt by its habit of building mud nests amongst them. An individual of this wasp was found by Dr. Birch when unrolling a mummy—"There being every reason to believe that the Insect had remained in the position in which it was found ever since the last rites were paid to the ancient Egyptian."

Fam. 2. Vespidae—Social Wasps.

Claws of the feet simple, neither toothed nor bifid, middle tibiae with two spurs at the tip. Insects living in societies, forming a common dwelling of a papery or card-like material; each generation consists of males and females and of workers—imperfect females—that assist the reproductive female by carrying on the industrial occupations.

The anterior wing possesses four submarginal cells, as in the Eumenidae. The attention of entomologists has been more directed to the habits and architecture than to the taxonomy of these Insects, so that the external structure of the Insects themselves has not been so minutely or extensively scrutinised as is desirable; de Saussure, the most important authority, bases his classification of the Insects themselves on the nature of the nests they form. These habitations consist of an envelope, protecting cells similar in form to the comb of the honey-bee, but there is this important difference between the two, that while the bee forms its comb of wax that it secretes, the wasps make use of paper or card that they form from fragments of vegetable tissue,—more particularly woody fibre—amalgamated by means of cement secreted by glands; the vegetable fragments are obtained by means of the mandibles, the front legs playing a much less important part in the economy of the Vespidæ than they do in that of the bees and fossorial Hymenoptera.

In most of the nests of Vespidæ the comb is placed in stages or stories one above the other, and separated by an intervening space, but in many cases there is only one mass of comb. It is the rule that, when the cells of the comb are only partially formed, eggs are deposited in them, and that the larva resulting from the egg is fed and tended by the mother, or by her assistants, the workers; as the larvae grow, the cells are increased in correspondence with the size of the larva; the subsequent metamorphosis to pupa and imago taking place in the cells after they have been entirely closed. The food supplied is of a varied nature according to the species, being either animal or vegetable, or both.

Fig. 31—Section of the subterranean nest of the common wasp, Vespa germanica, in position. (After Janet.) a, One of the chambers of an ant's nest, Lasius flavus, placed above the wasps' nest; b, root to which the first attachment of the nest was made; c, secondary attachments; d, the first-made attachment; e, a flint within the envelopes of the nest; f, the chief suspensory pillar of the second layer of comb; g, lateral galleries; h, one of the secondary pillars of suspension between two layers of comb; i, the layers of wasp-paper forming the envelope of nest; j, vacant space round the nest; k, flints that fell to the bottom during the work of excavation; l, numerous larvae of a fly, Pegomyia inanis (?) placed vertically in ground beneath the nest; m¹ to m⁷, the layers of comb, in m² the cells are indicated, in m⁸ (above the main figure) the arrangement of the three cells forming the commencement of the new layer of comb, m⁷, is shown; n, gallery of access from surface; o, burrow of a mole; p, interval of 90 mm. between top of nest and surface; q, height of the nest, 163 mm.

Although the nests of the social wasps are very elaborate constructions, yet they serve the purposes of the Insects for only a single season. This is certainly the case in our own country. Here each nest is commenced by a single female or queen; she at first performs unaided all the duties for the inauguration of the colony; she lays the foundation of the cells, deposits the eggs in them, feeds the young, and thus rears a brood of workers that at once assist her, and for the future relieve her of a considerable portion of her former occupations; the nest is by them added to and increased, till the cold weather of the autumn is at hand; at this time many males and females are produced; the cold weather either destroys the inhabitants of the nest, or reduces their vitality so that it is impossible for them to pursue successfully the avocations necessary for their subsistence, and they succumb to adversity. The young females, however, hibernate, and each one that lives through the winter is the potential founder of a new nest in the way we have already described. It might be supposed that in tropical countries where no cold season occurs the phenomena would be different, that the colonies would be permanent, and that the nests would be inhabited until they were worn out. De Saussure, however, informs us that this is not the case, but that in the tropics also the colonies die off annually. "The nests are abandoned," he says, "without it being possible to discover the reason, for apparently neither diminution of temperature nor scarcity of food cause them (the Insects) to suffer. One is tempted to suppose that the death of the Insects is the result of a physiological necessity."

Nests of Social Wasps.—In Europe wasps' nests disappear very soon after they are deserted. As it would appear from de Saussure's conclusions that in the tropics as well as in the temperate regions the rule is that the colonies endure only a portion of one year, and that a new nest is commenced by a single founder once in twelve months, it is a somewhat remarkable fact that some tropical wasp-nests are much more durable than the lives of the inhabitants require, so that solidly constructed nests are often found hanging to the trees long after they have been deserted, and are sometimes overgrown with moss. Cuming has recorded the fact that he found in South America an old wasp-nest that had been taken possession of by swallows. We do not assign, however, much importance to the views of de Saussure, because we may anticipate that enquiry will reveal much variety in the habits of tropical and sub-tropical wasps. It is known that species exist that store up honey, after the fashion of bees, and von Ihering has recently shown[^[41]] that in Brazil, species of several genera form new colonies by swarming, after the manner of bees. So that it is possible that certain colonies may remain for a long period in the same nest.

Much more variety exists in wasps' nests than would be supposed probable; those formed by some of the tropical species of Vespidae are enveloped in so solid and beautifully constructed an envelope of papier-maché, that they resist with complete success the torrential rains of the tropics; while some of those found in our own country are made of extremely soft and delicate paper, which is probably chiefly glandular products. Our British Vespidae number only eight species, all belonging to the one genus Vespa, and yet they exhibit three different modes of nidification. Vespa vulgaris, V. germanica and V. rufa form subterranean nests, while V. arborea, V. sylvestris and V. norvegica suspend their habitations from the branches of trees, bushes, or strong annual plants. Vespa crabro, the hornet, usually adopts an intermediate course, forming its nest above ground, but in a spot where it is protected and concealed. The favourite habitat of this formidable Insect is the interior of an old tree, but the hornet will sometimes avail itself of the protection of a thatched roof. Both it and other arboreal species are said, however, to occasionally make subterranean nests. It is ascertained that V. austriaca, the eighth species, is an inquiline.

Fig. 32—Nest of (?) Polybia sp. The envelope partly cut open; o, entrance. (After de Saussure.)

De Saussure,[^[42]] the monographer of the social wasps, classifies them according to the architecture of their nests. He establishes three groups: (1) Stelocyttares, in which the layers of comb are not connected with the envelope, but are supported by pillars made by the wasps (Fig. 31); (2) Poecilocyttares, an unsatisfactory group of which the chief characteristics appear to be that the nest is always covered by an envelope, and the comb is supported by an object such as the branch of a tree, round, or on, which the envelope is placed (Fig. 32); (3) Phragmocyttares, in which the layers of comb are supported, in part or entirely, by the envelope of the nest, communication being effected by a hole in each layer of the comb (Fig. 33). de Saussure's classification is far from satisfactory. There are many social wasps that construct nests destitute of any proper envelope; as an example of this, we may mention the species of the abundant genus Polistes; these Insects make hexagonal cells, of paper-like material, forming an irregular comb, or mass, attached to bushes by a stalk near its centre; these nests are placed so that the mouths of the open cells look downwards. The species of Ischnogaster (Fig. 34) make layers of comb, connected by a pedicel, but without any envelope; these Insects form a section of Stelocyttares called Gymnodomes.

Most of the nests of the Poecilocyttares have only a single layer of comb. The wasps of the genera Synoeca and Polybia have the habit of spreading a layer of cells on a leaf, or on the bark of a tree, and of covering this with an envelope that is pierced by a single orifice only, but that does not rest on the cells, and so allows circulation of the Insects between the cells and the envelope. This appears to be the arrangement in a nest of Synoeca cyanea preserved in the British Museum; in this construction a large layer of cells is moulded on the branch of a tree, whose contour, for a length of two or three feet, it consequently follows; while outside the mass there is placed a continuous envelope, leaving a considerable distance between it and the cells.

It would be impossible in the space at our disposal to give a satisfactory account of all the forms of wasp-nests, and we must therefore refer the student to de Saussure's work, confining ourselves to a brief notice of some specially interesting forms. The habitation of the Brazilian Polybia (Myrapetra) scutellaris is a very solid, closed structure, covered externally with rough knobs or angular projections. Although of very large size—it may be upwards of two feet in length—it is suspended from a branch, and has but one orifice; the arrangement of the combs in the interior is that of the Phragmocyttares, they being firmly attached to the outer envelope, and so placed as to form a curved surface, the convexity of which is downwards: the number of wasps in a well-developed nest of this kind must be very great. This species is said to be a honey-gathering wasp.

One of the best known of the South American wasps' nests is the construction (Fig. 33) of Chartergus chartarius; these nests are so regularly shaped, and formed of papier-maché so compact and solid, as to look like stone: this edifice is attached in a very firm manner to the branch of a tree, and has a single portal of entry beneath; its interior arrangement is much like that of Myrapetra scutellaris.

A very remarkable wasp's nest is preserved in the British Museum of Natural History; it is considered to be the work of Montezumia dimidiata Sauss. an Eumenid wasp; it is a large mass of cells encircling the branch of a tree, which therefore projects somewhat after the manner of an axle through the middle: the cells are very numerous, and are quite as regular as those of the most perfect of the combs of bees: the mass is covered with a very thick layer of paper, the nest having somewhat the external appearance of half a cocoa-nut of twice the usual size.

Fig. 33—Section of nest of Chartergus chartarius. South America, o, Entrance. (After de Saussure.)

Apoica pallida, a South American Insect, forms a nest in a somewhat similar manner to Polistes, but it is covered on its outer aspect by a beautiful paper skin, so that the nest looks somewhat like a toadstool of large size attached to the branch of a tree.

The nests of the Insects of the genus Polybia—which we have already mentioned as located by de Saussure in his unsatisfactory group Poecilocyttares—usually have somewhat the form and size of pears or apples suspended to twigs of trees or bushes; these little habitations consist of masses of cells, wrapped in wasp-paper, in which there are one or more orifices for ingress and egress. Smith says that the combs in the nest of P. pygmaea are of the most exquisite construction, and that it is by no means an uncommon circumstance to find the outer envelope of the nest ornamented with patches of delicate hexagonal tracery. This nest is about the size of an orange.

We have already noticed the variety of nests formed by our British species of the genus Vespa; in other parts of the world the edifices formed by species of Vespa attain a very large size. V. crabroniformis in China, and V. velutina in India, make nests several feet or even yards in length, inhabited by an enormous number of individuals; they are apparently constructed of a material like brittle paper, and are arranged much like the nests of our British hornet, V. crabro. Vespa orientalis mixes a considerable quantity of earth with the paper it uses for its constructive efforts. In the British Museum collection there is a nest said to be that of the Japanese hornet, V. japonica. This is completely covered by a paper envelope, and has apparently only a single small orifice for ingress and egress. In the same collection there is a nest from Bahia (believed to be that of a social wasp, though of what species is unknown), the outer wall of which is apparently formed entirely of earth, and is a quarter or half an inch thick: the comb inside appears also to be formed of clay, the whole forming an elaborate construction in pottery. One is tempted to believe it may prove to be the production of a social Eumenid.

Habits of Social Wasps.—We have already briefly noticed the way in which a colony of wasps is founded, but some further particulars as to the mode in which the society is increased and developed may be mentioned. The queen-wasp makes at first only a very small group of three or four incomplete cells; each cell is at first circular, or nearly so, and moreover is of smaller diameter than it will afterwards be. In each of the first three or four incomplete cells an egg is laid, and more cells are commenced; but as the eggs soon hatch and produce larvae that grow rapidly, the labours of the queen-wasp are chiefly directed to feeding the young. At first she supplies them with saccharine matter, which she procures from flowers or fruits, but soon gives them a stronger diet of insect meat. This is procured by chasing living Insects of various kinds. Some species of wasps prefer particular kinds of Insects, and the hornet is said to be very fond of the honey-bee, but as a rule Diptera are the prey selected. When an Insect has been secured, the hard and innutritious parts are bitten off, and the succulent parts, more especially the thorax which contains chiefly muscular tissue, are reduced to a pulp by means of the mandibles; this is offered to the larvae, which are said to stretch out their heads to the mother to receive the food, after the manner of nestling birds. When a larva is full grown it spins a cocoon in the cell and changes to a pupa. It is said by some entomologists that the queen-wasp closes the cell for the purpose of the larval metamorphosis; but this is contradicted by others, and is probably erroneous. In about a month, or a little less, from the time of deposition of the egg, the perfect Insect is ready for issue, and almost immediately after leaving its cell it assists in the work that is going on for the development of the society. The Insects produced at this early period of the colony are exclusively workers, i.e. imperfect females. They relieve the queen of the task of supplying the larvae with food, and she henceforth remains within the nest, being, it is said, herself fed by her workers; the society now rapidly increases in numbers, and fresh combs are formed, the upper layer being always the oldest. About the month of August, cells of larger size than those that have previously been constructed are formed, and in these males and perfect females are produced; in a few weeks after this the colony languishes and becomes extinct. When it is no longer possible for the enfeebled wasps to carry out their tasks of feeding the brood, they drag the larvae out of the cells and destroy them. An uncertain number of queen-wasps seek protected nooks in which to pass the winter, and each of these queens may be the founder of a nest in the ensuing spring. It should be remarked that de Saussure states that all the intermediate grades between perfect and imperfect females exist, and Marchal's recent observations confirm this. There is in fact no line of demarcation between worker and queen in the wasps as there is in the honey-bee. Von Siebold long since drew attention to the existence of parthenogenesis in certain species of wasps, and it appears probable that it is of common occurrence.

Our knowledge of the social life of European wasps has recently been much increased by the observations of two French naturalists, P. Marchal and C. Janet. The latter has given an elaborate history of a nest of the hornet, showing the rate and variations of increase in numbers. His observations on this and other species indicate that warmth is of the utmost importance to wasps; the Insects themselves create a considerable amount of heat, so that the temperature of their abodes is much greater than that of the air. He considers that in Europe an elevated temperature is essential for the development of the individual,[^[43]] and that the chief object of the various wrappers of paper with which the Insects surround their nests is to keep up this high temperature. These envelopes give a great deal of trouble to the Insects, for they have to be repeatedly destroyed and reformed, as the combs they contain increase in size. Marchal's observations[^[44]] relate chiefly to the production of the sexes and worker-forms, in the subterranean species, Vespa germanica and V. vulgaris. The layers of comb include cells of two sizes. The upper layers, which are the first formed, consist of small cells only: the lower combs are constructed (at Paris) early in August, and consist of larger cells from which males and large females are reared. The males are, however, reared also in large numbers in the small cells. If the queen be removed, the workers become fertile, and produce parthenogenetically many eggs, but all of the male sex. He entertains no doubt that even when the queen is in full vigour the workers produce males if there is an abundant food supply.

The social wasps at present known number 500 or 600 species. Polistes is a very extensive genus, and it has also a very wide geographical distribution; some of the species—and those found in widely-distant parts of the world—are remarkable on account of their excessive variation in colour, and it is worthy of note that the extreme forms have been more than once taken from the same nest.

Next to Polistes, Vespa is the most numerous in species, about 150 being known, and it is to this genus that all our British social wasps belong. No Insects are better known in our islands than these wasps, owing to the great numbers of individuals that occur in certain seasons, as well as to their frequently entering our habitations and partaking of our food, and to the terror that is occasioned by their supposed ferocity and desire to sting. This last feature is a complete mistake; wasps never sting unless they are roused to do so by attacks, or by considerable interference with their work. The only real danger arises from the fact that a wasp may be occasionally taken into the mouth with fruit, or may be handled unawares. When they are flying about they are perfectly harmless unless attacked or irritated, and even if they settle on the person no danger of their stinging exists unless movement is made. Sichel correctly states that a person may station himself close to a wasp's nest and remain there without any risk at all, provided that he makes no movement; indeed, it is more than probable that if no movement, or if only gentle movement, be made, the wasps are unaware of the presence of an intruder. It is, however, well ascertained that if they are molested at their work, more especially when they are actually engaged in the duties of the nest, they are then extremely vindictive, and follow for a considerable distance those who have irritated them. The East Indian V. velutina is specially fierce when aroused, and is said by Horne to have followed a party through dense jungle for miles, and on some occasions to have stung animals, and even human beings, to death.

Fig. 34—Ischnogaster mellyi. Java. A, Female imago (the line at the side shows its length); B, nest, C, maxilla; D, labium; E, mandible (tip downwards). The nest is probably upside down, although shown here as by de Saussure.

This vindictiveness is, however, only an exceptional mood due to some interference with the colony. Even the hornet, notwithstanding its threatening appearance, is harmless unless unduly provoked; its nests and their inhabitants can be kept in domesticity, exhibited to strangers, even moved from place to place, yet the hornets will not take offence if due gentleness be observed. It is said that wasps will rear the progeny of a neighbour in circumstances where this assistance is necessary. Hess has related a case in which a queen-hornet had commenced a nest, and was killed by an accident, leaving young brood in the comb unprovided for: as a result many of the helpless grubs died, and others were in a state of starvation, when a strange queen-hornet appeared, associated itself with the comb, and, adopting the orphan brood, nourished them and brought them to their full size.

We have already alluded to the fact that, so far as external structure is concerned, there is no great difference between the social and the solitary wasps. Both, too, run through analogous series of forms and colours, and the genus Ischnogaster (Fig. 34) seems to connect the two groups by both its structure and mode of life. The social habits are in many species only inferred, and with greater knowledge will probably prove fallacious as a guide to classification; indeed we have already said that in the genus Vespa—perhaps the most perfectly social of all the wasps—there is one species that has no worker, and that lives, it is supposed, as a parasite, in the nests of its congeners. For this species, V. austriaca, it has been proposed to create a separate genus, Pseudovespa, on account of this peculiarity of habit, although no structural character has been detected that could distinguish it. De Saussure has stated his conviction that workers do not exist in some of the exotic genera, so that it appears highly probable that with the progress of knowledge the present division between social and solitary wasps will prove untenable.

Remains of Insects referred to the genera Polistes and Vespa have been found in tertiary strata in various parts of Europe and in North America.

Fig. 35—Masaris vespiformis. A, male; B, female. Egypt. (After Schaum.)

Fam. 3. Masaridae.

Anterior wing with two complete sub-marginal cells. Antennae usually incrassate or clubbed at the extremity. Claws distinctly or obsoletely dentate.

This is a group of fifty or sixty species with but few genera, and most of its components appear to be Insects of the greatest rarity. In their appearance the Insects of this Family differ considerably from the other Diploptera, and as the wings are only imperfectly, or not at all, plicate, it must be admitted that the systematic affinities of the group require reconsideration. The pronotal structure is, however, completely that of Diploptera. The typical form of the Family, Masaris vespiformis, though described a hundred years since, is a species of such extreme rarity, and its sexes are so different, that entomologists have only recently been able to agree about it. It has been found in Egypt and Algeria. The genera Ceramius, Jugurthia, Quartenia and Coelonites are also members of the Mediterranean fauna, while Paragia is Australian, and Trimeria South American. Several species of the genus Masaris inhabit North America, and Cresson has recently described another Masarid genus from the same country, under the name of Euparagia.

The little that is known of their natural history is almost limited to an account given by Giraud of the habits of Ceramius lusitanicus, of which species he found a colony near Briançon. The Insect makes nests in the earth; they are entered by means of a chimney-like passage analogous to what is formed by certain Odynerus; the gallery when completed is about six centimetres long, and at its extremity is an earthen cell in which the larva lives; this is fed by the mother, who brings to it from time to time a supply of a paste, described as being somewhat like dried honey. The growth of the larva is believed to be rapid.

Fig. 36—Cells constructed by Coelonites abbreviatus. (After André.)

Some fragmentary observations made by Lichtenstein on Coelonites abbreviatus have also been recorded. This species, near Montpellier, constructs earthen cells; they are not, however, subterranean, but are placed side by side on the dry stems of plants (Fig. 36); these cells are stored with a material similar to that supplied by Ceramius lusitanicus to its young.

CHAPTER III

HYMENOPTERA ACULEATA CONTINUED—DIVISION III. FOSSORES OR FOSSORIAL SOLITARY WASPS—FAMILY SCOLIIDAE OR SUBTERRANEAN FOSSORS—FAMILY POMPILIDAE OR RUNNERS—FAMILY SPHEGIDAE OR PERFECT-STINGERS

Division III. Fossores.

Aculeate Hymenoptera, in which the abdomen, though very diverse in form, does not bear prominences on the upper aspect of the basal segments; front wing without longitudinal fold along the middle; hairs of body not plumose. Only two forms (male and female) of each species.

Fossorial Hymenoptera are distinguished from other Aculeates at present only by negative characters, i.e. they are Aculeates, but are not ants, bees or wasps. According to their habits they fall into four, by no means sharply distinguished, groups—(1) those that form no special receptacles for their young, but are either of parasitic or sub-parasitic habits, or take advantage of the abodes of other Insects, holes, etc.; (2) constructors of cells of clay formed into pottery by the saliva of the Insect, and by drying; (3) excavators of burrows in the ground; (4) makers of tunnels in wood or stems of plants. Several species make use of both of the last two methods. The habits are carnivorous; the structures formed are not for the benefit of the makers, but are constructed and stored with food for the next generation. Their remarkable habits attracted some attention even 2000 years or more ago, and were to some extent observed by Aristotle. The great variety in the habits of the species, the extreme industry, skill, and self-denial they display in carrying out their voluntary labours, render them one of the most instructive groups of the animal kingdom. There are no social or gregarious forms, they are true individualists, and their lives and instincts offer many subjects for reflection. Unlike the social Insects they can learn nothing whatever from either example or precept. The skill of each individual is prompted by no imitation. The life is short, the later stages of the individual life are totally different from the earlier: the individuals of one generation only in rare cases see even the commencement of the life of the next; the progeny, for the benefit of which they labour with unsurpassable skill and industry, being unknown to them. Were such a solicitude displayed by ourselves we should connect it with a high sense of duty, and poets and moralists would vie in its laudation. But having dubbed ourselves the higher animals, we ascribe the eagerness of the solitary wasp to impulse or instinct, and we exterminate their numerous species from the face of the earth for ever, without even seeking to make a prior acquaintance with them. Meanwhile our economists and moralists devote their volumes to admiration of the progress of the civilisation that effects this destruction and tolerates this negligence.

Fig. 37.—Sceliphron nigripes ♀ (Sub-Fam. Sphegides). Amazons. × 3⁄2.

It should be noted that in the solitary as in the social Insects the males take no part whatever in these industrial occupations, and apparently are even unaware of them. It is remarkable that, notwithstanding this, the sexual differences are in the majority less than is usual in Insects. It is true that the various forms of Scoliidae exhibit sexual distinctions which, in the case of Thynnides and Mutillides are carried to an extreme degree, but these are precisely the forms in which skill and ingenuity are comparatively absent, the habits being rather of the parasitic than of the industrial kind, while the structure is what is usually called degraded (i.e. wingless). The great difference between the habits of the sexes, coupled with the fact that there is little or no difference in their appearance, has given rise to a curious Chinese tradition with regard to these Insects, dating back to Confucius at least.[^[45]] The habit of stinging and storing caterpillars in a cell, from which a fly similar to itself afterwards proceeds having been noticed, it was supposed to be the male that performed these operations; and that when burying the caterpillars he addressed to them a spell, the burden of which is "mimic me." In obedience the caterpillars produce the wasp, which is called to this day "Jiga," that is in English "mimic me." The idea was probably to the effect that the male, not being able to produce eggs, used charmed caterpillars to continue the species.

Summary of the Prey of Fossores.

Great diversity of opinion exists as to the classification of the Fossores. This arises chiefly from the incomplete state of the collections studied, and from the fact that the larger part of the works published are limited to local faunae. Opinions as to the families vary; some admitting only three or four, others upwards of twenty. After consideration of the various views, the writer thinks it best to admit at present only three families, which speaking broadly, correspond with habits, viz. (1) Scoliidae, subterranean stingers; (2) Pompilidae, runners; (3) Sphegidae, stingers above ground.

1. Scoliidae. Pronotum and tegulae in contact. Abdomen with the plane of the ventral surface interrupted by a chink between the first and second segments. Numerous wingless forms.

2. Pompilidae. Pronotum and tegulae in contact. Abdomen with the plane of the ventral surface not interrupted by a chink. Legs very long. No wingless forms.

3. Sphegidae. Pronotum and tegulae not in contact. No wingless forms.

We shall treat as sub-families those divisions of Scoliidae and Sphegidae considered by many as families.

Fam. 1. Scoliidae.

The members of this family, so far as is known, display less perfect instincts than the Sphegidae and Pompilidae, and do not construct cells or form burrows. Information as to the habits is almost confined to European forms. We adopt five sub-families.

Sub-Fam. 1. Mutillides.—The sides of the pronotum reach the tegulae: the female is destitute of wings and ocelli, frequently having the parts of the thorax so closely soldered that the divisions between them are obliterated: the males are winged, furnished with ocelli, and having the thoracic divisions distinct; intermediate tibiae with two apical spurs. Front wing with two or three sub-marginal cells. The larvae live parasitically at the expense of other Hymenoptera Aculeata.

The Mutillides have some resemblance to ants, though, as they are usually covered with hair, and there is never any node at the base of the abdomen, they are readily distinguished from the Formicidae. The great difference between the sexes is their most striking character. Their system of coloration is often very remarkable, the velvet-like pubescence clothing their bodies being variegated with patches of sharply contrasted vivid colour; in other cases the contrast of colour is due to bare, ivory-like spaces. They have the faculty of stridulating, the position and nature of the organ for the purpose being the same as in ants.

Very little exact information exists as to the habits and life-histories of the species. Christ and Drewsen, forty or fifty years ago, recorded that M. europaea lives in the nests of bees of the genus Bombus, and Hoffer has since made some observations on the natural history of the same species in South East Europe, where this Mutilla is found in the nests of ten or eleven species of Bombus, being most abundant in those of B. agrorum and B. variabilis; occasionally more individuals of Mutilla than of bees may be found in a nest. He supposes that the egg of the Mutilla is placed in the young larva of the Bombus, and hatches in about three days; the larva feeds inside the bee-larva, and when growth is completed a cocoon is spun in the interior of the pupa-case of the bee. When the perfect Insects emerge, the males leave the nest very speedily, but the females remain for some time feeding on the bees' honey. Females are usually produced in greater numbers than males. This account leaves much to be desired. From the observations of Radoszkowsky it is clear that other species of Mutillides are by no means confined to the nests of Bombus but live at the expense of Aculeate Hymenoptera of various groups. This naturalist asserts that the basal abdominal segment of the parasite resembles in form that of the species on which it preys.

The apterous condition of the females of Mutillides and Thynnides is very anomalous in the Fossors; this sex being in the other families distinguished for activity and intelligence. The difference between the sexes is also highly remarkable. The males differ from the females by the possession of wings and by the structural characters we have mentioned, and also in a most striking manner in both colour and form; Burmeister, indeed, says that in South America—the metropolis of Mutillides—there is not a single species in which the males and females are alike in appearance; this difference becomes in some cases so extreme that the two sexes of one species have been described as Insects of different families.

Fig. 38—Mutilla stridula. Europe. A, Male; B, female.

Upwards of one thousand species are assigned to the genus Mutilla, which is distributed over the larger part of the world; there is so much difference in these species as to the nervuration of the wings in the males, that several genera would be formed for them were it not that no corresponding distinctions can be detected in the females. Three or four species of Mutilla are described as being apterous in the male as well as in the female sex; they are very rare, and little is known about them. Only three species of Mutillides occur in Britain, and they are but rarely seen, except by those who are acquainted with their habits. The African and East Indian genus, Apterogyna, includes some extremely peculiar Hymenoptera; the males have the wing nervuration very much reduced, and the females are very ant-like owing to the deep constriction behind the first abdominal ring.

Sub-Fam. 2. Thynnides.—Males and females very different in form; the male winged, the front wing with three, or only two, sub-marginal cells; the female wingless and with the thorax divided into three sub-equal parts.

Fig. 39—Methoca ichneumonides. A, Male; B, female. Britain.

The Thynnides are by some entomologists not separated from the Mutillides; but the distinction in the structure of the thorax of the females is very striking. In the Thynnides the nervuration of the wing appears always to extend to the outer margin, and in the Mutillides not to do so. This family is represented in Britain by a single very rare Insect, Methoca ichneumonides: to the unskilled observer the female would appear to be without doubt an ant. This Insect is by some considered as the type of a family distinct from the Thynnides proper. Thynnides are numerous in Australia. Very little is really known as to their habits, though it has been stated that they are parasitic on Lepidoptera, Bakewell having obtained specimens from subterranean cocoons of that Order. Those who are interested in differences between the sexes of one species should examine the extraordinary examples of that phenomenon presented by the Thynnides; the dissimilarity throughout the group—which is now of considerable extent—being so extreme that no entomologist would from simple inspection believe the two sexes to have any connection; but the fact that they are so connected has been demonstrated beyond doubt. In very few cases, however, have the sexes been matched, so that at present males are no doubt standing in the lists of Hymenoptera as one species and their females as other species.

Sub-Fam. 3. Scoliides.—Pronotum reaching back to the tegulae; legs stout; intermediate tibiae with one apical spur; both sexes winged; the nervures not extending to the posterior (i.e. distal) margin.

This group includes some of the largest and most powerful of the Aculeate Hymenoptera. Its members are usually hairy Insects with thick legs, the colour being black, more or less variegated with bands or spots of red or yellow; the hind body is elongate, has only a very short pedicel, and in the male is usually terminated by three projecting spines. The pronotum is of variable dimensions, but its front angles are always co-adapted with the points of insertion of the front wings. The nervuration of the front wings is confined to the basal part, the extensive apical or outer area possessing no nervures. There is frequently a great difference in the size of the two sexes of the same species, the female being very much larger than the other sex. The larvae, so far as is known, devour those of Lamellicorn Coleoptera.

Fig. 40.—Scolia haemorrhoidalis ♀. Europe.

Fabre has investigated the habits of some of the species of Scoliides found in France, and has informed us that their means of subsistence consists of larvae of the larger Lamellicorn beetles, Cetonia, Oryctes, Anoxia, and Euchlora; these beetles belong to very different divisions of the Lamellicornia, but they have in common the fact that their larvae are of subterranean habits, living in the earth or in accumulations of débris in which there is a large proportion of vegetable matter or roots. The female Scolia penetrates into the ground in order to find the Lamellicorn larvae necessary as food for its progeny. Scolia bifasciata attacks the larvae of several species of Cetonia, and S. (Colpa) interrupta chooses the larvae of the chafers Anoxia villosa and A. matutinalis. The mother Scolia enters the ground in August or September, and having found a suitable larva stings it and deposits an egg on the ventral surface of the prey; the paralysed larva is left where it was found, no attempt being made to place it in a special receptacle. The egg is placed on the ventral surface, well behind the feet, under a mass of matter in the alimentary canal. Shortly after being hatched the young destroyer penetrates with its head the skin of the victim, and in this position commences to feed; it is necessary that it should obtain its food without killing the Cetonia larva, for it cannot prosper on decaying food, so that if the Cetonia larva die the Scolia larva likewise perishes; the latter, accordingly, does not withdraw its head from the interior of the victim, but remains always in the same position, as it grows larger extending its head forwards into the front part of the interior of its victim; the internal organs of the latter are consumed in a systematic order so as to delay bringing about its death till the last moment, and thus all the interior of the Cetonia larva is appropriated till nothing remains but an empty skin. By a series of experiments, Fabre showed how essential it is that this apparently revolting operation should be carried on with all details strictly en règle. If the head of the Scolia larva be taken out from the victim and applied to another part of the body of the Cetonia, the result is that it cannot eat; even if it be replaced in the original situation, after being taken away, it frequently happens that the Cetonia larva dies, its death involving also that of the destroyer. It is necessary, too, that the victim should be paralysed, for if an intact Cetonia larva be taken and bound down in such a position that it cannot move, and if a small orifice in its skin be made in the proper spot and a young Scolia larva be placed on it, the little parasite will avail itself of the opportunity and commence to feed on the larva provided for it, but the latter will speedily die, and the Scolia necessarily perishes with it. Thus both the paralysis of the victim and the special mode of eating are essential to the life of the Scolia. The operation of stinging the larva so as to produce the necessary paralysis, or rather insensibility, is a difficult one, and requires great skill and patience. The Cetonia larva is of large size, and must be pierced in one particular spot; in order to reach this the Scolia mounts on its victim, and is frequently dislodged by its struggles; sooner or later, however, the proper position is obtained by the wasp, and the larva is then stung in the exact spot necessary to allow the sting (and the poison introduced by it) to reach the most important of the nervous ganglia that control the movements of the body, this spot being, in the case of the Cetonia, the line of demarcation between the pro- and meso-thorax, on the middle line of the ventral surface of the body. The Scolia gives but one sting to the victim, and this it will not administer until it can do so exactly in the proper place. This practice of devouring the victim slowly, without killing it till all is eaten, is very widely spread in the Hymenoptera, and it is satisfactory to find that we may infer from Fabre's observations that it is not so horrible as it would at first appear; for it is probable that the stinging prevents decomposition of the victim, not by reason, as some have supposed, of the poison injected by the wasp having an antiseptic effect, but rather by means of destroying sensibility, so that the creature does not die from the pain, as it is believed it did in certain cases where Fabre induced the young Scolia larva to feed on a victim that had not been stung. We may here remark that very little exact information exists as to the operation of stinging. Fabre attaches great importance to the sting being inflicted on a nerve-ganglion. Whether a sting that did not reach this part might not have a sufficient effect appears, however, doubtful.[^[46]]

A remarkable form of Scoliides, with wings of smaller size than usual and deeply divided, has been described by Saunders under the name Pseudomeria graeca. Still more remarkable is Komarovia victoriosa found in Central Asia; in this Insect the male retains the appearance of a slender, pallid Scolia, but the female differs totally in form, and has the peculiar wings so reduced in size as to be useless for flight.

Sub-Fam. 4. Sapygides.—Closely allied to the Scoliides, but possessing slender legs and antennae; also the first abdominal segment is less disconnected from the second, so that the outline is less interrupted; the eyes are deeply emarginate; the hind body is not spinose at the apex.

Fig. 41.—Sapyga 5-punctata ♀, Britain.

The economy of Sapyga, the only genus, has been the subject of difference of opinion. The views of Latreille and others that these species are parasitic upon bees is confirmed by the observations of Fabre, from which it appears that S. 5-punctata lives in the burrows of species of the bee-genus Osmia, consuming the store of provisions, consisting of honey-paste, that the bee has laid up for its young. According to the same distinguished observer, the Sapyga larva exhibits hypermetamorphosis (i.e. two consecutive forms), and in its young state destroys the egg of the bee; but his observations on this point are incomplete and need repetition. We have two species of Sapyga in Britain; they differ in colour, and the sexes of S. 5-punctata also differ in this respect; the abdomen, spotted with white in both sexes is in the female variegate with red. Smith found our British Sapyga 5-punctata carrying caterpillars.

Sub-Fam. 5. Rhopalosomides.—Antennae elongate, spinigerous; ocelli very prominent; tarsi of peculiar structure, their claws bifid.

Fig. 42—Rhopalosoma poeyi. A, female imago; B, front of head. Cuba. (After Westwood.)

This sub-family has recently been proposed by Ashmead[^[47]] for an extremely rare American Insect that had previously been placed by Cresson among parasitic Hymenoptera. Westwood classed Rhopalosoma among Diploptera, saying of it "animal quoad affinitates excrucians." We reproduce Westwood's figure, but not being acquainted with the Insect we can express no opinion as to whether it is allied to the Scoliidae or to the Sphegidae. The habits are, we believe, quite unknown.

Fam. 2. Pompilidae.

Pronotum at the sides reaching the tegulae; hind body never definitely pedicellate, though the first segment is sometimes elongate and conical; hind legs long; eyes elliptic in form, not emarginate.

The Pompilidae are perhaps the most extensive and important of the groups of Fossores, and are distributed over all the lands of the globe, with the exception of some islands and of the inclement arctic regions. The sting of the Pompilidae, unlike that of most of the Fossores, inflicts a burning and painful wound; the creatures sometimes attain a length of two or three inches, and a sting from one of these giants may have serious results. Although there is considerable variety in the external form of the members of the group, the characters given above will enable a Pompilid to be recognised with approximate certainty. The elongation of the hind legs includes all the parts, so that while the femur extends nearly as far back as the extremity of the body—in dried examples at any rate—the tibiae and the long tarsi extend far beyond it; thus these Insects have great powers of running; they are indeed remarkable for extreme activity and vivacity. They may frequently be seen running rapidly on the surface of the ground, with quivering wings and vibrating antennae, and are probably then employed in the search for prey, or some other of the operations connected with providing a store of food for their young. Spiders appear to be their special, if not their only, prey. Several authors have recorded details as to the various ways in which the prey is attacked. Fabre has observed the habits of several species, and we select his account of the modus operandi of species of the genera Pompilus and Calicurgus, in their attacks on poisonous spiders that inhabit holes in the ground or in walls. The wasp goes to the mouth of the spider's burrow, and the latter then dashes to the entry, apparently enraged at the audacity of its persecutor.

Fig. 43.—Calicurgus hyalinatus ♀. Britain.

The Calicurgus will not actually enter a burrow when there is a spider in it, because if it did so the spider would speedily dispose of the aggressor by the aid of its poisonous fangs. The Calicurgus, therefore, has recourse to strategy with the object of getting the spider out of its nest; the wasp seizes its redoubtable foe by one foot and pulls; probably it fails to extract the spider, and in that case rapidly passes to another burrow to repeat its tactics; sooner or later a spider is in some moment of inattention or incapacity dragged from its stronghold, and, being then comparatively helpless, feels itself at a disadvantage and offers but a feeble resistance to the wasp, which now pounces on its body and immediately inflicts a sting between the fangs of the foe, and thus at once paralyses these dangerous weapons; thereafter it stings the body of the spider near to the junction of the abdomen and cephalothorax, and so produces complete inactivity. Having secured its prey, the wasp then seeks a suitable hole in which to deposit it; probably an empty burrow of a spider is selected for the purpose, and it may be at a height of several feet in a wall; the Hymenopteron, walking backwards, drags its heavy prey up the wall to bring it to the den. When this is accomplished an egg is deposited on the spider, and the wasp goes in search of a fragment or two of mortar, with which the mouth of the burrow is finally blocked. Fabre's accounts refer to the habits of several species, and give a good insight into some points of the instincts of both the spider and the wasp. It seems that a sense of superiority is produced in one or other of the foes, according as it feels itself in suitable conditions; so that though a spider out of its burrow and on the ground is speedily vanquished by the Pompilid, yet if the two be confined together in a vase, both are shy and inclined to adopt defensive or even evasive tactics, the result probably being that the wasp will be killed by the spider during the night, that being the period in which the attacking powers of the spider are more usually brought into play.

It seems to be the habit of some Pompilus to procure a victim before they have secured a place for its reception; and Fabre took advantage of this fact, and made very interesting observations on some points of the instinct of these wasps. Having found a Pompilus that, after having caught a spider and paralysed it, was engaged in making a retreat for its reception, he abstracted the booty, which was deposited at the top of a small tuft of vegetation near to where the Pompilus was at work. In this case the burrow in course of preparation was subterranean, and was formed by the Pompilus itself, which therefore could not, while it was engaged underground, see what took place near it. It is the habit of the wasp to leave its work of excavation from time to time, and to visit the prey as if to assure itself of the safety of this object, and to enjoy the satisfaction of touching it with the mouth and palping it. Desirous of testing the wasp's memory of locality, Fabre took the opportunity, while the Insect was working at the formation of its burrow, of removing, as we have said, the booty from the place where it had been deposited, and putting it in another spot some half-yard off. In a short time the Pompilus suspended work and went straight to the spot where it had deposited its property, and finding this absent, entered on a series of marches, counter-marches, and circles round the spot where it had left the prey, as if quite sure that this was really the place where the desired object ought to be. At last convinced that the paralysed prey was no longer where it had been placed, the Pompilus made investigations at a greater distance and soon discovered the spider. Fabre recounts that its movements then appeared to indicate astonishment at the change of position that it thus ascertained to have occurred. The wasp, however, soon satisfied itself that this was really the very object it was seeking, and seizing the spider by the leg slightly altered its position by placing it on the summit of a small tuft of vegetation; this latter proceeding being apparently always carried out by this species of Pompilus. Then it returned to its excavation, and Fabre again removed the spider to a third spot; the wasp when it next rested from its work made its way immediately to the second spot, where it had last left the spider, thus showing that it possessed an accurate memory for locality; the wasp was very much surprised at the absence of the valued prize and persisted in seeking it in the immediate vicinity without once returning to the place where it had been first located. Fabre repeated this manoeuvre five times, and the Pompilus invariably returned at once to the spot where it had last left its prey. The acute memory for localities displayed by this Insect seems to be more or less general throughout the Aculeate Hymenoptera, and is of very great importance to them. The power of finding the object appears to depend on sight, for when Fabre, after removing the spider to a fresh spot, made a slight depression in the ground, placed the spider in it and covered it over with a leaf, the wasp did not find it. At the same time, the Insect's sight must be a very different sense from our own, for the wasp, when seeking its lost booty, frequently passed within a couple of inches of it without perceiving it, though it was not concealed.

Belt gives an example of the habits of the Mexican Pompilus polistoides. He noticed it, when hunting for spiders, make a dart at a web in the centre of which a spider was stationed; by this movement the creature was frightened and fell to the ground, where it was seized by the wasp and stung. The Pompilus then dragged its prisoner up a tree and afterwards flew off with it, the burden being probably too heavy for conveyance to the nest without the vantage of an elevation to start from.

Several modifications adopted by Pompilidae in their mode of stinging their spider-victims have been recorded by Ferton; these we cannot allude to in detail, but will nevertheless mention that one species stings the body of its spider-prey at random, and that in other cases it would appear that the paralysis of the spider is evanescent. In short, there are various degrees of perfection in the details of the art of stinging.

The most remarkable of the forms of Pompilidae are the numerous species of Pepsis, a genus peculiar to America, whence upwards of 200 species are already known.[^[48]] Some of them attain a length of two inches or more, and are able to conquer the largest spiders; even the formidable Mygale avicularis succumbs to their agility and skill. Some of these Pepsis have beautifully coloured wings; according to Cameron, this may be due to scales. P. formosus, Say, is called in Texas the tarantula-killer; according to Buckley, its mode of attack on the huge spider is different from that made use of by its European ally. When it discovers a tarantula it flies "in circles in the air, around its victim. The spider, as if knowing its fate, stands up and makes a show of fighting, but the resistance is very feeble and of no avail. The spider's foe soon discovers a favourable moment and darts upon the tarantula, whom it wounds with its sting, and again commences flying in circles." The natural retreat of this huge spider, Mygale hentzii, is in holes in the ground, and this account does not inform us whether the spider allows itself to be overcome when in its nest, or is only attacked when out of its retreat.

The genus Mygnimia includes a very large number of species, and has a wider geographical distribution than Pepsis, being found in the tropical regions of both the Old and New Worlds, some of them rivalling in size and ferocity the larger specimens of the genus Pepsis. In the Insects of this genus there is usually a more or less distinct small space of more pallid colour on the middle of each front wing. Parapompilus is a curious genus consisting of Insects of a great variety of peculiar coloration, and having the wings short, so as to be of little use for flight. P. gravesii is an inhabitant of Chili.

Agenia carbonaria and A. hyalipennis are small and feeble Insects inhabiting the south of Europe. A. carbonaria extends to the south of England. They construct, as nests for their offspring, small earthenware vessels, differing in form according to the species, those of A. hyalipennis being vase-like in shape, while those of A. carbonaria are contracted near the mouth, something after the fashion of a wide-mouthed bottle. The Insect is able by some means—Fabre thinks by the use of saliva—to varnish the interior of the vessel so that it will not absorb water; the outside of the cells is, however, not so protected, and speedily crumbles away when exposed to the action of water; hence the vessel is placed in a protected situation, such as in a tree-stump, or a hole in a wall, or even in an empty snail-shell under a heap of stones. The cells are stored with spiders that have been paralysed by stinging and that serve as food for the larva of the Agenia. The larva of A. carbonaria has been described, and some particulars as to its habits have been given by Verhoeff. It has been stated that this wasp does not paralyse its prey by stinging, but substitutes a process of biting to prevent the spider from hurting the larva that is to feed on it; and Verhoeff's observations seem to show that the legs of the spider are broken by some proceeding of the kind. The Agenia larva is of peculiar shape, the head not being inflexed, while the pleurae of each segment, from the second onwards, are prominent, so as to give the outline of the body a scalloped appearance. This larva is much infested by an Ichneumon that devours, it appears, not only the larva itself, but also the spider that was destined to be food for the larva. Verhoeff seems to have found some evidence that Pompilus sericeus may also be a parasite on the Agenia.

The construction of earthenware cells, instead of the burrows usual in Pompilidae, by the species of this genus is one of the cases alluded to in our introductory remarks as to allied Fossores exhibiting different habits. Mr. Pride has recently sent us from Brazil similar earthen vessels constructed by some Pompilid.

The habits of Pompilids of the genus Ceropales are analogous to those of the parasitic bees. Pérez has recently given us information as to a very curious form of parasitism in this genus; he says that when a Pompilus has obtained a spider as provision for its young, it is pursued by a Ceropales, which lays an egg on the spider, thus as it were substituting in advance its own young for that of the Pompilus. Information as to the subsequent course of events in this case is not at present forthcoming. In another case a Ceropales was observed to oviposit on the spider, not while this is being carried in, but subsequently by entering the nest for the purpose; a habit quite similar to that of some parasitic bees. Ferton has recently made the unexpected discovery that some Pompilus act as robbers; one individual taking away by force the spider that another has captured and is carrying off.

Lichtenstein described a Pompilid larva, that he afterwards ascertained to be Calicurgus hyalinatus, as possessing the extraordinary habit of feeding as an external parasite fixed to the dorsal surface of a spider; thus repeating, it would appear, the habits of some of the Ichnemonidae, though the perfect Insect (Fig. 143) does not differ in structure from its congeners. Emery has given an account of some Pompilids that do not bury their prey, but after stinging it and depositing an egg, simply leave the spider on the spot.

Buller has described the habits of a Pompilid in New Zealand; his account is interesting because it shows a remarkable similarity in the proceedings of this antipodean wasp to those of its congeners on our own side of the world. The species is not scientifically named, but it appears that it is known in New Zealand as "the Mason-bee." It forms a nest of yellow clay consisting apparently of about eight cells, each of which is filled with one or more spiders in a paralysed condition. The figure given of the larva of this Insect by Buller shows it to possess a peculiarly formed head.

It is pleasing to find that Pompilidae do not make use of cruel methods when others will serve their purpose. We are informed that a large Australian Pompilid—Priocnemis bicolor—may find a Cicada sucking sap from a hole it has pierced in a tree. The Priocnemis has not the art of making the puncture necessary to procure sap, so the wasp seizes the Cicada, and shakes it till it leaves its hold and flies away, when the Priocnemis takes its place and sips the sap. It is added that the wasp never hurts the Cicada.

Fam. 3. Sphegidae.

Pronotum free from the tegulae; when the stigmatic lobes extend as far back as the wing-insertion, they are placed below it and separated by a space from it.

This large assemblage of Fossores is the one about which the greatest difference of opinion prevails. It is based entirely on the prothoracic characters mentioned above, and cannot be looked on as natural. We shall, however, follow Kohl[^[49]] in treating for the present as only one family the divisions considered by many as distinct families. They are ten in number.

Sub-Fam. 1. Sphegides.—Hind body with a slender pedicel of variable length; two spurs on the middle tibia. The propodeum usually horizontally elongate.[^[50]]

This group includes a great number of species, about 200 of which are referred to the genus Sphex.

The habits of one species of this genus have been fully described by Fabre; he assigns to the species the name of S. flavipennis, but Kohl considers that it is more probably S. maxillosus. This Insect forms its nests, in the South of France, in the ground, excavating a main shaft with which are connected cells intended for the reception of the provisions for the young. The entrance to the burrow is formed by piercing a hole in the side of a very slight elevation of the soil. Thus the entrance to the construction consists of a horizontal gallery, playing the part of a vestibule, and this is used by the Sphex as a place of retreat and shelter for itself; at the end of the vestibule, which may be two or three inches long, the excavation takes an abrupt turn downwards, extending in this manner another two or three inches, and terminating in an oval cell the larger diameter of which is situate in a horizontal plane. When this first cell has been completed, stored with food, and an egg laid in it, the entrance to it is blocked up, and another similar cell is formed on one side; a third and sometimes a fourth are afterwards made and provisioned, then the Insect commences anew, and a fresh tunnel is formed; ten such constructions being the number usually prepared by each wasp. The Insect works with extreme energy, and as the period of its constructive activity endures only about a month, it can give but two or three days to the construction and provisioning of each of its ten subterranean works. The provisions, according to Fabre, consist of a large species of field-cricket, of which three or four individuals are placed in each cell. Kohl states, however, that in Eastern Europe an Insect that he considers to be the same species as Fabre's Sphex, makes use of locusts as provisions, and he thinks that the habit may vary according to the locality or to the species of Orthoptera that may be available in the neighbourhood. However that may be, it is clear from Fabre's account that this part of the Sphex's duties do not give rise to much difficulty. The cricket, having been caught, is paralysed so that it may not by its movements destroy the young larva for whose benefit it is destined. The Sphex then carries it to the burrow to store it in one of the cells; before entering the cell the Insect is in the habit of depositing its prey on the ground, then of turning round, entering the burrow backwards, seizing as it does so the cricket by the antennae, and so dragging it into the cell, itself going backwards. The habit of depositing its prey on the ground enabled Fabre to observe the process of stinging; this he did by himself capturing a cricket, and when the wasp had momentarily quitted its prey, substituting the sound cricket for the paralysed one. The Sphex, on finding this new and lively victim, proceeds at once to sting it, and pounces on the cricket, which, after a brief struggle, is overcome by the wasp; this holds it supine, and then administers three stings, one in the neck, one in the joint between the pro- and meso-thorax, and a third at the base of the abdomen, these three spots corresponding with the situation of the three chief nervous centres governing the movements of the body. The cricket is thus completely paralysed, without, however, being killed. Fabre proved that an Insect so treated would survive for several weeks, though deprived of all power of movement. Three or four crickets are placed by the wasp in each cell, 100 individuals or upwards being thus destroyed by a single wasp. Although the sting has such an immediate and powerful effect on the cricket, it occasions but a slight and evanescent pain to a human being; the sting is not barbed, as it is in many bees and true wasps, and appears to be rarely used by the Insect for any other purpose than that of paralysing its victims. The egg is laid by the Sphex on the ventral surface of the victim between the second and third pairs of legs. In three or four days the young larva makes its appearance in the form of a feeble little worm, as transparent as crystal; this larva does not change its place, but there, where it was hatched, pierces the skin of the cricket with its tiny head, and thus begins the process of feeding; it does not leave the spot where it first commenced to feed, but gradually enters by the orifice it has made, into the interior of the cricket. This is completely emptied in the course of six or seven days, nothing but its integument remaining; the wasp-larva has by this time attained a length of about 12 millimetres, and makes its exit through the orifice it entered by, changing its skin as it does so. Another cricket is then attacked and rapidly consumed, the whole stock being devoured in ten or twelve days from the commencement of the feeding operations; the consumption of the later-eaten crickets is not performed in so delicate a manner as is the eating of the first victim. When full-grown, the process of forming a cocoon commences: this is a very elaborate operation, for the encasement consists of three layers, in addition to the rough silk that serves as a sort of scaffolding on the exterior: the internal coat is polished and is of a dark colour, owing to its being coloured with a matter from the alimentary canal: the other layers of the cocoon are white or pale yellow. Fabre considers that the outer layers of the cocoon are formed by matter from the silk-glands, while the interior dark coat is furnished by the alimentary canal and applied by the mouth of the larva: the object of this varnish is believed to be the exclusion of moisture from the interior of the cocoon, the subterranean tunnels being insufficient for keeping their contents dry throughout the long months of winter. During the whole of the process of devouring the four crickets, nothing is ejected from the alimentary canal of the larva, but after the cocoon is formed the larva ejects in it, once for all, the surplus contents of the intestine. Nine months are passed by the Insect in the cocoon, the pupal state being assumed only towards the close of this period. The pupa is at first quite colourless, but gradually assumes the black and red colour characteristic of the perfect wasp. Fabre exposed some specimens of the pupa to the light in glass tubes, and found that they went through the pupal metamorphosis in just the same manner as the pupae that remained in the darkness natural to them during this stage of their existence.

Sphex coeruleus is frequently stated to have the habit of provisioning its nests with both Orthoptera and Spiders; but Kohl considers with reason that this record is, as regards spiders, a mistake, arising probably from a confusion with some other Insect of similar appearance, such as Pelopaeus (Sceliphron) coeruleus. S. coeruleus is no doubt the same as S. (Chlorion) lobatus, which Rothney observed in East India, provisioning its nests with Orthoptera. He discovered a nest in process of construction, and during the absence of the mother-wasp abstracted from the burrow a large field-cricket that she had placed in it; he then deposited the Orthopteron near the cell; the parent Sphex on returning to work entered the tunnel and found the provision placed therein had disappeared; she came out in a state of excitement, looked for the missing cricket, soon discovered it, submitted it to the process of malaxation or kneading, and again placed it in the nest, after having cleared it from some ants that had commenced to infest it. She then disappeared, and Rothney repeated the experiment; in due course the same series of operations was performed, and were repeated many times, the Sphex evidently acting in each case as if either the cricket had disappeared owing to its being incompletely stunned, or to its having been stolen by ants. Finally, the observer placed the cricket at a greater distance from the nest, when it recovered from the ill-treatment it had received sufficiently to make its escape. The points of interest in this account are the fact that the cricket was only temporarily paralysed, and that the wasp was quite able to cope with the two special difficulties that must frequently occur to the species in its usual round of occupations.

The genus Ammophila is of wide distribution, and its species make vertical tunnels in the ground. The habits of some of the species found in France have been described by Fabre. The Insect does not inhabit the burrow while it is in process of formation, but quits it; and some of the species temporarily close the entry to the incomplete nest with a stone. The tunnel is a simple shaft with a single cell at its termination; this is stored with caterpillars, the different species of Ammophila selecting different grubs for the purpose. A. hirsuta hibernates in the perfect state, and carries on its work in the spring; it chooses a single larva of considerable size belonging to one of the nocturnal Lepidoptera, and this it paralyses by a series of about nine stings, of which one is implanted in each segment from the first thoracic ring backwards; it forms the burrow only after the food to be placed therein has been obtained. The caterpillar used is subterranean in habit, and the Ammophila detects the larva by some sense, the nature of which appears at present quite uncertain. A. holosericea chooses smaller larvae of the family Geometridae, and uses only one or two stingings to paralyse each larva; several caterpillars are used to provision a single cell, and they are often selected of different colours.

Marchal has also published an important account of the proceedings of A. affinis; he confirms Fabre's observations, and even adds to their interest by suggesting that the Ammophila administers special stings for the purpose of paralysing the mandibles of the caterpillar and depriving it of any power of afterwards injuring the larva that will feed on it. He thinks the mother-Ammophila herself profits by appropriating an exudation from the victim.

Some species of Sphegides have the curious habit of choosing the interiors of human habitations as the spots most suitable for the formation of their own domestic establishments. Fabre has given a charming account of the habits of Pelopaeus (Sceliphron) spirifex, a species that inhabits the South of Europe, and that forms its nests in the cottages of the peasants. The spot usually selected is a nook in the broad, open fireplace, out of reach of the flames, though not of the smoke; here the Pelopaeus forms a nest of earth, consisting of ten to fifty cells, the material being mud or clay brought in little balls by the aid of the Insect's mandibles; about twenty visits are required in order to complete one cell, so that for the construction of a large nest of fifty cells, about one thousand visits must be made by the Insect. It flies in and out of the house apparently not at all incommoded by the human habitants, or by the fact that the peasant's potage may be simmering on the fire quite close to where the fearless little creature is carrying on its architectural operations. The cells are stored with spiders, of which the wasp has to bring a plentiful supply, so that its operations extend over a considerable period. The prey is captured by the Pelopaeus whilst on the wing, and carried off at once, being probably stung by the wasp during the process of transit; apparently it is killed by the operation, not merely paralysed. Only small spiders are taken by this species, and the larva of the Pelopaeus consumes them in a short time, one by one, before the process of decomposition sets in; the egg, too, is laid on the first spider introduced, and this is of course at the bottom of the cell, so that the spiders are eaten by the wasp's larva in the order in which they were brought to the cell. The cell is sealed up when full, the number of spiders placed in it being on the average about eight. The larva completes its task of consuming the store in about ten days, and then forms a cocoon for its metamorphosis. Two or three generations are produced in a single year, the autumnal one passing eight or nine months in the clay cells, which are lodged in a nook of the peasant's hearth, and exposed to the smoke of his fire during all the months of winter. Pelopaeus (Sceliphron) is a genus including many species;[^[51]] several of them are known to be specially attached to the habitations of human beings. Roth has given an account of the habits of P. (Sceliphron) laetus in Australia; he says that in some parts it is very difficult to keep these wasps out of the houses; the nest is formed of mud, and constructed on the furniture or in any part of a room that suits the fancy of the Insect. This it must be admitted is, according to human ideas, liable to the charge of being very capricious. Roth timed a wasp building its nest, and found that it brought a fresh load of mud every two or three minutes. If the wasp be allowed to complete the nest undisturbed, she does so by adding to the exterior diagonal streaks of mud, so giving to the nest the look of a small piece of the bark of a common acacia. The construction consists of from ten to twenty cells, and when completed is provisioned with spiders for the use of the young. This wasp is much pestered by parasites, some of which prevent the development of the larvae by consuming the spiders intended by the mother-wasp for its young. A fly, of the Order Diptera, is said to follow the wasp when carrying a spider, and to deposit also an egg on the food; as the Dipterous larvae have more rapid powers of assimilation, the Pelopaeus larvae are starved to death; and their mildewed remains may be found in the cell, after their enemies have become fully developed and have flown away. Another parasite is said to eat the wasp-larva, and attains this end by introducing an egg through the mud wall and the cocoon of the wasp—a habit that seems to indicate a Leucospid parasite. Tachytes australis, a wasp of the sub-family Larrides also dispossesses this Pelopaeus in a manner we shall subsequently describe. This fragment of natural history from Australia has a special interest, for we find repeated there similar complex biological relations to those existing in the case of the European congeners.

P. (Sceliphron) madraspatanus is common in the north-west provinces of Hindostan, and is called the "mud-dauber" by the European residents. According to Horne it constructs its cells in the oddest places, but chiefly about the inhabited apartments in houses. It is perfectly fearless when engaged in building: the cells are four to six in number, and are usually provisioned with spiders to the number of about twenty. On one occasion it was observed that green caterpillars were stored instead of spiders. The species is said to be protected by a peculiar odour as well as by its sting; it is also stated that it disguises its edifice when completed by making it look like a dab of mud, and on one occasion "rays of mud were observed round the nest, even more exactly imitating a lump of mud thrown with some force." P. (Sceliphron) bilineatus, formerly thought to be a variety of P. madraspatanus, builds its nests in hedges and trees.

Sub-Fam. 2. Ampulicides.—Prothorax long and narrow, forming a neck in front; clypeus beak-like; four submarginal cells, the outer one being complete; metathorax elongate, the posterior part of the metasternum deeply divided to allow a perfect inflection of the abdomen.

Fig. 44—Ampulex compressa. Male. East India.

This is one of the smallest of the divisions of the Sphegidae, but has a very wide distribution, being represented in both the Eastern and Western Hemispheres. It is allied to the Sphegides, but differs by the prolongation of the neck and of the head, and by the articulation between the petiole and thorax being placed on the under surface of the body; the wing-nervures are said to be of inferior importance owing to their frequently differing in individuals of the same species. These Insects appear to be rare in individuals, as well as few in species, and but little has been recorded as to their habits; but it is known that they live on cockroaches. Perkins has given a brief sketch of the habits of Ampulex sibirica that is of great interest, but requires confirmation. He says that this Insect, in West Africa, enters apartments where cockroaches abound, and attacking one, that may probably be four times its own size, succeeds, after a struggle, in stinging it; the cockroach instantly becomes quiet and submissive, and suffers itself to be led away and placed in confinement in some spot such as a keyhole, and in one case was apparently prevented from afterwards escaping, by the wasp carrying some heavy nails into the keyhole. The larva of the Ampulex may be presumed to live on the Blattid, as it is added that dead bodies of the cockroaches are frequently found with the empty cocoon protruding from them. This account, if correct, points to some features in the habits of this Insect that are unique. A remark made by Rothney in reference to the habits of A. (Rhinopsis) ruficornis seems to indicate some similar instinct on the part of that species; he says, "I also saw two or three of these wasps collar a peculiar cockroach by the antennae and lead it off into a crack in the bark, but as the cockroach reappeared smiling each time, I don't know what was up." The same observer records that this species associates with Sima rufonigra, an ant it greatly resembles in appearance, as well as with a spider that is also of similar appearance (Fig. 72). Schurr has given a brief account of the proceedings of Ampulex compressa, and his statements also tend to confirm the correctness of Perkins' report. The habits of a species of Ampulex were partially known to Réaumur, who described them on the authority of M. Cossigni. The species is believed to be A. compressa, which occurs not only in East India, but also in the island of Bourbon, the locality where M. Cossigni made his observation: his account is, like the others, a mere sketch of certain points observed, the most important of which is that when Ampulex cannot introduce the cockroach into a hole that it has selected as suitable, it bites off some portions of the body in order to reduce the poor Insect to the necessary extent.

From these fragmentary observations it would appear that the sting of the Ampulex has not so powerful a paralysing effect as that of most other Fossores; and that the Ampulex does not form any nest, but takes advantage of suitable holes and crevices to store the victim in; also that it displays considerable ingenuity in the selection of materials with which to block up the cavity in which it has placed the partially incapacitated creature.

The genus Dolichurus is by some entomologists considered the type of a sub-family allied to the Ampulicides; it long consisted of a small and rare European Insect, but some exotic species have recently been added to it. It will probably prove not sufficiently distinct from Ampulicides, although the pronotum is much shorter, but Handlirsch has recently observed that the European species attacks Blattidae as do the normal Ampulicides; and Ferton has recorded that D. haemorrhous lives at the expense of Loboptera decipiens, the wasp depositing its egg on the left intermediate femur of the prey. This is placed in a solitary cell, and is entirely consumed by the larva, life being preserved till within a few hours of the end of the repast, which occupies altogether eight days.

Sub-Fam. 3. Larrides.—Hind body not pedicellate, or with only a short pedicel; one spur on the middle tibia; labrum inconspicuous. Marginal cell of the front wings appendiculate,[^[52]] or mandibles excised externally, or both.

This group is by some writers called Tachytides instead of Larrides, as owing to a change of nomenclature Tachytes may now be considered its principal genus. It is in connection with this and the neighbouring sub-families of Sphegidae that some of the greatest taxonomical difficulties exist. We include in Larrides the "Miscophus group" of Kohl.

The species of the genus Tachytes seem to have habits very similar to those of the genus Sphex; they form shafts in the earth and provision them with Orthoptera; like the Sphex and other Fossores, they have the habit, when they fly to their tunnel with a victim, of depositing it for a short time on the ground close to the mouth of the burrow while they turn round and enter backwards; and, after doing this they again seize their prey and drag it into the burrow. Fabre availed himself of an opportunity to remove the prey while the Hymenopteron was entering the hole alone; as a result it had to come out again to seek the object; this it soon found, and carried to the hole, relinquishing it again as usual while it turned round; Fabre repeated the operation several times, and always with the same result; the wasp, though it might have kept hold of the victim while it turned, and thus have saved itself from losing the precious object, never did so.

One species of Tachytes in the south of France selects as its prey Orthoptera of the family Mantidae, Insects of a highly ferocious disposition, and provided with most powerful front legs, capable of cutting in two by a single act the body of an aggressor like the Tachytes; the latter is, however, by no means dismayed by the arms of its future victim, but hovering above the latter for some time, as if to confuse it, and causing it repeatedly to turn its very mobile head, the Tachytes at last pounces down and instantaneously stings the Mantis in the nerve centre between the formidable arms, which at once are reduced to incapacity; subsequently the Tachytes paralyses each of the other pairs of legs, and then carries off its victim.

Fig. 45.—Tachytes pectinipes ♀. Britain.

Larra anathema chooses mole-crickets as the viand for its young, and Tachysphex panzeri selects grasshoppers of the family Acridiidae. Larra pompiliformis (= Tachytes niger, Fabre) sometimes associates itself with Sphex flavipennis (? S. maxillosus, according to Kohl), forming its burrow amidst the works of a colony of that species, and making use, like the Sphex, of crickets for provender. This led Fabre to believe that the Larra stole its prey from the Sphex, but he has since withdrawn this indictment, and declares that the Larra obtains its crickets by the more honourable, if not more humane, process of catching and stinging them itself. Smith has informed us, on the faith of his own observation, that L. pompiliformis uses both Lepidopterous larvae and grasshoppers for its stores.

T. (Larrada) australis, according to Whittell, plays the part of a burglar, breaking open the cells of Pelopaeus (Sceliphron) laetus after they have been completed and stored with spiders; it then takes possession of the cell, and curiously enough the Pelopaeus permits this, although the cell contains its egg and the store of food that is intended for the use of its own young. To us this seems very strange, but it is probable that the Pelopaeus has no idea of the consequences of the intruder's operations; it being one of the strange facts of nature that these highly endowed creatures never even see the offspring for whose welfare they labour with such extraordinary ingenuity and perseverance. Neither can we suppose that they have a conception of it derived from a knowledge of their own individual history; for their very complete metamorphosis is scarcely reconcilable with any such recollection on their part. It may possibly therefore be the case that, having no idea whatever of the offspring, they are equally destitute of any conception that it will be destroyed by the operations of the Larrada. However this may be, Whittell informs us that both wasps skirmish about for a little as if each were mistrustful and somewhat afraid of the other; this ends by the Pelopaeus withdrawing its opposition and by the Larrada taking possession of the cell, which it then proceeds to divide into two, using for the purpose of the partition portions of the material of the nest itself; possibly it is only a contraction of the size of the cell, not a true division, that is effected; however this may be, after it is accomplished the Larrada deposits its own egg in the cell, having, it is believed by Whittell, previously destroyed that of the Pelopaeus. Judging from what occurs in other species it is, however, more probable that the destruction of the egg or young of the Pelopaeus is carried out by the larva of the Larrada and not by the parent-wasp. From a remark made by Maindron as to the proceedings of Larrada modesta, in Ternate, it seems probable that its habits may prove to be similar to those of L. australis, for it frequents the nests of Pelopaeus after they have been completed.

Sub-Fam. 4. Trypoxylonides.—Differ from Larrides by the inner margin of the eyes being concave, and the marginal cell not appendiculate. (In Trypoxylon there is only one distinct submarginal and one distinct discoidal cell, a second of each being indicated faintly.)

The nervuration of Trypoxylon is very peculiar, and differs from that of the widely-distributed genus Pison, though according to Kohl's views the two may be correctly associated to form this sub-family. The species of Trypoxylon are apparently rather fond of human propinquity, and build clay- or mud-nests in or near houses. T. albitarse has this habit, and is well known in Southern Brazil under the name of "Marimbouda da casa"; this Insect, like Pelopaeus, stores its nest with spiders, and Peckholt has remarked that however great may be the number of spiders placed by the mother-wasp in a cell, they are all consumed by the larva, none ever being found in the cell after the perfect Insect escapes therefrom. The European T. figulus forms a nest either in bramble-stems or in sandy soil or walls; it makes use of spiders as provisions.

Sub-Fam. 5. Astatides.—Eyes very large in the male, meeting broadly on the vertex; two spurs on the middle tibia.

Fig. 46—Astata boops, male. Britain.

We have two species of the genus Astata in Britain: one of them—A. boops—is known to form burrows in the ground, each of which contains only a single cell; this, it appears, is usually provisioned with bugs of the genus Pentatoma, Insects remarkable for their strong and offensive odour. St. Fargeau records that this species also makes use of a small cockroach for forming the food-store: thus exhibiting an unique catholicity in the toleration of the disagreeable; almost the only point of connection between bugs and cockroaches being their disagreeable character. According to Smith, Oxybelus, another genus of Fossores, is also used. Authorities are far from agreement as to the validity and relations of the sub-family Astatides. It consists only of the widely-distributed genus Astata, with which the North American Diploplectron (with one species) is doubtfully associated.

Sub-Fam. 6. Bembecides.—Labrum frequently elongate; wing-nervures extending very near to the outer margin; marginal cell of front wing not appendiculate; mandibles not emarginate externally; hind body stout, not pedicellate.

The elongation of the labrum, though one of the most trustworthy of the characters of the Bembecides, cannot be altogether relied on owing to the variation it presents both in this and the allied sub-families. The Bembecides carry their prey to their young tucked underneath their own bodies and hugged to the breast; they affect loose, sandy soils for nidification; make use, in the great majority of the cases where the habits are known, of Diptera for provisions, and give these dead to the young; making repeated visits to supply fresh food to the progeny, which notwithstanding this fact, are distributed in isolated burrows.

Fig. 47.—Bembex rostrata ♂. Europe.

One of the most interesting of Fabre's studies of the instincts of Hymenoptera is devoted to Bembex rostrata. The Bembecides have the habit of forming their nests in the ground in wide expanses of sand, and of covering them up, they leave them so that there appears to be absolutely nothing by which the exact position of the nest can be traced; nevertheless the Bembex flies direct to it without any hesitation. How necessary it is to these Insects to possess this faculty of finding their nests will be understood when we recall that the Bembex does not provision its nest once and for all, but supplies the young at first with only insufficient food, and has therefore to return at daily, or other intervals, with a fresh store of provisions. The burrow is made in the sand by means of the fore-legs; these work with such rapidity and skill that a constant stream of sand flows out behind the Insect while it is engaged in the act of excavation. The nest or cell in which the larva is to live, is formed by this process of digging; but no fastening together of the material occurs, nor does any expedient seem to be resorted to, other than that of making a way through the sand by clearing out all the pieces of stick or stone that might diminish facility of access. The cell being formed, the Bembex leaves the spot in search of prey, and when it has secured a victim in the shape of a two-winged fly, it returns therewith to the burrow, and the booty is placed therein, an egg being deposited on it. The wasp then leaves the burrow, disguising, however, the spot where it is situate, and flies away; to proceed possibly with the formation of other burrows.[^[53]] In the course of twenty-four hours the egg hatches, and the larva in two or three days completely devours the stock provided for it. The mother-wasp then returns with another fly—this time probably a larger one—penetrates rapidly to the bottom of the burrow, and again retreats, leaving the second stock of provisions for the benefit of the greedy larva. These visits of supply are repeated with increased frequency, as the appetite of the larva for the benefit of which they are made increases with its growth. During the fourteen or fifteen days that form this portion of the life-cycle, the single larva is supplied with no less than fifty to eighty flies for food. To furnish this quantum, numerous visits are made to each burrow, and as the mother Bembex has several burrows—though how many does not appear to be known—her industry at this time must be very great. All the while, too, a great danger has to be avoided, for there is an enemy that sees in the booty brought by the Bembex to its young, a rich store for its own progeny. This enemy is a feeble, two-winged fly of the family Tachinidae and the genus Miltogramma; it hangs about the neighbourhood of the nests, and sooner or later finds its opportunity of descending on the prey the Bembex is carrying, choosing for its purpose a moment when the Bembex makes a brief delay just at the mouth of the burrow; then down comes the Miltogramma and lays one, two, or three eggs on some portion of the booty that may be projecting from beneath the body of the wasp. This latter carries in the food for its own young, but thus introduces to the latter the source of its destruction, for the Miltogramma larvae eat up the supply of food intended for the Bembex larvae, and if there be not enough of this provender they satisfy their voracity by eating the Bembex larva itself. It is a remarkable fact that notwithstanding the presence of these strange larvae in the nest the mother Bembex continues to bring food at proper intervals, and, what is stranger still, makes no effort to rid the nest of the intruders: returning to the burrow with a supply of food she finds therein not only her legitimate offspring, a single tenant, but several others, strangers, it may be to the number of twelve; although she would have no difficulty in freeing the nest from this band of little brigands, she makes no attempt to do so, but continues to bring the supplies. In doing so she is fulfilling her duty; what matters it that she is nourishing the enemies of her race? Both race and enemies have existed for long, perhaps for untold periods of time, why then should she disturb herself, or deviate from her accustomed range of duties? Some of us will see in such proceedings only gross stupidity, while others may look on them as sublime toleration.

The peculiar habits of Bembex rostrata are evidently closely connected with the fact that it actually kills, instead of merely paralysing, its prey; hence the frequent visits of supply are necessary that the larvae may have fresh, not putrefying, food; it may also be because of this that the burrow is made in a place of loose sand, so that rapid ingress may be possible to the Bembex itself, while the contents of the burrow are at the same time protected from the inroads of other creatures by the burrow being filled up with the light sand. Fabre informs us that the Bembex larva constructs a very remarkable cocoon in connection with the peculiar nature of the soil. The unprotected creature has to pass a long period in its cocoon, and the sandy, shifting soil renders it necessary that the protecting case shall be solid and capable of keeping its contents dry and sound. The larva, however, appears to have but a scanty supply of silk available for the purpose of constructing the cocoon, and therefore adopts the device of selecting grains of sand, and using the silk as a sort of cement to connect them together. For a full account of the ingenious way in which this difficult task is accomplished the reader should refer to the pages of Fabre himself. Bembecides appear to be specially fond of members of the Tabanidae (or Gad-fly family) as provender for their young. These flies infest mammals for the purpose of feasting on the blood they can draw by their bites, and the Bembecides do not hesitate to capture them while engaged in gratifying their blood-thirsty propensities. In North America a large species of Bembecid sometimes accompanies horsemen, and catches the flies that come to attack the horses; and Bates relates that on the Amazons a Bembecid as large as a hornet swooped down and captured one of the large blood-sucking Motuca flies that had settled on his neck. This naturalist has given an account of some of the Bembecides of the Amazons Valley, showing that the habits there are similar to those of their European congeners.

Sphecius speciosus is a member of the Stizinae, a group recognised by some as a distinct sub-family. It makes use, in North America, of Insects of the genus Cicada as food for its young. Burrows in the ground are made by the parent Insect; the egg is deposited on the Cicada, and the duration of the feeding-time of the larva is believed to be not more than a week; the pupa is contained in a silken cocoon, with which much earth is incorporated. Riley states that dry earth is essential to the well-being of this Insect, as the Cicada become mouldy if the earth is at all damp. As the Cicada is about twice as heavy as the Sphecius itself, this latter, when about to take the captured burden to the nest, adopts the plan of climbing with it to the top of a tree, or some similar point of vantage, so that during its flight it has to descend with its heavy burden instead of having to rise with it, as would be necessary if the start were made from the ground.

Sub-Fam. 7. Nyssonides.—Labrum short; mandibles entire on the outer edge; hind body usually not pedicellate; wing with the marginal cell not appendiculate.

This group has been but little studied, and there is not much knowledge as to the habits of the species. It is admitted to be impossible to define it accurately. It is by some entomologists considered to include Mellinus, in which the abdomen is pedicellate (Fig. 48), while others treat that genus as forming a distinct sub-family, Mellinides. Kohl leaves Mellinus unclassified. Gerstaecker has called attention to the fact that many of the Insects in this group have the trochanters of the hind and middle legs divided: the division is, as a rule, not so complete as it usually is in Hymenoptera Parasitica; but it is even more marked in some of these Nyssonides than it is in certain of the parasitic groups.

Mellinus arvensis is one of our commonest British Fossores, and we are indebted to the late F. Smith for the following account of its habits: "It preys upon flies, and may be commonly observed resorting to the droppings of cows in search of its prey; it is one of the most wary and talented of all its fraternity; were it at once to attempt, by a sudden leap, to dart upon its victim, ten to one it would fail to secure it; no, it does no such thing, it wanders about in a sort of innocent, unconcerned way, amongst the deluded flies, until a safe opportunity presents itself, when its prey is taken without any chance of failure; such is its ordinary mode of proceeding. At Bournemouth the flies are more active, more difficult to capture, or have they unmasked the treacherous Mellinus? and is it found necessary to adopt some fresh contrivance in order to accomplish its ends? if so, it is not deficient in devices. I noticed once or twice, what I took to be a dead specimen of Mellinus, lying on patches of cow-dung; but on attempting to pick them up off they flew; I at once suspected the creature, and had not long to wait before my suspicions were confirmed. Another, apparently dead fellow, was observed; and there, neither moving head or foot, the treacherous creature lay, until a fine specimen of a Bluebottle ventured within its grasp, when, active as any puss, the Mellinus started into life, and pounced upon its victim."

Fig. 48.—Mellinus arvensis ♀. Britain.

Lucas states that in the north of France Mellinus sabulosus provisions its nest with Diptera, which it searches for on the flowers of Umbelliferae, and then carries to its nest. This is a burrow in the earth, and when it is reached the Hymenopteron deposits its Insect burden for a moment on the ground while it turns round in order to enter the burrow backwards. The same writer states that two varieties of this Insect live together—or rather in the same colonies—and make use of different species of Diptera, even of different genera, as food for their young. These Diptera are stung before being placed in the nest. The stinging does not kill the Insect, however, for Lucas was able to keep one specimen alive for six weeks after it had passed this trying ordeal.

Sub-Fam. 8. Philanthides.—Labrum small; anterior wings with three complete submarginal cells; hind body constricted at the base but not so as to form a slender pedicel.

This sub-family contains Insects resembling wasps or Crabronides in appearance, and is, as regards the pronotal structure, intermediate between the two great divisions of the Fossores, for the pronotal lobe extends nearly or quite as far back as the tegulae, and in Philanthus the two come into almost actual contiguity.

The species of the genus Cerceris are numerous in Europe, and several of them are known to make burrows in the ground, and store them with beetles for the benefit of the future larvae. The beetles chosen differ in family according to the species of Cerceris; but it appears from the observations of Fabre and Dufour that one kind of Cerceris never in its selection goes out of the limits of a particular family of beetles, but, curiously enough, will take Insects most dissimilar in form and colour provided they belong to the proper family. This choice, so wide in one direction and so limited in another, seems to point to the existence of some sense, of the nature of which we are unaware, that determines the selection made by the Insect. In the case of our British species of Cerceris, Smith observed C. arenaria carrying to its nest Curculionidae of very diverse forms; while C. labiata used a beetle—Haltica tabida—of the family Chrysomelidae.

Fig. 49.—Philanthus triangulum ♂. Britain.

The beetles, after being caught, are stung in the chief articulation of the body, that, namely, between the pro- and mesothorax. Cerceris bupresticida confines itself exclusively to beetles of the family Buprestidae. It was by observations on this Insect that Dufour first discovered the fact that the Insects stored up do not decay: he thought, however, that this was due to the liquid injected by the wasp exercising some antiseptic power; but the observations of Fabre have shown that the preservation in a fresh state is due to life not being extinguished; the stillness, almost as if of death, being due to the destruction of the functional activity of the nerve centres that govern the movements of the limbs.

It has long been known that some species of Cerceris prey on bees of the genus Halictus, and Marchal has recently described in detail the proceedings of C. ornata. This Insect catches a Halictus on the wing, and, holding its neck with the mandibles, bends her body beneath it, and paralyses it by a sting administered at the front articulation of the neck. The Halictus is subsequently more completely stunned or bruised by a process of kneading by means of the mandibles of the Cerceris. Marchal attaches great importance to this "malaxation"; indeed, he is of opinion that it takes as great a part in producing or prolonging the paralysis as the stinging does. Whether the malaxation would be sufficient of itself to produce the paralysis he could not decide, for it appears to be impossible to induce the Cerceris to undertake the kneading until after it has reduced the Halictus to quietude by stinging.

Fabre made some very interesting observations on Cerceris tuberculata, their object being to obtain some definite facts as to the power of these Insects to find their way home when removed to a distance. He captured twelve examples of the female, marked each individual on the thorax with a spot of white paint, placed it in a paper roll, and then put all the rolls, with their prisoners, in a box; in this they were removed to a distance of two kilometres from the home and then released. He visited the home five hours afterwards, and was speedily able to assure himself that at any rate four out of the twelve had returned to the spot from whence they had been transported, and he entertained no doubt that others he did not wait to capture had been equally successful in home-finding. He then commenced a second experiment by capturing nine examples, marking each with two spots on the thorax, and confining them in a dark box. They were then transported to the town of Carpentras, a distance of three kilometres, and released in the public street, "in the centre of a populous quarter," from their dark prison. Each Cerceris on being released rose vertically between the houses to a sufficient height, and then at once passed over the roofs in a southerly direction—the direction of home. After some hours he went back to the homes of the little wasps, but could not find that any of them had then returned; the next day he went again, and found that at any rate five of the Cerceris liberated the previous day were then at home. This record is of considerable interest owing to two facts, viz. that it is not considered that the Cerceris as a rule extends its range far from home, and that the specimens were liberated in a public street, and took the direction of home at once.

Philanthus apivorus is one of the best known of the members of this sub-family owing to its habit of using the domestic honey-bee as the food for its offspring. In many respects its habits resemble those of Cerceris ornata, except that the Philanthus apparently kills the bee at once, while in the case of the Cerceris, the Halictus it entombs does not perish for several days. The honey-bee, when attacked by the Philanthus, seems to be almost incapable of defending itself, for it appears to have no power of finding with its sting the weak places in the armour of its assailant. According to Fabre, it has no idea of the Philanthus being the enemy of its race, and associates with its destroyer on amicable terms previous to the attack being made on it. The Philanthus stings the bee on the under-surface of the mentum; afterwards the poor bee is subjected to a violent process of kneading, by which the honey is forced from it, and this the destroyer greedily imbibes. The bee is then carried to the nest of the Philanthus. This is a burrow in the ground; it is of unusual depth—about a yard according to Fabre—and at its termination are placed the cells for the reception of the young; in one of these cells the bee is placed, and an egg laid on it: as the food in this case is really dead, not merely in a state of anæsthesia, the Philanthus does not complete the store of food for its larvae all at once, but waits until the latter has consumed its first stock, and then the mother-wasp supplies a fresh store of food. In this case, therefore, as in Bembex, the mother really tends the offspring.

Sub-Fam. 9. Mimesides.—Small Insects with pedicellate hind body, the pedicel not cylindric; mandibles not excised externally; inner margin of eyes not concave; middle tibia with one spur; wings with two, or three, submarginal cells.

Mimesides is here considered to include the Pemphredonides of some authors. Mimesides proper comprises but few forms, and those known are small Insects. Psen concolor and P. atratus form their nests in hollow stems, and the former provisions its nest with Homopterous Insects of the family Psyllidae. Little information exists as to their habits; but Verhoeff states that the species of Psen—like members of the Pemphredoninae—do not form cocoons.

The Pemphredonine subdivision includes numerous small and obscure Insects found chiefly in Europe and North America (Fig. 51, P. lugubris); they resemble the smaller black species of Crabronides, and are distinguished from them chiefly by the existence of at least two complete, submarginal cells on the anterior wing instead of one.

Fig. 50.—Mimesa bicolor ♂. Britain.

The species of Passaloecus live in the burrows that they form in the stems of plants; Pemphredon lugubris frequents the decayed wood of the beech. The larva and pupa of the latter have been described by Verhoeff; no cocoon is formed for the metamorphosis. Both these genera provision their nests with Aphidae. This is also the case with Stigmus pendulus, but the burrows of this species form a complex system of diverticula proceeding from an irregular main channel formed in the pithy stems of bushes. Cemonus unicolor, according to Giraud, forms its burrows in bramble-stems, but it also takes advantage, for the purposes of nidification, of the abandoned galls of Cynips, and also of a peculiar swelling formed by a fly—Lipara lucens—on the common reed, Arundo phragmites. This species also makes use of Aphidae, and Verhoeff states that it has only an imperfect instinct as to the amount of food it stores.

Fig. 51.—Pemphredon lugubris ♀. Britain.

Sub-Fam. 10. Crabronides.—Pronotum short, front wing with one complete submarginal and two discoidal cells: hind body variable in form, pedicellate in some abnormal forms, but more usually not stalked.

The Crabronides (Vespa crabro, the hornet, is not of this sub-family) are wasp-like little Insects, with unusually robust and quadrangular head. They frequently have the hind tibiae more or less thickened, and the clypeus covered with metallic hair. It appears at present that they are specially attached to the temperate regions of the northern hemisphere, but this may possibly be in part due to their having escaped attention elsewhere. In Britain they form the most important part of the fossorial Hymenoptera, the genus Crabro (with numerous sub-genera) itself comprising thirty species. The males of some of the forms have the front tibiae and tarsi of most extraordinary shapes. They form burrows in dead wood, or in pithy stems, (occasionally in the earth of cliffs), and usually store them with Diptera as food for the larvae: the wings and dried portions of the bodies of the flies consumed by Crabronides are often exposed to view when portions of old wood are broken from trees.

Fig. 52.—Crabro cephalotes ♀. Britain.

The genus Oxybelus is included by some systematists, but with doubt, in this sub-family; if not placed here, it must form a distinct sub-family. It has the metathorax spinose, and the sub-marginal and first discoidal cells are not, or are scarcely, separated.

Crabro leucostomus has been observed by Fletcher to form cells for its larvae in the soft wood of broken willows: the food stored therein consists of two-winged flies of the family Dolichopodidae. This Crabro is parasitised by an Ichneumonid of the genus Tryphon, and by a two-winged fly of uncertain genus, but belonging to the family Tachinidae. The metamorphoses of Crabro chrysostomus have been briefly described by Verhoeff: the food stored consists of Diptera, usually of the family Syrphidae; the larva spins an orange-red cocoon, passes the winter therein, and assumes the pupal form in the spring; there is, he says, a segment more in the female pupa than there is in the male.

The species of the sub-genus Crossocerus provision their nests with Aphididae, but C. wesmaeli makes use, for the purpose, according to Ferton, of an elegant little fly of the family Tipulidae; according to Pissot this same wasp also makes use of a species of Typhlocyba, a genus of the Homopterous division of Rhynchota. Supposing there to be no mistake as to this latter observation, the choice of Diptera and of Homoptera by the same species indicates a very peculiar habit.

Fertonius (Crossocerus) luteicollis in Algeria forms cells at a slight depth in sandy soil, and provisions them with ants. The ant selected is Tapinoma erraticum, and the individuals captured are the wingless workers. The mode of hunting has been described by Ferton; the wasp hovers over one of the ant-paths at a distance of a few millimetres only above the surface, and when an ant that is considered suitable passes, the Fertonius pounces on it, stings it, and carries it off to the burrow; forty or fifty ants are accumulated in a cell, the egg is laid in the heap of victims about one-third of the depth from the bottom; the resulting larva sucks the ants one by one, by attaching itself to the thorax behind the first pair of legs. There is a very interesting point in connection with the habits of this species, viz. that the ants are not only alive, but lively; they have, however, lost the power of co-ordinating the movements of the limbs, and are thus unable to direct any attack against the feeble larva. Ferton thinks there are three generations of this species in a single year.

* * * * *

Note.—In a note on p. [99] we have mentioned the new publication of Mr. and Mrs. Peckham on the habits of Fossores. We may here add that it contains much fresh information on these Insects, together with criticisms of the views of Fabre and others. One of the points most noteworthy is that they have observed Crabro stirpicola working night and day for a period of forty-two consecutive hours. They made experiments on Bembex spinolae with a view of ascertaining whether the female provisions two nests simultaneously; as the result they think this improbable. If the female Bembecid make nests only consecutively, it is clear it must have but a small fecundity. The larval life extends over about fifteen days; and if we allow three months as the duration of life of a female, it is evident that only about six young can be produced in a season.

CHAPTER IV

HYMENOPTERA ACULEATA CONTINUED—DIVISION IV. FORMICIDAE OR ANTS

Division IV. Heterogyna or Formicidae—Ants.

The segment, or the two segments, behind the propodeum, either small or of irregular form, so that if not throughout of small diameter, the articulation with the segment behind is slender, and there is great mobility. The trochanters undivided. The individuals of each species are usually of three kinds, males, females and workers; the latter have no wings, but the males and females are usually winged, though the females soon lose the flying organs. They live in communities of various numbers, the majority being workers. The larvae are helpless maggots fed and tended by the workers or by the female.

Fig. 53—Abdomens of ants. A, Of Camponotus rubripes (Formicides); B, of Ectatomma auratum (Ponerides); C, of Aphaenogaster barbara (Myrmicides). a, Propodeum; b, first abdominal segment forming a scale or node; c, second; d, third abdominal segment.

In ants the distinction between the three great regions of the body is very marked. The abdomen is connected with the propodeum in a peculiar manner, one or two segments being detached from the main mass to form a very mobile articulation. This is the most distinctive of the characters of ants. The structure and form of these parts varies greatly in the family: and the Amblyoponides do not differ in a marked manner from the Scoliidae in fossorial Hymenoptera.

Fig. 54—Front of head of Dinoponera grandis. A, Mouth closed; B, open.

The arrangement of the parts of the mouth is remarkable, and results in leaving the mandibles quite free and unconnected with the other trophi; the mouth itself is, except during feeding, closed completely by the lower lip and maxilla assuming an ascending vertical direction, while the upper lip hangs down and overlaps the lower lip, being closely applied to it; so that in Ponerides the palpi, except the apices of the maxillary pair, are enclosed between the upper and lower lips (Fig. 54, A). In Cryptocerini the palpi are not covered by the closed lips, but are protected by being placed in chinks at the outsides of the parts closing the mouth. The mandibles of ants can thus be used in the freest manner without the other parts of the mouth being opened or even moved. The mandibles close transversely over the rest of the mouth, and when shut are very firmly locked. There are, however, some ants in which the lips remain in the position usual in mandibulate Insects.

The antennae, except in the males of some species, have a long basal joint and are abruptly elbowed at its extremity. The eyes and ocelli vary excessively, and may be totally absent or very highly developed in the same species. The winged forms are, however, never blind. The size of the head varies extremely in the same species; it is frequently very small in the males, and largest in the workers. In some ants the worker-caste consists of large-headed and small-headed individuals; the former are called soldiers, and it has been supposed that some of them may act the part of superior officers to the others. It should be clearly understood that there is no definite distinction between soldiers and workers; so that in this respect they are widely different from Termites.

Fig. 55—Oecodoma cephalotes. South America. A, Worker major; B, female after casting the wings.

Fig. 56.—Stridulating organ of an ant, Myrmica rubra, var. laevinodis. Sagittal section of part of the 6th and 7th post-cephalic segments. (After Janet.) a, a¹, muscles; b, connecting membrane (corrugated) between 6th and 7th segments: c, 6th segment; d, its edge or scraper; e, striate area, or file on 7th segment; f, posterior part of 7th segment; g, cells, inside body; h, trachea.

The complex mass forming the thorax is subject to great change of structure in the same species, according as the individuals are winged or wingless. The sutures between the dorsal (notal) pieces are frequently obliterated in the workers, while they are distinct in the males and females, and the pieces themselves are also much larger in size in these sexed individuals. The pro-mesothoracic stigma is apparently always distinct; the meso-metathoracic one is distinct in the male Dorylus, but can scarcely be detected in the winged forms of other ants, owing to its being enclosed within, and covered by, the suture between the two segments: in the workers, however, it is usually quite conspicuous. The posterior part of the thoracic mass, the propodeum or median segment, is of considerable size; no transverse suture between the component pieces of this part can be seen, but its stigma is always very distinct. The peduncle, or pedicel, formed by the extremely mobile segment or segments at the base of the abdomen (already noticed as forming the most conspicuous character of the family), exhibits much variety. Sometimes the first segment bears a plate or shield called a scale (Fig. 53, A, b); at other times there are two small segments (Fig. 53, B, C, b, c) forming nodes or knots, of almost any shape. The articulations between these segments are of the most perfect description. In many ants these parts bear highly developed stridulating organs, and the delicacy and perfection of the articulations allow the parts to be moved either with or without producing stridulation. In the male sex the peduncle and its nodes are much less perfect, and possess comparatively little capacity for movement; in the male of Dorylus (Figs. 79, A, and 80, f) the single node is only imperfectly formed. The eyes and ocelli of the males are usually more largely developed than they are in the female, though the head is much smaller.

The legs of ants are elongate, except in a few forms; the Cryptocerini and the males of Dorylides being the most conspicuous exceptions. The tarsi are five-jointed, the basal joint being disproportionately elongate, so that in use it acts in many species as if it were a portion of the tibia, the other four joints forming the functional foot. The front tibiae are furnished with a beautiful combing apparatus (Fig. 57).

Fig. 57.—Combs and brushes on front leg of an ant, Dinoponera grandis (tip of tibia, bearing the comb-like spur, and the base of the first joint of the tarsus; cf. fig. 75). A, Inner, B, outer aspect.

Features of Ant-life.—In order that the reader may realise the nature of ant-life we may briefly recount its more usual and general features. Numerous eggs are produced in a nest by one or more queens, and are taken care of by workers. These eggs hatch and produce helpless maggots, of which great care is taken by the workers. These nurses feed their charges from their own mouths, and keep the helpless creatures in a fitting state by transporting them to various chambers in conformity with changes of temperature, humidity, and so on. When full grown the maggots change to pupae. In some species the maggots form cocoons for themselves, but in others this is not the case, and the pupae are naked.[^[54]] After a brief period of pupal life a metamorphosis into the perfect Insect occurs. The creatures then disclosed may be either winged or wingless; the wingless are the workers and soldiers—imperfect females—the winged are males or females fully developed. The workers remain in or near the nest they were produced in, but the winged individuals rise into the air for a nuptial flight, often in great numbers, and couple. When this is accomplished the male speedily dies, but the females cast their wings and are ready to enter on a long life devoted to the production of eggs. From this account it will be gathered that males are only found in the nests for a very short time; the great communities consisting at other periods entirely of the two kinds of females and of young. The imperfect females are themselves in some species of various kinds; each kind being restricted, more or less completely, to a distinct kind of duty.

No Insects are more familiar to us than ants; in warm countries some of them even invade the habitations of man, or establish their communities in immediate proximity to his dwellings. Their industry and pertinacity have, even in remote ages, attracted the attention and admiration of serious men; some of whom—we need scarcely mention Solomon as amongst them—have not hesitated to point out these little creatures as worthy of imitation by that most self-complacent of all the species of animals, Homo sapiens.

Observation has revealed most remarkable phenomena in the lives of these Insects. Indeed, we can scarcely avoid the conclusion that they have acquired in many respects the art of living together in societies more perfectly than our own species has, and that they have anticipated us in the acquisition of some of the industries and arts that greatly facilitate social life. The lives of individual ants extend over a considerable number of years—in the case of certain species at any rate—so that the competence of the individual may be developed to a considerable extent by exercise; and one generation may communicate to a younger one by example the arts of living by which it has itself profited. The prolonged life of ants, their existence in the perfect state at all seasons, and the highly social life they lead are facts of the greatest biological importance, and are those that we should expect to be accompanied by greater and wider competence than is usually exhibited by Insects. There can indeed be little doubt that ants are really not only the "highest" structurally or mechanically of all Insects, but also the most efficient. There is an American saying to the effect that the ant is the ruler of Brazil. We must add a word of qualification; the competence of the ant is not like that of man. It is devoted to the welfare of the species rather than to that of the individual, which is, as it were, sacrificed or specialised for the benefit of the community. The distinctions between the sexes in their powers or capacities are astonishing, and those between the various forms of one sex are also great. The difference between different species is extreme; we have, in fact, the most imperfect forms of social evolution coexisting, even locally, with the most evolute.

These facts render it extremely difficult for us to appreciate the ant; the limitations of efficiency displayed by the individual being in some cases extreme, while observation seems to elicit contradictory facts. About two thousand species are already known, and it is pretty certain that the number will reach at least five thousand. Before passing to the consideration of a selection from what has been ascertained as to the varieties of form and of habits of ants we will deal briefly with their habitations and polymorphism, reserving some remarks as to their associations with other Insects to the conclusion of this chapter.

Nests.—Ants differ greatly from the other Social Hymenoptera in the nature of their habitations. The social bees construct cells of wax crowded together in large numbers, and the wasps do the like with paper; the eggs and young being placed, each one in a separate cell, the combinations of which form a comb. Ants have, however, a totally different system; no comb is constructed, and the larvae are not placed in cells, but are kept in masses and are moved about from place to place as the necessities of temperature, air, humidity and other requirements prompt. The habitations of ants are in all cases irregular chambers, of which there is often a multiplicity connected by galleries, and they sometimes form a large system extending over a considerable area. Thus the habitations of ants are more like those of the Termites than those of their own allies among the Hymenoptera. They are chiefly remarkable for their great variety, and for the skilful manner in which they are adapted by their little artificers to particular conditions. The most usual form in Europe, is a number of subterranean chambers, often under the shelter of a stone, and connected by galleries. It is of course very difficult to trace exactly the details of such a work, because when excavations are made for the purposes of examination, the construction becomes destroyed; it is known, however, that some of these systems extend to a considerable depth in the earth, it is said to as much as nine feet, and it is thought the object of this is to have access to sufficiently moist earth, for ants are most sensitive to variations in the amount of moisture; a quite dry atmosphere is in the case of many species very speedily fatal. This system of underground labyrinths is sometimes accompanied by above-ground buildings consisting of earth more or less firmly cemented together by the ants; this sort of dwelling is most frequently adopted when the soil in which the nests are placed is sandy; it is probable that the earth is in such cases fastened together by means of a cement produced by the salivary glands of the ants, but this has not been determined with certainty; vaulted galleries or tunnels of this kind are constructed by many species of ants in order to enable them to approach desired objects.

Fig. 58—Portion of combined nest of Formica fusca and Solenopsis fugax. (After Forel.) × ⅔. f, f′, Chambers of Formica, recognisable by the coarser shading; s, s′, chambers of the Solenopsis (with finer shading); s″, opening in one of the chambers, the entrance to one of the galleries that connects the chambers of the Solenopsis; w, walls forming the foundations of the nest and the limits of the chambers.

In South America Camponotus rufipes and other species that habitually dwell in stumps, in certain districts where they are liable to inundations, build also nests of a different nature on trees for refuge during the floods. In Europe, a little robber-ant, Solenopsis fugax, constructs its dwelling in combination with that of Formica fusca (Fig. 58), in such a manner that its chambers cannot, on account of the small size of the orifices, be entered by the much larger Formica. Hence the robber obtains an easy living at the expense of the larger species. The Sauba or Sauva ants of South America (the genus Atta of some, Oecodoma of other authors) appear to be most proficient in the art of subterranean mining. Their systems of tunnels and nests are known to extend through many square yards of earth, and it is said on the authority of Hamlet Clark that one species tunnelled under the bed of the river Parahyba at a spot where it was as broad as the Thames at London Bridge.

A considerable number of ants, instead of mining in the ground, form chambers in wood; these are usually very close to one another, because, the space being limited, galleries cannot be indulged in. Camponotus ligniperdus in Europe, and C. pennsylvanicus in North America, work in this way.

Our British Lasius fuliginosus lives in decayed wood. Its chambers are said by Forel to consist of a paper-like substance made from small fragments of wood. Cryptocerus burrows in branches. Colobopsis lives in a similar manner, and Forel informs us that a worker with a large head is kept stationed within the entrance, its great head acting as a stopper; when it sees a nest-fellow desirous of entering the nest, this animated and intelligent front-door then retreats a little so as to make room for ingress of the friend. Forel has observed that in the tropics of America a large number of species of ants live in the stems of grass. There is also quite a fauna of ants dwelling in hollow thorns, in spines, on trees or bushes, or in dried parts of pithy plants; and the tropics also furnish a number of species that make nests of delicate paper, or that spin together by means of silk the leaves of trees. One eastern species—Polyrhachis spinigera—fabricates a gauze-like web of silk, with which it lines a subterranean chamber after the manner of a trap-door spider.

Fig. 59—Ant-plant, Hydnophytum montanum. Java. (After Forel.)

Some species of ants appear to find both food and shelter entirely on the tree they inhabit, the food being usually sweet stuff secreted by glands of the plant. It is thought that the ants in return are of considerable benefit to the plant by defending it from various small enemies, and this kind of symbiosis has received much attention from naturalists. A very curious condition exists in the epiphytic plants of the genera Myrmecodia and Hydnophytum; these plants form large bulb-like (Fig. 59) excrescences which, when cut into, are found to be divided into chambers quite similar to those frequently made by ants. Though these structures are usually actually inhabited by ants, it appears that they are really produced by the plant independent of the Insects.

Variability and Polymorphism of Ants.—Throughout the Hymenoptera there are scattered cases in which one of the sexes appears in dimorphic form. In the social kinds of bees and wasps the female sex exists in two conditions, a reproductive one called queen, and an infertile one called worker, the limits between the two forms seeming in some cases (honey-bee) to be absolute as regards certain structures. This sharp distinction in structure is rare; while as regards fertility intermediate conditions are numerous, and may indeed be induced by changing the social state of a community.[^[55]] In ants the phenomena of the kind we are alluding to are very much more complex. There are no solitary ants; associations are the rule (we shall see there are one or two cases in which the association is with individuals of other species). In correlation with great proclivity to socialism we find an extraordinary increase in the variety of the forms of which species are made up. In addition to the male and female individuals of which the species of Insects usually consist, there are in ants workers of various kinds, and soldiers, all of which are modified infertile females. But in addition to the existence of these castes of infertile females, we find also numerous cases of variability or of dimorphism of the sexual individuals; and this in both sexes, though more usually in the female. Thus there exists in ants an extraordinary variety in the polymorphism of forms, as shown by the table on p. 141, where several very peculiar conditions are recorded.

The complex nature of these phenomena has only recently become known, and as yet has been but little inquired into. The difference between the thoracic structure in the case of the winged and wingless females of certain species (Fig. 55, and in vol. v. fig. 339) is enormous, but in other species this difference appears to be much less. The ordinary distinctions between the queen-female and worker-females appear to be of two kinds; firstly, that the former is winged, the latter wingless;[^[56]] and secondly, that the former possesses a receptaculum seminis, the latter does not. In a few cases it would seem that the dimorphism of winged and wingless forms is not complete, but that variability exists. Intermediate conditions between the winged and wingless forms are necessarily rare; nevertheless a certain number have already been detected, and specimens of Lasius alienus have been found with short wings. In rather numerous species some or all of the fertile females depart from the usual state and have no wings; (a similar condition is seen, it will be recollected, in Mutillides and Thynnides of the neighbouring family Scoliidae). A dimorphism as regards wings also exists in the male sex, though it is only extremely rarely in ants that the males are wingless. Neverless a few species exist of which only wingless males have been found, and a few others in which both winged and wingless individuals of this sex are known to occur. The wingless males of course approach the ordinary workers (= infertile wingless females) in appearance, but there is not at present any reason for supposing that they show any diminution in their male sexual characters. The distinction between workers and females as based on the existence or non-existence of a receptaculum seminis has only recently become known, and its importance cannot yet be estimated. The adult, sexually capable, though wingless forms, are called ergatoid, because they are similar to workers (Ἐργατης, a worker).[^[57]]

Table of the Chief Forms of Polymorphism in Ants.

In addition to the above there are apparently cases of females with post-metamorphic growth in Dorylides, but these have not yet been the subject of investigation.

Much has been written about the mode in which the variety of forms of a single species of ant is produced. As to this there exists but little actual observation or experiment, and the subject has been much complicated by the anxiety of the writers to display the facts in a manner that will support some general theory. Dewitz was of opinion that workers and queens of ants were produced from different kinds of eggs. This view finds but little support among recent writers. Hart in recording the results of his observations on the parasol ant (of the genus Atta)—one of the species in which polymorphism is greatest—says[^[58]] that these observations prove that "ants can manufacture at will, male, female, soldier, worker or nurse," but he has not determined the method of production, and he doubts it being "the character of the food." There is, however, a considerable body of evidence suggesting that the quality or quantity of the food, or both combined, are important factors in the treatment by which the differences are produced. The fact that the social Insects in which the phenomena of caste or polymorphism occur, though belonging to very diverse groups, all feed their young, is of itself very suggestive. When we add to this the fact that in ants, where the phenomena of polymorphism reach their highest complexity, the food is elaborated in their own organs by the feeders that administer it, it appears probable that the means of producing the diversity may be found herein. Wasmann has pointed out that the ants'-nest beetle, Lomechusa, takes much food from the ants, and itself destroys their young, and that in nests where Lomechusa is abundant a large percentage of ergatogynous forms of the ants are produced. He attributes this to the fact that the destruction of the larvae of the ant by the beetle brings into play the instinct of the ants, which seek to atone for the destruction by endeavouring to produce an increased number of fertile forms; many ergatogynous individuals being the result. This may or may not be the case, but it is clear that the ants' instinct cannot operate without some material means, and his observation adds to the probability that this means is the food supply, modified either qualitatively or quantitatively.

The existence of these polymorphic forms led Herbert Spencer to argue that the form of an animal is not absolutely determined by those "Anlagen" or rudiments that Weismann and his school consider to be all important in determining the nature or form of the individual, for if this were the case, how can it be, he asked, that one egg may produce either a worker, nurse, soldier or female ant? To this Wasmann (who continued the discussion) replied by postulating the existence of double, triple or numerous rudiments in each egg, the treatment the egg receives merely determining which of these rudiments shall undergo development.[^[59]] Forel seems to have adopted this explanation as being the most simple. The probability of Weismann's hypothesis being correct is much diminished by the fact that the limit between the castes is by no means absolute. In many species intermediate forms are common, and even in those in which the castes are believed to be quite distinct, intermediate forms occur as very rare exceptions.[^[60]] Emery accounts[^[61]] for the polymorphism, without the assistance of the Weismannian hypothetical compound rudiments, by another set of assumptions; viz. that the phenomenon has been gradually acquired by numerous species, and that we see it in various stages of development; also that variation in nutrition does not affect all the parts of the body equally, but may be such as to carry on the development of certain portions of the organisation while that of other parts is arrested. Speaking broadly we may accept this view as consistent with what we know to be the case in other Insects, and with the phenomena of post-embryonic development in the class. But it must be admitted that our knowledge is at present quite inadequate to justify the formulation of any final conclusions.

The geological record of Formicidae is not quite what we should have expected. They are amongst the earliest Hymenoptera; remains referred to the family have been found in the Lias of Switzerland and in the English Purbecks. In Tertiary times Formicidae appear to have been about the most abundant of all Insects. At Florissant they occur in thousands and form in individuals about one-fourth of all the Insects found there. They have also been met with numerously in the European Tertiaries, and Mayr studied no less than 1500 specimens found in amber. Formicides and Myrmicides are more abundant than Ponerides, but this latter group has the larger proportion of extinct genera; conditions but little dissimilar to those existing at present.

Classification of Ants.—Ants are considered by many entomologists to form a series called Heterogyna. They can, however, be scarcely considered as more than a single family, Formicidae, so that the serial name is superfluous. Their nearest approach to other Aculeates is apparently made, by Amblyopone, to certain Mutillides (e.g. Apterogyna) and to the Thynnides, two divisions of Scoliidae. Emery considers Dorylides rather than Amblyoponides to be the most primitive form of ants, but we are disposed to consider Forel's view to the effect we have above mentioned as more probably correct. The point is, however, very doubtful. The condition of the peduncle is in both the sub-families we have mentioned very imperfect compared with that of other ants. Both these sub-families are of very small extent and very imperfectly known. We shall also follow Forel in adopting six sub-families, Camponotides, Dolichoderides, Myrmicides, Ponerides, Dorylides, and Amblyoponides. Emery rejects the Amblyoponides as being merely a division of the Ponerides. This latter group displays the widest relations of all the sub-families, and may be looked on as a sort of central form. The Camponotides and Dolichoderides are closely allied, and represent the highest differentiation of the families in one direction. The Myrmicides are also highly differentiated, but are not allied to the Camponotides and Dolichoderides.[^[62]]

Sub-Fam. 1. Camponotides.—Hind body furnished with but one constriction, so that only a single scale or node exists on the pedicel. Poison-sac forming a cushion of convolutions, on which is situate the modified sting, which forms merely an ejaculatory orifice for the poison.

The members of this very extensive division of ants can be readily distinguished from all others, except the Dolichoderides, by the absence of a true sting, and by the peculiar form of the hind body; this possesses only a single scale at the base, and has no constriction at all on the oval, convex and compact mass of the abdomen behind this. The cloacal orifice is circular, not, as in other ants, transverse. These characters are accompanied by a difference in habits. The Camponotides, though they do not sting, produce poison in large quantity, and eject it to some distance. Hence, if two specimens are confined in a tube they are apt to kill one another by the random discharges they make. Janet suggests that in order to neutralise the effect of this very acid poison, they may have some means of using, when they are in their natural abodes, the alkaline contents of a second gland with which they are provided. We shall mention the characters by which the Camponotides are distinguished from the small sub-family Dolichoderides when we deal with the latter.

The sub-family includes 800 or more species. Camponotus itself is one of the most numerous in species of all the genera of Formicidae, and is distributed over most parts of the earth. We have no species of it in Britain, but in the south of Europe the Camponotus become very conspicuous, and may be seen almost everywhere stalking about, after the fashion of our British wood-ant, Formica rufa, which in general appearance Camponotus much resembles.

Until recently, the manner in which fresh nests of ants were founded was unknown. In established nests the queen-ant is fed and tended by the workers, and the care of the helpless larvae and pupae also devolves entirely on the workers, so that the queens are relieved of all functions except that of producing eggs. It seemed therefore impossible that a fresh nest could be established by a single female ant unless she were assisted by workers. The mode in which nests are founded has, however, been recently demonstrated by the observations of Lubbock, M‘Cook, Adlerz, and more particularly by those of Blochmann, who was successful in observing the formation of new nests by Camponotus ligniperdus at Heidelberg. He found under stones in the spring many examples of females, either solitary or accompanied only by a few eggs, larvae or pupae. Further, he was successful in getting isolated females to commence nesting in confinement, and observed that the ant that afterwards becomes the queen, at first carries out by herself all the duties of the nest: beginning by making a small burrow, she lays some eggs, and when these hatch, feeds and tends the larvae and pupae; the first specimens of these latter that become perfect Insects are workers of all sizes, and at once undertake the duties of tending the young and feeding the mother, who, being thus freed from the duties of nursing and of providing food while she is herself tended and fed, becomes a true queen-ant. Thus it seems established that in the case of this species the division of labour found in the complex community, does not at first exist, but is correlative with increasing numbers of the society. Further observations as to the growth of one of these nascent communities, and the times and conditions under which the various forms of individuals composing a complete society first appear, would be of considerable interest.

An American species of the same genus, C. pennsylvanicus, the carpenter-ant, establishes its nests in the stumps of trees. Leidy observed that solitary females constructed for themselves cells in the wood and closed the entrances, and that each one in its solitary confinement reared a small brood of larvae. The first young produced in this case are said to be of the dwarf caste, and it was thought by the observer that the ant remained not only without assistance but also without food during a period of some weeks, and this although she was herself giving food to the larvae she was rearing.

Adlerz states that the females or young queens take no food while engaged in doing their early work, and that the large quantity of fat-body they possess enables them to undergo several months of hunger. In order to feed the young larvae they use their own eggs or even the younger larvae. It is to the small quantity of food rather than to its nature that he attributes the small size of the first brood of perfect workers. M. Janet[^[63]] has recently designed an ingenious and simple apparatus for keeping ants in captivity. In one of these he placed a solitary female of Lasius alienus, unaccompanied by any workers or other assistants, and he found at the end of 98 days that she was taking care of a progeny consisting of 50 eggs, 2 larvae, 5 pupae in cocoons, 5 without cocoons. On the 102nd day workers began to emerge from the cocoons.[^[64]] From these observations it is evident that the queen-ant, when she begins her nest, lives under conditions extremely different from those of the royal state she afterwards reaches.

In many kinds of ants the full-grown individuals are known to feed not only the larvae by disgorging food from their own mouths into those of the little grubs, but also to feed one another. This has been repeatedly observed, and Forel made the fact the subject of experiment in the case of Camponotus ligniperdus. He took some specimens and shut them up without food for several days, and thereafter supplied some of them with honey, stained with Prussian blue; being very hungry, they fed so greedily on this that in a few hours their hind bodies were distended to three times their previous size. He then took one of these gorged individuals and placed it amongst those that had not been fed. The replete ant was at once explored by the touches of the other ants and surrounded, and food was begged from it. It responded to the demands by feeding copiously a small specimen from its mouth: when this little one had received a good supply, it in turn communicated some thereof to other specimens, while the original well-fed one also supplied others, and thus the food was speedily distributed. This habit of receiving and giving food is of the greatest importance in the life-history of ants, and appears to be the basis of some of the associations that, as we shall subsequently see, are formed with ants by numerous other Insects.

Fig. 60—Oecophylla smaragdina. Worker using a larva for spinning.

Oecophylla smaragdina, a common ant in Eastern Asia, forms shelters on the leaves of trees by curling the edges of leaves and joining them together. In doing this it makes use of an expedient that would not be believed had it not been testified by several competent and independent witnesses. The perfect ant has no material with which to fasten together the edges it curls; its larva, however, possesses glands that secrete a supply of material for it to form a cocoon with, and the ants utilise the larvae to effect their purpose. Several of them combine to hold the foliage in the desired position, and while they do so, other ants come up, each one of which carries a larva in its jaws, applies the mouth of the larva to the parts where the cement is required, and makes it disgorge the sticky material. Our figure is taken from a specimen (for which we are indebted to Mr. E. E. Green) that was captured in the act of bearing a larva.

Formica rufa, the Red-ant, Wood-ant, or Hill-ant, is in this country one of the best-known members of the Formicidae. It frequents woods, especially such as are composed, in whole or part, of conifers, where it forms large mounds of small sticks, straws, portions of leaves, and similar material. Although at first sight such a nest may appear to be a chaotic agglomeration, yet examination reveals that it is arranged so as to leave many spaces, and is penetrated by galleries ramifying throughout its structure. These mound-nests attain a considerable size when the operations of the industrious creatures are not interfered with, or their work destroyed, as it too often is, by ignorant or mischievous persons. They may reach a height of three feet or near it, and a diameter of twice that extent. The galleries by which the heaps are penetrated lead down to the earth below. From the mounds extend in various directions paths constantly traversed by the indefatigable ants. M‘Cook observed such paths in the Trossachs; they proceed towards the objects aimed at in lines so straight that he considers they must be the result of some sense of direction possessed by the ants; as it is impossible to suppose they could perceive by the sense of sight the distant objects towards which the paths were directed: these objects in the case M‘Cook describes were oak-trees up which the ants ascended in search of Aphides.

M‘Cook further observed that one of the oak-trees was reached by individuals from another nest, and that each of the two parties was limited to its own side of the tree, sentinels being placed on the limits to prevent the trespassing of an intruder; he also noticed that the ants saw an object when the distance became reduced to about an inch and a half from them. This species is considered to be wanting in individual courage; but when acting in combination of vast numbers it does so with intelligence and success. It does not make slaves, but it has been observed by Bignell and others that it sometimes recruits its numbers by kidnapping individuals from other colonies of its own species. Its nests are inhabited by forty or fifty species of guests of various kinds, but chiefly Insects. Another ant, Myrmica laevinodis, sometimes lives with it in perfect harmony, and Formicoxenus nitidulus lives only with F. rufa. Amongst the most peculiar of its dependants we may mention large beetles of the genera Cetonia and Clythra, which in their larval state live in the hills of the wood-ant. It is probable that they subsist on some of the vegetable matter of which the mounds are formed. Adlerz has given some attention to the division of labour amongst the different forms of the workers of ants, and says that in F. rufa it is only the bigger workers that carry building and other materials, the smaller individuals being specially occupied in the discovery of honey-dew and other Aphid products. In Camponotus it would appear, on the other hand, that the big individuals leave the heavy work to be performed by their smaller fellows.

The wood-ant and its near allies have been, and indeed still are, a source of great difficulty to systematists on account of the variation that occurs in the same species, and because this differs according to locality. Our European F. rufa has been supposed to inhabit North America, and the interesting accounts published by M‘Cook of the mound-making ant of the Alleghanies were considered to refer to it. This Insect, however, is not F. rufa, as was supposed by M‘Cook, but F. exsectoides, Forel. It forms colonies of enormous extent, and including an almost incredible number of individuals. In one district of about fifty acres there was an Ant City containing no less than 1700 of these large ant-hills, each one teeming with life. It was found by transferring ants from one hill to another that no hostility whatever existed between the denizens of different hills; the specimens placed on a strange hill entered it without the least hesitation. Its habits differ in some particulars from those of its European congener; the North American Insect does not close the formicary at night, and the inquilines found in its nest are very different from those that live with F. rufa in Europe. Whether the typical wood-ant occurs in North America is doubtful, but there are races there that doubtless belong to the species.

F. sanguinea is very similar in appearance to its commoner congener F. rufa, and is the only slave-making ant we possess in Britain. This species constructs its galleries in banks, and is of very courageous character, carrying out its military operations with much tactical ability. It is perfectly able to live without the assistance of slaves, and very frequently does so; indeed it has been asserted that it is in our own islands (where, however, it is comparatively rare) less of a slave-owner than it is in Southern Europe, but this conclusion is very doubtful. It appears when fighting to be rather desirous of conquering its opponents by inspiring terror and making them aware of its superiority than by killing them; having gained a victory it will carry off the pupae from the nest it has conquered to its own abode, and the ants of the stranger-species that develop from these pupae serve the conquerors faithfully, and relieve them of much of their domestic duties. The species that F. sanguinea utilises in this way in England are F. fusca, F. cunicularia, and possibly Lasius flavus. Huber and Forel have given graphic accounts of the expeditions of this soldier-ant. In the mixed colonies of F. sanguinea and F. fusca the slaves do most of the house-work, and are more skilful at it than their masters. Adlerz says that one of the slaves will accomplish twice as much work of excavation in the same time as the slave-owner; these latter being lazy and fond of enjoyment, while the slaves are very industrious.

Fig. 61—Head of Polyergus rufescens. (After André.)

Polyergus rufescens, an European ant allied to Formica, is renowned since the time of Huber (1810) as the slave-making or Amazon ant. This creature is absolutely dependent on its auxiliaries for its existence, and will starve, it is said, in the midst of food unless its servitors are there to feed it. Wasmann, however, states that Polyergus does possess the power of feeding itself to a certain extent. Be this as it may, the qualities of this ant as warrior are superb. When an individual is fighting alone its audacity is splendid, and it will yield to no superiority of numbers; when the creatures are acting as part of an army the individual boldness gives place to courage of a more suitable sort, the ants then exhibiting the act of retreating or making flank movements when necessary. If a Polyergus that is acting as a member of a troop finds itself isolated, and in danger of being overpowered, it has then no hesitation in seeking safety even by flight. This species is provided with mandibles of a peculiar nature; they are not armed with teeth, but are pointed and curved; they are therefore used after the manner of poignards, and when the ant attacks a foe it seizes the head between the points of these curved mandibles, and driving them with great force into the brain instantly paralyses the victim.

Mandibles of this shape are evidently unfitted for the purposes of general work, they can neither cut, crush, nor saw, and it is not impossible that in their peculiar shape is to be found the origin of the peculiar life of Polyergus: we find similar mandibles reappearing amongst the aberrant Dorylides, and attaining a maximum of development in the ferocious Eciton; they also occur, or something like them, in a few aberrant Myrmicides; and in the male sex of many other ants, which sex exercises no industrial arts, this sort of mandible is present.

The ants that Polyergus usually attacks in order to procure slaves are Formica fusca and F. fusca, race auricularia; after it has routed a colony of one of these species, P. rufescens pillages the nest and carries off pupae and some of the larger larvae to its own abode. When the captives thus deported assume the imago state, they are said to commence working just as if they were in their own houses among their brothers and sisters, and they tend their captors as faithfully as if these were their own relatives: possibly they do not recognise that they are in unnatural conditions, and may be quite as happy as if they had never been enslaved. The servitors are by no means deficient in courage, and if the place of their enforced abode should be attacked by other ant-enemies they defend it bravely. The fact that P. rufescens does not feed its larvae has been considered evidence of moral degeneration, but it is quite possible that the Insect may be unable to do so on account of some deficiency in the mouth-parts, or other similar cause. The larvae of ants are fed by nutriment regurgitated from the crop of a worker (or female), and applied to the excessively minute mouth of the helpless grub: for so delicate a process to be successfully accomplished, it is evident that a highly elaborated and specialised arrangement of the mouth-parts must exist, and it is by no means improbable that the capacity of feeding its young in true ant-fashion is absent in Polyergus for purely mechanical reasons.

M‘Cook states that the North American ant, Polyergus lucidus, which some entomologists consider to be merely a variety of the European species, makes slaves of Formica schaufussi, itself does no work, and partakes of food only when fed by its servitors. He did not, however, actually witness the process of feeding. When a migration takes place the servitors deport both the males and females of P. lucidus. M‘Cook adds that the servitors appear to be really mistresses of the situation, though they avail themselves of their power only by working for the advantage of the other species.

Fig. 62—Myrmecocystus mexicanus. Honey-pot ant, dorsal view.

Fig. 63—Myrmecocystus mexicanus. Lateral view.

The honey-ant of the United States and Mexico has been investigated by M‘Cook and others; the chief peculiarity of the species is that certain individuals are charged with a sort of honey till they become enormously distended, and in fact serve as leather bottles for the storage of the fluid. The species Myrmecocystus hortideorum and M. melliger, are moderate-sized Insects of subterranean habits, the entrance to the nest of M. hortideorum being placed in a small raised mound. The honey is the product of a small gall found on oak leaves, and is obtained by the worker-ants during nocturnal expeditions, from which they return much distended; they feed such workers left at home as may be hungry, and then apparently communicate the remainder of the sweet stuff they have brought back to already partly charged "honey-bearers" left in the nest. The details of the process have not been observed, but the result is that the abdomens of the bearers become distended to an enormous extent (Figs. 62, 63), and the creatures move but little, and remain suspended to the roof of a special chamber. It is considered by M‘Cook that these living honey-tubs preserve the food till a time when it is required for the purposes of feeding the community. The distension is produced entirely by the overcharging of the honey-crop, the other contents of the abdomen being forced by the distention to the posterior part of the body. Lubbock has since described an Australian ant, Melophorus inflatus, having a similar peculiar habit, but belonging to the allied tribe Plagiolepisii. Quite recently a South African honey-tub ant belonging to the distinct genus Plagiolepis (Ptrimeni For.) has been discovered, affording a proof that an extremely specialised habit may arise independently of relation between the Insects, and in very different parts of the world.

Species of the genus Lasius are amongst the most abundant of the ant-tribe in Britain. They are remarkable for their constructive powers. L. niger, the common little black garden-ant, forms extensive subterranean galleries, and is extremely successful in the cultivation of various forms of Aphidae, from the products of which the species derives a large part of its subsistence. The ants even transport the Aphidae to suitable situations, and thus increase their stock of this sugary kind of cattle, and are said to take the eggs into their own dwellings in the autumn so that these minute and fragile objects may be kept safe from the storms and rigours of winter. These little creatures are brave, but when attacked by other ants they defend themselves chiefly by staying in their extensive subterranean galleries, and blocking up and securing these against their assailants.

L. fuliginosus, another of our British species, has very different habits, preferring old trees and stumps for its habitation; in the hollows of these it forms dwellings of a sort of card; this it makes from the mixture of the secretions of its salivary glands with comminuted fragments of wood, after the fashion of wasps. It is a moderate-sized ant, much larger than the little L. niger, and is of a black colour and remarkably shining; it gives off a very strong but by no means disagreeable odour, and may be seen on the hollow trees it frequents, stalking about in large numbers in a slow and aimless manner that contrasts strikingly with the active, bustling movements of so many of its congeners. When this species finds suitable trees near one another, a colony is established in each; the number of individuals thus associated becomes very large, and as the different colonies keep up intercommunication, this habit is very useful for purposes of defence. Forel relates that he once brought a very large number of Formica pratensis and liberated them at the base of a tree in which was a nest of L. fuliginosus; these latter, finding themselves thus assaulted and besieged, communicated in some way, information of the fact to the neighbouring colonies, and Forel soon saw large columns of the black creatures issuing from the trees near by and coming with their measured paces to the assistance of their confrères, so that the invaders were soon discomfited and destroyed. Although the European and North American representatives of the sub-family Camponotides live together in assemblies comprising as a rule a great number of individuals, and although the separate nests or formicaries which have their origin from the natural increase of a single original nest keep up by some means a connection between the members, and so form a colony of nests whose inhabitants live together on amicable terms, yet there is no definite information as to how long such association lasts, as to what is the nature of the ties that connect the members of the different nests, nor as to the means by which the colonies become dissociated. It is known that individual nests last a long time. Charles Darwin has mentioned in a letter to Forel that an old man of eighty told him he had noticed one very large nest of Formica rufa in the same place ever since he was a boy. But what period they usually endure is not known, and all these points probably vary greatly according to the species concerned. It has been well ascertained that when some ants find their nests, for some unknown reason, to be unsuitable the inhabitants leave their abodes, carrying with them their young and immature forms, and being accompanied or followed by the various parasites or commensals that are living with them. Wasmann and other entomologists have observed that the ants carry bodily some of their favourite beetle-companions, as well as members of their own species. Forel observed that after a nest of Formica pratensis had been separated into two nests placed at a considerable distance from one another so as to have no intercommunication, the members yet recognised one another as parts of the same family after the lapse of more than a month; but another observation showed that after some years of separation they were no longer so recognised.

Although it is now well ascertained that ants are able to distinguish the individuals belonging to their own nests and colonies from those that, though of their own species, are not so related to them, yet it is not known by what means the recognition is effected; there is, however, some reason to suppose that it is by something of the nature of odour. One observer has noticed that if an ant fall into water it is on emerging at first treated as if it were a stranger by its own friends; but other naturalists have found this not to be the case in other species. Contact with corrosive sublimate deprives ants for a time of this power of recognising friends, and under its influence they attack one another in the most ferocious mariner.

The nests and colonies of the species of Camponotides we have considered are all constructed by societies comprising a great number of individuals; there are, however, in the tropics numerous species that form their nests on foliage, and some of these contain only a few individuals. The leaf-nests (Fig. 64) of certain species of Polyrhachis are said to be formed of a paper-like material, and to contain each a female and about 8 or 10 worker ants. Forel[^[65]] has examined nests of several Indian species, and finds they differ from those of other ants in consisting of a single cavity, lined with silk like that of a spider. These nests are further said to be constructed so as to render them either inconspicuous or like other objects on the leaves; P. argentea covers its small dwelling with little bits of vegetable matter, and a nest of P. rastella was placed between two leaves in such a manner as to be entirely hidden. All the species of the genus do not, however, share these habits, P. mayri making a card-nest, like Dolichoderus and some other ants. The species of the genus Polyrhachis are numerous in the tropics of the Old World.

Fig. 64—Nest of Polyrhachis sp. (After Smith.)

Forbes noticed that a species of this genus, that makes its paper-like nest on the underside of bamboo-leaves produces a noise by striking the leaf with its head in a series of spasmodic taps. The same observer has recorded a still more interesting fact in the case of another species of this genus—a large brown ant—found in Sumatra. The individuals were "spread over a space, perhaps a couple of yards in diameter, on the stem, leaves, and branches of a great tree which had fallen, and not within sight of each other; yet the tapping was set up at the same moment, continued exactly the same space of time, and stopped at the same instant; after the lapse of a few seconds all recommenced at the same instant. The interval was always of about the same duration, though I did not time it; each ant did not, however, beat synchronously with every other in the congeries nearest to me; there were independent tappings, so that a sort of tune was played, each congeries dotting out its own music, yet the beginnings and endings of the musical parties were strictly synchronous."

Fig. 65—Polyrhachis pandurus, worker. Singapore.

Mr. Peal has also recorded that an ant—the name is not mentioned, but it may be presumed to be an Assamese species—makes a concerted noise loud enough to be heard by a human being at twenty or thirty feet distance, the sound being produced by each ant scraping the horny apex of the abdomen three times in rapid succession on the dry, crisp leaves of which the nest is usually composed. These records suggest that these foliage-ants keep up a connection between the members of different nests somewhat after the same fashion as do so many of the terrestrial Camponotides. Although the species of Camponotides have no special organ for the production of sound in the position in which one is found in Myrmicides and Ponerides, yet it is probable that they are able to produce a sound by rubbing together other parts of the abdomen.

Sub-Fam. 2. Dolichoderides.—Hind body furnished with but one constriction so that only a single scale or node is formed; Sting rudimentary; the poison-sac without cushion.

Fig. 66—Tapinoma erraticum, worker. Britain. Upper side and profile.

The Dolichoderides are similar to Camponotides in appearance, and are distinguished chiefly by the structure of the sting and the poison apparatus. To this we may add that Forel also considers the gizzard to be different in the two sub-families, there being no visible calyx in the Dolichoderides, while this part is largely developed in the Camponotides. This is one of the least extensive of the sub-families of ants, not more than 150 species being yet discovered. Comparatively little is known of the natural history of its members, only a very small number of species of Dolichoderides being found in Europe. The best known of these (and the only British Dolichoderid) is Tapinoma erraticum, a little ant of about the size of Lasius niger, and somewhat similar in appearance, but very different in its habits. T. erraticum does not cultivate or appreciate Aphides, but is chiefly carnivorous in its tastes. Our knowledge of it is due to Forel, who has noticed that it is very fond of attending the fights between other ants. Here it plays the part of an interested spectator, and watching its opportunity drags off the dead body of one of the combatants in order to use it as food. Although destitute of all power of stinging, this Insect has a very useful means of defence in the anal glands with which it is provided; these secrete a fluid having a strong characteristic odour, and possessing apparently very noxious qualities when applied to other ants. The Tapinoma has no power of ejecting the fluid to a distance, but is very skilful in placing this odorous matter on the body of an opponent by touching the latter with the tip of the abdomen; on this being done its adversary is usually discomfited. This Insect is subterranean in its habits, and is said to change its abode very frequently. T. erraticum occurs somewhat rarely in Britain. Forel has also noted the habits of Liometopum microcephalum, another small European species of Dolichoderides. It is a tree-ant, and by preference adopts, and adapts for its use, the burrows made by wood-boring beetles. It forms extremely populous colonies which may extend over several large trees, the inhabitants keeping up intercommunication by means of numerous workers. No less than twelve mighty oaks were found to be thus united into a colony of this ant in one of the Bulgarian forests. The species is very warlike, and compensates for the extreme minuteness of its individuals by the skilful and rapid rushes made by combined numbers on their ant-foes of larger size.

Fritz Müller has given a brief account, under the name of the Imbauba ant, of a Brazilian arboreal ant, that forms small nests in the interior of plants. The species referred to is no doubt an Azteca, and either A. instabilis, or A. mulleri. The nests are founded by fertilised females which may frequently be found in the cells on young Cecropia plants. Each internode, he says, has on the outside, near its upper part, a small pit where the wall is much thinner, and in this the female makes a hole by which she enters. Soon afterwards the hole is completely closed by a luxuriant excrescence from its margins, and it remains thus closed until about a dozen workers have developed from the eggs of the female, when the hole is opened anew from within by the workers. It is said that many of the larvae of these ants are devoured by the grubs of a parasite of the family Chalcididae. This Insect is thought to protect the plant from the attacks of leaf-cutting ants of the genus Atta.

We may here briefly remark that much has been written about the benefits conferred on plants by the protection given to them in various ways by ants: but there is reason to suppose that a critical view of the subject will not support the idea of the association being of supreme importance to the trees.[^[66]]

Sub-Fam. 3. Myrmicides.—Pedicel of abdomen formed of two well-marked nodes (knot-like segments). Sting present (absent in the Cryptocerini and Attini). (It should be noted that the workers of the genera Eciton and Aenictus of the subfamily Dorylides have, like the Myrmicides, two nodes in the pedicel.)

This sub-family consists of about 1000 species, and includes a great variety of forms, but, as they are most of them of small size, they are less known than the Camponotides, and much less attention has been paid to their habits and intelligence. Forel, until recently, adopted four groups: Myrmicini, Attini, Pseudomyrmini and Cryptocerini; but he is now disposed to increase this number to eight.[^[67]] They are distinguished by differences in the clypeus, and in the form of the head; but it must be noted that the characters by which the groups are defined are not in all cases fully applicable to the males. The Cryptocerini are in external structure the most highly modified of Hymenoptera, if not of all the tribes of Insecta.

Fig. 67—Pheidologeton laboriosus, large and small workers. East India.

i. The Myrmicini proper are defined by Forel as having the antennae inserted near the middle, a little behind the front, of the head, which has carinae on the inner sides, but none on the outer sides, of the insertions of the antennae; the clypeus extends between the antennae.

Fig. 68—Formicoxenus nitidulus, male. (After Adlerz.)

Certain genera of small European ants of the group Myrmicini display some most anomalous phenomena. This is especially the case in Formicoxenus, Anergates and Tomognathus. The facts known have, however, been most of them only recently discovered, and some obscurity still exists as to many of even the more important points in these extraordinary life-histories.

Fig. 69.—Anergates atratulus. Europe. A, male, with part of hind leg broken off; B, female, with wings: C, female, after casting the wings and becoming a queen.

It has long been known that the little Formicoxenus nitidulus lives as a guest in the nests of Formica rufa, the wood-ant; and another similar ant, Stenamma westwoodi, which shares the same life, was declared by Nylander and Smith to be its male; it was however shown some years ago by André that this is a mistake, and that S. westwoodi is really the male of another ant that had till then been called Asemorhoptrum lippulum. This correction left the workers and females of Formicoxenus nitidulus destitute of a male, but Adlerz has recently discovered that the male of this species is wingless and similar to the worker, the female being a winged Insect as usual. It is very curious that the characters by which the male is distinguished from the worker should vary in this species; but according to Adlerz this is the case, individuals intermediate in several points between the males and workers having been discovered. This phenomenon of quite wingless males in species where the female is winged is most exceptional, and is extremely rare in Insects; but it occurs, as we shall see, in one or two other Myrmicides. Charles Darwin made the very reasonable suggestion that winged males may be developed occasionally as an exceptional phenomenon, and it is very probable that this may be the case, though it has not yet been demonstrated. Formicoxenus nitidulus occurs in England in the nests of Formica rufa and of F. congerens, but we are not aware that the male has ever been found in this country. The genus Anergates is allied to Formicoxenus, and occurs in Central Europe, but has not been found in Britain; the female, as in Formicoxenus, is winged and the male wingless, but there is no worker-caste; the male is a rather helpless creature, and incapable of leaving the nest. The species lives in company with Tetramorium caespitum a little ant very like Myrmica, and not uncommon in South-East England. The female Anergates is at first an active little creature with wings, but after these are lost the body of the Insect becomes extremely distended as shown in Fig. 69, C; the creature is in this state entirely helpless, and as there are no workers, the Anergates is completely dependent, for the existence of itself and its larvae, on the friendly offices of the Tetramorium that lives with it. The mode of the association of these two Insects is at present both unparalleled and inexplicable, for only workers of the Tetramorium are found in company with the ♂ and ♀ Anergates; the community, in fact, consisting of males and females of one species and workers of another. The nests of Anergates are so rare that only a few naturalists have been able to observe them (Schenk, von Hagens, and Forel may be specially mentioned), but in the spots where they occur, nests of the Tetramorium, containing all the forms of that species, are numerous, and it therefore seems probable that a young fertile female of the Anergates may leave a nest in which it was born, enter a nest of the Tetramorium, and, destroying the queen thereof, substitute herself in the place of the victim; but if this be really the case, the larvae and pupae of the Tetramorium must also be destroyed, for no young of the Tetramorium are ever found in these strange associations. It is very difficult to believe that the Tetramorium workers should be willing to accept as their queen a creature that commenced her acquaintance with them by destroying their own queen or queens and a number of their young sisters; especially as the Tetramorium is a more powerful ant than the Anergates, and could readily dispose of the murderous intruder if it were disposed to do so. It is known, however, that colonies of Tetramorium completely destitute of queens sometimes occur, and Wasmann has suggested that the female Anergates may seek out one of these, and installing herself therein as a substitute, may be accepted by the orphaned colony. This plausible hypothesis has still to be verified.

The genus Cardiocondyla also exhibits the phenomenon of apterous, worker-like males, while in one species, C. emeryi, a winged male is also known to exist.

Tomognathus sublaevis is a little Myrmicid ant, found rarely in Denmark and Sweden, where its habits have recently been studied by Adlerz. A band of the Tomognathus attack the nest of another little Myrmicid, Leptothorax acervorum, and succeed by their own pertinacity and the fears of the Leptothorax in obtaining possession of it; the legitimate owners disappear, leaving the Tomognathus in possession of their larvae and pupae; these complete their development only to find themselves the slaves of Tomognathus. The subsequent relations of the two ants are friendly, the slaves even preventing their masters from wandering from the nest when they wish to do so. If an established mixed community of this nature is in want of additional servitors, the Tomognathus secure a supply by raids after the fashion of the Amazon-ant, bringing back to their abode larvae and pupae of Leptothorax to be developed as slaves. It was formerly supposed that the Tomognathus continued its species by perpetual parthenogenesis of the workers, for neither males nor females could be found. Adlerz[^[68]] has, however, now discovered the sexual individuals. The male is an ordinary winged ant, and is so like that sex of the Leptothorax, that Adlerz had failed to distinguish the two before he reared them. The females are apterous, and in fact like the workers. It would perhaps be more correct to say that the workers of this species vary greatly but never become winged; some of them have ocelli and a structure of the thorax more or less similar to that of winged females, though none have been found with wings. Certain of these females possess a receptaculum seminis, and Adlerz treats this as the true distinction between female and worker. In accordance with this view the female of Tomognathus may be described as a worker-like individual possessing a receptaculum seminis, and having more or less of the external structures of winged females, though never being actually winged. It is probable that other workers reproduce parthenogenetically. The males of this species will not unite with females from the same nest, thus differing from many other ants, in which union between individuals of the same nest is the rule. Finally, to complete this curious history, we should remark that the larvae of the Tomognathus are so similar to those of the Leptothorax that Adlerz is quite unable to distinguish the two.

Strongylognathus testaceus and S. huberi live in association with Tetramorium caespitum, and are cared for by these latter ants; it is notable that as in the case of the slave-making Polyergus rufescens the mandibles of the Strongylognathus are cylindrical and pointed, and therefore unsuitable for industrial occupations. S. testaceus is a weak little ant, and lives in small numbers in the nests with T. caespitum, which it is said to greatly resemble in appearance. The proportions of the forms of the two species usually associated is peculiar, there being a great many workers of T. caespitum both in the perfect and pupal states, and also all the sexes of the Strongylognathus, of which, however, only a few are workers. This would seem to suggest that S. testaceus attacks and pillages the nests of T. caespitum in order to carry off worker pupae, just as Polyergus rufescens does. But the facts that S. testaceus is a weaker Insect than the Tetramorium, and that only a few of its worker-caste are present in a community where there are many workers of the Tetramorium, seem to negative the view that the latter were captured by the former; and the mode in which the associated communities of these two species are started and kept up is still therefore in need of explanation.

Strongylognathus huberi is a much stronger Insect than its congener, S. testaceus, and Forel has witnessed its attack on Tetramorium caespitum. Here the raid is made in a similar manner to that of Polyergus rufescens on Formica; the Tetramorium is attacked, and its pupae carried off to the abode of the Strongylognathus to serve in due time as its slaves. The mandibles of S. huberi, being similar in form to those of Polyergus rufescens are used in a similar manner.

Although T. caespitum is common enough in South-East England, it is to be regretted that none of the guests or associates we have mentioned in connection with it occur in this country. It is a most variable species, and is distributed over a large part of the globe.

Our British species of Myrmicides, about ten in number, all belong to the group Myrmicini; none of them are generally common except Myrmica rubra, which is a most abundant Insect, and forms numerous races that have been considered by some entomologists to be distinct species; the two most abundant of these races are M. ruginodis and M. scabrinodis, which sometimes, at the time of the appearance of the winged individuals, form vast swarms.

The tiny Monomorium pharaonis is a species that has been introduced into Britain, but now occurs in houses in certain towns; it sometimes accumulates on provisions in such numbers as to be a serious nuisance. Seventeen thousand individuals weigh 1 gramme, and it is probable that a nest may include millions of specimens.

The genus Aphaenogaster[^[69]] and its immediate allies include the harvesting ants of Europe and North America: they form subterranean nests consisting of isolated chambers connected by galleries; some of the chambers are used as store-houses or granaries, considerable quantities of corn, grass, and other seeds being placed in them. A. structor and A. barbarus have been observed to do this in Southern Europe, by Lespès, Moggridge, and others.

Fig. 70—Aphaenogaster (Messor) barbarus. Algeria. A, male; B, winged female; C, large worker or soldier; D, small worker. × 3⁄2.

In the deserts about Algeria and Tunis a harvesting ant, Aphaenogaster (Messor) arenarius, is an important creature: its subterranean dwellings are very extensive, and are placed at a depth of several feet from the surface. Entrance to these dwellings is obtained by small holes, which are the orifices of galleries many feet in length: the holes are surrounded by pellets of sand projecting somewhat above the general surface, and consequently making the places conspicuous. The subterranean works occupy an area of fifty or a hundred square yards excavated at a depth of three to six feet. In these immense nests there exists a form of worker, of very small size, that never comes to the surface.[^[70]]

Pogonomyrmex barbatus and other species have been observed to do harvesting in North America. After the workers of P. barbatus have taken the seeds into the nest they separate the husks and carry them out, depositing them on a heap or kitchen-midden, formed near by. M‘Cook has witnessed and described the process of stripping the seeds.

Certain genera—e.g. Aphaenogaster, Pheidole—exhibit great disparity in the forms of the workers, some of which are of size much superior to the others, and possess disproportionately large heads; these large individuals are found in the same nest as the smaller forms. All the intermediate forms may frequently be found, and at the same time, in the genus Aphaenogaster; but in Pheidole intermediates are of the utmost rarity.

The genus Cremastogaster is remarkable on account of the shape of the hind body and its articulation, which give the abdomen the appearance of being put on upside down. This mode of articulation may allow the Insect to threaten its enemies when they are in front of it; but it is doubtful whether the Cremastogaster possesses an effective sting.

Fig. 71—Cremastogaster tricolor, worker. A, with abdomen extended; B, uplifted.

ii. The group Attini is distinguished by the presence of a carina near the eye, by the antennae being inserted at a moderate distance from one another, by the clypeus being prolonged backwards between them; and by the absence of a sting. The group is not represented in Europe, but in Tropical America the ants belonging to it are amongst the most important of natural objects. The species of the genus Atta (usually styled Oecodoma) are the formidable leaf-cutting ants of America. They occur in enormous colonies in certain places, and will in a short time completely strip a tree of its leaves. As they appear to prefer trees of a useful kind, especially those planted by man, their ravages are often of the most serious nature. The natives, feeling it hopeless to contend with these Insect hordes, only too frequently abandon all attempts to cultivate the trees and vegetables the Insects are fond of. Both Bates and Belt have given accounts of some points in the economy of these ants. They are amongst the largest of the Formicidae, the females in some cases measuring about two and a half inches across their expanded wings; the males are much smaller, but are less dissimilar to their partners than is usual among ants. The workers, on the other hand, are so extremely different, that no one would suppose them to be at all related to the males and females (see vol. v. Fig. 339).

The mode of operation of these ants is to form paths from their formicary extending for a considerable distance in various directions, so that they have a ready access to any spot in a district of considerable extent; when a tree or bush is found bearing leaves suitable for their purposes, the worker ants ascend it in large numbers and cut up the leaves by biting out of them pieces similar in size and shape to a small coin; these pieces are then carried back in the jaws of the ants to their nests; the ant-paths are therefore constantly traversed by bands of little creatures carrying burdens homewards, or hurrying outwards in search of suitable trees.

The formicaries are of considerable size, and are described as consisting of low mounds of bare earth of considerable extent. Bates speaks of as great a circumference as forty yards; these accumulations of earth have frequently an appearance different from the adjoining soil, owing to their being formed of subsoil brought up from below; they are kept bare by the ants constantly bringing to and depositing on the surface fresh material resulting from their subterranean excavations. The true abodes, beneath the earth, are of greater extent than the mounds themselves, and extend to a considerable depth; they consist of chambers connected by galleries.

The leaf-cutting ants extend their range to North America, and M‘Cook has recently called attention to a case there in which A. fervens made an underground route at an average depth of 18 inches, and at an occasional depth of 6 feet, extending 448 feet entirely beneath the earth, after which it was continued for 185 feet to reach a tree which the ants were engaged in defoliating. This route, extending altogether to a length of more than 600 feet, presented only a very slight deviation from a straight line drawn between the point of departure and the object to be attained. By what sense this ant was enabled to make a subterranean tunnel in a straight line to a desired object situated at so great a distance, we know not.

The use the leaf-cutting ants make of the enormous amount of material they gather was for long a subject of debate, and has only recently been ascertained by the observations of Möller. After being carried to the nest the pieces of leaves are cut into small fragments by another set of workers and formed into balls, which are packed in various parts of the nest, and amongst which the mycelium of a fungus—Rozites gongylophora—ramifies. This fungus the ants cultivate in the most skilful manner: they manage to keep it clear from mouldiness and bacterial agents, and to make it produce a modified form of growth in the shape of little white masses, each one formed by an agglomeration of swellings of the mycelium. These form the chief food of the colony. Möller ascertained by experiment that the results were due to a true cultivation on the part of the ants: when they were taken away from the nests, the mycelium produced two kinds of conidia instead of the ant-food.

Many details of the economy of these leaf-cutting ants are still very imperfectly known. The large-headed forms, called soldiers, have been the subject of contradictory statements; Bates having concluded from his own observations that they are harmless, while Mr. J. H. Hart assures us that they are very fierce and vindictive, and inflict very serious wounds by biting (the Attini do not sting). We anticipate that the observations of both these naturalists will prove to be substantially correct, and that the differences in habits will be found to be owing to distinctions in the conditions of the community. In connection with this point we may remark that the function of the excessively large heads of certain kinds of soldier-ants is still obscure. In the East Indian Pheidologeton diversus the big soldiers are quite one hundred times as large as the smaller workers. As these latter bite viciously it would naturally be supposed that their gigantic confrères with enormous heads would be warriors of a most formidable nature; but, as a matter of fact, the giants are unable to bite even when they try to do so. Aitken has somewhere suggested that these enormous individuals play the part of state elephants; and we have been informed by Colonel Bingham that the small ants may frequently be seen riding in numbers on their unwieldy fellows. We anticipate however, that some other function will be found to exist for these forms with enormous heads. An examination of their organs of sense and of voice is very desirable.

Details of the modes in which the great communities of the leaf-cutting Attidae are maintained, are still wanting. The females do not, we have been informed by Mr. Hart, possess any considerable powers of aftergrowth, so that there is no reason to suppose them to be unusually prolific. At certain seasons great swarms of winged individuals are produced, and after leaving the nests pair in the manner of our European Myrmica. Possibly the females may, after losing their wings, again enter the large communities. Von Ihering states that the workers of Atta lundi are fertile.

iii. The group Pseudomyrmini includes the genera Pseudomyrma and Sima, which are by some entomologists treated as but a single genus. The antennae are inserted near together on the front of the head; there is no carina on the head external to their insertion, and the clypeus does not extend forwards between them. The Insects are usually of elongate form, possess a sting, and have a naked pupa. The group occurs in both hemispheres, but is exclusively exotic, and but little is known of the habits of its members. Forel has recently observed that numerous species live inside dried stems of grass or in hollow twigs, and are beautifully adapted for this mode of life by their elongate form, some of them being as slender as needles. Some interesting observations have been made in Nicaragua by Belt on Pseudomyrma bicolor and its relations with an acacia-tree, in the thorns of which it lives. The acacia in question is called the bull's-horn thorn, because the branches and trunk are armed with strong curved spines, set in pairs, and much resembling the horns of the quadruped whose name they bear. The ant takes possession of a thorn by boring a small hole near the distal extremity, and forms its nest inside. The leaves of this plant are provided with glands that secrete a honey-like fluid, which it appears forms the chief, if not the sole, subsistence of the ant. Belt considers that the presence of the ant is beneficial to the acacia; he supposes that the ants assume the rights of proprietors, and will not allow caterpillars or leaf-cutting ants to meddle with their property; the leaves are, he thinks, so preserved to the benefit of the tree.

Fig. 72—Sima rufo-nigra and its associates. A, winged female; B, worker, of the ant; C, Rhinopsis ruficornis; a fossorial wasp of the sub-family Ampulicides; D, a spider, Salticus sp. The coloration is extremely similar in all these creatures.

Rothney has given some particulars of the habits of Sima rufo-nigra, an ant of this group that appears to be not uncommon near Calcutta, where it lives on the trunks of trees in company with a spider and a wasp that greatly resemble it in form and in colour. The three creatures seem to associate together on amicable terms; indeed the wasp and the ant occasionally indulge in wrestling matches without doing one another any serious harm. In connection with this fact we may observe that other species of ants have been observed to indulge in sports and feats of agility.

S. leviceps, an Australian species of this genus, is furnished with a stridulating file that has the appearance of being constructed so as to produce two very different kinds of sounds.

Fig. 73—Stridulating file of Sima leviceps.

iv. The Cryptocerini are distinguished from other ants by their antennae being inserted at the sides of the head, where they are placed between ridges or in a groove into which they can be withdrawn; when in some cases they are entirely concealed. These ants assume a great variety of shapes and forms, some of which look almost as if they were the results of an extravagant imagination. The skeleton is usually much harder than it is in other ants; the abdomen consists almost entirely of one very large segment, there being, however, three others visible at its extremity; these segments can be only slightly protruded, and the ants have no power of stinging. They are probably most of them arboreal in their habits. Nearly all of the known forms are exotic. According to the observations of Bates the species of the genus Cryptocerus in the Amazons Valley may frequently be observed in dry open places on low trees and bushes, or running on branches of newly felled trees; they also visit flowers abundantly. The species generally are wood-borers, usually perforating the dead branches of trees. C. atratus has been observed to construct its nests in the dead, suspended branches of woody climbers; a number of neatly drilled holes are all that can be seen externally; but, inside, the wood is freely perforated with intercommunicating galleries. Each community appears to consist of a single female and two kinds of workers; the latter in some species are quite unlike each other, differing in the form of the head, and in the armature of the thorax and nodes of the peduncle. The species of Cryptocerus appear to be omnivorous, and are frequently attracted by the excrement of birds. The pupae are not enclosed in a cocoon. In the South of Europe two very minute ants, of the genera Strumigenys and Epitritus, belonging to this family, are met with under very large stones partly embedded in the earth. They are of the greatest rarity.

Fig. 74—Cyrptocerus atratus, worker. Amazons. The compressed first joint of the hind foot is shown at a and b in different positions.

Sub-fam. 4. Ponerides.—Hind body elongate, furnished with one node at the base, and having also great capacity of movement between the first and second segments, between which there is usually a slight constriction. Sting well developed.

This sub-family includes numerous genera and about 400 species. The Ponerides have an elongate hind-body; the second segment behind the node is capable of great movement in and out of the preceding segment, and for this purpose is furnished with a basal portion slightly more slender than the apical part; this basal part is usually concealed within the more anterior segment, the hind margin of which embraces it very closely. On the middle of the dorsal aspect of this articulation there is usually placed a stridulating organ, consisting of an elongate band or patch of very fine lines; this gives out a sound when the second segment is moved in and out of the first at a time when the posterior edge of the latter is slightly depressed.

We follow Forel in including the Australian bull-dog ants—Myrmecia—in Ponerides, as well as the Odontomachi. The former have, however, a definite pedicel, consisting of two nodes (Fig. 76). In the Odontomachi the mandibles are approximate at their bases, being inserted on the middle of the front of the head (Fig. 77).

This sub-family includes a considerable number of species, and is found in all parts of the world. Extremely little is known as to the habits, but the true Ponerides do not, so far as is known, occur in large communities, and it seems probable that they are destitute of the powers of combined action that are so remarkable in the Camponotides, and in some of the Myrmicides and Dorylides. Most of the species that have been described are known by only one sex, so that very little knowledge exists as to the sexual distinctions; but from the little that is known it would appear that the three sexual forms are not so differentiated as they are in most of the Camponotides and Myrmicides.

Fig. 75—Dinoponera grandis, worker. Amazons.

The species of the genus Leptogenys are believed by Emery and Forel to possess an apterous female. Mr. Perkins has observed that the Hawaiian L. falcigera has workers with different kinds of sting, but no true female. Males of this species are, however, abundant. Wroughton has recently discovered that one member of this genus is of Termitophagous habits, but this is not the case with L. falcigera. Dinoponera grandis (Fig. 75) is the largest of the Ponerides, its workers attaining an inch and a quarter in length. This Insect, according to Bates, marches in single file in the thickets at Pará; its colonies consist of a small number of individuals, and are established at the roots of slender trees. The effects of its powerful sting are not so serious as is the case with some of the smaller ants.

In Britain we have only two representatives of the sub-family, viz. Ponera contracta, a small ant of dirty-yellow colour, found rarely in the Southern counties, living in moss or under stones. Its colonies consist of only a few individuals; Forel giving fifty as the highest number he has observed. The second species, P. punctatissima, presents the almost unique peculiarity of possessing two forms of the male sex, one of them resembling the worker in most of its peculiarities, and in being destitute of wings, while the other is winged, as is usual in male ants. In the island of St. Vincent another species of Ponera has been discovered having an apterous and worker-like male, and was named by Forel P. ergatandria.[^[71]] The discovery of this form has led him to express some doubt as to whether Ponera punctatissima has two forms of males; but it seems probable that it really is so, the ergatoid males being produced under somewhat different circumstances from the normal males. We shall subsequently see that Cardiocondyla and a few other Myrmicides exhibit an analogous peculiarity.

The genus Myrmecia is confined to the Australian continent and Tasmania, and includes a considerable number of species of large and moderate-sized ants, the classification of which has been a subject of difference of opinion. This has arisen from the fact that the nodes of the abdominal pedicel are more similar to those existing in the Myrmicides than to those of the typical Ponerides. There are, however, some American members of the latter sub-family (Paraponera clavata, e.g.) that differ but little in this point from Myrmecia, and, moreover, the pupae of Myrmecia are enclosed in a cocoon, while in the Myrmicides they are usually naked. On the other hand the nests are, it appears, very large and populous, more like what exists in the Myrmicides; there is no true stridulating organ on the first abdominal segment. The genus is therefore one of those interesting anomalies that form so large a proportion of the Australian fauna, and will probably be ultimately treated as a distinct sub-family. There are about thirty species.

The ants of this genus are well known to the residents in Australia, where they are called "bull-dog ants." They form large mounds of earth for their nests. The workers, and females (Fig. 76) are much alike except during the period when the latter are still carrying their wings. The males, however, differ considerably, being of more slender form, and possessing only insignificant mandibles, and straight antennae with a quite short basal joint.

Fig. 76—Myrmecia pyriformis. Australia. Female after casting wings.

Forel considers Myrmecia to be the most formidable of all the ants; the hills are said to be sometimes five feet high, and the colonies are immense in numbers, while the Insect is an inch or more in length, and armed with a very powerful sting, the use of which on the human body is said to give rise in some cases to serious symptoms. On the other hand, we have seen statements to the effect that the sting of Myrmecia has only very evanescent sequelae; it is also said that the ant-hills have only a slight elevation, so that probably both these points differ according to the species. It appears from a communication of Miss Shepherd's that the formidable Myrmecia forficata has its larvae destroyed by a parasitic Hymenopteron (Eucharis myrmeciae) of brilliant colour and considerable size, so that we have the curious fact of the hordes of this most formidably armed ant, which possesses also large eyes, falling a victim to a brilliant and very conspicuous Insect. Particulars of this case of parasitic attack are still wanting. There are other cases known of the larvae of ants being destroyed by parasitic Diptera and Hymenoptera, but in none of them have any sufficient observations been made as to the mode in which the attack is made. Lowne says that M. gulosa itself attacks large beetles of the genus Anoplognathus and buries them; and he also adds the very curious statement that M. nigrocincta, when running, is able to take leaps of a foot in length.

The Odontomachi were formerly considered a distinct sub-family, distinguished by the peculiar mandibles (Fig. 77). Many of the Ponerides have elongate mandibles, but they are inserted at the sides of the front of the head, not in the middle of the front. These organs in some species of Odontomachi serve as levers, by aid of which the Insect can execute considerable leaps. In only a few species are the males known; Mayr and Forel state that they are destitute of the peculiar mandibles characteristic of the worker.

Fig. 77—Anochetus ghiliani, worker. Tangier.

The unique European representative of the Odontomachi, Anochetus ghiliani, occurs in Andalusia. Near Tangier Mr. George Lewis found it to be not uncommon; but the sexes are not known, and it even appears doubtful whether there exists any well-marked division between workers and female. Lewis observed, among the ordinary forms, individuals with longer bodies, usually one in a nest, and he supposed these to be females; Saunders, on examining these examples, found them to possess distinct ocelli, and therefore agreed with Lewis as to their being the female sex. Dr. Emery subsequently examined these same specimens, and took what is scarcely a different view, viz. that they are not females but an intermediate form; and he also expressed the opinion that "the true female may not exist." The male of Anochetus is not known. The female of A. mayri, a Neotropical species, has rudimentary wings.

Sub-fam. 5. Dorylides.—Clypeus extremely small, the antennae inserted very near the front margin of the head. Hind body usually elongate and subcylindrical, with an imperfect pedicel formed by the constriction of the back of the first segment, but occasionally there are two nodes in the workers. Distinctions between the two sexes, and between the workers and sexed forms, enormous, the queens truly wingless. The females and workers usually blind, or at any rate destitute of facetted eyes. (In Ecitonini the antennae are not inserted quite at the front of the head, and there are two nodes in the pedicel.)

We have reserved to the end of the ants the consideration of the two groups Dorylides and Amblyoponides, recent investigations having rendered it somewhat doubtful whether they can be maintained as distinct from Ponerides. The chief character of the Dorylides is that the males are much less ant-like in form than they are in the other groups, and that the distinction between the females and workers are enormous. The little that is known as to the males and females of this group suggests the view that these sexes may offer sufficient reason for keeping the Dorylides as a group distinct from the other ants; but it must be admitted that it is very difficult to find satisfactory characters to distinguish the workers of the Dorylides in some cases from the Ponerides, in others (Eciton) from the Myrmicides.[^[72]]

Fig. 78—Various forms of worker of Eciton hamatum. Guatemala.

The Dorylides are of great interest, for they exhibit the remarkable phenomenon of a nomadic social life, accompanied by imperfect sight in the wanderers. The sub-family includes two apparently distinct groups: (1) the Ecitonini, peculiar to the New World, and having a close relationship with the Myrmicides; and (2) the Dorylini existing chiefly in the eastern hemisphere, and related closely by its workers to the Ponerides and Amblyoponides. (i.) The Ecitonini consist of the species of the genus Eciton, the wandering ants of America, and of Labidus, which there is now good reason for believing to consist of the males of Eciton. The female is still uncertain. The Eciton are nomad ants having no fixed abode, but wandering from place to place in search of prey, and forming temporary resting-places. The species are rather numerous, and the habits of several have been described by Bates, who, however, was not acquainted with some of the most peculiar features in their biology, these having been since revealed by Belt and W. Müller.

These ants are predaceous in their habits, and some of the species travel in vast hordes; they occasionally enter houses and clear them of much of the vermin with which they may be infested. They have no facetted eyes, some of the forms being quite blind, while others have a pair of peculiar lenses in the position normally occupied by the compound eyes. Usually there are two castes of the workers, and in some species these are very different from one another, the mandibles being in the larger form very elongate, cylindrical and unfit for industrial purposes, while the individuals of the smaller caste have the outer jaws shorter, with their edges apposed and coadapted: in other species individuals with mandibles differentiated from the normal form do not exist. The nomad habits of these ants were described by Bates, but the detection of their temporary resting-places was reserved for Belt, who found that, after their plundering raids, they retired to a place of concealment, and there clustered together in a compact mass like a swarm of bees. Belt says: "They make their temporary habitations in hollow trees and sometimes underneath large fallen trunks that offer suitable hollows. A nest that I came across in the latter situation was open at one side. The ants were clustered together in a dense mass, like a great swarm of bees, hanging from the roof, but reaching to the ground below. Their innumerable long legs looked like brown threads binding together the mass, which must have been at least a cubic yard in bulk, and contained hundreds of thousands of individuals, although many columns were outside, some bringing in the pupae of ants, others the legs and dissected bodies of various Insects. I was surprised to see in this living nest tubular passages leading down to the centre of the mass, kept open, just as if it had been formed of inorganic materials. Down these holes the ants who were bringing in booty passed with their prey. I thrust a long stick down to the centre of the cluster and brought out clinging to it many ants holding larvae and pupae."

Turning now to the Labidus question: many American species of this genus have long been known, though all of them by the male sex only. The discoveries (to be subsequently alluded to) made in the Old World as to the relations between the driver ants and Dorylus raised a suspicion that Labidus might be the male of Eciton, the distinctions in the two cases being very analogous: this conjecture has been almost proved to be correct by the recent observations of Hetschko and W. Müller. The latter, who observed the temporary nests of Eciton hamatum, confirms Belt's statements as to the ants hanging together in clumps, like swarms of bees; he also states that the change from one temporary abode to another takes place at night, though, as is well known, the hunting forays of this ant are carried on in the daytime. The periods of migration appear to be determined by the time at which all the larvae have assumed the pupal state, this at any rate being the time chosen in the case observed by Müller. This naturalist bagged a part of one of the nests by the aid of ether, and found the larger portion to consist of pupae; there were also some larvae and eggs; a specimen of Labidus (L. burchelli) was also found on friendly terms with the Eciton-workers; and myrmecophilous Coleoptera were discovered. The pupae are enclosed in cocoons. Persistent search failed to reveal any female, but the examination was made under great difficulties. Müller also states that the earliest pupated larvae yield soldiers, the latest the smallest forms of workers. From observations made by Forel on a pupa, it seems probable that a wingless form of male may be found to exist. If therefore, as appears practically certain, Labidus is the winged male of Eciton, it is probable also that males of more or less worker-like form exist, as is now known to be the case in some other Formicidae.

We may here notice a peculiar apterous female ant recently described by André under the name of Pseudodicthadia incerta. He thought this might prove to be the female of Eciton-Labidus; but his description and figure are imperfect, and do not greatly support his idea of a connection between Eciton and Pseudodicthadia.

ii. The group Dorylini includes the genus Dorylus, which was founded many years ago for Insects very like Labidus. As in the case of the American Insect named, males only were known; two or three allied genera, consisting exclusively of individuals of the sex mentioned, were subsequently described. In the regions inhabited by these males numerous species of blind ants are known, but only in the worker form, and were, or still are, referred to genera called Typhlopone and Anomma. Nothing that could be considered to be a female pertaining to any of these Insects was discovered until Gerstaecker described under the generic name Dicthadia an extraordinary apterous female ant found in Java, and it was suspected that it might be the long-expected female of the male Dorylus and of the worker Typhlopone or Anomma. This remained for many years without confirmation, but in 1880 Trimen announced the discovery in South Africa of an enormous apterous female ant, allied to Dicthadia; it had been disinterred from a nest of small red ants believed (wrongly) to be Anomma. As Dorylus had been previously found in connection with allied worker ants it has since then been clear that notwithstanding the enormous differences existing between these three forms they may all pertain to one (or to closely allied) species. From this summary the student should understand that he will find in myrmecological literature many references to two or three genera that really belong to one species.

Fig. 79—Dorylus helvolus. Africa. A, male; B, female (Dicthadia); C, worker major (Typhlopone); D, worker minor. (After Emery.)

The workers of the Dorylini at present known are without exception quite blind, and are believed to be all of predaceous habits; it is thought by some that they have no fixed abodes, but, like the Ecitonini, frequently change their residence, and it has been suggested that in doing so they make use of the nests of other ants as temporary abodes; all these points are, however, still unsettled, and as there are several genera it is not unlikely that considerable variety will be found to prevail. The driver ants of Africa, belonging to the genus Anomma, are in some respects similar to Eciton in habits, as they enter human habitations and cause nearly everything else to quit; it is probable that they are also exclusively carnivorous. Savage detected the nests of A. arcens, but the account he has given of them is too vague to permit one to decide whether the assemblages he saw were of a nomad kind. The workers of this species vary greatly in size, and Emery has recently stated that he believes all the supposed species of the genus to be merely varieties of A. burmeisteri. The female of the driver ants is still quite unknown. A Dorylus has been ascertained to be the male of Typhlopone. The male Dorylus (Figs. 79, A, and 80) is of great interest, for the propodeum is in a more primitive form than it is in any other petiolate Hymenopteron known to us, while at the same time the pronotum and mesonotum are very highly developed. The genus Typhlatta Sm. has been recently identified by Wroughton and Forel as the worker-condition of which Aenictus is the winged male. The genus Alaopone will probably be found to have some species of Dorylus as its male.

Fig. 80—Body of male of Dorylus sp. Delagoa Bay. a, pronotum; b, c, divisions of mesonotum; d, metanotum; e, propodeum; f, first abdominal segment; g, h, points of insertion of anterior and posterior wings.

The females of the Dorylides are amongst the rarest of Insects, and are also amongst the greatest of natural curiosities. Although worker ants and female ants are merely forms of one sex—the female—yet in this sub-family of ants they have become so totally different from one another in size, form, structure, and habits that it is difficult to persuade oneself they can possibly issue from similar eggs. In the Insect world there are but few cases in which males differ from females so greatly as the workers of Dorylides do from the females, the phenomena finding their only parallel in the soldiers and females of Termites; the mode in which this difference is introduced into the life of the individuals of one sex is unknown. The largest of all the Dorylides are the African Insects of the genus Rhogmus. Only the male is known.

The specimens of female Dorylides that have been detected may, after fifty or sixty years of research, be still counted on the fingers. As the greatest confusion exists in entomological literature owing to the forms of a single species having been described as two or three genera, the following summary of the principal names of genera of Dorylides may be useful:—

Eciton = the workers, Labidus = male: ♀ unknown.

Pseudodicthadia: female only known, possibly that of Eciton.

Cheliomyrmex: workers and soldiers only known.

Aenictus = the male, Typhlatta = worker: ♀ unknown.

Rhogmus: male; female unknown. (According to Emery the worker is very small and like Alaopone.)

Anomma: only worker known; male probably a Dorylus.

Dorylus = male; Dicthadia = ♀: Alaopone and Typhlopone = workers.

Sub-Fam. 6. Amblyoponides.—Abdomen destitute of distinct pedicel; the articulation between the first and second segments behind the true petiole being broad.

Fig. 81—Amblyopone, worker. Tasmania.

We follow Forel in separating Amblyopone and a few allies from the Ponerides, because the abdominal pedicel is more imperfect than in any other ants. It is, indeed, very difficult to frame a definition that will include the Amblyoponides among ants, and at the same time separate Formicidae and Scoliidae. Forel considers the Amblyoponides to approach closely to certain divisions of the Scoliidae (Thynnides, e.g.). Little is known of these Insects, though they are widely distributed. Amblyopone is found in Australia and New Zealand; the allied genus Stigmatomma has a wide distribution, occurring even in Europe. The social life is believed to be imperfect, and the habits subterranean and sedentary. The males and females are winged; the latter much resemble the workers, which are nearly blind, and have a considerable general resemblance to Anomma in Dorylides.

Association of Ants with other kinds of Insects.—We have already alluded to the fact that a few species of ants are used by other species as attendants, and that the two kinds then live together quite amicably; and we have also seen that a few ants live in association with other species on terms that are not yet understood. One little ant, Formicoxenus nitidulus, lives only in the large nests of Formica rufa; these ants tolerate the little Formicoxenus, which so far as is known does them neither good nor harm. There are also a considerable number of species of small ants that are in the habit of choosing the neighbourhood of larger species for their dwelling-places; in some cases the nests are constructed actually within a portion of the edifice of the more powerful species, and the rule then appears to be that these neighbours do not molest one another. Notwithstanding the militant lives that many of them lead, ants cannot be considered as of generally ferocious disposition.

But the most remarkable point in connection with their toleration consists in the fact that the nests of many species are inhabited by quite a colony of foreign Insects of various Orders; many of these, being found nowhere else, are spoken of as ants'-nest or Myrmecophilous Insects.[^[73]] The relations of ants with other Insects are of the most varied and complex character; some of their guests live with them on terms of the most intimate association, being indeed absolutely dependent for their existence on the good offices of their hosts; others of the ants'-nest Insects are enemies, while others are neutral or indifferent to the ants. We have already mentioned that the guests migrate in company with their hosts.

Many species of ants derive a considerable portion of their sustenance from the sweet substances excreted by Aphidae. Ants may constantly be seen occupied with clusters of Aphidae, and it is said that the ingenious little creatures defend from enemies the manufacturers of the sweet-stuff they are so fond of, even going so far as to form barricades and covered places for the isolation and protection of this peculiar kind of cattle; a few ants keep some of the root-feeding Aphidae in their nests. Coccidae and other Homoptera, which also excrete much matter of a sugary nature, are likewise consorted with by ants; as are also the larvae of some butterflies of the family Lycaenidae; these latter being believed to furnish to the ants some substance of a nutritious kind.

Fig. 82—The beetle, Atemeles, soliciting food from an ant. (After Wasmann.)

The Insects we have spoken of are, however, rather of the nature of ant-cattle, and the fondness of the ants for them is not very remarkable. The relations of the ants to the peculiar species of Insects that live only in or around their nests are much more extraordinary. The greater number of these guests belong to the Order Coleoptera, and of these there are many hundreds—probably many thousands—of species that depend on ants for their existence. The family Pselaphidae furnishes a large number of ants'-nest beetles, and it appears probable that most of them excrete some sugary substance of which the ants are fond. Many of these Pselaphidae are of the most fantastic shapes, more especially the members of the sub-family Clavigerides. But the most curious of all the ant's-nest beetles are the Paussidae, a family exclusively dependent on ants, and having the curious faculty, when disturbed, of bombarding—that is, of discharging a small quantity of vapour or liquid in a state of minute subdivision accompanied by a detonation. Many species of Staphylinidae are peculiar to ant's-nests, and most of them are indifferent or inimical to their hosts, but some of them, such as Atemeles (Fig. 82) and Lomechusa, are doubtless producers of sweet stuff that is liked by the ants. The ants feed some of their special favourites amongst these guests in the same manner as they feed one another, viz. by opening the mouth, causing a drop of liquid to appear on the lip, and remaining passive while the guest partakes of the proffered bonne bouche. This way of giving food to other individuals is a most remarkable feature in the character of ants; it is not the same system that they adopt in feeding the larvae, for they then make a series of actual movements, and force the nutriment into the mouths of the grubs. Besides the Insects we have mentioned there are also Orthoptera, Hemiptera, Poduridae and Thysanura, Acari, and small Isopod crustaceans that live exclusively in company with ants. We have mentioned that a few Hymenopterous and Dipterous parasites have been detected living at the expense of ants; it is probable that closer observation of the ant larvae and pupae in their nests will disclose a greater number of the parasites of this latter class.

Much attention has been given to the relations between ants and their guests by Wasmann.[^[74]] He arranges them in four categories; 1, "Symphily" for the true guests, which are fed and tended by the ants, the guests often affording some substance the ants delight in; 2, "Metochy," the class of tolerated guests, being so far as is known not disagreeable to the hosts; 3, "Synecthry," including those Insects, etc., to which the ants are hostile, but which nevertheless maintain themselves in the midst of their foes; 4, Parasites, dwelling in the bodies of the adult, or of the young ants. Many of these ants'-nest Insects present a more or less perfect resemblance to the ants in one or more points, such as sculpture, colour, size, or form. To these resemblances Wasmann attaches great importance. We should, too, notice that some of the inquilines[^[75]] have become acquainted with the movements and habits of the ants, and stroke them (as the ants do one another) to induce them to disgorge food in the manner we have alluded to. According to Janet, ants of the genus Lasius are infested by Acari of the genus Antennophorus. The ants carry the mites, which assume positions so as not to cause greater inconvenience than is inevitable. Moreover, the ants give food to the mites when requested, and behave in a most obliging way to them, though there is not any reason for supposing that in this case the ants derive any benefit from the Symphily.

The relations between ants and plants have been of late years much discussed. We have already briefly alluded to the subject when speaking of the Pseudomyrmini. We will here only remark that ants frequent plants not only for the purpose of securing the sweet stuff excreted by the Aphidae that live on them, but also for the sake of getting the sweet products the plants themselves afford. Mr. Aitken, speaking of ants in India, says: "I have come to the conclusion that one of the most important sources of food-supply which ants have is the sacchariferous glands to be found at the bases of so many leaves." It is supposed that the ants are on the whole beneficial to the plants that thus afford them supply; and this fact is considered by many to afford an adequate explanation of the existence of these interesting relations.

CHAPTER V

COLEOPTERA—OR BEETLES

Order V. Coleoptera.

Apparently wingless Insects when at rest, but really with four wings; the elytra, or anterior pair, shell-like, reposing on the back of the body and fitted together accurately along the middle by a straight suture; the posterior pair membranous, folded together under the elytra. Mouth with mandibles; lower lip not divided along the middle. Metamorphosis great and very abrupt; the larva being a grub or maggot, which changes to a pupa (usually soft) in which the external structure of the perfect Insect is conspicuous.

Coleoptera—or Beetles—are chiefly distinguished from other Insects by the solidity of their external integument, and by the peculiar nature of the first pair of their alar organs, which do not serve as instruments of flight, but as shells for protecting the upper face of the after-body, which, unlike the other parts, remains as a rule soft and membranous. These modifications of structure, though apparently slight, must be really extremely advantageous, for beetles are the predominant Order of Insects in the existing epoch. They depart from most other Insects in being less aerial in their habits; therefore, notwithstanding their enormous numbers, they do not meet the eye so frequently as flies, bees, or butterflies. The parts of the hard outer skeleton are beautifully fitted together, and as their modifications are easily appreciated they offer as fascinating a subject for study as do the skeletons of Vertebrata. The habits of beetles are so extremely varied that it is but little exaggeration to say that Coleoptera are to be found everywhere, when looked for. The number of species at present known is probably about 150,000. Of these somewhere about 3300 have been found in Britain.

Fig. 83—Under-surface of a beetle, Harpalus caliginosus; legs and antenna of one side, and some parts of the mouth removed. A, antenna; B, mandible; C, labrum; D, ligula; E, paraglossa; F, labial palp; G, inner lobe of maxilla; H, outer lobe (palpiform) of maxilla; I, maxillary palp; K, mentum; L, gena; M, gula; N, buccal fissure; V, plates of ventral segments. 1, Prosternum; 2, prosternal episternum; 3, prosternal epimeron; 4, anterior and middle coxal cavities; 5, inflexed side of pronotum; 6, mesosternum; 7, mesosternal episternum; 8, mesosternal epimeron; 9, metasternum; 10, posterior division of metasternum or ante-coxal piece; 11, metasternal episternum; 12, metasternal epimeron; 13, epipleuron or inflexed margin of elytron; 14, ventral or ambulatory setae; 15, trochanter; 16, posterior coxa; 17, femur; 18, tibia; 19, tarsus. (Modified from Leconte and Horn.)

The structure of the hard parts of the skeleton is of importance, as the classification of this enormous number of species is entirely based thereon; it will be readily understood from the accompanying diagram (Fig. 83). The general proportions of the chief parts of the body call for a few remarks. The prothorax is remarkably free, and is therefore capable of a much greater amount of movement independent of the after-body than it is in other Insects. The mesothorax is, on the other hand, much reduced; its chief function in the higher forms is to support the elytra, and to help to keep them together by means of its scutellum. The metathorax, on the contrary, is largely developed, except in the rather numerous forms that are entirely deprived of powers of flight. The composition of the abdomen has been a subject of great difference of opinion. Its upper surface is usually entirely covered by the elytra; the parts visible on the lower surface are called ventral segments, and are usually five in number. Although these five plates may constitute all that is superficially visible of the abdomen, yet if the elytra are taken off it is found that a larger number of segments—usually seven or eight—are visible on the dorsum. This seeming discrepancy of number between the dorsal and ventral plates is due to two facts; 1, that the hind coxae have a great and complex development, so that they conceal the true base of the venter, which, moreover, remains membranous to a greater or less extent, and thus allows much mobility, and at the same time a very accurate coadaptation between the hard parts of the venter and the metasternum[^[76]]; 2, that the terminal segments are withdrawn into the interior of the body, and are correspondingly much modified, the modification being greater in the case of the ventral than in that of the dorsal plates. The anatomy of the parts of the abdomen that are not externally visible has not been adequately studied by coleopterists, but Verhoeff has inaugurated a careful study of the comparative anatomy of the terminal segments[^[77]]; unfortunately, however, he has not so thoroughly studied the modifications at the base, and as it is not clear that these are so uniform as he has taken for granted, it is possible that his numbering of the segments may have to be in some cases modified. The retracted plates or segments are so intimately connected with the internal copulatory organs that it is no easy matter to interpret them. For the nomenclature of these parts we must refer the student to Verhoeff's later works. He considers the abdomen as composed of ten segments, the dorsal plates being demonstrable, while the tenth ventral plate is usually absent. The anal orifice is placed immediately beneath the tenth dorsal plate, and above the genital orifice, which lies behind and above the ninth ventral plate. Peytoureau admits a diversity in both the number of segments and the position of the orifice. These studies in comparative anatomy are surrounded with difficulties, and no morphological conclusions based on them can be considered as final until they have been confirmed by observation of the development of the parts.

The elytra—or wing-cases—frequently have a remarkable sculpture, the use of which is unknown. According to Hofbauer there are between the outer and inner layers, glands secreting a fluid that reaches the surface through small pores. Hicks supposed that he detected nerve cells. Meinert is of opinion that the elytra correspond to the tegulæ of Hymenoptera rather than to the wings of other Insects, but the little evidence that exists is not favourable to this view. The two elytra are usually, in repose, very perfectly fitted together by a complete coadaptation along the middle of the body, so that it is difficult to separate them; this line of junction is called the suture. There are forms in which the coadaptation is quite imperfect (Malacodermidae) and some in which it does not exist at all (Meloë). The wings proper of beetles correspond to the posterior pair in other Insects, and are much more irregular in nervuration than those of most other Insects, correlative, it is supposed, with the folding they are subjected to in order to get them beneath the wing-cases. There are large numbers of species, genera, and groups of genera, all the members of which have the wings so much reduced in size as to be quite useless for purposes of flight. These forms are called apterous, though they are not really so, for the elytra (which are really the anterior wings) are present, and even the posterior wings are not truly absent in these cases, though they are sometimes so extremely rudimentary as to elude all but the most careful observation. The number of forms in which the elytra are absent is extremely small; this condition occurs only in the female sex; it is usually confined to cases in which the female is larva-like in form; but in the extraordinary Mediterranean Lamellicorn genus, Pachypus, the females are destitute of wings and elytra, though the anterior parts of the body are normally formed: these individuals live underground and rarely or never emerge. When the wings are absent the elytra are frequently soldered; that is to say, united together along the suture by some sort of secondary exudation; this union occurs in every degree of firmness, and appears to be variable in the individuals of one species; probably in accordance with the age of the individual. In most beetles the elytra are not only themselves closely connected, but are also very accurately coadapted with the sides of the body, except at the tip. Sometimes a coadaptation occurs between the tips of the elytra and the body, but not at the tip of the latter. In such cases one or more dorsal plates are left exposed: the last of such exposed dorsal plates is termed pygidium; a similar plate anterior to the pygidium is called propygidium.

Larvae.—Owing to the difficulty of rearing Coleoptera, less is perhaps known of their life-histories than of those of other Insects. They exhibit, however, extreme diversity correlative with the great specialisation of so many beetles to particular kinds of life. Most beetles must have exactly the right conditions to live in. The larvae of many forms are known. They are composed of a head, three thoracic segments (usually very distinct), and a number of abdominal segments varying from eight to ten. Coleopterous larvae are usually described as having nine abdominal segments; and it is but rarely that ten can be readily detected; they are, however, visible in various forms, as is the case in the form figured (Fig. 84). A great many of them possess a peculiar pseudopod at the underside of the body near or at the extremity; it can in many cases be entirely retracted into the body, and is generally described as being the protruded termination of the alimentary canal. Inspection of a series of larvae shows that it represents a body segment: it is sometimes armed with hooks. Three pairs of small thoracic legs are often present, but are very often completely absent. These thoracic legs may be present in the young larva, but not in the older (Bruchus). The usual number of spiracles is nine pairs, one prothoracic, eight abdominal; but this is subject to many exceptions, and mesothoracic and metathoracic stigmata are occasionally found. The figures we give in the following pages will enable the student to form some idea of the variety of form exhibited by beetle larvae.

Fig. 84—Larva of a beetle, Family Cerambycidae (? Aromia moschata). The first spiracle is placed just at the hind margin of the large prothoracic segment. (From La Massane.)

Pupation usually takes place in a cavity in the earth, or near the feeding-place, but a great many species form a cocoon, composed either of fragments of earth or of wood, and slightly cemented together. A few suspend themselves by the tail after the manner of butterfly caterpillars (Cassididae, Coccinellidae). The pupae are usually extremely soft, their appendages not being fastened to the body. But some pupae (Staphylinides) are truly obtected, having a hard shell and the rudimentary appendages fastened by exudation to the body, like Lepidopterous pupae, and others (Coccinellidae) are intermediate between this state and the normal soft pupa. The pupal state lasts but a short time, from one to three weeks being the usual period. The perfect Insect is at first soft and almost colourless, and it is often some days before it attains its complete coloration and hardness.

Classification.—Owing to the hardness of the skeleton, beetles shrivel but little after death, so that the form and relations of the various sclerites can usually be detected with ease. These sclerites seem to be remarkably constant (except in the case of sexual distinctions) within the limits of each species, and are very useful for the formation of genera and groups of genera; but they vary so much outside the limits mentioned that it is very difficult to make use of them for defining the larger groups. Hence it is not easy to frame accurate definitions of the families, and still less so to arrange these families in more comprehensive series. The natural difficulty has been much increased by the habit coleopterists have of framing their definitions on what is visible without the aid of dissection. Nevertheless considerable progress has been made. We are obliged at present to adopt upwards of eighty families; and we are able to distinguish on positive characters five series; this leaves a large number of forms still unclassified, and these we have here associated as a sixth series, which we have called Coleoptera Polymorpha. This series corresponds with the two series called in books Clavicornia and Serricornia. As it is admitted to be impossible to define these two series, we think it much better to act accordingly, and to establish for the present a great group that can only be characterised by the fact that its members do not belong to any of the other five series. No doubt a larger knowledge of development, coupled with the advance of comparative anatomy, will ultimately bring about a better state of affairs. The Strepsiptera, with one family Stylopidae, are only provisionally included among the Coleoptera. These six series are fairly equal as regards extent. Though the Polymorpha includes the larger number of forms, yet a large part of them belong to four great families (Staphylinidae, Buprestidae, Elateridae, Malacodermidae), which are easily recognisable, so that the number of unmanageable forms is not really great. Indeed, an acquaintance with the external anatomy of two or three dozen species, selected as typical, would enable a student to classify with tolerable certainty the vast majority of species that he would subsequently meet with.

Series 1. Lamellicornia.—Antennae with the terminal joints leaf-like (or broader than the others, if not actually leaf-like), and capable of separation and of accurate apposition. Tarsi five-jointed.

Series 2. Adephaga—(Caraboidea of some authors).—Antennae never lamelliform, thin at the end; all the tarsi five-jointed, with the fourth joint quite distinct. Maxillae highly developed, with the outer lobe slender and divided into two segments so as to be palpiform. Abdomen with six (or more) ventral segments visible.

Series 3. Polymorpha.—Antennae frequently with either a club, i.e. the distal joints broader [Clavicorn series of authors], or the joints from the third onwards more or less saw-like, the serrations being on the inner face [Serricorn series of authors]; but these and all the other characters, including the number of joints in the feet, very variable.

Series 4. Heteromera.—Front and middle tarsi five-jointed, hind tarsi four-jointed. Other characters very variable.

Series 5. Phytophaga.—Tarsi four-jointed [apparently], but with a small additional joint at the base of the fourth joint: sole usually densely pubescent [sometimes the feet are bare beneath or bristly, and occasionally the small joint at the base of the fourth joint is more distinct].

Series 6. Rhynchophora.—Head prolonged in front to form a beak; gula indistinguishable. [Palpi usually not evident.] Tarsi four-jointed [apparently], but with a very minute additional joint at the extreme base of the fourth joint.

Strepsiptera (see p. [298]).

The first and second series, with much of the third, form the Pentamera, the fifth and sixth the Tetramera [or Pseudotetramera[^[78]]]. The term Isomera was applied by Leconte and Horn to a combination of series 1, 2, 3, and 5.

Series 1. Lamellicornia.

Tarsi five-jointed; antennae with the terminal joints (usually three, sometimes more), broader on one side, so as to form a peculiar club, the leaves of which are movable, and in repose are more or less perfectly coadapted so as to have the appearance of being but one piece.

This series includes three families, Passalidae, Lucanidae, and Scarabaeidae; the latter includes an enormous majority of the species, and in them the structure of the antennae characteristic of the series is well developed; but in the other two families the form of the antennae is not so widely different from that of other Coleoptera. The larvae live on decaying vegetable matter, roots or dung. They have three pairs of legs, and are thick clumsy grubs with curved bodies, the last two segments being of larger size than usual. Many of them possess organs of stridulation, and the structure of their spiracles is very peculiar, each one being more or less completely surrounded by a chitinous plate. The spiracles usually form a system entirely closed, except at the moment when the skin is shed and the tracheal exuviae are detached. Meinert[^[79]] considers these spiracles to be organs of hearing. The life of the larvae is passed underground or in the decaying wood on which the Insect feeds.

Fig. 85.—Antennae of Lamellicorns. 1, Neleus interruptus; 2, Lucanus cervus ♂; 3, Phanaeus splendidulus ♀; 4, Phileurus didymus ♀; 5, Polyphylla fullo ♂.

Most of the members of this series are remarkable on account of the great concentration of the nerve-centres. This is extreme in Rhizotrogus, where there are only two great ganglia, viz. the supra-oesophageal and a great ganglion situate in the thorax, and consisting of the conjoined infra-oesophageal, thoracic, and abdominal ganglia. According to Brandt[^[80]] there are several distinct forms of concentration in the series; the Lucanidae only participate in it to the extent that the perfect Insects exhibit fewer ganglia than the larvae; the latter possess two cephalic, three thoracic, and eight abdominal ganglia, while the perfect Insect has the abdominal ganglia reduced in number to six, and they are placed partially in the thorax. The diminution in number takes place in this case by the amalgamation of the first two abdominal with the last thoracic ganglia.

Fig. 86—View of one side of meso- and metathorax of a Passalid larva from Borneo showing the stridulating organs. a, b, Portions of the metathorax; c, coxa of 2nd leg; d, striate or stridulating area thereon; e, basal part of femur of middle leg; f, hairs with chitinous process at base of each; g, the diminutive 3rd leg modified for scratching the striated area. × 20.

Fam. 1. Passalidae.—Labrum large, mobile; mentum deeply cut out in the middle for the accommodation of the ligula; the lamellae of the antenna brought together by the curling up of the antenna. The elytra entirely cover the dorsal surface of the abdomen. There are four or five hundred species of this family known; they are usually shining-black, unattractive beetles, of large size, and are abundant in the decaying wood of tropical forests. They are quite absent from Europe, and there is only one species in the United States of North America. The larvae are very interesting, from the fact that they appear to have only four legs. This arises from the posterior pair being present only as very short processes, the function of which is to scrape striated areas on the preceding pair of legs and so produce sound. In the species figured (Fig. 86) this short leg is a paw-like structure, bearing several hard digits; but in other species it is more simple, and without the digits. The perfect Insect has no sound-producing organs, and it is very remarkable therefore to find the larvae provided with highly-developed stridulatory structures. No auditory organ is known, unless the peculiar spiracles be such.

Fig. 87—Head and prothorax of forms of the male of a stag-beetle; Homocoderus mellyi (Africa). A, Large, B, intermediate, C, small forms. (From a photograph by R. Oberthür.)

Fam. 2. Lucanidae (Stag-beetles).—Labrum indistinct, fixed; mentum not excised; antennae not curled in repose, with but little coadaptation of the terminal joints; the elytra entirely cover the dorsal surface of the abdomen. The Stag-beetles are well known on account of the extraordinary development of the mandibles in the male sex, these organs being in some cases nearly as long as the whole of the rest of the Insect, and armed with projections or teeth that give the Insects a most formidable appearance. So far as we have been able to discover, these structures are put to very little use, and in many cases are not capable of being of service even as weapons of offence. The males are usually very much larger than the females, and are remarkable on account of the great variation in the stature of different individuals of the same species; correlative with these distinctions of individual size we find extreme differences in the development of the head and mandibles. Moreover, the small male specimens exhibit not merely reductions in the size of the mandibles, but also show considerable differences in the form of these parts, due, in some cases, apparently to the fact that only when a certain length of the mandible is attained is there any development of certain of the minor projections: in other cases it is not possible to adopt this view, as the small mandibles bear as many projections as the large forms do, or even more. In each species these variations fall, in the majority of cases, into distinct states, so that entomologists describe them as "forms," the largest developments being called teleodont, the smallest priodont; the terms mesodont and amphiodont being applied to intermediate states. Leuthner, who has examined many specimens, states that in Odontolabis sinensis, no intermediates between the teleodont and mesodont forms occur, and as the two forms are very different they are liable to be mistaken for distinct species.

There are at present between 500 and 600 species of stag-beetles known; the Indo-Malayan and Austro-Malayan regions being richest in them. Australia possesses many remarkable and aberrant forms. In the Ceratognathini—a group well represented in New Zealand as well as in Australia—the structure of the antennae is like that of the Scarabaeidae, rather than of the Lucanidae. The most aberrant form known is, however, our common Sinodendron cylindricum; this departs in numerous features from other Lucanidae, and instead of the mandibles of the male being more largely developed, there is a horn on the head; it is very doubtful whether this Insect should be allowed to remain in the family. Little is known of the habits and development of Lucanidæ, except in the case of three or four species that are common in Europe.

Fig. 88—Sinodendron cylindricum. A, Larva; B, pupa. New Forest.

The common stag-beetle, Lucanus cervus, is our largest British beetle. The larva much resembles that of Melolontha vulgaris, but attains a larger size, and the anal aperture is placed longitudinally instead of transversely; it lives in decaying wood, or eats the roots of trees without being injurious; its life in this state lasts about four years; the pupal period is passed through rapidly, and the perfect Insect may remain for months underground before it becomes active; this occurs in June and July. This larva stridulates by scraping certain hard tubercular ridges on the third pair of legs, over a specially adapted rough area at the base of the second pair.

The Passalidae and Lucanidae are united by some authorities as a group called Pectinicornia; the term Lamellicornia being then confined to the Scarabaeidae. The Passalidae appear, however, to be really more nearly allied to the Scarabaeidae than to the Lucanidae.

Fam. 3. Scarabaeidae (Chafers).—The leaflets of the antennae are well coadapted, and are susceptible of separation. The elytra usually leave the pygidium uncovered. The number of visible ventral segments is usually six, or at the sides seven, not five, as in Lucanidae and Passalidae. This is one of the most important families of Insects. About 13,000 species are already known; as some of them are highly remarkable creatures on account of the males being armed with horns, they are figured in many works on natural history. There is great variety of form, and the following five sub-families may be adopted, though authorities are by no means agreed as to the classification of this extensive family, which, moreover, be it remarked, is increasing by the discovery of about 300 new species every year.

Abdominal spiracles placed in a line on the connecting membranes, and entirely covered by the wing-cases (Laparosticti). .......... Sub-fam. 1. Coprides.[^[81]]

Abdominal spiracles placed almost in a line, but only the basal three on the connecting membranes; the terminal one usually not covered by the wing-cases. .......... Sub-fam. 2. Melolonthides.

Abdominal spiracles placed in two lines, the basal three on the connecting membranes, the others on the ventral segments (Pleurosticti).

The claws of the tarsi unequal. .......... Sub-fam. 3. Rutelides.

The claws of the tarsi equal; the front coxae transverse, but little prominent in the descending axis. .......... Sub-fam. 4. Dynastides.

The claws of the tarsi equal; the front coxae more prominent, shorter transversely. .......... Sub-fam. 5. Cetoniides.

i. The Coprides form an immense group of about 5000 species; they differ somewhat in habits from other Lamellicorns, inasmuch as most of them live on dung, or decaying animal matter; the sub-family connects with the Lucanidae, so far as superficial characters go, by means of two of its groups, Trogini and Nicagini, the latter being very near to the Ceratognathini in Lucanidae. So little is known as to the morphology and development of these groups that it is not possible to pronounce an opinion as to the validity of this apparent alliance. Trox stridulates by rubbing two raised lines on the penultimate dorsal segment across two striate ribs on the inner face of the elytra; Geotrupes, on the other hand, produces an audible sound by rubbing together a file on the posterior coxa and a fine ridge on the contiguous ventral segment. The larva in this genus has a different organ for stridulation from the imago; it is placed on the second and third pairs of legs, the latter pair being much reduced in size.

The most interesting division of the Coprides is the group Scarabaeini. No member of this group inhabits the British islands, but in Southern Europe, and in still warmer lands, these Insects are well known from the curious habit many of the species have of rolling about balls of dung and earth. The long hind legs are chiefly used for this purpose, and it is on the peculiar structure of these limbs that the group has been established. Many of the stone Scarabaei found in Egyptian tombs represent some kind of Scarabaeini, and it has been said that the ancient Egyptians looked on these Insects as sacred because of their movements. These must certainly appear very strange to those who see them and are unacquainted with their object. It is stated that the dwellers in the valley of the Nile thought the actions of these Insects, when rolling their balls, were typical of the planetary and lunar revolutions; and that the sudden appearance of the beetles after a period of complete absence was emblematic of a future life. Many accounts have been given of the habits of members of this group, but according to Fabre all are more or less erroneous; and he has described the habits and life-history of Scarabaeus sacer (Fig. 89), as observed by him in Southern France. These Insects act the part of scavengers by breaking up and burying the droppings of cattle and other animals. The female Scarabaeus detaches a portion of the dung and forms it into a ball, sometimes as large as the fist; this it rolls along by means of its hind legs, by pushing when necessary with its broad head, or by walking backwards and dragging the ball with its front legs. The strength and patience displayed by the creature in the execution of this task are admirable. Frequently the owner of this small spherical property is joined, so Fabre informs us, by a friend, who is usually of the same sex and assists her in pushing along the ball till a suitable place is reached. When this is attained, the owner commences to excavate a chamber for the reception of the ball; sometimes the false friend takes advantage of the opportunity thus offered and carries off the ball for her own use. Should no disappointment of this sort occur, the Scarabaeus accomplishes the burying of the ball in its subterranean chamber, and accompanies it for the purpose of devouring it; the feast is continued without intermission till the food is entirely exhausted, when the Scarabaeus seeks a fresh store of provender to be treated in a similar manner. According to M. Fabre's account these events occur in the spring of the year, and when the hot weather sets in the Scarabaeus passes through a period of quiescence, emerging again in the autumn to recommence its labours, which are now, however, directed immediately to the continuance of the species; a larger subterranean chamber is formed, and to this retreat the beetle carries dung till it has accumulated a mass of the size of a moderate apple; this material is carefully arranged, previous to the laying of the egg, in such a manner that the grub to be hatched from the egg shall find the softest and most nutritive portions close to it, while the coarser and more innutritious parts are arranged so as to be reached by the grub only after it has acquired some strength; lastly, a still more delicate and nutritive paste is prepared by the mother beetle for a first meal for the newly-hatched grub, by some of the food being submitted to a partial digestion in her organs; finally, the egg is deposited in the selected spot, and the chamber closed. Certain of the Coprides exhibit, according to Fabre, some extremely exceptional features in their life-histories. The mother, instead of dying after oviposition, survives, and sees the growth of her young to the perfect state, and then produces another generation. No similar case can be pointed out in Insects, except in the Social kinds; but from these the Coprides observed by Fabre differ profoundly, inasmuch as the number of eggs produced by the mother is extremely small; Copris hispanus, for instance, producing in each of its acts of oviposition only one, two, or three eggs.

Fig. 89—Scarabaeus sacer. Portugal.

ii. The Melolonthides are probably almost as numerous as the Coprides, some 4000 species being already known. The larvae are believed to feed chiefly on roots. Melolontha vulgaris, the common cockchafer, is very abundant in some parts of Europe, and owing to this and to the great damage it causes, has attracted much attention. The memoir by Straus-Durckheim[^[82]] on its anatomy is one of the classical works of Entomology. This Insect is so injurious in some parts of France that money is paid by the local authorities for its destruction. M. Reiset informs us that under this arrangement 867,173,000 perfect cockchafers, and 647,000,000 larvae were destroyed in the Seine-inférieure in the four years from the middle of 1866 to 1870. Unlike the Coprides, the larval life in Melolonthides is prolonged, and that of the imago is of brief duration. In Central Europe the life-cycle of the individual in M. vulgaris occupies three years, though in dry periods it may be extended to four years. In Scandinavia the time occupied by the development appears to be usually five years. The fertile female enters the ground and deposits its eggs in two or three successive batches of about fifteen each. The eggs swell as the development of the embryo progresses; the larva emerges about five weeks after the eggs have been deposited, and is of relatively large size. When young the larvae can straighten themselves out and crawl, but when older they lose this power, and when above ground rest helplessly on their sides. In the winter they descend deeply into the earth to protect themselves from frost. The pupa state lasts only a few days, but after the final transformation the perfect Insect may remain motionless for as much as eight months underground before commencing its active life in the air.[^[83]] In the perfect state the Insect is sometimes injurious from the large quantity of foliage it destroys. Schiödte[^[84]] considered that these larvae (and those of numerous other Scarabaeidae) stridulate by rubbing certain projections on the stipes of the maxilla against the under-surface of the mandible. These surfaces appear, however, but little adapted for the purpose of producing sound.

iii. The Rutelides number about 1500 species; there are many Insects of brilliant metallic colours amongst them, but very little is known as to their life-histories. The larvae are very much like those of Melolonthides.

iv. The Dynastides are the smallest division in number of species, there being scarcely 1000 known; but amongst them we find in the genera Dynastes and Megasoma some of the largest of existing Insects. The horns and projections on the heads and prothoraces of some of the males of these Insects are truly extraordinary, and it does not appear possible to explain their existence by any use they are to their possessors. These structures are but little used for fighting. Baron von Hügel informs the writer that in Java he has observed large numbers of Xylotrupes gideon; he noticed that the males sometimes carry the females by the aid of their horns; but this must be an exceptional case, for the shape of these instruments, in the majority of Dynastides, would not allow of their being put to this use. The development of these horns varies greatly in most of the species, but he did not find that the females exhibited any preference for the highly armed males. The fact that the males are very much larger than the females, and that the armature is usually confined to them, suggests, however, that some sexual reason exists for these remarkable projections. Many Dynastides possess organs of stridulation, consisting of lines of sculpture placed so as to form one or two bands on the middle of the propygidium, and brought into play by being rubbed by the extremities of the wing-cases. This apparatus is of a less perfect nature than the structures for the same purpose found in numerous other beetles. We have no member of this sub-family in Britain, and there are scarcely a dozen in all Europe. Decaying vegetable matter is believed to be the nutriment of Dynastides. The European Oryctes nasicornis is sometimes found in numbers in spent tan. The growth and development of the individual is believed to be but slow.

v. The Cetoniides are renowned for the beauty of their colours and the elegance of their forms; hence they are a favourite group, and about 1600 species have been catalogued. They are specially fond of warm regions, but it is a peculiarity of the sub-family that a large majority of the species are found in the Old World; South America is inexplicably poor in these Insects, notwithstanding its extensive forests. In this sub-family the mode of flight is peculiar; the elytra do not extend down the sides of the body, so that, if they are elevated a little, the wings can be protruded. This is the mode of flight adopted by most Cetoniides, but the members of the group Trichiini fly in the usual manner. In Britain we have only four kinds of Cetoniides; they are called Rose-chafers. The larvae of C. floricola and some other species live in ants' nests made of vegetable refuse, and it is said that they eat the ants' progeny. Two North American species of Euphoria have similar habits. The group Cremastochilini includes numerous peculiar Insects that apparently have still closer relations with ants. Most of them are very aberrant as well as rare forms, and it has been several times observed in North America that species of Cremastochilus not only live in the nests of the ants, but are forcibly detained therein by the owners, who clearly derive some kind of satisfaction from the companionship of the beetles. The species of the genus Lomaptera stridulate in a peculiar manner, by rubbing the edges of the hind femora over a striate area on the ventral segments.

Series II. Adephaga or Caraboidea.

All the tarsi five-jointed; antennae filiform, or nearly so; mouth-parts highly developed, the outer lobe of the maxilla nearly always divided into a two-jointed palpus; supports of the labial palpi developed as joints of the palpi, and in some cases approximate at their bases. Abdomen with the exposed segments one more in number at the sides than along the middle, the number being usually five along the middle, six at each side.

This extensive series includes the tiger-beetles, ground-beetles, and true water-beetles; it consists of six families, and forms a natural assemblage. It is sometimes called Carnivora or Filicornia. The exceptions to the characters we have mentioned are but few. The supports of the labial palpi are frequently covered by the mentum, and then the palpi appear three-jointed; but when the joint-like palpiger is not covered these palps appear four-jointed. As a rule, approximation of these supports is indicative of high development. In some of the lower forms the trophi remain at a lower stage of development than is usual. This is especially the case with the genus Amphizoa, which forms of itself the family Amphizoidae. The Bombardier-beetles make an exception as regards the abdominal structure, for in some of them no less than eight segments are visible, either along the middle line or at the sides of the venter. In Hydroporides (one of the divisions of Dytiscidae) the front and middle feet have each only four joints. Many naturalists unite the Gyrinidae with the Adephaga, and a few also associate with them the Paussidae and Rhysodidae; but we think it better at present to exclude all these, though we believe that both Paussidae and Rhysodidae will ultimately be assigned to the series. The larvae are usually very active, and have a higher development of the legs than is usual in this Order. Their tarsi possess two claws.

Fam. 4. Cicindelidae (Tiger-beetles).—Clypeus extending laterally in front of the insertion of the antennae. Lower lip with the palpi usually greatly developed, but with the ligula and paraglossae very much reduced, often scarcely to be detected. Maxillae with the outer lobe forming a two-jointed palp,[^[85]] the inner lobe elongate, furnished at the tip with a hook-like process, which is usually articulated by a joint with the lobe itself. The tiger-beetles are very active Insects, running with extreme speed, and sometimes flying in a similar manner; they are all predaceous, and amongst the most voracious and fierce of the carnivorous beetles, so that they well deserve their name. Bates, speaking of the Amazonian Megacephala, says "their powers of running exceed anything I have ever observed in this style of Insect locomotion; they run in a serpentine course over the smooth sand, and when closely pursued by the hand they are apt to turn suddenly back and thus baffle the most practised hand and eye." He further says that the species he observed (being of diverse colours) agreed in colour with the general colours of the "locale they inhabit." The larvae of Cicindelidae live in deep burrows, sinking more or less vertically into the ground, and in these they take up a peculiar position, for which their shape is specially adapted; the head and prothorax are broad, the rest of the body slender, the fifth segment of the abdomen is furnished on the back with a pair of strong hooks; the ocelli on the sides of the head are very perfect. Supporting itself at the top of the burrow by means of these hooks and of its terminal tube, the larva blocks the mouth of the burrow with its large head and prothorax, and in this position waits for its prey. This consists of Insects that may alight on the spot or run over it. When an Insect ventures within reach, the head of the larva is thrown back with a rapid jerk, the prey is seized by the long sharp mandibles, dragged to the bottom of the burrow and devoured. The burrows are often more than a foot deep, and are said to be excavated by the larva itself, which carries up the earth on the shovel-like upper surface of its head. The female tiger-beetle is endowed with powerful and elongate excavating instruments at the termination of the body, and it is probable that when placing the egg in the earth she facilitates the future operations of the larva by forming the outlines of the burrow. Extremely few larvae of Cicindelidae are known, but they all exhibit the type of structure mentioned above, and apparently have similar habits. Our little British Cicindela, most of which are so active on the wing, agree in these respects with the African species of Manticora, which are entirely apterous, and are the largest of the Cicindelidae. Péringuey found a breeding-ground of M. tuberculata near Kimberley; the larvae were living in the usual Cicindelid manner; but the ground was so hard that he was not able to investigate the burrows, and there were but few Insects that could serve as food in the neighbourhood.

Fig. 90—Cicindela hybrida. Britain. A, larva (after Schiödte); B, imago, male.

The Cicindelidae, although one of the smaller families of Coleoptera, now number about 1400 species; of these about one-half belong to the great genus Cicindela, to which our four British representatives of the Cicindelidae are all assigned. There is no general work of much consequence on this important family, and its classification is not thoroughly established.[^[86]]

Fig. 91—Mouth-parts of tiger-beetles. A, Profile of Pogonostoma sp. (Madagascar): a, antenna; b, labial palp; c, maxillary palp; d, palpiform lobe of maxilla; e, mandible; f, labrum. B, Section of head of Manticora maxillosa (South Africa): a, front of upper part of head-capsule; b, gula; c, tentorium; d, eye; e, labrum; f, left mandible; g, maxilla; h, maxillary palp; i, labial palp; k, support of this palp; l, labium.

Tiger-beetles display considerable variety of structure, especially as regards the mouth, which exhibits very remarkable developments of the palpi and labrum (Fig. 91). The tiger-beetles, like most other Insects that capture living prey, do not consume their victims entire, but subsist chiefly on the juices they squeeze out of them; the hard and innutritious parts are rejected after the victim has been thoroughly lacerated and squeezed; the mouth forms both trap and press; the palpi spread out in order to facilitate the rapid engulfing of a victim, then close up under it and help to support it in the mouth; while the labrum above closes the cavity in the other direction. The mouth itself is a large cavity communicating very freely with the exterior, but so completely shut off from the following parts of the alimentary canal that it is difficult to find the orifice of communication; the labium being much modified to form the posterior wall. For the capture of the prey, always living but of various kinds, a mechanism with great holding power and capable of rapid action is required. The mouth of the terrestrial Manticora (Fig. 91, B), exhibits great strength; some of the chitinous parts are extremely thick, the mandibles are enormous, the palpi, however, are comparatively low in development. In the arboreal genus Pogonostoma the palpary structures (Fig. 91, A) attain a development scarcely equalled elsewhere in the Insect world. The great majority of the Cicindelidae are inhabitants of the warmer, or of the tropical regions of the world, and very little is known as to their life-histories; they show great diversity in their modes of hunting their prey. Some are wingless; others are active on the wing; and of both of these divisions there are forms that are found only on trees or bushes. Some, it is believed, frequent only the mounds of Termites. The characteristic feature common to all is great activity and excessive wariness. The genus Pogonostoma, to which we have already alluded, is confined to Madagascar, where the species are numerous, but are rare in collections on account of the difficulty of securing them. Raffray informs us that certain species frequent the trunks of trees, up which they run in a spiral manner on the least alarm. The only way he could obtain specimens was by the aid of an assistant; the two approached a tree very quietly from opposite sides, and when near it, made a rush, and joined hands as high up the trunk as they could, so as to embrace the tree, when the Pogonostoma fell to the ground and was captured.

Fig. 92—Leistus spinibarbis. A, Larva (after Schiödte); B, imago. Britain.

Fam. 5. Carabidae (Ground-beetles).—Clypeus not extending laterally in front of the antennae. Maxillae with the outer lobe destitute of an articulated hook at the tip. Antennae covered (except the basal joints) with a minute pubescence. Hind legs not very different from the middle pair, formed for running, as usual in beetles. This is one of the largest and most important of the families of Coleoptera, including as it does 12,000 or 13,000 described species. In this country Carabidae are nearly entirely terrestrial in habits, and are scarcely ever seen on the wing; many of the species indeed have merely rudimentary wings; in the tropics there are, however, many arboreal forms that take wing with more or less alertness. The larvae (Fig. 92, A) are usually elongate in form and run freely; they may be known by their tarsi ending in two claws, by the exserted, sharp, calliper-like mandibles, by the body ending in two processes (sometimes jointed) and a tube of varying length projecting backwards. The pupae usually have the hind pair of legs so arranged that the tips of the tarsi project behind, beyond the extremity of the body. The Carabidae are carnivorous and predaceous both as larvae and perfect Insects; they attack living Insects, worms, or other small, soft creatures, but do not disdain dead specimens. Some species of Carabus, found in North Africa where snails abound, are specially formed for attacking these molluscs, having the head long and slender so that it can be thrust into the shell of the snail. A few species have been detected eating growing corn, and even the young seeds of some Umbelliferae; these belong chiefly to the genera Harpalus, Zabrus, and Amara. Some species of the abundant genera Pterostichus and Harpalus, are said to be fond of ripe strawberries. The most anomalous forms of Carabidae are the Pseudomorphides, a sub-family almost peculiar to Australia, the members of which live under bark, and have but little resemblance to other Carabids owing to their compact forms and continuous outlines. The genus Mormolyce is one of the wonders of the Insect world on account of the extraordinary shape of its members; the sides of the elytra form large crinkled expansions, and the head is unusually elongate. These Insects live on the underside of fallen trees in the Malay Archipelago and Peninsula; no reason whatever can be at present assigned for their remarkable shape.

There are a considerable number of blind members of this family: some of them live in caverns; these belong chiefly to the genus Anophthalmus, species of which have been detected in the caves of the Pyrenees, of Austria, and of North America. It has been shown that the optic nerves and lobes, as well as the external organs of vision, are entirely wanting in some of these cave Carabidae; the tactile setae have, however, a larger development than usual, and the Insects are as skilful in running as if they possessed eyes. Anophthalmus is closely related to our British genus Trechus, the species of which are very much given to living in deep crevices in the earth, or under large stones, and have some of them very small eyes. In addition to these cavernicolous Anophthalmus, other blind Carabidae have been discovered during recent years in various parts of the world, where they live under great stones deeply embedded in the earth; these blind lapidicolous Carabidae are of extremely minute size and of most sluggish habits; the situations in which they are found suggest that many successive generations are probably passed under the same stone. Not a single specimen has ever been found above ground. The minute Carabids of the genus Aëpus, that pass a large part of their lives under stones below high-water mark (emerging only when the tide uncovers them), on the borders of the English Channel and elsewhere, are very closely allied to these blind Insects, and have themselves only very small eyes, which, moreover, according to Hammond and Miall, are covered in larger part by a peculiar shield.[^[87]] A few Carabidae, of the genera Glyptus and Orthogonius, are believed to live in the nests of Termites. Savage found the larva of G. sculptilis in the nests of Termes bellicosus; it has been described by Horn, and is said to bear so great a resemblance to young queens of the Termites as to have been mistaken for them.[^[88]] Mr. Haviland found Rhopalomelus angusticollis in Termites' nests in South Africa. Péringuey states that it emits a very strong and disagreeable odour. It is probable that it preys on the Termites, and this also is believed to be the habit of the Ceylonese Helluodes taprobanae. Some species of the Mediterranean genus Siagona stridulate by means of a file on the under surface of the prothorax, rubbed by a striate area, adapted in form, on the anterior femora.

A valuable memoir on the classification of this important family is due to the late Dr. G. H. Horn;[^[89]] he arranges Carabidae in three sub-families; we think it necessary to add a fourth for Mormolyce:

1. Middle coxal cavities enclosed externally by the junction of the meso- and meta-sternum; neither epimeron nor episternum attaining the cavity.

Head beneath, with a deep groove on each side near the eye for the reception of the antennae or a part thereof. .......... Sub-fam. 3. Pseudomorphides.

Head without antennal grooves. .......... Sub-fam. 2. Harpalides.

2. Middle coxal cavities attained on the outside by the tips of the episterna and epimera. .......... Sub-fam. 4. Mormolycides.

3. Middle coxal cavities attained on the outside by the tips of the epimera, but not by those of the episterna. .......... Sub-fam. 1. Carabides.

These four sub-families are of extremely different extent and nature. The Harpalides are the dominant forms, and include upwards of 10,000 known species; while the various tribes into which the sub-family is divided include, as a rule, each many genera; the Carabides are next in importance, with upwards of 2000 species, but are divided into a comparatively large number of tribes, each of which averages a much smaller number of genera than do the tribes of Harpalides; Pseudomorphides includes only about 100 species; and Mormolycides consists of the single genus Mormolyce with three species.

Fam. 6. Amphizoidae.—Antennae destitute of pubescence: outer lobe of maxilla not jointed; metasternum with a short transverse impressed line on the middle behind. Hind legs slender, not formed for swimming. This family is limited to the genus Amphizoa; the species of which may be briefly described as lowly organised Carabidae that lead an aquatic life. The geographical distribution is highly remarkable, there being but three species, two of which live in Western North America, the third in Eastern Tibet. The habits of American Amphizoa are known; they pass a life of little activity in very cold, rapid streams; they do not swim, but cling to stones and timber. The larva was recently discovered in Utah by Messrs. Hubbard and Schwarz:[^[90]] it has the same habits as the perfect Insect, and in general form resembles the larvae of the genus Carabus; but it has no terminal tube to the body, the abdomen consisting of eight segments and a pair of short terminal appendages; the spiracles are obsolete, with the exception of a pair placed near to one another at the termination of the eighth abdominal segment. As regards the mouth this larva is Carabid, as regards the abdomen and stigmata Dytiscid of a primitive type.

Fig. 93—Amphizoa lecontei. North America. A, Larva; B, imago.

Fig. 94—Pelobius tardus. Britain. A, Young larva; B, adult larva; C, imago. (A and B after Schiödte.)

Fam. 7. Pelobiidae.—Antennae destitute of pubescence: outer lobe of maxilla jointed, metasternum with a short transverse impressed line on the middle behind. Hind legs rather slender, formed for swimming, the tarsi longer than the tibiae. This family is limited to the one genus Pelobius (Hygrobia of some authors). Like Amphizoa, to which it is in several respects analogous, it has a singular geographical distribution; there are only four known species, one lives in Britain and the Mediterranean region, one in Chinese Tibet, two in Australia. Pelobius may be briefly described as a Carabid adapted to a considerable extent for living in and swimming about in water; differing thus from Amphizoa, which has no special adaptation for swimming. The larva of Pelobius is remarkable; it breathes by means of branchial filaments on the under surface of the body, the spiracles being present, though those of the abdomen are very minute and the others small. The head is very large, the mandibles are not tube-like, the food being taken after the manner of the Carabidae; the 8th abdominal segment ends in three long processes; the small 9th segment is retracted beneath them. The adult Pelobius tardus is remarkable for its loud stridulation. The sound is produced by an apparatus described correctly by Charles Darwin;[^[91]] there is a file on the inside of the wing-cases, and the Insect turns up the tip of the abdomen and scrapes the file therewith. The Insects are called squeakers in the Covent Garden market, where they are sold.

Fig. 95—Cnemidotus caesus. England. A, Imago; B, larva, highly magnified. (After Schiödte.)

Fam. 8. Haliplidae.—Antennae bare, ten-jointed; metasternum marked by a transverse line; posterior coxae prolonged as plates, covering a large part of the lower surface of the abdomen; the slender, but clubbed, hind femora move between these plates and the abdomen. The Haliplidae are aquatic, and are all small, not exceeding four or five millimetres in length. The ventral plates are peculiar to the Insects of this family, but their function is not known. The larvae are remarkable on account of the fleshy processes disposed on their bodies; but they exhibit considerable variety in this respect; their mandibles are grooved so that they suck their prey. In the larva of Haliplus, according to Schiödte, there are eight pairs of abdominal spiracles, but in Cnemidotus (Fig. 95, B), there are no spiracles, and air is obtained by means of a trachea traversing each of the long filaments. The Insects of these two genera are so similar in the imaginal instar that it is well worthy of note that their larvae should be distinguished by such important characters. Haliplidae is a small family consisting of three genera, having about 100 species; it is very widely distributed. We have 13 species in Britain, all the genera being represented.

Fig. 96.—Cybister roeseli (= laterimarginalis De G.) Europe. A, Larva (after Schiödte); B, ♂ imago.

Fam. 9. Dytiscidae (Water-beetles).—Antennae bare; hind legs formed for swimming, not capable of ordinary walking: metasternum without a transverse line across it; behind closely united with the extremely large coxae. Outer lobe of maxilla forming a two-jointed palpus. The Dytiscidae, or true water-beetles, are of interest because—unlike the aquatic Neuroptera—they exist in water in both the larval and imaginal instars; nevertheless there is reason for supposing that they are modified terrestrial Insects: these reasons are (1) that in their general organisation they are similar to the Carabidae, and they drown more quickly than the majority of land beetles do; (2) though the larvae are very different from the larvae of terrestrial beetles, yet the imaginal instars are much less profoundly changed, and are capable of existing perfectly well on land, and of taking prolonged flights through the air; (3) the pupa is, so far as known, always terrestrial. The larvae and imagos are perfectly at home in the water, except that they must come to the surface to get air. Some of them are capable, however, when quiescent, of living for hours together beneath the water, but there appears to be great diversity in this respect.[^[92]] The hind pair of legs is the chief means of locomotion. These swimming-legs (Fig. 97) are deserving of admiration on account of their mechanical perfection; this, however, is exhibited in various degrees, the legs in the genera Dytiscus and Hydroporus being but slender, while those of Cybister are so broad and powerful, that a single stroke propels the Insect for a considerable distance.

Fig. 97—Hind- or swimming-leg of Cybister tripunctatus. A, The whole leg detached; B, the movable parts in the striking position. a, Coxa; b, trochanter; c, femur; d, tibia; e, last joint of tarsus.

The wing-cases fit perfectly to the body, except at the tip, so as to form an air-tight space between themselves and the back of the Insect; this space is utilised as a reservoir for air. When the Dytiscus feels the necessity for air it rises to the surface and exposes the tip of the body exactly at the level of the water, separating at the same time the abdomen from the wing-cases so as to open a broad chink at the spot where the parts were, during the Insect's submersion, so well held together as to be air- and water-tight. The terminal two pairs of spiracles are much enlarged, and by curving the abdomen the beetle brings them into contact with the atmosphere; respiration is effected by this means as well as by the store of air carried about under the wing cases. The air that enters the space between the elytra and body is shut in there when the Insect closes the chink and again dives beneath the water. The enlargement of the terminal stigmata in Dytiscus is exceptional, and in forms more highly organised in other respects, such as Cybister, these spiracles remain minute; the presumption being that in this case respiration is carried on almost entirely by means of the supply the Insect carries in the space between the elytra and the base of the abdomen.[^[93]] The structure of the front foot of the male Dytiscus, and of many other water-beetles, is highly remarkable, the foot being dilated to form a palette or saucer, covered beneath by sucker-like structures of great delicacy and beauty; by the aid of these the male is enabled to retain a position on the female for many hours, or even days, together. Lowne has shown that the suckers communicate with a sac in the interior of the foot containing fluid, which exudes under pressure. As the portions of the skeleton of the female on which these suckers are brought to bear is frequently covered with pores, or minute pits, it is probable that some correlation between the two organisms is brought about by these structures. The females in many groups of Dytiscidae bear on the upper surface of the body a peculiar sculpture of various kinds, the exact use of which is unknown; in many species there are two forms of the female, one possessing this peculiar sculpture, the other nearly, or quite, without it. The larvae of Dytiscidae differ from those of Carabidae chiefly by the structure of the mouth and of the abdomen. They are excessively rapacious, and are indeed almost constantly engaged in sucking the juices of soft and small aquatic animals, by no means excluding their own kind. The mode of suction is not thoroughly known, but so far as the details have been ascertained they are correctly described, in the work on aquatic Insects, by Professor Miall, we have previously referred to; the mandibles are hollow, with a hole near the tip and another at the base, and being sharp at the tips are thrust into the body of a victim, and then by their closure the other parts of the mouth, which are very beautifully constructed for the purpose, are brought into fitting mechanical positions for completing the work of emptying the victim. Nagel states that the larva of Dytiscus injects a digestive fluid into the body of its victim, and that this fluid rapidly dissolves all the more solid parts of the prey, so that the rapacious larva can easily absorb all its victim except the insoluble outer skin. The abdomen consists of only eight segments, and a pair of terminal processes; the stigmata are all more or less completely obsolete—according to species—with the exception of the pair on the eighth segment at the tip of the body; the terminal segments are frequently fringed with hairs, that serve not only as means of locomotion, but also to float the pair of active stigmata at the surface when the creature rises to get air. Although the larvae of Dytiscidae are but little known, yet considerable diversity has already been found. Those of Hyphydrus and some species of Hydroporus have the front of the head produced into a horn, which is touched by the tips of the mandibles.

Dytiscidae are peculiar inasmuch as they appear to flourish best in the cooler waters of the earth. Lapland is one of the parts of Europe richest in Dytiscidae, and the profusion of species in the tropics compared with those of Europe is not nearly so great as it is in the case of most of the other families of Coleoptera. About 1800 species are at present known, and we have rather more than 100 species in Britain.[^[94]]

Series III. Polymorpha.

Antennae frequently either thicker at the tip (clavicorn) or serrate along their inner edge (serricorn); but these characters, as well as the number of joints in the feet and other points, are very variable.

Upwards of fifty families are placed in this series; many of these families are of very small extent, consisting of only a few species; other families of the series are much larger, so that altogether about 40,000 species—speaking broadly, about one-fourth of the Coleoptera—are included in the series. We have already (p. [189]) alluded to the fact that it is formed by certain conventional series, Clavicornia, Serricornia, etc. united, because it has hitherto proved impossible to define them.

Fam. 10. Paussidae.—Antennae of extraordinary form, usually two-jointed, sometimes six- or ten-jointed. Elytra elongate, but truncate behind, leaving the pygidium exposed. Tarsi five-jointed. The Paussidae have always been recognised as amongst the most remarkable of beetles, although they are of small size, the largest attaining scarcely half an inch in length. They are found only in two ways; either in ants' nests, or on the wing at night. They apparently live exclusively in ants' nests, but migrate much. Paussidae usually live in the nests of terrestrial ants, but they have been found in nests of Cremastogaster in the spines of Acacia fistulosa. They have the power of discharging, in an explosive manner, a volatile caustic fluid from the anus, which is said by Loman to contain free iodine. Their relations to the ants are at present unexplained, though much attention has been given to the subject. When observed in the nests they frequently appear as if asleep, and the ants do not take much notice of them.

Fig. 98.—Paussus cephalotes ♂. El Hedjaz. (After Raffray.)

On other occasions the ants endeavour to drag them into the interior of the nest, as if desirous of retaining their company: the Paussus then makes no resistance to its hosts; if, however, it be touched, even very slightly, by an observer, it immediately bombards: the ants, as may be imagined, do not approve of this, and run away. Nothing has ever been observed that would lead to the belief that the ants derive any benefit from the presence of the Paussi, except that these guests bear on some part of the body—frequently the great impressions on the pronotum—patches of the peculiar kind of pubescence that exists in many other kinds of ants'-nest beetles, and is known in some of them to secrete a substance the ants are fond of, and that the ants have been seen to lick the beetles. On the other hand, the Paussi have been observed to eat the eggs and larvae of the ants. The larva of Paussus is not known,[^[95]] and Raffray doubts whether it lives in the ants' nests. There are about 200 species of Paussidae known, Africa, Asia and Australia being their chief countries; one species, P. favieri, is not uncommon in the Iberian peninsula and South France, and a single species was formerly found in Brazil. The position the family should occupy has been much discussed; the only forms to which they make any real approximation are Carabidae, of the group Ozaenides, a group of ground beetles that also crepitate. Burmeister and others have therefore placed the Paussidae in the series Adephaga, but we follow Raffray's view (he being the most recent authority on the family),[^[96]] who concludes that this is an anomalous group not intimately connected with any other family of Coleoptera, though having more affinity to Carabidae than to anything else. The recently discovered genus Protopaussus has eleven joints to the antennae, and is said to come nearer to Carabidae than the previously known forms did, and we may anticipate that a more extensive knowledge will show that the family may find a natural place in the Adephaga. The description of the abdomen given by Raffray is erroneous; in a specimen of the genus Arthropterus the writer has dissected, he finds that there are five ventral segments visible along the middle, six at the sides, as in the families of Adephaga generally. There is said to be a great difference in the nervous systems of Carabidae and Paussidae, but so little is known on this point that we cannot judge whether it is really of importance.

Fig. 99.—A, Larva of Gyrinus (after Schiödte); B, under side of Gyrinus sp. (after Ganglbauer). 1, Prosternum; 2, anterior coxal cavity; 3. mesothoracic episternum; 4, mesoepimeron; 5, mesosternum; 6, metathoracic episternum; 7, middle coxal cavity; 8, metasternum; 9, hind coxa; 10, ventral segments. [N.B.—The first ventral segment really consists, at each side, of two segments united; this may be distinctly seen in many Gyrinidae.]

Fam. 11. Gyrinidae (Whirligig beetles).—Antennae very short; four eyes; middle and hind legs forming short broad paddles; abdomen with six segments visible along the middle, seven along each side. These Insects are known to all from their habit of floating lightly on the surface of water, and performing graceful complex curves round one another without colliding; sometimes they may be met with in great congregations. They are admirably constructed for this mode of life, which is comparatively rare in the Insect world; the Hydrometridae amongst the bugs, and a small number of different kinds of Diptera, being the only other Insects that are devoted to a life on the surface of the waters. Of all these, Gyrinidae are in their construction the most adapted for such a career. They are able to dive to escape danger, and they then carry with them a small supply of air, but do not stay long beneath the surface. Their two hind pairs of legs are beautifully constructed as paddles, expanding mechanically when moved in the backward direction, and collapsing into an extremely small space directly the resistance they meet with is in the other direction. The front legs of these Insects are articulated to the thorax in a peculiar direction so that their soles do not look downwards but towards one another; hence the sensitive adhesive surface used during coupling is placed on the side of the foot, forming thus a false sole: a remarkable modification otherwise unknown in Insects. They breathe chiefly by means of the very large metathoracic spiracles.

The larvae (Fig. 99, A) are purely aquatic, and are highly modified for this life, being elongate creatures, with sharp, mandibles and nine abdominal segments, each segment bearing on each side a trachea branchia; these gills assist to some extent in locomotion. The stigmata are quite obsolete, but the terminal segment bears four processes, one pair of which may be looked on as cerci, the other as a pair of gills corresponding with the pair on each of the preceding segments. The mandibles are not suctorial, but, according to Meinert, possess an orifice for the discharge of the secretion of a mandibular gland. Gyrinidae are chiefly carnivorous in both the larval and imaginal instars. Fully 300 species are known; they are generally distributed, though wanting in most of the islands of the world except those of large size. The finest forms are the Brazilian Enhydrus and the Porrorhynchus of tropical Asia.[^[97]] In Britain we have nine species, eight of Gyrinus, one of Orectochilus; the latter form is rarely seen, as it hides during the day, and performs its rapid gyrations at night.

The Gyrinidae are one of the most distinct of all the families of Coleoptera: by some they are associated in the Adephagous series; but they have little or no affinity with the other members thereof. Without them the Adephaga form a natural series of evidently allied families, and we consider it a mistake to force the Gyrinidae therein because an objection is felt by many taxonomists to the maintenance of isolated families. Surely if there are in nature some families allied and others isolated, it is better for us to recognise the fact, though it makes our classifications look less neat and precise, and increases the difficulty of constructing "tables."

Fam. 12. Hydrophilidae.—Tarsi five-jointed, the first joint in many cases so small as to be scarcely evident: antennae short, of less than eleven joints, not filiform, but consisting of three parts, a basal part of one or two elongate joints, an intermediate part of two or more small joints, and an apical part of larger (or at any rate broader) joints, which are pubescent, the others being bare. Outer lobe of maxillae usually complex, but not at all palpiform, maxillary palpi often very long; the parts of the labium much concealed behind the mentum, the labial palpi very widely separated. Hind coxae extending the width of the body, short, the lamina interior small in comparison with the lamina exterior. Abdomen of five visible segments. The Hydrophilidae are an extensive family of beetles, unattractive in colours and appearance, and much neglected by collectors. A large part of the family live in water, though most of them have only feeble powers of aquatic locomotion, and the beetles appear chiefly to devote their attention to economising the stock of air each individual carries about. The best known forms of the family are the species of Hydrophilus. They are, however, very exceptional in many respects, and are far more active and predaceous than most of the other forms. Much has been written about Hydrophilus piceus, one of the largest of British beetles. This Insect breathes in a most peculiar manner: the spiracles are placed near bands of delicate pubescence, forming tracts that extend the whole length of the body, and in this particular species cover most of the under surface of the body; these velvety tracts retain a coating of air even when the Insect is submerged and moves quickly through the water. It would appear rather difficult to invent a mechanism to supply these tracts with fresh air without the Insect leaving the water; but nevertheless such a mechanism is provided by the antennae of the beetle, the terminal joints of which form a pubescent scoop, made by some longer hairs into a funnel sufficiently large to convey a bubble of air. The Insect therefore rises to the surface, and by means of the antennae, which it exposes to the air, obtains a supply with which it surrounds a large part of its body; for, according to Miall, it carries a supply on its back, under the elytra, as well as on its ventral surface. From the writer's own observations, made many years ago, he inclines to the opinion that the way in which the Hydrophilus uses the antennae to obtain air varies somewhat according to circumstances.

Many of the members of the sub-family Hydrophilides construct egg-cocoons. In the case of Hydrophilus piceus, the boat-like structure is provided with a little mast, which is supposed by some to be for the purpose of securing air for the eggs. Helochares and Spercheus (Fig. 100) carry the cocoon of eggs attached to their own bodies. Philydrus constructs, one after the other, a number of these egg-bags, each containing about fifteen eggs, and fixes each bag to the leaf of some aquatic plant; the larvae as a rule hatch speedily, so that the advantage of the bag is somewhat problematic.

Fig. 100.—Spercheus emarginatus ♀. Britain. A, Upper surface of beetle; B, under surface of abdomen, with the egg-sac ruptured and some of the eggs escaping.

The larvae of the aquatic division of the family have been to a certain extent studied by Schiödte and others; those of the Sphaeridiides—the terrestrial group of the family—are but little known. All the larvae seem to be predaceous and carnivorous, even when the imago is of vegetable-feeding habits; and Duméril states that in Hydrous caraboides the alimentary canal undergoes a great change at the period of metamorphosis, becoming very elongate in the adult, though in the larva it was short. The legs are never so well developed as they are in the Adephaga, the tarsi being merely claw-like or altogether wanting; the mandibles are never suctorial. The respiratory arrangements show much diversity. In most of the Hydrophilides the process is carried on by a pair of terminal spiracles on the eighth abdominal segment, as in Dytiscidae, and these are either exposed or placed in a respiratory chamber. In Berosus the terminal stigmata are obsolete, and the sides of the body bear long branchial filaments. Cussac says that in Spercheus (Fig. 101) there are seven pairs of abdominal spiracles, and that the larva breathes by presenting these to the air;[^[98]] but Schiödte states that in this form there are neither thoracic nor abdominal spiracles, except a pair placed in a respiratory chamber on the eighth segment of the abdomen, after the manner described by Miall as existing in Hydrobius. No doubt Cussac was wrong in supposing the peculiar lateral abdominal processes to be stigmatiferous. In Berosus there are patches of aëriferous, minute pubescence on the body. The pupae of Hydrophilides repose on the dorsal surface, which is protected by spinous processes on the pronotum, and on the sides of the abdomen.

We have already remarked that this is one of the most neglected of the families of Coleoptera, and its classification is not satisfactory. It is usually divided into Hydrophilides and Sphaeridiides. The Sphaeridiides are in large part terrestrial, but their separation from the purely aquatic Hydrophilides cannot be maintained on any grounds yet pointed out. Altogether about 1000 species of Hydrophilidae are known, but this probably is not a tenth part of those existing. In Britain we have nearly ninety species. Some taxonomists treat the family as a series with the name Palpicornia. The series Philhydrida of older authors included these Insects and the Parnidae and Heteroceridae.

Fig. 101—Larva of Spercheus emarginatus. (After Schiödte).

Fam. 13. Platypsyllidae.—This consists of a single species. It will be readily recognised from Fig. 102, attention being given to the peculiar antennae, and to the fact that the mentum is trilobed behind. This curious species has been found only on the beaver. It was first found by Ritsema on American beavers (Castor canadensis) in the Zoological Gardens at Amsterdam, but it has since been found on wild beavers in the Rhone in France; in America it appears to be commonly distributed on these animals from Alaska to Texas. It is very remarkable that a wingless parasite of this kind should be found in both hemispheres. The Insect was considered by Westwood to be a separate Order called Achreioptera, but there can be no doubt that it is a beetle. It is also admitted that it shows some points of resemblance with Mallophaga, the habits of which are similar. Its Coleopterous nature is confirmed by the larva, which has been described by both Horn and Riley.[^[99]] Little is known as to the food and life-history. Horn states that the eggs are placed on the skin of the beaver amongst the densest hair; the larvae move with a sinuous motion, like those of Staphylinidae. It has been suggested that the Insect feeds on an Acarid, Schizocarpus mingaudi; others have supposed that it eats scales of epithelium or hairs of the beaver.

Fig. 102—Platypsyllus castoris. A, Upper side; B, lower side, with legs of one side removed; C, antenna. (After Westwood.)

Fig. 103—Leptinus testaceus. Britain.

Fam. 14. Leptinidae.—Antennae rather long, eleven-jointed, without club, but a little thicker at the extremity. Eyes absent or imperfect. Tarsi five-jointed. Elytra quite covering abdomen. Mentum with the posterior angles spinously prolonged. A family of only two genera and two species. Their natural history is obscure, but is apparently of an anomalous nature; the inference that may be drawn from the little that is known being that they are parasitic on mammals. There is little or nothing in their structure to indicate this, except the condition of blindness; and until recently the Insects were classified amongst Silphidae. Leptinus testaceus (Fig. 103) is a British Insect, and besides occurring in Europe is well known in North America. In Europe it has been found in curious places, including the nests of mice and bumble-bees. In America it has been found on the mice themselves by Dr. Ryder, and by Riley in the nests of a common field-mouse, together with its larva, which, however, has not been described. The allied genus Leptinillus is said by Riley to live on the beaver, in company with Platypsyllus.[^[100]] It has been suggested that the natural home of the Leptinus is the bee's nest, and that perhaps the beetle merely makes use of the mouse as a means of getting from one nest of a bumble-bee to another.

Fam. 15. Silphidae.—The mentum is usually a transverse plate, having in front a membranous hypoglottis, which bears the exposed labial palpi, and immediately behind them the so-called bilobed ligula. The anterior coxae are conical and contiguous: prothoracic epimera and episterna not distinct. Visible abdominal segments usually five, but sometimes only four, or as many as seven. Tarsi frequently five-jointed, but often with one joint less. Elytra usually covering the body and free at the tips, but occasionally shorter than the body, and even truncate behind so as to expose from one to four of the dorsal plates; but there are at least three dorsal plates in a membranous condition at the base of the abdomen. These beetles are extremely diverse in size and form, some being very minute, others upwards of an inch long, and there is also considerable range of structure. In this family are included the burying-beetles (Necrophorus), so well known from their habit of making excavations under the corpses of small Vertebrates, so as to bury them. Besides these and Silpha, the roving carrion-beetles, the family includes many other very different forms, amongst them being the larger part of the cave-beetles of Europe and North America. These belong mostly to the genera Bathyscia in Europe, and Adelops in North America; but of late years quite a crowd of these eyeless cave-beetles of the group Leptoderini have been discovered, so that the European catalogue now includes about 20 genera and 150 species. The species of the genus Catopomorphus are found in the nests of ants of the genus Aphaenogaster in the Mediterranean region. Scarcely anything is known as to the lives of either the cave-Silphidae or the myrmecophilous forms.

The larvae of several of the larger forms of Silphidae are well known, but very little has been ascertained as to the smaller forms. Those of the burying-beetles have spiny plates on the back of the body, and do not resemble the other known forms of the family. The rule is that the three thoracic segments are well developed, and that ten abdominal segments are also distinct; the ninth abdominal segment bears a pair of cerci, which are sometimes elongate. Often the dorsal plates are harder and better developed than is usual in Coleopterous larvae. This is especially the case with some that are endowed with great powers of locomotion, such as S. obscura (Fig. 104). The food of the larvae is as a rule decomposing animal or vegetable matter, but some are predaceous, and attack living objects. The larger Silpha larvae live, like the Necrophorus, on decomposing animal matter, but run about to seek it; hence many specimens of some of these large larvae may sometimes be found amongst the bones of a very small dead bird. We have found the larva and imago of S. thoracica in birds' nests containing dead nestlings. S. atrata and S. laevigata make war on snails. S. lapponica enters the houses in Lapland and ravages the stores of animal provisions. S. opaca departs in a very decided manner from the habits of its congeners, as it attacks beetroot and other similar crops in the growing state; it is sometimes the cause of serious loss to the growers of beet. The larvae of the group Anisotomides are believed to be chiefly subterranean in habits; that of A. cinnamomea feeds on the truffle, and the beetle is known as the truffle-beetle.

Fig. 104—A, Larva of Silpha obscura. Europe. (After Schiödte). B, Ptomaphila lacrymosa, Australia.

The number of species of Silphidae known must be at present nearer 900 than 800. Of these an unusually large proportion belong to the European and North American regions; Silphidae being apparently far from numerous in the tropics. Rather more than 100 species are natives of Britain. The family reappears in considerable force in New Zealand, and is probably well represented in South Australia and Tasmania. The most remarkable form known is perhaps the Australian genus Ptomaphila (Fig. 104, B). The classification of the family is due to Dr. Horn.[^[101]] The only change of importance that has since been suggested is the removal of Sphaerites from this family to Synteliidae. Anisotomidae and Clambidae have been considered distinct families, but are now included in Silphidae.

Fam. 16. Scydmaenidae.—Minute Insects allied to Silphidae, but with the hind coxae separated, and the facets of the eyes coarser; the tarsi are five-jointed; the number of visible abdominal segments is six. These small beetles are widely spread over the earth's surface, and about 700 species are now known, of which we have about a score in Britain; many live in ants' nests, but probably usually rather as intruders than as guests that have friendly relations with their hosts. Nothing is known as to their life-histories, but the food of the imago, so far as is known, consists of Acari. Mastigus is a very aberrant form, found in moss and dead leaves in Southern Europe. By means of Brathinus the family is brought very near to Silphidae; Casey, however, considers Brathinus to belong to Staphylinidae rather than to Scydmaenidae. The South European Leptomastax is remarkable on account of the slender, long, sickle-shaped mandibles. The Oriental genus Clidicus is the largest and most remarkable form of the family; it has a very slender neck to its broad head, and is more than a quarter of an inch long.

Fam. 17. Gnostidae.—Minute Insects with three-jointed antennae, five-jointed tarsi, and three apparent ventral segments, the first of which, however, is elongate, and consists of three united plates. Elytra entirely covering the after-body. The family consists of two species which have been found in the nests of ants, of the genus Cremastogaster, in Brazil.[^[102]]

Fam. 18. Pselaphidae.—Very small Insects; the elytra much abbreviated, usually leaving as much as half the abdomen uncovered; the maxillary palpi usually greatly developed, and of a variety of remarkable forms; the segments of the abdomen not more than seven in number, with little or no power of movement. Tarsi with not more than three joints. These small Coleoptera mostly live in the nests of ants, and present a great diversity of extraordinary shapes, and very peculiar structures of the antennae and maxillary palpi. Owing to the consolidation of some of its segments, the abdomen frequently appears to have less than the usual number. In the curious sub-family Clavigerides, the antennae may have the joints reduced to two or even, to all appearance, to one; the tarsi suffer a similar reduction. There are about 2500 species of Pselaphidae known; many of them have never been found outside the ants' nests; very little, however, is known as to their natural history. It is certain that some of them excrete, from little tufts of peculiar pubescence, a substance that the ants are fond of. The secretory patches are found on very different parts of the body and appendages. Claviger testaceus is fed by the ants in the same way as these social Insects feed one another; the Claviger has also been seen to eat the larvae of the ants. They ride about on the backs of the ants when so inclined. The family is allied to Staphylinidae, but is easily distinguished by the rigid abdomen. Only one larva—that of Chennium bituberculatum—is known. It appears to be very similar to the larvae of Staphylinidae. The best account of classification and structure is that given by M. Achille Raffray,[^[103]] who has himself discovered and described a large part of the known species.

Fam. 19. Staphylinidae.—Elytra very short, leaving always some of the abdominal segments exposed, and covering usually only two of the segments. Abdomen usually elongate, with ten dorsal, and seven or eight ventral segments; of the latter six or seven are usually exposed; the dorsal plates as hard as the ventral, except sometimes in the case of the first two segments; the segments very mobile, so that the abdomen can be curled upwards. The number of tarsal joints very variable, often five, but frequently as few as three, and not always the same on all the feet. Staphylinidae (formerly called Brachelytra or Microptera) is one of the most extensive of even the great families of Coleoptera; notwithstanding their diversity, they may in nearly all cases be recognised by the more than usually mobile and uncovered abdomen, combined with the fact that the parts of the mouth are of the kind we have mentioned in Silphidae. The present state of the classification of this family has been recently discussed by Ganglbauer.[^[104]]

Fig. 105—Staphylinidae. A, Larva of Philonthus nitidus. Britain. (After Schiödte.) B, Ocypus olens, Britain; C, tip of abdomen, of O. olens with stink-vessels.

At present about 9000 species are known, some of which are minute, while scarcely any attain a size of more than an inch in length, our common British black cock-tail, or "devil's coach-horse beetle," Ocypus olens, being amongst the largest. Though the elytra are short, the wings in many forms are as large as those of the majority of beetles; indeed many Staphylinidae are more apt at taking flight than is usual with Coleoptera; the wings when not in use are packed away under the short elytra, being transversely folded, and otherwise crumpled, in a complicated but orderly manner. It is thought that the power of curling up the abdomen is connected with the packing away of the wings after flight; but this is not the case: for though the Insect sometimes experiences a difficulty in folding the wings under the elytra after they have been expanded, yet it overcomes this difficulty by slight movements of the base of the abdomen, rather than by touching the wings with the tip. What the value of this exceptional condition of short elytra and corneous dorsal abdominal segments to the Insect may be is at present quite mysterious. The habits of the members of the family are very varied; many run with great activity; the food is very often small Insects, living or dead; a great many are found in fungi of various kinds, and perhaps eat them. It is in this family that we meet with some of the most remarkable cases of symbiosis, i.e. lives of two kinds of creatures mutually accommodated with good will. The relations between the Staphylinidae of the genera Atemeles and Lomechusa, and certain ants, in the habitations of which they dwell, are very interesting. The beetles are never found out of the ants' nests, or at any rate not very far from them. The most friendly relations exist between them and the ants: they have patches of yellow hairs, and these apparently secrete some substance with a flavour agreeable to the ants, which lick the beetles from time to time. On the other hand, the ants feed the beetles; this they do by regurgitating food, at the request of the beetle, on to their lower lip, from which it is then taken by the beetle (Fig. 82). The beetles in many of their movements exactly resemble the ants, and their mode of requesting food, by stroking the ants in certain ways, is quite ant-like. So reciprocal is the friendship that if an ant is in want of food, the Lomechusa will in its turn disgorge for the benefit of its host. The young of the beetles are reared in the nests by the ants, who attend to them as carefully as they do to their own young. The beetles have a great fondness for the ants, and prefer to sit amongst a crowd thereof; they are fond of the ants' larvae as food, and indeed eat them to a very large extent, even when their own young are receiving food from the ants. The larva of Lomechusa, as described by Wasmann (to whom we are indebted for most of our knowledge of this subject),[^[105]] when not fully grown, is very similar to the larvae of the ants; although it possesses legs it scarcely uses them: its development takes place with extraordinary rapidity, two days, at most, being occupied in the egg, and the larva completing its growth in fourteen days. Wasmann seems to be of opinion that the ants scarcely distinguish between the beetle-larvae and their own young; one unfortunate result for the beetle follows from this, viz. that in the pupal state the treatment that is suitable for the ant-larvae does not agree with the beetle-larvae: the ants are in the habit of digging up their own kind and lifting them out and cleaning them during their metamorphosis; they also do this with the beetle-larvae, with fatal results; so that only those that have the good fortune to be forgotten by the ants complete their development. Thus from thirty Lomechusa larvae Wasmann obtained a single imago, and from fifty Atemeles larvae not even one.

Many other Staphylinidae are exclusively attached to ants' nests, but most of them are either robbers, at warfare with the ants—as is the case with many species of Myrmedonia that lurk about the outskirts of the nests—or are merely tolerated by the ants, not receiving any direct support from them. The most remarkable Staphylinidae yet discovered are some viviparous species, forming the genera Corotoca and Spirachtha, that have very swollen abdomens, and live in the nests of Termites in Brazil:[^[106]] very little is, however, known about them. A very large and powerful Staphylinid, Velleius dilatatus, lives only in the nests of hornets and wasps. It has been supposed to be a defender of the Hymenoptera, but the recent observations of Janet and Wasmann make it clear that this is not the case: the Velleius has the power of making itself disagreeable to the hornets by some odour, and they do not seriously attack it. The Velleius finds its nutriment in larvae or pupae of the wasps that have fallen from their cells, or in other organic refuse.

The larvae of Staphylinidae are very similar to those of Carabidae, but their legs are less perfect, and are terminated only by a single claw; there is no distinct labrum. The pupae of some are obtected, i.e. covered by a secondary exudation that glues all the appendages together, and forms a hard coat, as in Lepidoptera. We have about 800 species of Staphylinidae in Britain, and it is probable that the family will prove one of the most extensive of the Order. It is probable that one hundred thousand species or even more are at present in existence.

Fam. 20. Sphaeriidae.—Very minute. Antennae eleven-jointed, clubbed. Tarsi three-jointed. Abdomen with only three visible ventral segments. This family includes only three or four species of Insects about 1⁄50 of an inch long. They are very convex, and be found walking on mud. S. acaroides occurs in our fens. Mr. Matthews considers that they are most nearly allied to Hydrophilidae.[^[107]]

Fig. 106—Trichopteryx fascicularis. Britain. A, Outline of perfect Insect; B, part of upper surface; C, larva from side; D, from above; E, pupa; F, wing; G, natural size of imago.

Fam. 21. Trichopterygidae.—Extremely minute: antennae clavicorn (basal and apical joints thicker than middle joints); tarsi three-jointed; elytra sometimes covering abdomen, in other cases leaving a variable number of segments exposed; wings fringed. This family comprises the smallest Insects; Nanosella fungi being only 1⁄100 of an inch long, while the largest Trichopterygid is only 1⁄12 of an inch. The small size is not accompanied by any degeneration of structure, the minute, almost invisible forms, having as much anatomical complexity as the largest Insects. Very little is known as to the natural history. Probably these Insects exist in all parts of the world, for we have about eighty species in England, and Trichopterygidae are apparently numerous in the tropics.[^[108]]

Fam. 22. Hydroscaphidae.—Extremely minute aquatic Insects, with elongate abdomen. Antennae eight-jointed. The other characters are much the same as those we have mentioned for Trichopterygidae. The family is not likely to come before the student, as only three or four species from Southern Europe and North America are known.[^[109]]

Fig. 107—A, Larva of Orthoperus brunnipes (after Perris); B, O. atomarius, perfect Insect. Britain.

Fam. 23. Corylophidae.—Minute beetles. Tarsi four-jointed, but appearing only three-jointed, owing to the hind joint being concealed by the emarginate (or notched) second joint. Six free ventral segments. Maxillae with only one lobe. Antennae of peculiar form. There are about 200 species of these little Insects, but the family is apparently represented all over the world, and will probably prove to be much more extensive. The peculiar larva of Orthoperus brunnipes was found abundantly by Perris in thatch in France. Mr. Matthews proposes to separate the genus Aphanocephalus as a distinct family, Pseudocorylophidae.[^[110]] In Corylophidae the wings are fringed with long hairs, as is the case in so many small Insects: the species of Aphanocephalus are rather larger Insects, and the wings are not fringed; the tarsi are only three-jointed.

Fam. 24. Scaphidiidae.—Front coxae small, conical; prothorax very closely applied to the after-body; hind coxae transverse, widely separated: abdomen with six or seven visible ventral plates; antennae at the extremity with about five joints that become gradually broader. Tarsi five-jointed. This family consists of a few beetles that live in fungi, and run with extreme rapidity; they are all small, and usually rare in collections. Some of the exotic forms are remarkable for the extreme tenuity and fragility of the long antennae, which bear fine hairs. The number of described species does not at present reach 200, but the family is very widely distributed. We have three or four species in Britain. All we know of the larvae is a description of that of Scaphisoma agaricinum by Perris;[^[111]] it is like the larva of Staphylinidae, there are nine abdominal segments in addition to a very short, broad pseudopod, and very short cerci. This larva feeds on agarics; it goes through its development in about three weeks; unlike the adult it is not very active.

Fig. 108—Scaphisoma agaricinum. Britain. A Larva (after Perris); B perfect Insect.

Fam. 25. Synteliidae.—Antennae clavicorn, with very large club: labium, with hypoglottis and the parts beyond it, exposed. Front coxae transverse. Abdomen with five visible ventral segments, and eight or nine dorsal, the basal four of which are semi-corneous. This family includes only five species; its classification has given rise to much difference of opinion. We have, after consideration of all its characters, established it as a distinct family[^[112]] allied to Silphidae. The perfect Insects live on the sap running from trees: but nothing else is known of their natural history. Like so many others of the very small families of aberrant Coleoptera, it has a very wide distribution; Syntelia being found in Eastern Asia and Mexico, while the sub-family Sphaeritides occurs, as a single species, in Europe and North America. The earlier instars are unknown.

Fig. 109—Syntelia westwoodi. Mexico. (From Biol. Centr. Amer.)

Fig. 110—Platysoma depressum. Europe. A, Larva (after Schiödte); B, perfect Insect.

Fam. 26. Histeridae.—Very compact beetles, with very hard integument, short, bent antennae, with a very compact club: no hypoglottis. Elytra closely applied to body, but straight behind, leaving two segments exposed. Abdomen with five visible ventral segments; with seven dorsal segments, all hard. Front coxae strongly transverse, hind coxae widely separated. The extremely compact form, and hard integument, combined with the peculiar antennae—consisting of a long basal joint, six or seven small joints, and then a very solid club of three joints covered with minute pubescence—render these Insects unmistakable. The colour is usually shining black, but there are numerous departures from this. The way in which these Insects are put together so as to leave no chink in their hard exterior armour when in repose, is very remarkable. The mouth-parts are rather highly developed, and the family is entitled to a high rank; it consists at present of about 2000 species;[^[113]] in Britain we have about 40. The larvae are without ocelli or labrum, but have well-developed mandibles, the second and third thoracic segments being short, the ninth segment of the abdomen terminal, with two distinctly jointed cerci.[^[114]] Histeridae are common in dung, in carcases, decaying fungi, etc., and some live under bark—these being, in the case of the genus Hololepta, very flat. Some are small cylinders, elaborately constructed, for entering the burrows of Insects in wood (Trypanaeus); a certain number are peculiar to ants' nests. Formerly it was supposed that the Insects were nourished on the decaying substances, but it is now believed, with good reason, that they are eminently predaceous, in both larval and imaginal instars, and devour the larvae of Diptera, etc. The relations of the ants'-nest forms to the ants is not made out, but it is highly probable that they eat the ants' larvae, and furnish the ants with some dainty relish. A few species live in company with Termites.

Fam. 27. Phalacridae.—Body very compact; elytra entirely covering it; apical joints of antennae rather broader, usually long; front coxae globular; posterior coxae contiguous; abdomen with five visible ventral segments; tarsi five-jointed, fourth joint usually small and obscure. This family consists entirely of small Insects: the tarsal structure is very aberrant, and is also diverse, so that the student may without careful observation pass the Insects over as having only four-jointed tarsi; their structure, so far as the front two pairs are concerned, being very nearly that of many Phytophaga. The larvae live in the heads of flowers, especially of the flowers of Compositae. Having bored their way down the stems, they pupate in earthen cocoons. Heeger[^[115]] says that he has observed in favourable seasons six generations; but the larvae die readily in unfavourable seasons, and are destroyed in vast numbers when the meadows are mowed. Seven years ago very little was known as to the family, and the list of their species scarcely amounted to 100, but now probably 300 are described. They occur in all parts of the world; we have fourteen in Britain.

Fig. 111—Olibrus bicolor. Europe. A, Larva (after Heeger); B, perfect Insect.

Fam. 28. Nitidulidae.—Antennae with a three-jointed club; all the coxae separated, and each with an external prolongation; tarsi five-jointed, the fourth joint smaller than any of the others; abdomen with five visible plates. These Insects are numerous, about 1600 species being at present known; in many of them the elytra nearly or quite cover the hind body, but in many others they are more or less abbreviated; in this case the Insects may be distinguished from Staphylinidae by the form of their antennae, and the smaller number of visible ventral segments. The habits are very varied, a great many are found on flowers, others are attracted by the sap of trees; some live in carcases. We have about 90 species in Britain; several forms of the genera Meligethes and Epuraea are among the most abundant of our beetles. Most of what is known as to the larvae is due to Perris; several have been found living in flowers; that of Pria haunts the flower of Solanum dulcamara at the junction of the stamens with the corolla; the larva of Meligethes aeneus sometimes occasions much loss by preventing the formation of seed in cultivated Cruciferae, such as Rape. These floricolous larvae grow with great rapidity, and then leave the flowers to pupate in the ground. The larva of Nitidula lives in carcases, though it is not very different from that of Pria. The larva of Soronia lives in fermenting sap, and has four hooks curving upwards at the extremity of the body. The curious genus Cybocephalus consists of some very small, extremely convex Insects that live in flowers in Southern Europe; they have only four joints to the tarsi. The perfect Insects of the group Ipides are remarkable from having a stridulating organ on the front of the head. The classification of the well-known genus Rhizophagus has given rise to much discussion; although now usually placed in Nitidulidae, we think it undoubtedly belongs to Cucujidae.

Fig. 112—Pria dulcamarae. Britain. A, Larva (after Perris); B, perfect Insect.

Fig. 113—Temnochila coerulea. Europe. A, Larva (after Perris); B, perfect Insect.

Fam. 29. Trogositidae.—Differs from Nitidulidae in the structure of the tarsi; these appear to be four-jointed, with the third joint similar in size and form to the preceding; they are, however, really five-jointed, an extremely short basal joint being present. Hind coxae contiguous. The club of each antenna is bilaterally asymmetric, and the sensitive surface is confined to certain parts of the joints. There are some 400 or 500 species of Trogositidae, but nearly all of them are exotic. The larvae (Fig. 113, A), are predaceous, destroying other larvae in large numbers, and it is probable that the imagos do the same. The larva of Tenebroides (better known as Trogosita) mauritanica is found in corn and flour, and is said to have sometimes been very injurious by eating the embryo of the corn, but it is ascertained that it also devours certain other larvae that live on the corn. This beetle has been carried about by commerce, and is now nearly cosmopolitan. Our three British species of Trogositidae represent the three chief divisions of the family, viz. Nemosomides, Temnochilides, Peltides; they are very dissimilar in form, the Peltides being oval, with retracted head. It is doubtful whether the members of the latter group are carnivorous in any of their stages; it is more probable that they live on the fungi they frequent. Peltidae stand as a distinct family in many works.[^[116]]

Fig. 114—Bitoma crenata. Britain. A, Larva (after Perris); B, perfect Insect.

Fam. 30. Colydiidae.—Antennae with a terminal club, tarsi four-jointed, none of the joints broad; front and middle coxae small, globose, embedded; hind coxae transverse, either contiguous or separated; five visible ventral segments, several of which have no movement. This is a family of interest, owing to the great diversity of form, to the extraordinary sculpture and clothing exhibited by many of its members, and to the fact that most of its members are attached to the primitive forests, and disappear entirely when these are destroyed. We have fifteen species in Britain, but about half of them are of the greatest rarity. There are about 600 species known at present; New Zealand has produced the greatest variety of forms; the forests of Teneriffe are rich in the genus Tarphius. The sedentary lives of many of these beetles are very remarkable; the creatures concealing themselves in the crannies of fungus-covered wood, and scarcely ever leaving their retreats, so that it is the rarest circumstance to find them at any distance from their homes. Langelandia anophthalma lives entirely underground and is quite blind, the optic lobes being absent. Some Colydiidae are more active, and enter the burrows of wood-boring Insects to destroy the larvae (Colydium). Few of the larvae are known; but all appear to have the body terminated by peculiar hard corneous processes, as is the case with a great variety of Coleopterous larvae that live in wood.[^[117]]

Fam. 31. Rhysodidae.—Tarsi four-jointed; mouth-parts covered by the large mentum; front tibiae notched on the inner edge. This family consists only of a few species, but is found nearly all over the world in the warm and temperate regions. In many of their characters they resemble the Adephaga, but are very different in appearance and in the mouth. The larvae are not known. Some authorities think these Insects should be placed in the series Adephaga,[^[118]] but it is more probable that they will prove to be amongst the numerous aberrant forms of Coleoptera that approach the various large natural series, without really belonging to them. The three families, Colydiidae, Cucujidae, and Rhysodidae, exhibit relations not only with other families of the Coleoptera Polymorpha, but also with most of the great series; Adephaga, Rhynchophora, Phytophaga, and Heteromera, being each closely approached.

Fam. 32. Cucujidae.—Tarsi five- or four-jointed, the first joint often short: antennae sometimes clubbed, but more often quite thin at the tip; front and middle coxae deeply embedded, globular, but with an angular prolongation externally; abdomen with five visible ventral segments, all movable. This family and the Cryptophagidae are amongst the most difficult families to define; indeed it is in this portion of the Clavicorns that an extended and thorough study is most urgently required. The Cucujidae include a great diversity of forms; they are mostly found under the bark of trees, and many of them are very flat. Many of the larvae are also very flat, but Ferris says there is great diversity in their structure: they are probably chiefly carnivorous. There are about 400 species described; we have nearly a score in Britain.

Fig. 115—Brontes planatus. Britain. A, Larva; B, pupa; C, perfect Insect. (A and B after Perris.)

The family Cupesidae of certain taxonomists must be at present associated with Cucujidae, though the first joint of the tarsus is elongate.

Fam. 33. Cryptophagidae.—Front and middle coxae very small and deeply embedded; antennae with enlarged terminal joints; tarsi five-jointed, the posterior sometimes in the male only four-jointed; abdomen with five visible ventral segments, capable of movement, the first much longer than any of the others. A small family composed of obscure forms of minute size, which apparently have mould-eating habits, though very little is known on this point and several of the genera (Antherophagus, Telmatophilus) are found chiefly on growing plants, especially in flowers. Although the imago of Antherophagus lives in flowers, yet the larva has only been found in the nests of bumble-bees; there is reason for believing that the imago makes use of the bee to transport it from the flowers it haunts to the nests in which it is to breed;[^[119]] this it does by catching hold of the bee with its mandibles when the bee visits the flower in which the beetle is concealed. It is strange the beetle should adopt such a mode of getting to its future home, for it has ample wings. We must presume that its senses and instinct permit it to recognise the bee, but do not suffice to enable it to find the bee's nest. Some of the larvae of the genus Cryptophagus are found abundantly in the nests of various wasps, where they are probably useful as scavengers, others occur in the nests of social caterpillars, and they are sometimes common in loose straw; this being the habitat in which Perris found the one we figure.

Fig. 116—Cryptophagus dentatus. Britain. A, Larva (after Perris); B, perfect Insect.

Fam. 34. Helotidae.—Front and middle coxal cavities round, with scarcely any angular prolongation externally; all the coxae widely separated; five visible ventral segments, all mobile. The Insects of this family are closely allied to Trogositidae and Nitidulidae, and have the tarsal structure of the former family; but the Helotidae are different in appearance from any members of either of these two families, and are readily distinguished by the coxal character. They are frequently classified with the Erotylidae, from which they differ by the differently shaped feet, especially by the diminished basal joint. There is but one genus, and for a long time only two or three species were known, and were great rarities in collections; in the last few years the number has been raised to nearly forty.[^[120]] They are remarkable beetles with oblong form, and a somewhat metallic upper surface, which is much sculptured, and possesses four yellow, smooth spots on the elytra. According to Mr. George Lewis they are found feeding at the running sap of trees, but the larvae are not known. Helotidae are peculiar to the Indo-Malayan region (including Japan) with one species in Eastern Africa.

Fam. 35. Thorictidae.—Tarsi five-jointed, none of the joints broad; front coxae small, rather prominent, but not at all transverse; five visible ventral plates, all mobile; metasternum very short; antennae short, with a solid club. This little family, consisting of the genus Thorictus, appears to be a distinct one, though the structure has only been very imperfectly studied. It is peculiar to the Mediterranean region, where the species live in ants' nests. They appear to be on terms of great intimacy with the ants; a favourite position of the beetle is on the scape of the antenna of an ant; here it hooks itself on firmly, and is carried about by the ant. Like so many other ants'-nest beetles, Thorictidae possess tufts of golden hair, which secrete some substance, the flavour of which is appreciated by the ants; these tufts in Thorictidae are situated either at the hind angles of the pronotum, or on the under surface of the body on each side of the breast; Wasmann thinks that when the beetles are riding about, as above described, the ants have then an opportunity of getting at the patches on the under surface.

Fig. 117—Tritoma bipustulata. Erotylidae. Britain. A, Larva (after Perris); B, perfect Insect.

Fam. 36. Erotylidae.—Tarsi five-jointed, but with the fourth usually very small, the first three more or less broad and pubescent beneath. Antennae strongly clubbed. Front and middle coxal acetabula round, without angular prolongation externally; five visible ventral segments. This is now a large and important family of about 1800 species, but it is chiefly exotic and tropical, its members haunting the fungoid growths in forests. We have only six species in Britain, and the whole of Europe has only about two dozen, most of them insignificant (and in the case of the Dacnides aberrant, approaching the Cryptophagidae very closely). The sub-family Languriides (quite wanting in Europe) consists of more elongate Insects, with front acetabula open behind; they have different habits from Erotylides proper; some are known to live as larvae in the stems of herbaceous plants. They possess a highly developed stridulating organ on the front of the head. The Clavicorn Polymorpha are very closely connected with the Phytophaga by Languriides.

Fam. 37. Mycetophagidae.—Tarsi four-jointed, slender, the front feet of the male only three-jointed; coxae oval, not deeply embedded; abdomen with five ventral segments, all movable. A small family, of interest chiefly because of the anomaly in the feet of the two sexes, for which it is impossible to assign any reason. The species are small, uninteresting Insects that live chiefly on Cryptogams of various kinds, especially in connection with timber; the larvae being also found there. There are about a dozen species in Britain, and scarcely 100 are described from all the world. The Diphyllides, placed by Leconte and Horn in this family, seem to go better in Cryptophagidae.

Fig. 118—Litargus bifasciatus. Mycetophagidae. Britain. A, Larva (after Perris); B, perfect Insect.

Fam. 38. Coccinellidae (Lady-birds).—Tarsi apparently three-jointed; the first two joints pubescent beneath; the third joint consisting really of two joints, the small true third joint being inserted near the base of the second joint, the upper surface of which is grooved to receive it. Head much concealed by the thorax. Antennae feebly clubbed. The lady-birds number fully 2000 species. The structure of their feet distinguishes them from nearly all other Coleoptera except Endomychidae, which are much less rotund in form, and have larger antennae. One genus of Endomychids—Panomoea—bears, however, a singular resemblance to lady-birds, both in form and style of coloration. Several species of Coccinellidae are remarkable on account of the numerous variations in colour they present. Coccinellidae frequently multiply to an enormous extent, and are of great value, as they destroy wholesale the plant-lice, scale-Insects, and Acari that are so injurious to cultivated plants. They also eat various other soft-bodied Insects that attack plants. As they are excessively voracious, and are themselves singularly free from enemies and multiply with great rapidity, all these features of their economy render them of inestimable value to the agriculturist and horticulturist. The species of the sub-family Epilachnides feed on plants, and one or two are occasionally injurious. The body-fluid of Coccinellidae has an unpleasant odour and taste. Many lady-birds have the power of exuding, when disturbed, small quantities of a yellow fluid. Lutz has shown that this is not a special secretion, but an exudation of the fluid of the body that takes place through a small orifice at the tip of the tibia, from pressure caused by contraction of the body and limb.[^[121]]

The larvae are much more active than beetle-larvae usually are, and many of them are very conspicuous when running about on plants to hunt their prey. They usually cast their skins three times, and sometimes concomitantly change a good deal in colour and form; the larval life does not usually exceed four or five weeks; at the end of which time the larva suspends itself by the posterior extremity, which is glued by a secretion to some object; the larval skin is pushed back to the anal extremity, disclosing the pupa; this differs in several respects from the usual pupa of beetles; it is harder, and is coloured, frequently conspicuously spotted, and dehisces to allow the escape of the beetle, so that the metamorphosis is altogether more like that of Lepidoptera than that of Coleoptera. There is much variety in the larvae; some of them bear large, complexly-spined, projections; those of the group Scymnites have small depressions on the surface, from which it has been ascertained that waxy secretions exude; but in Scymnus minimus no such excretions are formed. Certain species, when pupating, do not shuffle the skin to the extremity of the body, but retain it as a covering for the pupa. The larvae that feed on plants are much less active than the predaceous forms. We are well supplied with Coccinellidae in Britain, forty species being known here.

The systematic position of Coccinellidae amongst the Coleoptera has been for long a moot point. Formerly they were associated with various other beetles having three-jointed, or apparently three-jointed, feet, as a series with the name Trimera, or Pseudotrimera. But they are generally placed in the Clavicorn series, near Endomychidae. Verhoeff has recently made considerable morphological studies on the male genital organs of Coleoptera, and as the result, he concludes that Coccinellidae differ radically from all other Coleoptera as regards these structures, and he therefore treats them as a distinct series or sub-order, termed Siphonophora. The genus Lithophilus has been considered doubtfully a member of Coccinellidae, as the tarsi possess only in a slight degree the shape characteristic of the family: Verhoeff finds that they are truly Coccinellidae, forming a distinct division, Lithophilini; and our little species of Coccidula, which have somewhat the same appearance as Lithophilini, he treats as another separate group, Coccidulini.

Fam. 39. Endomychidae.[^[122]]—Tarsi apparently three-jointed, the first two joints broad, the terminal joint elongate; at the base of the terminal joint there is, however, a very small joint, so that the tarsi are pseudotetramerous; antennae rather large, with a large club; labium not at all retracted behind the mentum; front and middle coxae globose; abdomen with five movable ventral segments, and a sixth more or less visible at the tip. This family includes a considerable diversity of elegant Insects that frequent fungoid growths on wood. It comprises at present fully 500 species, but nearly the whole of them are exotic, and inhabit the tropical forests. We have only two British species, both of which are now rarities, but apparently were much commoner at the beginning of the century. The larvae are broader than is usual in Coleoptera; very few, however, are known.

Fig. 119—Mycetaea hirta. Britain. A, Larva (after Blisson); B, perfect Insect.

Fam. 40. Mycetaeidae.—Tarsi four-jointed, the first two joints not very different from the third, usually slender; abdomen with five visible ventral segments, which are movable; front and middle coxae globular. The little Insects composing this family are by many placed as a division of Endomychidae, and Verhoeff is of opinion that the group is an altogether artificial one; but we think, with Duval, it makes matters simpler to separate them. There are only some forty or fifty species, found chiefly in Europe and North America. We have three in Britain; one of these, Mycetaca hirta is very common, and may be found in abundance in cellars in the heart of London, as well as elsewhere; it is said to have injured the corks of wine-bottles, and to have caused leakage of the wine, but we think that it perhaps only increases some previous deficiency in the corkage, for its natural food is fungoid matters. The larva is remarkable on account of the clubbed hairs at the sides of the body.

Fam. 41. Latridiidae.—Tarsi three-jointed; anterior coxal cavities round, not prolonged externally; abdomen with five visible and mobile ventral segments. Very small Insects, species of which are numerous in most parts of the world, the individuals being sometimes very abundant. The larvae (Fig. 120, A) are said by Perris to have the mandibles replaced by fleshy appendages. The pupa of Latridius is remarkable, on account of the numerous long hairs with heads instead of points; the larva of Corticaria is very like that of Latridius, but some of the hairs are replaced by obconical projections. The sub-family Monotomides is by many treated as a distinct family; they have the elytra truncate behind, exposing the pygidium, and the coxae are very small and very deeply embedded. Most of the Latridiidae are believed to live on fungoid matters; species of Monotoma live in ants' nests, but probably have no relations with the ants. A few species of Latridiides proper also maintain a similar life; Coluocera formicaria is said to be fond of the stores laid up by Aphaenogaster structor in its nests. About 700 species are now known; scarcely any of the individuals are more than one-tenth of an inch long. We have about 40 species in Britain. The North American genus Stephostethus has the prosternum constructed behind the coxae, somewhat in the same manner as it is in the Rhynchophorous series of Coleoptera.

Fig. 120—Latridius minutus. Britain. A, Larva (after Perris); B, perfect Insect.

Fam. 42. Adimeridae.—Tarsi appearing only two-jointed, a broad basal joint and an elongate claw-bearing joint, but between the two there are two very small joints. This family consists only of the American genus Adimerus; nothing is known of the life-history of these small Insects. They are of some interest, as this structure of the foot is not found in any other beetles.

Fig. 121—Adimerus setosus. Adimeridae. A, the Insect; B, one foot more enlarged. Mexico. From Biol. Centr. Amer. Col. ii. pt. i.

Fig. 122—Tiresias serra. Larva. New Forest.

Fam. 43. Dermestidae.—Tarsi five-jointed; antennae usually short, with the club frequently very large in proportion, and with the under side of the thorax bearing a hollow for its reception. Front coxae rather long, oblique: hind coxa formed to receive the femur when in repose. A family of 300 or 400 species of small or moderate-sized beetles; the surface, usually covered with fine hair, forming a pattern, or with scales. Byturus, the position of which has long been disputed, has now been placed in this family; it has a more imperfectly formed prosternum, and the third and fourth joints of the tarsi are prolonged as membranous lobes beneath; the hind coxae leave the femora quite free. Dermestidae in the larval state nearly all live on dried animal matter, and are sometimes very destructive; some of them totally destroy zoological collections. They are very remarkable on account of the complex clothing of hairs they bear; they have good powers of locomotion, and many of them have a peculiar gait, running for a short distance, then stopping and vibrating some of their hairs with extreme rapidity. They exhibit great variety of form. Many of them are capable of supporting life for long periods on little or no food, and in such cases moult an increased number of times: pupation takes place in the larval skin. Anthrenus fasciatus has been reared in large numbers on a diet of dried horse-hair in furniture. The young larva of this species observed by the writer did not possess the remarkable, complex arrangement of hairs that appeared when it was further grown. The most curious of Dermestid larvae is that of Tiresias serra, which lives amongst cobwebs in old wood, and probably feeds on the remains of Insects therein, perhaps not disdaining the cobwebs themselves. Attention has been frequently called to the hairs of the larvae of these Insects, but they have never been adequately discussed, and their function is quite unknown.

Fam. 44. Byrrhidae (Pill-beetles).—Oval or round, convex beetles; tarsi five-jointed, front coxae not exserted, transverse; hind coxa shielding the retracted femur. The whole of the appendages capable of a complete apposition to the body. Although a small family of only 200 or 300 species, Byrrhidae are so heterogeneous that no characteristic definition that will apply to all the sub-families can be framed. Very little is known as to their life-histories. Byrrhus pilula is one of our commonest beetles, and may be found crawling on paths in early spring even in towns; it moves very slowly, and when disturbed, at once contracts the limbs so completely that it looks like an inanimate object. The larva is cylindrical, soft; the prothoracic and last two abdominal segments are larger than the others, the last bearing two pseudopods; its habits are unknown, and no good figure exists of it.

The chief groups of Byrrhidae are Nosodendrides, Byrrhides (including Amphicyrtides), Limnichides, and Chelonariides. The first consists of species frequenting the exuding sap of trees; they have an unusually large mentum, abruptly clubbed antennae, and the head cannot be retracted and concealed. The genus Nosodendron seems to be distributed over a large part of the world. The Byrrhides have the antennae gradually thicker towards the tip, the mentum small, and the head and thorax so formed that the former can be perfectly retracted. The species are rather numerous, and are found in the northern and antipodeal regions, being nearly completely absent from the tropics. The Limnichides are minute Insects living in very moist places; they have small delicate antennae, which are imperfectly clubbed. The group is very widely distributed.

The Chelonariides are a very peculiar form of Coleoptera: oval Insects of small size with the prothorax so formed that the head can be withdrawn under (rather than into) it, and then abruptly inflexed, so that the face then forms part of the under surface: the antennae have the basal three joints thicker than the others; these being not in the least clubbed, but having the joints so delicately connected that the organs are rarely unmutilated. The modifications of the head and prothorax are quite unlike those of other Byrrhidae, and if the Chelonariides do not form a distinct family, they should be associated with Dascillidae. Nothing is known as to the earlier stages. They are chiefly tropical Insects, though one species is found in North America.

Fam. 45. Cyathoceridae.—Minute Insects of broad form; parts of the mouth concealed; antennae four-jointed; tarsi not divided into joints; prosternum small. The only species of this aberrant family, Cyathocerus horni, has been found in Central America. Nothing is known as to its life-history.

Fam. 46. Georyssidae.—Antennae short, clubbed; tarsi four-jointed; prosternum very small; front coxae exserted, but not contiguous. There are about two dozen species of these small beetles known. Our British Georyssus pygmaeus lives in extremely wet places, and covers itself with a coating of mud or fine sand so that it can only be detected when in movement. Nothing further is known as to its life-history or habits. Members of the genus have been detected in widely-separated parts of the globe.

Fam. 47. Heteroceridae.—Labrum and mandibles projecting forwards; antennae short, the terminal seven joints broad and short, forming a sort of broad serrate club; legs armed with stout spines; tarsi four-jointed. The Heteroceridae are small beetles covered with very dense but minute pubescence; they live in burrows among mud or sand in wet places, and are found in many parts of the world. They possess a stridulating organ in the form of a slightly elevated curved line on each side of the base of the abdomen, rubbed by the posterior femur. The larvae live in the same places as the beetles; they have well-developed thoracic legs, the mandibles are porrect, the three thoracic segments rather large, and the body behind these becomes gradually narrower; they are believed to eat the mud amongst which they burrow. We have seven British species of Heteroceridae.

Fam. 48. Parnidae.—Prosternum distinct in front of the coxae, usually elongate, behind forming a process received into a definite cavity on the mesosternum; head retractile, the mouth protected by the prosternum. Tarsi five-jointed, terminal joint long. Although the characters of these Insects are not very different from those of Byrrhidae, of Dascillidae, and even of certain Elateridae, there is practically but little difficulty in distinguishing Parnidae. They are of aquatic habits, though many, in the perfect state, frequently desert the waters. There are about 300 or 400 species known, but the family is doubtless more extensive, as these small beetles attract but little notice. There are two groups:—1. Parnides, in which the front coxae have a considerable transverse extension, the antennae are frequently short and of peculiar structure, and the body is usually clothed with a peculiar, dense pubescence. 2. Elmides, with round front coxae, a bare, or feebly pubescent body, and simple antennae. Parnus is a genus commonly met with in Europe, and is less aquatic in habits than its congeners; it is said to enter the water carrying with it a coating of air attached to its pubescence. Its larvae are not well known; they live in damp earth near streams, and are said to much resemble the larvae of Elateridae. Potamophilus acuminatus has a very interesting larva, described by Dufour; it lives on decaying wood in the Adour. It is remarkable from the ocelli being arranged so as to form an almost true eye on each side of the head; there are eight pairs of abdominal spiracles, and also a pair on the mesothorax, though there are none on the pro- or meta-thorax; each of the stigmata has four elongate sacs between it and the main tracheal tube; the body is terminated by a process from which there can be protruded bunches of filamentous branchiae. The larvae of Macronychus quadrituberculatus is somewhat similar, though the features of its external structure are less remarkable. The Elmides live attached to stones in streams; the larva is rather broad, fringed at the sides of the body, and bears behind three elegant sets of fine filamentous branchiae. The North American genus Psephenus is placed in Parnidae, though instead of five, the male has seven, the female six, visible ventral segments; the larva is elliptical, with dilated margins to the body. Friederich, has given,[^[123]] without mentioning any names, a detailed account of Brazilian Parnid larvae, that may perhaps be allied to Psephenus.

Fam. 49. Derodontidae.—Tarsi five-jointed, slender, fourth joint rather small; front coxae prominent and transversely prolonged; middle coxae small; abdomen with five visible segments, all mobile, the first not elongated. One of the smallest and least known of the families of Coleoptera; it consists of four or five species of small Insects of the genera Derodontus and Peltasticta, found in North America, Europe, and Japan. The distinction of the family from Cleridae is by no means certain; our European Laricobius apparently possessing characters but little different. Nothing is known as to the life-histories.

Fam. 50. Cioidae.—Small or minute beetles; antennae short, terminal joints thicker; tarsi short, four-jointed; anterior and middle coxae small, oval, deeply embedded; abdomen with five ventral segments, all mobile. The position of these obscure little Insects seems to be near Colydiidae and Cryptophagidae, though they are usually placed near Bostrichidae. So far as known, they all live in fungi, or in wood penetrated by fungoid growths. The cylindrical larvae live also in similar matter; they usually have the body terminated behind by one or two hooks curved upwards; that of Cis melliei (Fig. 124) has, instead of these hooks, a curious chitinous tube. About 300 species of the family are now known; a score, or so, occurring in Britain. The Hawaiian Islands have a remarkably rich and varied fauna of Cioidae.

Fig. 123—Derodontus maculatus. North America.

Fig. 124—Cis melliei. Martinique. A, Perfect Insect; B, pupa; C, larva; D, terminal portion of body of larva. (After Coquerel.)

Fam. 51. Sphindidae.—This family of half a dozen species of rare and small Insects, differs from Cioidae by the tarsi being five-jointed at any rate on the front and middle feet, opinions differing as to whether the number of joints of the hind tarsi is four or five. These Insects live in fungi growing in wood, e.g. Reticularia hortensis, that are at first pulpy and afterwards become powder. The larvae of both of our British genera, Sphindus and Aspidiphorus, have been described by Perris, who considers them allied to the fungivorous Silphidae and Latridiidae. The systematic position of these Insects has been the subject of doubt since the days of Latreille.

Fam. 52. Bostrichidae (Apatidae of some authors).—Tarsi five-jointed, but the first joint very short and imperfectly separated from the second; front coxae prominent, contiguous, very little extended transversely; five visible ventral segments. The Bostrichidae attack dry wood, and sometimes in such large numbers that timber is entirely destroyed by them; most of them make cylindrical burrows into the wood. The larvae have the posterior part of the body incurved, and resemble the wood-boring larvae of Anobiidae rather than the predaceous larvae of Cleridae. We follow Leconte and Horn in placing Lyctides as a division of Bostrichidae; although differing very much in appearance, they have similar habits and larvae. The typical Bostrichides are remarkable for their variety of sculpture and for the shapes of the posterior part of the body; this part is more or less conspicuously truncate, and furnished with small prominences. Dinapate wrightii, found in the stems of a species of Yucca in the Mojave desert of California, attains a length of nearly two inches; its larva is extremely similar to that of A. capucina. Some of the forms (Phonapate) stridulate in a manner peculiar to themselves, by rubbing the front leg against some projections at the hind angle of the prothorax. Upwards of 200 species of the family are known. In Britain we have only four small and aberrant forms.

Fig. 125—Apate capucina. Europe. A, Larva (after Perris); B, perfect Insect.

Fam. 53. Ptinidae.—Tarsi five-jointed, first joint not reduced in size, often longer than second; front and middle coxae small, not transversely extended, the former slightly prominent; five visible ventral segments; prosternum very short. Here are included two sub-families, Ptinides and Anobiides; they are considered as distinct families by many authors, but in the present imperfect state of knowledge[^[124]] it is not necessary to treat them separately.

Fig. 126—"Biscuit-weevil." Anobium paniceum.

Fig. 127.—Early stages of Anobium paniceum. A, Eggs, variable in form; B, larva; C, pupa; D, asymmetrical processes terminating body of pupa. [This larva is probably the "book-worm" of librarians].

Ptinidae are sometimes very destructive to dried animal matter, and attack specimens in museums; Anobiides bore into wood, and apparently emerge as perfect Insects only for a very brief period; Anobium (Sitodrepa) paniceum is, however, by no means restricted in its tastes; it must possess extraordinary powers of digestion, as we have known it to pass several consecutive generations on a diet of opium; it has also been reported to thrive on tablets of dried compressed meat; in India it is said to disintegrate books; a more usual food of the Insect is, however, hard biscuits; weevilly biscuits are known to every sailor, and the so-called "weevil" is usually the larva of A. paniceum (Fig. 127, B). In the case of this Insect we have not detected more than one spiracle (situate on the first thoracic segment); the other known larvae of Anobiides are said to possess eight abdominal spiracles. The skeleton in some of this sub-family is extremely modified, so as to allow the Insects to pack themselves up in repose; the head is folded in over the chest, and a cavity existing on the breast is thus closed by the head; in this cavity the antennae and the prominent mouth-parts are received and protected; the legs shut together in an equally perfect manner, so that no roughness or chink remains, and the creature looks like a little hard seed. Anobium striatum is a common Insect in houses, and makes little round holes in furniture, which is then said to be "worm-eaten." A. (Xestobium) tessellatum, a much larger Insect, has proved very destructive to beams in churches, libraries, etc. These species are the "death-watches" or "greater death-watches" that have been associated with the most ridiculous superstitions (as we have mentioned in Volume V., when speaking of the lesser death-watches, or Psocidae). The ticking of these Insects is really connected with sex, and is made by striking the head rapidly against the wood on which the Insect is standing.

The very anomalous genus Ectrephes (Fig. 128) is found in ants' nests in Australia. Westwood placed it in Ptinidae. Wasmann has recently treated it as a distinct family, Ectrephidae, associating it with Polyplocotes and Diplocotes, and treating them as allied to Scydmaenidae.

Fig. 128—Ectrephes kingi. West Australia. (After Westwood.)

Fam. 54. Malacodermidae.—Seven (or even eight) visible ventral segments, the basal one not co-adapted in form with the coxae; tarsi five-jointed. Integument softer than usual, the parts of the body not accurately co-adapted. This important family includes a variety of forms: viz. Lycides, Drilides, Lampyrides, Telephorides; though they are very different in appearance, classifiers have not yet agreed on separating them as families. Of these the Lampyrides, or glow-worms, are of special interest, as most of their members give off a phosphorescent light when alive; in many of them the female is apterous and like a larva, and then the light it gives is usually conspicuous, frequently much more so than that of its mate; in other cases the males are the most brilliant. The exact importance of these characters in the creatures' lives is not yet clear, but it appears probable that in the first class of cases the light of the female serves as an attraction to the male, while in the second class the very brilliant lights of the male serve as an amusement, or as an incitement to rivalry amongst the individuals of this sex.

Fig. 129—Phengodes hieronymi. Cordoba, South America. (After Haase.) A, Male; B, female. l, l, Positions of luminous spots; ls, spiracles. About × 3.

The well-known fire-flies (Luciola) of Southern Europe are an example of the latter condition. They are gregarious, and on calm, warm nights crowds of them may be seen moving and sparkling in a charming manner. These individuals are all, or nearly all, males; so rare indeed is the female that few entomologists have even noticed it. The writer once assisted in a large gathering of Luciola italica in the Val Anzasca, which consisted of many hundreds of specimens; all of those he caught, either on the wing or displaying their lights on the bushes, were males, but he found a solitary female on the ground. This sex possesses ordinary, small eyes instead of the large, convex organs of the male, and its antennae and legs are much more feeble, so that though provided with elytra and wings it is altogether a more imperfect creature. Emery has given an account of his observations and experiments on this Insect, but they do not give any clear idea as to the exact function of the light.[^[125]] In our British glow-worm the female is entirely apterous—hence the name glow-worm—but the male has elytra and ample wings, and frequently flies at night into lighted apartments. Although so little has been ascertained as to the light of Lampyridae, there are two facts that justify us in supposing that it is in some way of importance to the species. These are: (1) that in a great many species the eyes have a magnificent and unusual development; (2) that the habits of the creatures are in nearly all cases nocturnal. It is true that the little Phosphaenus hemipterus is said to be diurnal in habits, but it is altogether an exceptional form, being destitute of wings in both sexes, and possessed of only very feeble light-giving powers, and we have, moreover, very little real knowledge as to its natural history; it is said that the female is of the utmost rarity, though the male is not uncommon.

The nature of the luminosity of Lampyris has given rise to many contradictory statements; the light looks somewhat like that given off by phosphorus, and is frequently spoken of as phosphorescence; but luminescence is a better term. The egg, larva, pupa, and male are luminous as well as the female (at any rate in L. noctiluca); the luminescence is, however, most marked in the female imago, in which it is concentrated near the extremity of the abdomen; here there are two strata of cells, and many fine capillary tracheae are scattered through the luminous substance. Wielowiejski concludes that the light-producing power is inherent in the cells of the luminous organ, and is produced by the slow oxidation of a substance formed under the influence of the nervous system. The cells are considered to be essentially similar to those of the fat-body.[^[126]] The luminescence of Lampyridae is very intermittent, that is to say, it is subject to rapid diminutions and increases of its brilliancy; various reasons have been assigned for this, but all are guesses, and all that can be said is that the changes are possibly due to diminution or increase of the air-supply in the luminous organ, but of the way in which this is controlled there seems to be no evidence. Considerable difference of opinion has existed as to the luminescence of the eggs of Lampyris. If it exist in the matter contained in the egg, it is evident that it is independent of the existence of tracheae or of a nervous system. Newport and others believed that the light given by the egg depended merely on matter on its exterior. The observations of Dubois[^[127]] show, however, that it exists in the matter in the egg; he has even found it in the interior of eggs that had been deposited unfertilised.

From time to time, since the commencement of the nineteenth century, there have appeared imperfect accounts of extraordinary light-giving larvae found in South America, of various sizes, but attaining in some cases a length, it is said, of three inches; they are reported as giving a strong red light from the two extremities of the body, and a green light from numerous points along the sides of the body, and hence are called, it is said, in Paraguay the railway-beetle. We may refer the reader to Haase's paper[^[128]] on the subject of these "larvae," as we can here only say that it appears probable that most of these creatures may prove to be adult females of the extraordinary group Phengodini, in which it would appear that the imago of the female sex is in a more larva-like state than it is in any other Insects. The males, however, are well-developed beetles; unlike the males of Lampyrides, in general they have not peculiar eyes, but on the other hand they possess antennae which are amongst the most highly developed known, the joints being furnished on each side with a long appendage densely covered with pubescence of a remarkable character. There is no reason to doubt that Haase was correct in treating the Insect we figure (Fig. 129, B) as a perfect Insect; he is, indeed, corroborated by Riley.[^[129]] The distinctions between the larva and female imago are that the latter has two claws on the feet instead of one, a greater number of joints in the antennae, and less imperfect eyes; the female is in fact a larva, making a slightly greater change at the last ecdysis, than at those previous. It is much to be regretted that we have so very small a knowledge of these most interesting Insects. Malacodermidae are probably the most imperfect or primitive of all beetles, and it is a point of some interest to find that in one of them the phenomena of metamorphosis are reduced in one sex to a minimum, while in the other they are—presumably at least—normal in character.

Numerous larvae of most extraordinary, though diverse, shapes, bearing long processes at the sides of the body, and having a head capable of complete withdrawal into a slender cavity of the thorax, have long been known in several parts of the world, and Dr. Willey recently found in New Britain a species having these body-processes articulated. Though they are doubtless larvae of Lampyrides, none of them have ever been reared or exactly identified.

A very remarkable Ceylonese Insect, Dioptoma adamsi Pascoe, is placed in Lampyrides, but can scarcely belong there, as apparently it has but five or six visible ventral segments; this Insect has two pairs of eyes, a large pair, with coarse facets on the under side of the head, and a moderate-sized pair with fine facets on the upper side. Nothing is known as to the habits of this curiosity, not even whether it is luminous in one or both sexes.

It is believed that the perfect instar of Lampyrides takes no food at all. The larvae were formerly supposed to be vegetarian, but it appears probable that nearly all are carnivorous, the chief food being Mollusca either living or dead. The larvae are active, and in many species look almost as much like perfect Insects as do the imagos.

The other divisions of Malacodermidae—Lycides, Drilides, Telephorides—also have predaceous, carnivorous larvae. All these groups are extensive. Though much neglected by collectors and naturalists, some 1500 species of the family Malacodermidae have been detected. We have about 50 in Britain, and many of them are amongst the most widely distributed and abundant of our native Insects. Thus, however near they may be to the primitive condition of Coleoptera, it is highly probable that they will continue to exist alongside of the primitive Cockroaches and Aptera, long after the more highly endowed forms of Insect-life have been extinguished wholesale by the operations of mankind on the face of the earth.

Fig. 130—Malachius aeneus. Britain. A, Larva (after Perris); B, female imago.

Fam. 55. Melyridae (or Malachiidae).—Six visible and moveable ventral abdominal segments; the basal part more or less distinctly co-adapted with the coxae. These Insects are extremely numerous, but have been very little studied. In many works they are classified with Malacodermidae, but were correctly separated by Leconte and Horn, and this view is also taken by Dr. Verhoeff, the latest investigator. The smaller number of visible ventral segments appears to be due to a change at the base correlative with an adaptation between the base of the abdomen and the hind coxae. The characters are singularly parallel with those of Silphidae; but in Melyridae the antennae are filiform or serrate, not clavate. The habits in the two families are different, as the Melyridae are frequenters of flowers. Many of the Melyridae have the integument soft, but in the forms placed at the end of the family—e.g. Zygia—they are much firmer. Thus these Insects establish a transition from the Malacodermidae to ordinary Coleoptera. Although the imagos are believed to consume some products of the flowers they frequent, yet very little is really known, and it is not improbable that they are to some extent carnivorous. This is the case with the larvae that are known (Fig. 130, larva of Malachius aeneus). These are said by Perris to bear a great resemblance to those of the genus Telephorus, belonging to the Malacodermidae.

Fam. 56. Cleridae.—Tarsi five-jointed; but the basal joint of the posterior very indistinct, usually very small above, and closely united with the second by an oblique splice; the apices of joints two to four usually prolonged as membranous flaps; anterior coxae prominent, usually contiguous, rather large, but their cavities not prolonged externally; labial palpi usually with large hatchet-shaped terminal joint; ventral segments five or six, very mobile. The Cleridae are very varied in form and colours; the antennae are usually more or less clubbed at the tip, and not at all serrate, but in Cylidrus and a few others they are not clubbed, and in Cylidrus have seven flattened joints. The student should be very cautious in deciding as to the number of joints in the feet in this family, as the small basal joint is often scarcely distinguishable, owing to the obliteration of its suture with the second joint. The little Alpine Laricobius has the anterior coxal cavities prolonged externally, and the coxae receive the femora to some extent, so that it connects Cleridae and Derodontidae. The Cleridae are predaceous, and their larvae are very active; they are specially fond of wood-boring Insects; that of Tillus elongatus (Fig. 131) enters the burrows of Ptilinus pectinicornis in search of the larva. The members of the group Corynetides frequent animal matter, carcases, bones, etc., and, it is said, feed thereon, but Perris's recent investigations[^[130]] make it probable that the larvae really eat the innumerable Dipterous larvae found in such refuse; it is also said that the larvae of Cleridae spin cocoons for their metamorphosis; but Perris has also shown that the larvae of Necrobia ruficollis really use the puparia formed by Diptera. Some of the species of Necrobia have been spread by commercial intercourse, and N. rufipes appears to be now one of the most cosmopolitan of Insects. The beautifully coloured Corynetes coeruleus is often found in our houses, and is useful, as it destroys the death-watches (Anobium) that are sometimes very injurious. Trichodes apiarius, a very lively-coloured red and blue beetle, destroys the larvae of the honey-bee, and Lampert has reared Trichodes alvearius from the nests of Chalicodoma muraria, a mason-bee; he records that one of its larvae, after being full grown, remained twenty-two months quiescent and then transformed to a pupa. Still more remarkable is a case of fasting of the larva of Trichodes ammios recorded by Mayet;[^[131]] this Insect, in its immature form, destroys Acridium maroccanum; a larva sent from Algeria to M. Mayet refused such food as was offered to it for a period of two and a half years, and then accepted mutton and beef as food; after being fed for about a year and a half thereon, it died. Some Cleridae bear a great resemblance to Insects of other families, and it appears probable that they resemble in one or more points the Insects on which they feed. The species are now very numerous, about 1000 being known, but they are rare in collections; in Britain we have only nine species, and some of them are now scarcely ever met with.

Fig. 131—Larva of Tillus elongatus. (New Forest). A, Head; B, front leg; C, termination of the body, more magnified.

Fam. 57. Lymexylonidae.—Elongate beetles, with soft integuments, front and middle coxae exserted, longitudinal in position; tarsi slender, five-jointed; antennae short, serrate, but rather broad. Although there are only twenty or thirty species of this family, they occur in most parts of the world, and are remarkable on account of their habit of drilling cylindrical holes in hard wood, after the manner of Anobiidae. The larva of Lymexylon navale was formerly very injurious to timber used for constructing ships, but of late years its ravages appear to have been of little importance. The genus Atractocerus consists of a few species of very abnormal Coleoptera, the body being elongate and vermiform, the elytra reduced to small, functionless appendages, while the wings are ample, not folded, but traversed by strong longitudinal nervures, and with only one or two transverse nervures. Owing to the destruction of our forests the two British Lymexylonidae—L. navale and Hylecoetus dermestoides—are now very rarely met with.

Fig. 132—Hydrocyphon deflexicollis. Britain. A, Larva (after Tournier); B, imago.

Fam. 58. Dascillidae.—Small or moderate-sized beetles, with rather flimsy integuments, antennae either serrate, filiform, or even made flabellate by long appendages; front coxae elongate, greatly exserted; abdomen with five mobile ventral segments; tarsi five-jointed. This is one of the most neglected and least known of all the families of Coleoptera, and one of the most difficult to classify; though always placed amongst the Serricornia, it is more nearly allied to Parnidae and Byrrhidae, that are placed in Clavicornia, than it is to any of the ordinary families of Serricornia. It is probable that careful study will show that it is not natural as at present constituted, and that the old families, Dascillidae and Cyphonidae, now comprised in it, will have to be separated. Only about 400 species are at present known; but as nearly 100 of these have been detected in New Zealand, and 17 in Britain, doubtless the numbers in other parts of the world will prove very considerable, these Insects having been neglected on account of their unattractive exterior, and fragile structure. The few larvae known are of three or four kinds. That of Dascillus cervinus is subterranean, and is believed to live on roots; in form it is somewhat like a Lamellicorn larva, but is straight, and has a large head. Those of the Cyphonides are aquatic, and are remarkable for possessing antennae consisting of a great many joints (Fig. 132, A). Tournier describes the larva of Helodes as possessing abdominal but not thoracic spiracles, and as breathing by coming to the surface of the water and carrying down a bubble of air adhering to the posterior part of the body; the larva of Hydrocyphon (Fig. 132, A) possesses several finger-like pouches that can be exstulpated at the end of the body. It is probable that these larvae are carnivorous. The imago of this Insect abounds on the bushes along the banks of some of the rapid waters of Scotland; according to Tournier, when alarmed, it enters the water and goes beneath it for shelter. The third form of larva belongs to the genus Eucinetus, it lives on fungoid matter on wood, and has ordinary antennae of only four joints.[^[132]] It is very doubtful whether Eucinetus is related to other Dascillidae; some authorities indeed place it in Silphidae.

Fam. 59. Rhipiceridae.—Tarsi five-jointed, furnished with a robust onychium (a straight chitinous process bearing hairs) between the claws; antennae of the male bearing long processes, and sometimes consisting of a large number of joints. Mandibles robust, strongly curved, and almost calliper-like in form. This small family of less than 100 species is widely distributed, though confined to the warmer regions of the earth, a single species occurring in the extreme south of Eastern Europe. Very little is known as to the natural history. The larva of Callirhipis dejeani (Fig. 133, A) is described by Schiödte as hard, cylindrical in form, and peculiarly truncate behind, so that there appear to be only eight abdominal segments, the ninth segment being so short as to look like an operculum at the extremity of the body. It lives in wood.

Fig. 133—A, Larva of Callirhipis dejeani (after Schiödte); B, Rhipicera mystacina male, Australia; C, under side of its hind foot.

Fig. 134—Athous rhombeus. New Forest. A, Larva; B, female imago.

Fam. 60. Elateridae (Click-beetles).—Antennae more or less serrate along the inner margin, frequently pectinate, rarely filiform. Front coxae small, spherical. Thorax usually with hind angles more or less prolonged backwards; with a prosternal process that can be received in, and usually can move in, a mesosternal cavity. Hind coxa with a plate, above which the femur can be received. Visible ventral segments usually five, only the terminal one being mobile. Tarsi five-jointed. This large family of Coleoptera comprises about 7000 species. Most of them are readily known by their peculiar shape, and by their faculty of resting on the back, stretching themselves out flat, and then suddenly going off with a click, and thus jerking themselves into the air. Some, however, do not possess this faculty, and certain of these are extremely difficult to recognise from a definition of the family. According to Bertkau[^[133]] our British Lacon murinus is provided near the tip of the upper side of the abdomen with a pair of eversible glands, comparable with those that are better known in Lepidopterous larvae. He states that this Insect does not try to escape by leaping, but shams death and "stinks away" its enemy. The glands, it would appear, become exhausted after the operation has been repeated many times. The extent of the leap executed by click-beetles differs greatly; in some species it is very slight, and only just sufficient to turn the Insect right side up when it has been placed on its back. In some cases the Insects go through the clicking movements with little or no appreciable result in the way of consequent propulsion. Although it is difficult to look on this clicking power as of very great value to the Elateridae, yet their organisation is profoundly modified so as to permit its accomplishment. The junction of the prothorax with the after-body involves a large number of pieces which are all more or less changed, so that the joint is endowed with greater mobility than usual; while in the position of repose, on the other hand, the two parts are firmly locked together. The thoracic stigma is of a highly remarkable nature, and the extensive membrane in which it is placed appears to be elastic. Although the mechanics of the act of leaping are still obscure, yet certain points are clear; the prosternal process possesses a projection, or notch, on its upper surface near the tip; as a preliminary to leaping, this projection catches against the edge of the mesosternal cavity, and as long as this position is maintained the Insect is quiescent; suddenly, however, the projection slips over the catch, and the prosternal process is driven with force and rapidity into the mesosternal cavity pressing against the front wall thereof, and so giving rise to the leap.

Several larvae are well known; indeed the "wire-worms" that are sometimes so abundant in cultivated places are larvae of Elateridae. In this instar the form is usually elongate and nearly cylindrical; the thoracic segments differ but little from the others except that they bear rather short legs; the skin is rather hard, and usually bears punctuation or sculpture; the body frequently terminates in a very hard process, of irregular shape and bearing peculiar sculpture on its upper surface, while beneath it the prominent anal orifice is placed: this is sometimes furnished with hooks, the function of which has not yet been observed. The majority of these larvae live in decaying wood, but some are found in the earth; as a rule the growth is extremely slow, and the life of the larva may extend over two or more years. Some obscurity has prevailed as to their food; it is now considered to be chiefly flesh, though some species probably attack decaying roots; and it is understood that wire-worms destroy the living roots, or underground stems, of the crops they damage. Various kinds of Myriapods (see Vol. V. p. 29) are often called "wire-worm," but they may be recognised by possessing more than six legs. The larvae of the genus Cardiophorus are very different, being remarkably elongate without the peculiar terminal structure, but apparently composed of twenty-three segments.

The genus Pyrophorus includes some of the most remarkable of light-giving Insects. There are upwards of 100 species, exhibiting much diversity as to the luminous organs; some are not luminous at all; but all are peculiar to the New World, with the exception that there may possibly be luminous species, allied to the American forms, in the Fiji Islands and the New Hebrides. In the tropics of America the Pyrophorus, or Cucujos, form one of the most remarkable of the natural phenomena. The earliest European travellers in the New World were so impressed by these Insects that descriptions of their wondrous display occupy a prominent position in the accounts of writers like Oviedo, whose works are nearly 400 years old. Only one of the species has, however, been investigated. P. noctilucus is one of the most abundant and largest of the Pyrophorus, and possesses on each side of the thorax a round polished space from which light is given forth; these are the organs called eyes by the older writers. Besides these two eye-like lamps the Insect possesses a third source of light situate at the base of the ventral surface of the abdomen; there is no trace of this latter lamp when the Insect is in repose; but when on the wing the abdomen is bent away from the breast, and then this source of light is exposed; hence, when flying, this central luminous body can be alternately displayed and concealed by means of slight movements of the abdomen. The young larva of P. noctilucus is luminous, having a light-giving centre at the junction of the head and thorax; the older larva has also numerous luminous points along the sides of the body near the spiracles. It is remarkable that there should be three successive seats of luminescence in the life of the same individual. The eggs too are said to be luminous. The light given off by these Insects is extremely pleasing, and is used by the natives on nocturnal excursions, and by the women for ornament. The structure of the light-organs is essentially similar to that of the Lampyridae. The light is said to be the most economical known; all the energy that is used being converted into light, without any waste by the formation of heat or chemical rays. The subject has been investigated by Dubois,[^[134]] who comes, however, to conclusions as to the physiology of the luminous processes different from those that have been reached by Wielowiejski and others in their investigations on Glow-worms. He considers that the light is produced by the reactions of two special substances, luciferase and luciferine. Luciferase is of the nature of an enzyme, and exists only in the luminous organs, in the form, it is supposed, of extremely minute granules. Luciferine exists in the blood; and the light is actually evoked by the entry of blood into the luminous organ.

We have given to this family the extension assigned to it by Schiödte. Leconte and Horn also adopt this view, except that they treat Throscides as a distinct family. By most authors Eucnemides, Throscides, and Cebrionides are all considered distinct families, but at present it is almost impossible to separate them on satisfactory lines. The following table from Leconte and Horn exhibits the characters of the divisions so far as the imago is concerned:—

Posterior coxae laminate; trochanters small.

Labrum concealed; antennae somewhat distant from the eyes, their insertion narrowing the front .......... Eucnemides.

Labrum visible, free; antennae arising near the eyes under the frontal margin .......... Elaterides.

Labrum transverse, connate with the front.

Ventral segments six; claws simple; tibial spurs well developed. .......... Cebrionides.

Ventral segments five; claws serrate; tibial spurs moderate. .......... Perothopides.

Posterior coxae not laminate; trochanters of middle and posterior legs very long ..........Cerophytides.

Fig. 135—Larva of Fornax n. sp. Hawaii. A, Upper side; B, under side: s s, position of spiracles; C, head more enlarged; D, under side of terminal segment; a, anus.

Throscides are considered to be distinguished by the mesosternum being impressed on each side in front for the accommodation of the posterior face of the front coxae. The genus Throscus has the antennae clavate. The classification of the Elaterides and these forms is a matter of the greatest difficulty, and, if the larvae are also considered, becomes even more complex. Cebrionid larvae are different from those of any of the other divisions, and possess laminate, not calliper-like, mandibles. The larvae of Eucnemides (Fig. 135) are very little known, but are highly remarkable, inasmuch as it is very difficult to find any mouth-opening in some of them, and they have no legs. The other divisions possess very few species compared with Elaterides. In Britain we have about sixty species of Elaterides, four of Throscides and three of Eucnemides; Cerophytum was probably a native many years ago. Neither Perothopides nor Cebrionides are represented in our fauna; the former of these two groups consists only of four or five North American species, and the Cerophytides are scarcely more numerous.

Fam. 61. Buprestidae.—Antennae serrate, never elongate; prothorax fitting closely to the after-body, with a process received into a cavity of the mesosternum so as to permit of no movements of nutation. Five visible ventral segments, the first usually elongate, closely united with the second, the others mobile. Tarsi five-jointed, the first four joints usually with membranous pads beneath. This family is also of large extent, about 5000 species being known. Many of them are remarkable for the magnificence of their colour, which is usually metallic, and often of the greatest brilliancy; hence their wing-cases are used by our own species for adornment. The elytra of the eastern kinds of the genus Sternocera are of a very brilliant green colour, and are used extensively as embroidery for the dresses of ladies; the bronze elytra of Buprestis (Euchroma) gigantea were used by the native chieftains in South America as leg-ornaments, a large number being strung so as to form a circlet. The integument of the Buprestidae is very thick and hard, so as to increase the resemblance to metal. The dorsal plates of the abdomen are usually soft and colourless in beetles, but in Buprestidae they are often extremely brilliant. The metallic colour in these Insects is not due to pigment, but to the nature of the surface. Buprestidae appear to enjoy the hottest sunshine, and are found only where there is much summer heat. Australia and Madagascar are very rich in species and in remarkable forms of the family, while in Britain we possess only ten species, all of which are of small size, and nearly all are excessively rare. The family is remarkably rich in fossil forms; no less than 28 per cent of the Mesozoic beetles found by Heer in Switzerland are referred to Buprestidae.

Fig. 136—A, Larva of Euchroma goliath (after Schiödte); B, imago of Melanophila decostigma. Europe.

The larvae (Fig. 136, A) find nourishment in living vegetable matter, the rule being that they form galleries in or under the bark of trees and bushes, or in roots thereof; some inhabit the stems of herbaceous plants and one or two of the smaller forms have been discovered to live in the parenchyma of leaves. A few are said to inhabit dead wood, and in Australia species of Ethon dwell in galls on various plants. Buprestid larvae are of very remarkable shape, the small head being almost entirely withdrawn into the very broad thorax, while the abdomen is slender.[^[135]] A few, however, depart from this shape, and have the thoracic region but little or not at all broader than the other parts. The larvae of Julodis—a genus that inhabits desert or arid regions—are covered with hair; they have a great development of the mandibles; it is believed that they are of subterranean habits, and that the mandibles are used for burrowing in the earth. Only the newly hatched larva is, however, known.

Series IV. Heteromera.

Tarsi of the front and middle legs with five, those of the hind legs with four, joints.

This series consists of some 14,000 or 15,000 species. Twelve or more families are recognised in it, but the majority of the species are placed in the one great family, Tenebrionidae. The number of visible ventral segments is nearly always five. Several of the families of the series are of doubtful validity; indeed beyond that of Tenebrionidae the taxonomy of this series is scarcely more than a convention. The larvae may be considered as belonging to three classes; one in which the body is cylindrical and smooth and the integument harder than usual in larvae; a second in which it is softer, and frequently possesses more or less distinct pseudopods, in addition to the six thoracic legs; and a third group in which hypermetamorphosis prevails, the young larvae being the creatures long known as Triungulins, and living temporarily on the bodies of other Insects, so that they were formerly supposed to be parasites.

Fam. 62. Tenebrionidae.—Front coxae short, not projecting from the cavities, enclosed behind. Feet destitute of lobed joints. Claws smooth. This is one of the largest families of Coleoptera, about 10,000 species being already known. A very large portion of the Tenebrionidae are entirely terrestrial, wings suitable for flight being absent, and the elytra frequently more or less soldered. Such forms are described in systematic works as apterous. Unfortunately no comprehensive study has ever been made of the wings or their rudiments in these "apterous forms."[^[136]] it is probable that the wings, or their rudiments or vestiges, always exist, but in various degrees of development according to the species, and that they are never used by the great majority of the terrestrial forms. Many of the wood-feeding Tenebrionidae, and the genera usually placed at the end of the family, possess wings well adapted for flight. The apterous forms are chiefly ground-beetles, living in dry places; they are very numerous in Africa, California, and North Mexico. Their colour is nearly always black, and this is probably of some physiological importance; the integuments are thick and hard, and if the wing-cases are taken off, it will be found that they are usually more or less yellow on the inner face, even when jet-black externally; the external skeleton is very closely fitted together, the parts that are covered consisting of very delicate membrane; the transition between the hard and the membranous portions of the external skeleton is remarkably abrupt. These ground-Tenebrionidae form a very interesting study, though, on account of their unattractive appearance, they have not received the attention they deserve.

Fig. 137—Tenebrio molitor. Europe, etc. A, Larva (meal-worm); B, pupa (after Schiödte); C, imago.

Many of the Tenebrionidae, notwithstanding their dark colours, are diurnal in habits, and some of them run with extreme velocity in places so bare and desert that the means of existence of the Insects is a mystery. Most of the Tenebrionidae, however, shun the light. The food is usually vegetable matter, and it is apparently preferred in a very dry state. Mr. Gahan has recently recorded that in Praogena the under surface of the head has the gular region striate for stridulating purposes. This is the only instance known of a voice-organ in this situation, and moreover is the only case in all the Tenebrionidae in which any sound-producing organ has been discovered. The larvae exhibit but little variety, they are elongate and cylindrical, with harder integument than is usual in Coleopterous larvae; they have six thoracic legs, and at the under side of the posterior extremity the anus serves as a very short pseudopod. The resemblance of these larvae to those of Elateridae is considerable; but though the body is terminated by one or two small processes, these never attain the complexity of the terminal segment of Elateridae. The common meal-worm—i.e. the larva of Tenebrio molitor—is a very characteristic example of the group. The pupae are remarkable on account of peculiar projections, of varied and irregular form, that exist on the sides of the abdominal segments. Britain is very poor in these Insects; our list of them scarcely attains the number of thirty species.

Fam. 63. Cistelidae.—Claws comb-like. The very obscure beetles forming this family are only separated from Tenebrionidae on account of their pectinate claws. About 500 species of Cistelidae are recorded; the early instars, so far as known, do not differ from those of Tenebrionidae; the larvae are believed to live on dead wood.

Fam. 64. Lagriidae.—Anterior coxal cavities closed, tips of the front coxae free, claws smooth, penultimate joint of the tarsi broader, pubescent beneath. This family has very little to distinguish it from Tenebrionidae, and the group Heterotarsini appears to connect the two. It is a small family of about 200 species, widely distributed, and represented in Britain by one species, Lagria hirta. The early instars are similar to those of the Tenebrionidae, except that the larva is less retiring in its habits and wanders about on foliage: it is of broader form than that of most of the Tenebrionidae. The pupa has long projections at the sides of the abdominal segments.

Fam. 65. Othniidae.—Only about ten species are known of this dubious family. They are small Insects with weak integument, and are said by Leconte and Horn to be distinguished from "degraded Tenebrionidae" by the more mobile abdominal segments, the hind-margins of which are semi-membranous. The antennae are of the clubbed shape, characteristic of "Clavicornia," but this also occurs in numerous undoubted Tenebrionidae. Species of Othnius have been found in Japan and Borneo, as well as in North America. Nothing is known as to their metamorphoses.

Fam. 66. Ægialitidae.—All the coxae very widely separated; no co-adaptation between the sides of the abdomen and the edges of the wing-cases; five ventral segments and tip of a sixth visible. Two minute and rare Insects from North-West America constitute this family. It is distinguished from Pythidae by the minute front coxae, widely separated, completely closed in, and deeply embedded in the prosternum.

Fam. 67. Monommidae.—This is a small family of less than 100 species, the members of which have the details of their external structure much modified, permitting the Insect to pack itself up in repose in a very perfect manner. They are of small size and oval form; and are absent from Europe and the Antipodes. Nothing appears to be known as to the metamorphosis.

Fam. 68. Nilionidae.—Broad, circular Heteromera, of moderate size, with the front coxae but little separated, and the anterior acetabula closed, though having the appearance of being open in consequence of the tips of the epimera being free. The inflexed portion of the wing-cases remarkably broad. A small family of less than fifty species, found on fungi, chiefly in South America. The metamorphoses are not known. It is of very doubtful validity.

Fam. 69. Melandryidae.—Head not constricted behind the eyes; anterior acetabula not closed; claws smooth. Prothorax broad behind. These are loosely-fitted-together Insects, of moderate or small size, frequenting dry wood or fungi. About 200 species are known, found chiefly in temperate regions. The few described larvae are rather varied in their details and cannot be generalised at present. The characters of the members of this family require fresh investigation.

Fam. 70. Pythidae.—Distinguished from Melandryidae by the prothorax being narrow behind. This is a small family of about 100 species, found in temperate regions in connection with timber. The species of Rhinosimus have the head prolonged in front of the antennae so as to form a beak. The larva of Pytho depressus is flat and has parallel sides; the body is terminated by two widely-separated sharp processes. It is found occasionally under the bark of firs in Scotland.

Fam. 71. Pyrochroidae.—Differs from Melandryidae by the head forming a very narrow neck behind, and by the penultimate tarsal joints being broad. They are feeble Insects, though active on the wing. They are destitute of any of the various remarkable structures found in Mordellidae. Only about forty species are known, and the family is confined to the north temperate region, being best represented in Japan. Pyrochroa rubens is common in some parts of England; the larva is found under the bark of tree-stumps; it is remarkably flat, and has the eighth abdominal segment unusually long, while the ninth terminates the body in the form of two long sharp processes.

Fam. 72. Anthicidae.—Head with an abrupt narrow neck; prothorax narrower than the elytra. Middle and hind coxae placed in definite acetabula. Claws simple. These little Insects are numerous in species; they have little resemblance to Pyrochroidae, though the characters of the two families cause us to place them in proximity. There are about 1000 species known; though we have only about 12 in Britain, they are very numerous in the Mediterranean region. The family Pedilidae of Lacordaire and some others is now merged in Anthicidae. Thomson and Champion, on the other hand, separate some very minute Insects to form the family Xylophilidae, on account of certain differences in the form of the abdomen and tarsi. The Xylophilidae live in dead wood; the Anthicidae, on the surface of the earth, after the manner of ground-beetles; very little is, however, known as to their natural history.

Fam. 73. Oedemeridae.—Prothorax not forming sharp edges at the sides, head without a narrow neck. Penultimate tarsal joint broad; claws smooth. These Insects usually have a feeble integument, and bear a certain resemblance to Malacodermidae. Less than 500 species are known, but they are widely distributed, and occur in both temperate and tropical regions. The larvae live in old wood. Nacerdes melanura is common on our coasts, where its larva lives in timber cast up by the sea, or brought down by floods, and it is able to resist immersion by the tide. It is remarkable from the possession of five pairs of dorsal false feet on the anterior segments, and two pairs on the ventral aspect. In Asclera caerulea there are six dorsal and three ventral pairs of these remarkable pseudopods. We have six species of Oedemeridae in Britain, including Asclera as well as Nacerdes.

Fig. 138—Asclera caerulea. A, Larva; B, pupa (after Schiödte); C, imago. Cambridge.

Fam. 74. Mordellidae (incl. Rhipiphoridae).—Head peculiarly formed, vertex lobed or ridged behind, so that in extension it reposes on the front edge of the pronotum; capable of great inflection and then covering the prosternum; hind coxae with laminae forming a sharp edge behind, frequently very large. This family is a very distinct one, though it exhibits great variety. Lacordaire has pointed out that Rhipiphoridae cannot at present be satisfactorily distinguished from Mordellidae. Leconte and Horn separate the two by the fact that the sides of the prothorax form a sharp edge in Mordellidae, but not in Rhipiphoridae. A better character would perhaps be found by a study of the head, but as this would clearly result in a radical change in the composition of the two families it is preferable to treat them at present as only sub-families: if placed on a similar basis to the preceding families, the group would however form, not two, but several families. Besides the unusual shape of the head (Fig. 139, D) the ventral region of the body is remarkably formed, being very convex, and in many Mordellides terminating in a strong spinous process (Fig. 139, C). The elytra are, in several Rhipiphorids, of the groups Myoditini and Rhipidiini, reduced to a very small size, and the wings are not folded. The Mordellidae are remarkable for their activity; in the perfect state they usually frequent flowers, and fly and run with extreme rapidity. Mordellides are amongst the most numerous and abundant of the European Coleoptera, and in Britain the Anaspini swarm on the flowers of bushes and Umbelliferae. The life-histories appear to be singularly varied; but unfortunately they are incompletely known. The larvae of some of the Mordellids have been found in the stems of plants, and derive their nutriment therefrom. This is said by Schwarz to be undoubtedly the case with Mordellistena floridensis. Coquillett has found the larvae of M. pustulata in plant-stems under circumstances that render it highly probable that they were feeding on a Lepidopterous larva contained in the stems; and Osborn found a similar larva that was pretty certainly a Mordellistena, and fed voraciously on Dipterous larvae in the stems of a plant. The little that is known as to the metamorphoses of Mordella and Anaspis shows that they live in old wood, but does not make clear the nature of their food.

Fig. 139—Mordellistena floridensis. America. (After Riley.) A, Larva; B, pupa; C, imago; D, outline of detached head of imago of M. pumila, to show the neck.

Although it has been ascertained that the Rhipiphorides exhibit instances of remarkable metamorphosis, their life-histories are still very imperfectly known. Dr. Chapman has ascertained some particulars as to Metoecus paradoxus, which has long been known to prey in the larval state on the larvae of the common social wasps.[^[137]] The eggs are apparently not deposited in the nests of the wasps, but in old wood. The young larva is a triungulin, similar to that of the Cantharidae, we shall subsequently describe. It is not known how it makes its way to the wasps' nests, but it is possible that when a wasp visits some old wood haunted by these larvae, some of them may attach themselves to it and be carried to the wasps' nests. When access is gained to the cells the little Metoecus pierces the skin of one of the wasp-grubs, and entering in it feeds on the interior; after it has increased in size it emerges, changes its skin, and assumes a different form and habits; subsequently, as an external parasite, entirely devouring the wasp-larva, and then becoming a pupa, and finally a perfect Metoecus, in the cell of the wasp. The wasps, though they investigate the cells, do not apparently entertain any objection to the Metoecus, though there may be sometimes as many as twenty or thirty of the destroyers in a single nest. A few hours after the Metoecus has become a winged Insect and has escaped from the cells, it appears however, from the observations of Erné[^[138]] on nests of wasps in captivity, that the wasps become hostile to the foreigners, and it is probable that in a state of nature these leave the nest as quickly as possible. Emenadia flabellata, a genus allied to Metoecus, has been discovered by Chobaut to have a similar life-history, except that it attacks a solitary wasp of the genus Odynerus.[^[139]] An old record to the effect that a second species of Emenadia, E. bimaculata, lives in the stalks of Eryngium campestre, on the pith, is now thought to be erroneous. Fabre has found the larvae and pupae of another Rhipiphorid in the cells of a bee, Halictus sexcinctus.

The most remarkable of the Rhipiphorids, from the point of view of its habits, is certainly Symbius blattarum, which is now treated as the same as an Insect previously described by Thunberg from specimens found in amber and called Ripidius pectinicornis. This species is parasitic in cockroaches; the male and female are very different, the former being an active winged Insect, while the female is worm-like, differing but little from the larva, and never leaving the body of the cockroach. It is to be regretted that the life-history is not better known. The species has been found on board ship in vessels coming from India; the male has been met with in several European countries, but the female is excessively rare.

Fam. 75. Cantharidae or Meloidae (Blister-beetles, Oil-beetles).--Head with an abrupt neck; elytra and sides of the abdomen without any coadaptation; each claw of the feet with a long appendage closely applied beneath it. This distinct family consists of Heteromera with soft integument, and is remarkable for the fact that many of its members contain a substance that when extracted and applied to the human skin, possesses the power of raising blisters. The life-history is highly remarkable, the most complex forms of hyper-metamorphosis being exhibited. The species now known amount to about 1500; there can be no difficulty in recognising a member of the family by the above characters, except that in a very few cases each claw bears a projecting tooth, instead of an elongate appendage parallel with itself. The penultimate tarsal joint scarcely ever broader than the preceding; the colour and style of markings are extremely varied. There are two very distinct sub-families, Cantharides and Meloides; the former are winged Insects, and are frequently found on flowers or foliage. The Meloides are wingless, and consequently terrestrial; they have a very short metasternum, so that the middle coxae touch the hind; and they also have very peculiar wing-cases, one of the two overlapping the other at the base; in a few Meloids the wing-cases are merely rudiments.

The post-embryonic development of these Insects is amongst the most remarkable of modern entomological discoveries. The first steps were made by Newport in 1851,[^[140]] and the subject has since been greatly advanced by Fabre, Riley, and others. As an example of these peculiar histories, we may cite Riley's account[^[141]] of Epicauta vittata (Fig. 140), a blister-beetle living at the expense of North American locusts of the genus Caloptenus. The locust lays its eggs underground, in masses surrounded by an irregular capsule, and the Epicauta deposits its eggs in spots frequented by the locust, but not in special proximity to the eggs thereof. In a few days the eggs of the blister-beetle hatch, giving rise to little larvae of the kind called triungulin (Fig. 140, A), because each leg is terminated by three tarsal spines or claws. In warm, sunny weather these triungulins become very active; they run about on the surface of the ground exploring all its cracks, penetrating various spots and burrowing, till an egg-pod of the locust is met with; into this the triungulin at once eats its way, and commences to devour an egg. Should two or more triungulins enter the same egg-pod, battles occur till only one is left.

Fig. 140.—Hypermetamorphosis of Epicauta vittata. North America. (After Riley.) A, Young larva or triungulin; B, Caraboid instar or second larva; C, coarctate larva, or instar between the Scarabaeoid and Scolytoid larva; D, Scarabaeoid larva, from which the Scolytoid, or sixth, instar differs but little; E, pupa; F, imago.

After a few days passed in devouring a couple of eggs, the triungulin sheds its skin and appears as a different larva (Fig. 140, B), with soft skin, short legs, small eyes, and different form and proportions; a second moult takes place after about a week, but is not accompanied by any very great change of form, though the larva is now curved, less active, and in form like a larva of Scarabaeidae; when another moult occurs the fourth instar appears as a still more helpless form of larva (Fig. 140, D), which increases rapidly in size, and when full grown leaves the remains of the egg-pod it has been living on, and forms a small cavity near by; here it lies on one side motionless, but gradually contracting, till the skin separates and is pushed down to the end of the body, disclosing a completely helpless creature that has been variously called a semi-pupa, pseudo-pupa, or coarctate larva (Fig. 140, C); in this state the winter is passed. In spring the skin of the coarctate larva bursts, and there crawls out of it a sixth instar which resembles the fourth (Fig. 140, D), except in the somewhat reduced size and greater whiteness. It is worthy of remark that the skin it has deserted retains its original form almost intact. In this sixth instar the larva is rather active and burrows about, but does not take food, and in the course of a few days again moults and discloses the true pupa (Fig. 140, E). As usual in Coleoptera this instar lasts but a short time, and in five or six days the perfect beetle appears (Fig. 140, F). It is extremely difficult to frame any explanation of this complex development; there are, it will be noticed, no less than five stages interposed between the first larval instar and the pupal instar, and the creature assumes in the penultimate one a quasi-pupal state, to again quit it for a return to a previous state. It is possible to look on the triungulin and the pupal instars as special adaptations to external conditions; but it is not possible to account for the intermediate instars in this way, and we must look on them as necessitated by the physiological processes going on internally. Nothing, however, is known as to these. It may be well to mention that, after describing and figuring (loc. cit.) this series of instars, Riley changed his views as to their nomenclature.[^[142]] The following summary of the metamorphosis, to which we have added the two nomenclatures of Riley—the original one, when different from the amended one, being given in square brackets—may therefore be useful, viz.—Egg; 1, triungulin-larva—moult; 2, Caraboid larva [second larva, Caraboid stage]—moult; 3, Scarabaeoid larva [second larva, Scarabaeoid stage]—moult; 4, Scarabaeoid larva [second larva, ultimate stage] (large amount of food and much growth)—moult; 5, coarctate larva [pseudo-pupa, or semipupa]; 6, Scolytoid larva [third larva] (active, but little or no food taken)—moult; 7, pupa—moult; 8, perfect Insect.

M. Fabre has succeeded in elucidating the history of Sitaris humeralis, a Cantharid that lives at the expense of bees of the genus Anthophora.[^[143]] The eggs of the Sitaris are deposited in the earth in close proximity to the entrances to the bees' nests, about August. They are very numerous, a single female producing, it is believed, upwards of 2000 eggs. In about a month—towards the end of September—they hatch, producing a tiny triungulin of black colour; the larvae do not, however, move away, but, without taking any food, hibernate in a heap, remaining in this state till the following April or May, when they become active. Although they are close to the abodes of the bees they do not enter them, but seek to attach themselves to any hairy object that may come near them, and thus a certain number of them get on to the bodies of the Anthophora and are carried to its nest. They attach themselves with equal readiness to any other hairy Insect, and it is probable that very large numbers perish in consequence of attaching themselves to the wrong Insects. The bee in question is a species that nests in the ground and forms cells, in each of which it places honey and lays an egg, finally closing the receptacle. It is worthy of remark that in the case of the Anthophora observed by M. Fabre, the male appears about a month before the female, and it is probable that the vast majority of the predatory larvae attach themselves to the male, but afterwards seize a favourable opportunity, transfer themselves to the female, and so get carried to the cells of the bee. When she deposits an egg on the honey, the triungulin glides from the body of the bee on to the egg, and remains perched thereon as on a raft, floating on the honey, and is then shut in by the bee closing the cell. This remarkable act of slipping on to the egg cannot be actually witnessed, but the experiments and observations of the French naturalist leave little room for doubt as to the matter really happening in the way described. The egg of the bee forms the first nutriment of the tiny triungulin, which spends about eight days in consuming its contents; never quitting it, because contact with the surrounding honey is death to the little creature, which is entirely unfitted for living thereon. After this the triungulin undergoes a moult and appears as a very different creature, being now a sort of vesicle with the spiracles placed near the upper part; so that it is admirably fitted for floating on the honey (Vol. V. Fig. 86, 10). In about forty days, that is, towards the middle of July, the honey is consumed, and the vesicular larva after a few days of repose changes to a pseudo-pupa (11 of the fig. cited) within the larval skin. After remaining in this state for about a month, some of the specimens go through the subsequent changes, and appear as perfect Insects in August or September. The majority delay this subsequent metamorphosis till the following spring, wintering as pseudo-pupae and continuing the series of changes in June of the following year; at that time the pseudo-pupa returns to a larval form (12 of the fig. cited), differing comparatively little from the second instar. The skin, though detached, is again not shed, so that this ultimate larva is enclosed in two dead skins; in this curious envelope it turns round, and in a couple of days, having thus reversed its position, becomes lethargic and changes to the true pupa, and in about a month subsequent to this appears as a perfect Insect, at about the same time of the year as it would have done had only one year, instead of two, been occupied by its metamorphosis. M. Fabre employs the term, third larva, for the instar designated by Riley Scolytoid larva, but this is clearly an inconvenient mode of naming the instar. Sitaris humeralis is now very rare in Britain, but it seems formerly to have been more common, and it is not improbable that its triungulin may have been the "Pediculus melittae," that was believed by Kirby to be a sort of bee-louse. Some species of the genus Meloe are still common in Britain, and the Insects may be seen with heavy distended abdomen grazing on herbage in the spring. The females are enormously prolific, a single one producing, it is believed, about 10,000 eggs. Meloe is also dependent on Anthophora, and its life-history seems on the whole to be similar to that of Sitaris; the eggs are, however, not necessarily deposited in the neighbourhood of the bees' nests, and the triungulins distribute themselves on all sorts of unsuitable Insects, so that it is possible that not more than one in a thousand succeeds in getting access to the Anthophora nest. It would be supposed that it would be a much better course for these bee-frequenting triungulins to act like those of Epicauta, and hunt for the prey they are to live on; but it must be remembered that they cannot live on honey; the one tiny egg is their object, and this apparently can only be reached by the method indicated by Fabre. The history of these Insects certainly forms a most remarkably instructive chapter in the department of animal instinct, and it is a matter for surprise that it should not yet have attracted the attention of comparative psychologists. The series of actions, to be performed once and once only in a lifetime by an uninstructed, inexperienced atom, is such that we should a priori have denounced it as an impossible means of existence, were it not shown that it is constantly successful. It is no wonder that the female Meloe produces 5000 times more eggs than are necessary to continue the species without diminution in the number of its individuals, for the first and most important act in the complex series of this life-history is accomplished by an extremely indiscriminating instinct; the newly hatched Meloe has to get on to the body of the female of one species of bee; but it has no discrimination whatever of the kind of object it requires, and as a matter of fact, passes with surprising rapidity on to any hairy object that touches it; hence an enormous majority of the young are wasted by getting on to all sorts of other Insects; these larvae have been found in numbers on hairy Coleoptera as well as on flies and bees of wrong kinds; the writer has ascertained by experiment that a camel's-hair brush is as eagerly seized, and passed on to, by the young Meloe as a living Insect is.

The histories of several other Cantharids have been more or less completely discovered. Fabre has found the larva of Cerocoma schaefferi attacking the stores of provisions laid up by a fossorial wasp of the genus Tachytes, and consisting of Orthoptera of the family Mantidae. The student who wishes for further information may refer to M. Beauregard's work on this family.[^[144]]

Some half-dozen species of the genus Cephaloon found in Siberia, Japan, and North America, have, by some authorities, been separated as the family Cephaloidae. Nothing is known as to the metamorphosis of these rare beetles; and at present it is not necessary to distinguish them from Cantharidæ.

Fam. 76. Trictenotomidae.—Large Heteromera, with powerful free projecting mandibles; the antennae long, but with the terminal three joints short, with angular projections on one side. This family includes only two genera and seven or eight species. They are very remarkable Insects; Autocrates aenea being three inches long. The family is of considerable interest, as it seems to have no affinity with any other Coleoptera. The appearance of the species somewhat reminds one of Lucanidae, or Prionides; but Trictenotomidae have even less relation to those beetles than they have to the members of the Heteromerous series. The Trictenotomidae appear to be found only in the primitive forests of the Indian and Indo-Malayan regions. Nothing is known as to their life-histories.

Series V. Phytophaga.

Tarsi apparently four-jointed, the three basal joints usually densely set with cushion-like pubescence beneath; the third joint different in form, being divided into two lobes, or grooved on its upper surface so as to allow of the fourth joint being inserted near its base instead of at its extremity. Head not forming a definite prolonged beak; its labrum visible, the palpi rarely (and even then not completely) occluded in the mouth.

This great series of beetles includes something like 35,000 species. It approaches, like all the other series, the Polymorpha, especially the family Erotylidae placed therein, but in the great majority of cases there is no difficulty in recognising its members. The tarsi have never the Heteromerous formula, the head is not constructed like that of Rhynchophora, nor the mouth and feet like those of Adephaga; the antennae are different from those of the Lamellicorns. The tarsi are really five-jointed, for careful inspection shows that the long claw-joint has at its extreme base a small nodule, which is undoubtedly the fourth joint (Fig. 142, B). In speaking of the joints it is, however, customary not to refer to this small and functionally useless joint at all, and to call the claw-joint the fourth; when the little joint is referred to it may be called the true fourth joint.

Nearly the whole of the enormous number of species of this series are directly dependent on the vegetable kingdom for their nutriment; they are therefore well styled Phytophaga. This term is, however, restricted by some systematists to the family we have called Chrysomelidae. Although there is enormous variety in this series, three families only can be at all naturally distinguished, and this with difficulty. Of these the Bruchidae are seed-feeders, the Chrysomelidae, as a rule, leaf-feeders, the Cerambycidae wood and stem-feeders. The number of exceptions to this rule is but small, though certain Cerambycidae and certain Chrysomelidae live on roots.

Fam. 77. Bruchidae.—Prosternum extremely short; in front perpendicular; behind the coxae, forming merely a transverse lamina with pointed extremity. Hind femora more or less thickened. This comparatively small family includes about 700 species of small, unattractive beetles. The larvae live in seeds; hence some of the species are liable to be transported by means of commerce; some of them do considerable injury; peas and beans being specially subject to their attacks. They are able to complete their growth with a very small amount of nutriment, some of them consuming only a portion a little larger than themselves of a bean or pea. The larvae are fat maggots without legs, but Riley has discovered that the young larvae of Bruchus pisi and B. fabae have, when first hatched, three pairs of legs which are subsequently lost. They also have peculiar spinous processes on the pronotum. Both of these characteristics may be correlative with the transient differences in the activities of the larva, for the little creature is not at first located in the pea, but mines a gallery in the pod, in which it moves about, subsequently entering the pea and losing its legs. There is a good deal of difference in these respects between the two species—B. pisi and B. fabae—examined by Riley, and as but little is known of the life-histories of other Bruchidae it is probable that still greater variety prevails. Heeger has found that Bruchus lentis sometimes requires two seeds to enable it to complete its growth; it is, notwithstanding its legless state when half-grown, able to migrate by dropping to the earth, and dragging itself along by its mandibles till it comes to another pod into which it bites its way.

Fig. 141—Bruchus pisi or pea-weevil. A, Young larva; B, prothoracic spinous process; C, post-embryonic leg, greatly magnified; D, pea-pod, with tracks of entry; E, portion of pod, with egg, and the subsequently formed track, magnified; F, imago. (After Riley.)

The family has, until recently, been placed in the Rhynchophorous series, with which it has, however, no direct connection. On the other hand, it is so closely connected with Chrysomelidae that it is not possible to indicate good characters to distinguish the two at present. The Australian genus Carpophagus, and the large South American species of Caryoborus appear to be quite indistinguishable as families, though Lacordaire and Chapuis placed one in Bruchidae, the other in Chrysomelidae. The definition we have given applies, therefore, to the majority of the family, but not to the aberrant forms just mentioned. The European genus Urodon appears to belong to Anthribidae, not to Bruchidae. The family Bruchidae is called Mylabridae by some.

Fam. 78. Chrysomelidae.—Antennae moderately long; eyes moderately large, usually not at all surrounding the insertion of the antennae; upper surface usually bare, frequently brightly coloured and shining. This enormous family comprises about 18,000 species of beetles, in which the form and details of structure are very varied. No satisfactory character for distinguishing Chrysomelidae from Cerambycidae has yet been discovered, although the two families are certainly distinct and natural. Most of the Chrysomelidae live on foliage; few of them are more than half an inch long, whereas the Cerambycidae are wood-feeders and usually of more elongate form and larger size. The potato beetle, or Colorado beetle, that occasioned so much destruction in North America some thirty years ago, and the introduction of which into Europe was anticipated with much dread, is a good example of the Chrysomelidae. The turnip flea, a tiny hopping beetle, is among the smallest forms of the family, and is a member of another very extensive subdivision of Chrysomelidae, viz. Halticides. The term Phytophaga is by many naturalists limited to Chrysomelidae, the Cerambycidae being excluded. The classification of the family is but little advanced, but the enormous number of species of Chrysomelidae are placed in four divisions, viz.:—

Fig. 142—Doryphora decemlineata, the potato beetle. North America. A, Imago; B, hind-tarsus. 3, third joint; 4, true fourth joint; 5, so-called fourth joint.

Prothorax much narrower at the base than the elytra, and usually without side-margins (raised edges). Sub-fam. 1. Eupoda; with three divisions, Sagrides, Donaciides, Criocerides.

The basal ventral plates of the abdominal segments are somewhat shorter in the middle than at the sides, the fourth one being often invisible in the middle, while the fifth is very large. Sub-fam. 2. Camptosomes; with six divisions, Megascelides, Megalopides, Clythrides, Cryptocephalides, Chlamydes, Sphaerocarides.

In the other two groups there is no great disparity between the fourth and fifth ventral plates.

Prothorax not greatly narrower at the base than the elytra, and usually with distinct edges at the outsides. Sub-fam. 3. Cyclica; with four divisions, Lamprosomides, Eumolpides, Chrysomelides, Galerucides.

Front of the head bent downwards or inflexed, so that the mouth is on the lower aspect. Antennae inserted close together on the most anterior part of the head, so that they are more forward than the mouth. Sub-fam. 4. Cryptostomes; with two divisions Hispides, Cassidides.

In the other three divisions the mouth is placed as usual, but the insertion of the antennae varies a good deal.

The larvae of about 100 species of the family are known; they are arranged in accordance with their habits, by Chapuis,[^[145]] in six groups, viz.:

1. Elongate larvae, living under water, and there undergoing their metamorphosis. (Donaciides.)

2. Larvae mining in leaves, and undergoing their metamorphosis in the leaf. (Hispides and some Halticides.)

3. Short convex larvae, frequently with leathery and pigmented integuments, living exposed on plants. (Most of the Cyclica.)

4. Larvae of short form; covering the body with excrementitious matter. (Some Criocerides.)

5. Peculiar larvae of short form, spiny, and protecting their bodies by excrementitious matter attached by a special apparatus, the excrement itself being modified so as to be suitable for retention. (Cassidides.)

6. Elongate, pallid, larvae with curved abdomen; living in shell-like cases, and undergoing metamorphosis therein. (Most of the Camptosomes, the habits of which are known.)

Though our knowledge of these larvae extends to only about 100 out of 18,000 species, the above category by no means includes all the kinds of larvae; Captain Xambeu having recently discovered that the larva of Chrysochus pretiosus lives in the earth feeding on roots after the manner of a Rhizotrogus larva, which it resembles. The larva of Sagra splendida lives inside the stems of Dioscorea batatas, in swellings; the group Sagrides, to which it belongs, is a very anomalous one.

i. Eupoda. The beetles of the genus Donacia are of special interest. They form, with the genus Haemonia, a peculiar group, well represented in Europe, and also in our own country. They are all connected with aquatic plants, the species of Haemonia living entirely under water, while the Donacia live in the imago-state an aërial life, though many of them enter the water with great readiness, and, it is said, are able to take wing from the surface. The larvae live on the roots of aquatic plants, and derive not only nutriment but air therefrom; they pass several months as pupae (or as resting larvae waiting for pupation), under water in cocoons which they construct, and which, incredible as it may seem, are filled with air, not water. Exact details as to the construction of these cocoons are wanting. It was formerly absurdly supposed that the larva swelled itself out to the size of the cocoon it was about to make, and so served as a mould, subsequently contracting. The observations of Schmidt-Schwedt[^[146]] make it, however, more probable that the plant itself furnishes the air which, under pressure of the water (so he supposes), fills the cocoon; the larva wounds the root, piercing to an air-vessel and then constructs the cocoon on this spot, leaving to the last moment an orifice, according to Schmidt, as an exit for the water. The larva uses a similar artifice for obtaining air; it has no gills, but is provided near the extremity of the body with two sharp chitinous processes which it drives into the root of the plant till it penetrates an air-vessel. Schmidt thinks the processes serve as conduits to conduct the air to the tracheae, but Dewitz thinks the air enters the larva in a more normal manner, by means of a stigma placed at the base of the piercing process. A similar larva exists in Haemonia; which genus is additionally interesting from the fact that the imago lives entirely submerged. It is not known how it breathes. This genus is the only member of the Chrysomelidae that does not possess the structure of the feet that is characteristic of the Phytophaga. The late Professor Babington about sixty years ago found H. curtisi at Cley on the Norfolk coast on submerged Potamogeton pectinatus, but it has not been met with there for a great many years.

The larvae of Criocerides are of two kinds, in one of which the body is peculiarly shaped in conformity with the curious habit of using the excrement as a covering. The larva is less elongate than usual, and has the anus placed on the upper surface, and formed so that the excrement when voided is pushed forward on to the Insect; here it is retained by means of a slimy matter, and a thick coat entirely covering the creature, is ultimately formed. The larva of Lema melanopa is not uncommon about Cambridge, where it feeds on the leaves of growing corn. It is a remarkable fact that even in one genus the species have some of them this habit, but others not. The species of Crioceris living on lilies—C. merdigera, e.g.—are noted for possessing it; while C. asparagi does not protect itself in this way, but emits fluid from its mouth when disturbed. This larva is a serious nuisance in some localities to the cultivators of asparagus. The eggs are deposited on the stems of the plant—as shown in our figure—sometimes in great numbers.

The perfect Insects of many of the Criocerides possess a stridulating organ. Two contiguous areas at the base of the last dorsal segment, where they can be rubbed by the tips of the elytra, are slightly elevated and bear very close and fine straight lines.

Fig. 143—Crioceris asparagi. A, Eggs in position on stem of asparagus; B, one egg much enlarged; C, young larva. Cambridge.

ii. The Camptosomes, as we have already noticed, are distinguished by a peculiar structure of the abdomen. This character appears to be connected with a very remarkable habit, viz. the formation of a case to envelop the egg. The tip of the abdomen is somewhat curved downwards, and, in the female, bears a hollow near the extremity; when an egg is extruded the female holds it in this hollow by means of the hind legs, and envelops it with a covering said to be excrementitious. When the larva hatches, it remains within this case, and subsequently enlarges it by additions from its own body. The beautiful Insects of the genus Cryptocephalus, which is fairly well represented in Britain, belong to this division. The exotic group Megalopodes is incorrectly placed in Camptosomes; the side pieces of the prothorax meet in it behind the middle coxae, as they do in Rhynchophora. The species of Megalopodes stridulate by means of an area on the base of the meso-scutellum rubbed by a ridge inside the pronotum, as in the Cerambycidæ.

iii. The division Cyclica includes the great majority of Chryomelidae; we have not less than 170 species in Britain. The larvae live, like those of Lepidoptera, at the expense of foliage, and the species frequently multiply to such an extent as to be injurious. Some of them are destroyed in great numbers by Hymenopterous parasites, the Braconid genus Perilitus being one of the best known of these; in some cases the parasite deposits its eggs in either the larva or perfect Insect of the beetle, and the metamorphoses of the parasites in the latter case are sometimes, if not usually, completed, the larvae emerging from the living beetles for pupation.

iv. The Cryptostomes, though comparatively few in number of species, include some very remarkable beetles. There are two groups, Hispides and Cassidides. The former are almost peculiar to the tropics and are not represented by any species in the British fauna. The head in this group is not concealed; but in the Cassidides the margins of the upper surface are more or less expanded, so that the head is usually completely hidden by the expansion of the pronotum. Both the groups are characterised by the antennae being inserted very near together, and by the short claw-joint of the feet. Hispa is one of the most extensive of the numerous genera of Hispides, and is remarkable from the imago being covered on the surface with long, sharp spines. But little is known as to the metamorphosis, beyond the fact already alluded to, that the larvae of several species mine the interior of leaves. The larva of Hispa testacea, according to Perris,[^[147]] makes use of the leaves of Cistus salvifolius in Southern Europe; it is broad and flat, and possessed of six short legs. The eggs are not deposited by the parents inside the leaves, but are probably attached to various parts of the plant. After hatching, the young larva enters a leaf, and feeds on the parenchyma without rupturing the epidermis; but when it has consumed about three-fourths of the soft interior of the leaf it ruptures the epidermis of the upper surface, and seeks another leaf; this found, it places itself on the midrib, tears the upper epidermis, and lodges itself in the leaf. In the case of this second leaf it attacks the parenchyma in the neighbourhood of the petiole, and so forms an irregular tube which has an open mouth, the point of entry. In this tube it undergoes its metamorphosis. Each larva, it is said, always makes use of two leaves, and of two opposed leaves. A knowledge of the habits of some of the larger of the exotic Hispides would be of much interest.

Fig. 144—Pupa of Cassidid beetle (? Aspidomorpha sp.). A, With appendage extended; B, with the appendage reposing on the back. New Britain.

The Cassidides, in addition to the curious marginal expansion of their upper surface, have the power of withdrawing the head into the thorax, and hence they are often called shield or tortoise-beetles. They exhibit considerable variety in form and colour, and some of them display a peculiar metallic reflection of great delicacy and beauty; this disappears entirely after death, but it may be restored by thoroughly moistening the dead Insect. The colour, therefore, probably depends on the presence of water in the integument. The larvae of Cassidides are notorious on account of their habit of covering their bodies with dried excrement, for which purpose they are provided with a forked process at the posterior extremity; this serves to place the protecting matter in a proper position and to retain it there. The excrement assumes in various species forms so peculiar that they cannot be considered merely incidental. In several species this covering-matter is like lichen. This is the case with Dolichotoma palmarum, the larva of which has, in place of the usual fork, a more complex appendage on the back for the purpose of preparing and retaining its peculiar costume. The pupae, too, sometimes retain the larval skin. An extremely remarkable pupa of a Cassidid—possibly of the genus Aspidomorpha—was recently found by Dr. Arthur Willey in New Britain (Fig. 144). The back of the pupa is covered with a complex appendage, so that the creature has no resemblance to an Insect; this appendage is perhaps capable of being moved, or even extended (Fig. 144, A), during life. Whether it may be formed by the retention of portions of the moulted skins of the larva we cannot say with certainty.

Fig. 145—Nest of intestinally-made filaments under which the larva of Porphyraspis tristis lives.

The most remarkable of the Cassidid coverings yet discovered are those formed by certain small beetles of the tropical American genus Porphyraspis. P. tristis is apparently a common Insect at Bahia, where it lives on a cocoa-palm. The larva is short and broad, and completely covers itself with a very dense coat of fibres, each many times the length of the body, and elaborately curved so as to form a round nest under which the larva lives. On examination it is found that these long threads are all attached to the anal extremity of the Insect, and there seems no alternative to believing that each thread is formed by small pieces of fibre that have passed through the alimentary canal, and are subsequently stuck together, end to end. The process of forming these long fibres, each one from scores of pieces of excrement, and giving them the appropriate curve, is truly remarkable. The fibres nearest to the body of the larva are abruptly curled so as to fit exactly, and make an even surface; but the outside fibres stand out in a somewhat bushy fashion. The construction is much like that of a tiny bird's nest. Señor Lacerda informed the writer that the larva makes a nest as soon as it is hatched. Another Porphyraspis—P. palmarum—has been recorded as forming similar nests on a species of Thrinax in St. Domingo. Candèze says[^[148]] that when it has completed its growth the larva ejects on to the leaf a quantity of semi-liquid matter, and this, on drying, sticks the nest to the leaf, so that the metamorphosis is effected under shelter.

Fam. 79. Cerambycidae (Longicorns).—Form usually oblong, not much curved in outline at the sides; surface very frequently rendered dull by a very minute hairiness, which often forms a pattern; antennae usually long, and their insertion much embraced by the eyes. This great family of beetles includes some 12,000 or 13,000 known species. The elegance and variety of their forms and the charm of their colours have caused them to attract much attention, so that it is probable that a larger proportion of the existing species have been obtained than is the case in any other of the great families of Coleoptera. Still it is not likely that one-half of the living forms are known. It is not possible at present to point out any one character of importance to distinguish Cerambycidae from Chrysomelidae, though the members of the two families have, as a rule, but little resemblance in external appearance. Most of them live on, or in, wood, though many are nourished in the stems of herbaceous plants. The larvae live a life of concealment, and are soft, whitish grubs with powerful mandibles, and usually with a comparatively small head, which is not much exserted from the thorax. Most of them are without legs, but a good many have three pairs of small legs, and there are numerous cases in which the surface of the body is furnished above or below with swellings believed to act as pseudopods (Fig. 84), and help the larvae to move about in their galleries; but this is probably not the sole function of these organs, as their surface is varied in character, and often not of a kind that appears specially adapted to assist in locomotion. There is a slight general resemblance between the larvae of Cerambycidae and those of Buprestidae, and when the thorax of a Longicorn larva is unusually broad, e.g. Astynomus, this similarity is very pronounced.

Fig. 146—Saperda populnea. Britain.

The modes of life of Cerambycid larvae exhibit considerable variety, and much perfection of instinct is displayed by the larvae, as well as by the mother beetles. The larvae of Saperda populnea, are common in certain woods in the South of England in the stems of aspen; they consume only a small quantity of the interior of the stem, and are probably nourished by an afflux of sap to the spot where they are situated. Elaphidion villosum is called the oak-pruner in North America. The parent beetle lays an egg near the axilla of a leaf-stalk or small stem, and the young larva enters this and feeds on the tender material; as it grows it enters a larger limb, and makes an incision within this in such a manner that the wood falls to the ground with the larva within it, the dead wood serving subsequently as pabulum and as a shelter, within which the metamorphosis is completed. The species of the American genus Oncideres are called girdlers, because the parent beetle, after laying an egg in a small branch, girdles this round with a deep incision, so that the portion containing the larva sooner or later falls to the ground. The growth of a Longicorn larva frequently takes more than a year, and under certain circumstances it may be enormously prolonged. Monohammus confusus has been known to issue from wooden furniture in a dwelling-house when the furniture was fifteen years old. Individuals of another Longicorn have issued from the wood of a table, twenty and even twenty-eight years after the felling of the tree from which the furniture was made. Sereno Watson has related a case from which it appears probable that the life of a Longicorn beetle extended over at least forty-five years.[^[149]] It is generally assumed that the prolongation of life in these cases is due to the beetle resting quiescent for long after it has completed the metamorphosis. Recent knowledge, however, renders it more probable that it is the larval life that is prolonged; the larva continuing to feed, but gaining little or no nutriment from the dry wood in these unnatural conditions. Mr. C. O. Waterhouse had for some years a Longicorn larva under observation, feeding in this way in the wood of a boot-tree;[^[150]] the burrows in the wood contained a great deal of minute dust indicating that the larva passed much matter through the alimentary canal, probably with little result in the way of nutriment.

There are numerous Longicorns that bear a great resemblance in form and colour to Insects to which they are not related. Haensch[^[151]] has noticed that species of the genus Odontocera resemble various Hymenoptera, one species being called O. braconoides; he also observed that these Hymenoptera-like Longicorns, instead of withdrawing their underwings under the elytra as beetles generally do, vibrate them rapidly like Hymenoptera. A large number of Longicorns stridulate loudly by rubbing a ridge inside the pronotum on a highly specialised, striate surface at the base of the scutellum, and therefore covered up when the Insect is contracted in repose. A few produce noise by rubbing the hind femora against the edges of the elytra, somewhat after the fashion of grasshoppers. In this case there appears to be comparatively little speciality of structure, the femora bearing, however, more or less distinct small granules. The species of the Hawaiian genus Plagithmysus produce sound in both these manners, the thoracic stridulating organ being beautifully developed, while in some species the margin of the elytra and base of the femora are also well adapted for the purpose of sound-production, and in a few species of the genus there are also highly-developed stridulating surfaces on the hind and middle coxae. This is the only case in which a beetle is known to possess more than one set of sound-organs in the imago state.

Three divisions of this family are distinguished, viz.—

1. Front coxae large and transverse; prothorax with distinct side margins. .......... Sub-fam. 1. Prionides.

2. Front coxae not greatly extended transversely, thorax not margined; last joint of maxillary palpus not pointed, usually broader (more or less) than the preceding joint. .......... Sub-fam. 2. Cerambycides.

3. Front coxae usually round and deeply embedded; last joint of maxillary palpus pointed; front tibiae with a more or less distinct, slanting groove on the inner side. .......... Sub-fam. 3. Lamiides.

The Prionides are on the average considerably larger in size than the members of the other divisions, and they include some of the largest of Insects. The Amazonian Titanus giganteus and the Fijian Macrotoma heros are amongst the most gigantic. Some of the Prionides have a great development of the mandibles in the male sex analogous to that we have already noticed in Lucanidae. The larvae of the large Prionides appear in various parts of the world to have been a favourite food with native tribes, and Lumholz states that they are really good eating. In consequence of the destruction of forests that has progressed so largely of late years these gigantic Prionides have become much rarer.

Several aberrant forms are included in Prionides. The genus Parandra has five-jointed tarsi; the third joint being much smaller than usual, so that the fourth joint is not concealed by it. The Brazilian Hypocephalus armatus was for long a subject of dispute as to its natural position, and was placed by different authorities in widely-separated families of Coleoptera. The structure of this aberrant Longicorn seems to be only explicable on the hypothesis of warfare amongst the males.[^[152]] Nothing is, however, known as to the habits and history of the Insect, and only one or two specimens of the female have yet been obtained.

The family Spondylidae has been proposed for some of these aberrant Longicorns, but as it includes but very few, and highly discrepant, species, it is neither natural nor of much use for systematic purposes.

The Lamiides are the most highly specialised division of the Longicorns, and includes the larger number of the species. The front of the head is usually placed at right angles to the vertex, and in some cases (groups Hippopsini, Spalacopsini) it is strongly inflexed, so that the mouth is placed on the under side of the head. The extension of the eyes round the antennae is accompanied by very curious shapes of those organs, and not infrequently each eye is divided into two more or less widely-separated parts, so that the Insect has, on the external surface, four eyes.

Series VI. Rhynchophora.

Head more or less prolonged in front to form a snout or beak, called rostrum. Tarsi four-jointed, usually at least the third joint broad and densely pubescent beneath.

This enormous series includes about 25,000 species, and as may well be imagined shows a great variety of structure amongst its forms. The vast majority may, however, be readily recognised by the two characters mentioned above. There are some cases in which the beak is indistinct, and others in which the tarsi are five-jointed (Dryophthorus), and even slender (Platypides). In these cases a close examination shows that the gular region on the middle of the back of the under surface of the head cannot be detected, and that the back of the prosternum is very strongly consolidated by the side-pieces of the thorax meeting together and being very firmly joined behind the coxae. The beak is in the great majority perfectly distinct, though it varies so extremely in form that it can only be briefly described by saying that it is a prolongation of the head in front of the eyes, or that the antennae are inserted on its sides near to, or far from, the tip. It has been ascertained in many cases that the rostrum is used by the female to assist in placing the eggs in suitable places, a hole being bored with it; in some cases it is also used to push the egg far into the hole in which it has previously been placed by the ovipositor; but there are many forms in which it is fairly certain that it is not so used. What purpose it serves in the male is totally unknown. In many members of the series, the rostrum differs in form in the two sexes, and in most, if not in all, these cases it is clear that the distinctions tend in the direction of making the beak of the female more efficient for the mechanical purpose we have mentioned.

Fig. 147.—Eugnoristus monachus ♀. Madagascar. A, The imago; B, front of pronotum, head, and rostrum.

It was proposed by Leconte and Horn to separate this series from all the other Coleoptera as a primary division, and they looked on it as of lower or more imperfect structure. Packard has very properly protested against this interpretation; and there seems to be no reason whatever for considering the Rhynchophora as "lower" than other beetles; indeed we should be inclined to place such forms as Calandrides amongst the most perfect of Insects; their external structure (as shown by Eugnoristus monachus, Fig. 147) being truly admirable.

Only four families of Rhynchophora can be at present accepted as satisfactory; one of these—Curculionidae—includes an enormous majority of the whole series. Though it is probable that it will ultimately be divided into several families, the attempts to that end that have already been made are not satisfactory.

Fam. 80. Anthribidae.—Palpi usually not covered, but distinct and flexible. Antennae often long, not elbowed, the first joint not very long. Third joint of tarsus small, usually much concealed by being embraced by the second joint. Pygidium exposed; propygidium deeply grooved in the middle. This family includes 800 or more species, which are mostly tropical; it is very sparsely represented in the faunas of Europe and North America. It is quite distinct from Curculionidae with which it was formerly associated. It contains many graceful Insects having a certain resemblance with Longicorns on account of the large development of the antennae. The habits and metamorphoses are but little known. It seems probable that many species find their nutriment in old wood or boleti The larvae of some genera (Cratoparis and Araeocerus) have legs, but in others the legs are wanting, and the larvae are said to completely resemble those of Curculionidae. In the larva of our tiny British species, Choragus sheppardi, the legs are replaced by three pairs of thoracic, sac-like pseudopods. This Insect makes burrows in dead branches of hawthorn. The larvae of the genus Brachytarsus have been ascertained to prey on Coccidae.

Fig. 148—Platyrhinus latirostris, Anthribidae. Britain. A, the perfect Insect; B, tarsus and tip of tibia.

Fam. 81. Curculionidae (Weevils).—The beak of very variable length and thickness; the palpi small, nearly always concealed within the mouth, short, and rigid. Labrum absent. Antennae of the majority elbowed, i.e. with the basal joint longer, and so formed that when it is laterally extended the other joints can be placed in a forward direction. This enormous family includes about 20,000 known species, and yet a large portion of the species yearly brought from the tropics still prove to be new. The rostrum or beak exhibits excessive variety in form, and is in many cases different in the sexes; in this case it is usually longer and thinner in the female. As the rostrum is one of the chief characters by which a member of the family may be recognised, it is necessary to inform the student that in certain forms (the Australian Amycterides, e.g.) the organ in question may be so short and thick that it is almost absent. In these cases the Insect may be identified as a Curculionid by the gular area being absent on the under side of the head, and by the concealment of the palpi. The tarsi are usually of the same nature as those of Phytophaga, already described, but the true fourth joint is less visible. In the Brachycerides this joint is not present, and the third joint is not lobed. The palpi are flexible and more or less exserted in a very few species (Rhynchitides); in Rhinomacerides there is also present a minute labrum. The front coxae are deeply embedded, and in many forms the prosternum is peculiar in structure; the side-pieces (epimera) meeting at the back of the prosternum in the middle line. This, however, is not universal in the family, and it occurs in some other beetles (e.g., Megalopodides of the Phytophaga). The larvae are without legs. They are vegetarian, the eggs being deposited by the mother-beetle in the midst of the food. These larvae may be distinguished from those of Longicorns by the general form, which is sub-cylindric or rather convex, not flattened, and more particularly by the free, exserted head, the mouth being directed downwards; the attitude is generally a curve, and the anterior part of the body is a little the thicker. No part of plants is exempt from the attacks of the larvae of Curculionidae; buds, twigs, leaves, flowers, fruits, bark, pith, roots and galls may each be the special food of some Curculionid. Certain species of the sub-families Rhynchitides and Attelabides prepare leaves in an elaborate manner to serve as food and dwelling for their young. If young birches, or birch bushes from 5 to 10 feet in height, be looked at in the summer, one may often notice that some of the leaves are rolled so as to form, each one, a little funnel. This is the work of Rhynchites (or Deporaus) betulae, a little Curculionid beetle (Fig. 149). An inspection of one of these funnels will show that it is very skilfully constructed.

Fig. 149.—The leaf-rolling of Rhynchites betulae. Britain. A, Female beetle, magnified; B, the beetle forming the first incision on a leaf; C, the completed roll. (B and C after Debey.)

The whole of a leaf is not used in the formation of a funnel, cuts being made across the leaf in suitable directions. The beetle standing on a leaf, as shown in the figure, proceeds to cut with its mandibles an incision shaped like an erect S, commencing at a certain part of the circumference, and ending at the midrib of the leaf; the beetle then goes to the other side of the midrib, and continues its incision so as to form another S-like curve considerably different from the first; being prostrate and less abrupt. Thus the blade of the leaf is divided into two halves by certain curved incisions, the midrib remaining intact. The little funnel-twister now commences to roll up the leaf to form the funnel; and this part of the work is greatly facilitated by the shape of the incisions. Going back to the spot where it commenced work, by the aid of its legs it rolls one side of the leaf round an ideal axis, somewhat on the same plan as that adopted by a grocer in forming a paper-funnel for sugar. The incisions are found to be just of the right shape to make the overlaps in the rolling, and to retain them rolled-up with the least tendency to spring back. After some other operations destined to facilitate subsequent parts of its task, the beetle enters the rolled-up part of the leaf and brings it more perfectly together; it again comes out and, pursuing a different system, holds on with the legs of one side of the body to the roll, and with the other legs drags to it the portion of the leaf on the other side of the midrib so as to wrap this part (i.e. the result of its second incision) round the part of the funnel already constructed. This being done the Insect again enters the funnel, bites three or four small cavities on the inside of the leafy wall and deposits an egg in each. Afterwards it emerges and fits the overlaps together in a more perfect manner so as to somewhat contract the funnel and make it firmer; then proceeding to the tip, this is operated on by another series of engineering processes and made to close the orifice; this part of the operation being analogous to the closing by the grocer of his paper-funnel after the sugar has been put in. The operation of the beetle is, however, much more complex, for it actually makes a sort of second small funnel of the tip of the leaf, bends this in, and retains it by tucking in some little projections. The work, which has probably lasted about an hour, being now completed, the creature takes a longer or shorter rest before commencing another funnel. We have given only a sketch of the chief points of the work, omitting reference to smaller artifices of the craft master; but we may remark that the curved incisions made by the beetle have been examined by mathematicians and duly extolled as being conducted on highly satisfactory mathematical principles. It is impossible at present for us to form any conception as to the beetle's conceptions in carrying out this complex set of operations. Our perplexity is increased if we recollect its life-history, for we then see that neither precept or example can have initiated its proceedings, and that imitation is out of the question. The eggs hatch in their dark place, giving rise to an eyeless maggot, which ultimately leaves the funnel for the earth. The parts of this maggot subsequently undergo complete change to produce the motionless pupa of entirely different form, from which emerges the perfect Insect. Hence the beetle cannot be considered to have ever seen a funnel, and certainly has never witnessed the construction of one, though, when disclosed, it almost immediately sets to work to make funnels on the complex and perfect system we have so imperfectly described. More general considerations only add to the perplexity we must feel when reflecting on this subject. Why does the Insect construct the funnel at all? As a matter of protection it appears to be of little use, for the larvae are known to suffer from the attacks of parasites as other Insects do. We have not the least reason for supposing that this mode of life for a larva is, so far as utility is concerned, better than a more simple and usual one. Indeed, extraordinary as this may appear, it is well known that other species of the same genus adopt a simple mode of life, laying their eggs in young fruits or buds. We think it possible, however, that a knowledge of the mode of feeding of this larva may show that a more perfect nutrition is obtained from a well-constructed cylinder, and if so this would to a slight extent satisfy our longing for explanation, though throwing no light whatever on the physiology or psychology of the artificer, and leaving us hopelessly perplexed as to why a beetle in ages long gone by should or could adopt a mode of life that by long processes of evolution should, after enormous difficulties have been overcome, attain the perfection we admire.[^[153]]

Fam. 82. Scolytidae.—Rostrum extremely short, broad; tibiae frequently denticulate externally; antennae short, with a broad club. This family is not at all sharply distinguished from certain groups of Curculionidae (from Cossonides e.g.), but as the species have somewhat different habits, and in the majority of cases can be readily distinguished, it is an advantage to separate the two families. About 1400 species are at present known. Most of them are wood- and bark-feeders; some bore into hard wood; a few mine in twigs or small branches of trees, but the majority live in the inner layers of the bark; and this also serves as the nidus of the larvae. A small number of species have been found to inhabit the stems of herbaceous plants, or to live in dry fruits. Owing to their retiring habits they are rarely seen except by those who seek them in their abodes, when they may often be found in great profusion. The mother-beetle bores into the suitable layer of the bark, forming a sort of tunnel and depositing eggs therein. The young larvae start each one a tunnel of its own, diverging from the parent tunnel; hence each batch of larvae produces a system of tunnels, starting from the parents' burrow, and in many species these burrows are characteristic in form and direction, so that the work of particular Scolytids can be recognised by the initiated.

The Platypides bore into the wood of trees and stumps; they are chiefly exotic, and little is known about them. They are the most aberrant of all Rhynchophora, the head being remarkably short, flat in front, with the mouth placed on the under surface of the head, there being no trace of a rostrum: the tarsi are elongate and slender, the third joint not being at all lobed, while the true fourth joint is visible. Hence they have not the appearance of Rhynchophora. Some authorities treat the Platypides as a distinct family.

Some of the members of the group Tomicides also bore into the wood. Recent observations have shown that there is an important feature in the economy of certain of these wood-borers, inasmuch as they live gregariously in the burrow, and feed on peculiar fungi that develop there, and are called ambrosia. According to Hubbard[^[154]], some species cultivate these fungi, making elaborate preparations to start their growth. The fungi, however, sometimes increase to such an extent as to seal up the burrows, and kill the Insects by suffocation.

Scolytidae sometimes multiply to an enormous extent, attacking and destroying the trees in wooded regions. Much discussion has taken place as to whether or not they are really injurious. It is contended by one set of partisans that they attack only timber that is in an unhealthy, dying, or dead condition. It may be admitted that this is usually the case; yet when they occur in enormous numbers they may attack timber that is in a sort of neutral state of health, and so diminish its vigour, and finally cause its destruction. Hence it is of great importance that they should be watched by competent foresters.

The larvae of Scolytidae are said to completely resemble those of Curculionidae: except in the group Platypides, where the body is straight and almost cylindrical, and terminates in an oblique truncation bearing a short hard spine.[^[155]]

Fam. 83. Brenthidae.—Form elongate; rostrum straight, directly continuing the long axis of the body, often so thick as to form an elongate head; antennae not elbowed. The Brenthidae form a family of about 800 species, remarkable for the excessive length and slenderness of some of its forms, and for the extreme difference in the sexes that frequently exists. It is well represented in the tropics only, and very little is known as to the natural history and development. These beetles are stated to be wood-feeders, and no doubt this is correct in the case of the majority of the species; but Mr. Lewis observed in Japan that Zemioses celtis and Cyphagogus segnipes are predaceous, and enter the burrows of wood-boring Insects to search for larvae as prey: they are very much modified in structure to permit this; and as the other members of the group Taphroderides are similar in structure, it is probable that they are all predaceous. Nothing whatever is known as to the larval history of these carnivorous forms. Indeed an uncertainty, almost complete, prevails as to the early stages of this family. Riley has given a sketch of a larva which he had no doubt was that of Eupsalis minuta, the North American representative of the family; if he is correct the larva differs from those of Curculionidae by its elongate form, and by the possession of thoracic legs: these, though small, are three-jointed. Descriptions, supposed to be those of Brenthid larvae, were formerly published by Harris and Motschoulsky; but it is now clear that both were mistaken.

Fig. 150—Eupsalis minuta. North America. A, Larva; B, pupa; C, female imago; D, head of male. (After Riley.)

In the higher forms of Brenthidae the rostrum of the female is perfectly cylindrical and polished, and the mandibles are minute, hard, pointed processes placed at its tip. This organ is admirably adapted to its purpose; it being used for boring a hole in wood or bark, in which an egg is subsequently deposited. The males in these cases are extremely different, so that considerable curiosity is felt as to why this should be so. In some cases their head is thick, and there may be no rostrum, while large powerful mandibles are present.

In other cases the rostrum is slender, but of enormous length, so that it may surpass in this respect the rest of the body, although this itself is so drawn out as to be quite exceptional in the Insect world:[^[156]] the antennae are inserted near the tip of the rostrum instead of near its base, as they are in the female. The size of the males is in these cases usually much larger than that of the female.[^[157]] The males of some species fight; they do not, however, wound their opponent, but merely frighten him away. In Eupsalis it appears that the rostrum of the female is apt to become fixed in the wood during her boring operations; and the male then extricates her by pressing his heavy prosternum against the tip of her abdomen; the stout forelegs of the female serve as a fulcrum and her long body as a lever, so that the effort of the male, exerted at one extremity of the body of the female, produces the required result at the other end of her body. The New Zealand Brenthid, Lasiorhynchus barbicornis, exhibits sexual disparity in an extreme degree: the length of the male is usually nearly twice that of the female, and his rostrum is enormous. It is at present impossible to assign any reason for this; observations made at the request of the writer by Mr. Helms some years ago, elicited the information that the female is indefatigable in her boring efforts, and that the huge male stands near by as a witness, apparently of the most apathetic kind.

Coleoptera of uncertain position.

There are three small groups that it is impossible at present to place in any of the great series of beetles.

Fam. 84. Aglycyderidae.—Tarsi three-jointed, the second joint lobed; head not prolonged to form a beak. The two most important features of Rhynchophora are absent in these Insects, while the other structural characters are very imperfectly known, many parts of the external skeleton being so completely fused that the details of structure are difficult of appreciation. Westwood considered the tarsi to be really four-jointed, but it is not at all clear that the minute knot he considered the third joint is more than the articulation of the elongate terminal joint. The family consists only of two or three species of Aglycyderes, one of which occurs in the Canary Islands, and one or two in New Zealand and New Caledonia. The former is believed to live in the stems of Euphorbia canariensis; a New Zealand species has been found in connection with the tree-ferna. Cyathea dealbata

Fig. 151—Aglycyderes setifer. Canary Islands. A, Imago; B, tarsus according to Westwood; C, according to nature; D, maxilla; E, labium.

Fam. 85. Proterhinidae.—Tarsi three-jointed, the second joint lobed; head of the male scarcely prolonged, but that of the female forming a definite rostrum; maxillae and ligula entirely covered by the mentum. As in the preceding family the sutures on the under side of the head and prosternum cannot be detected. The minute palpi are entirely enclosed in the buccal cavity. There is a very minute true third joint of the tarsus, at the base of the terminal joint, concealed between the lobes of the second joint. The family consists of the genus Proterhinus; it is confined to the Hawaiian Islands, where these Insects live on dead wood in the native forests. The genus is numerous in species and individuals.

Fig. 152—Proterhinus lecontei. Hawaiian Islands. A, Male; B, female; C, front foot, more magnified.

Strepsiptera (or Rhipiptera, Stylopidae).—Male small or minute; prothorax extremely small; mesothorax moderate, the elytra reduced to small, free slips; metathorax and wings very large; nervuration of the latter radiating, without cross nervules. Female a mere sac, with one extremity smaller and forming a sort of neck or head.

Fig. 153.—Sexes of Strepsiptera. A, Male of Stylops dalii (after Curtis); B, female of Xenos rossii (after von Siebold).

These curious Insects are parasitic in the interior of other Insects, of the Orders Hymenoptera and Hemiptera. Their structure and their life-histories entitle them to be ranked as the most abnormal of all Insects, and entomologists are not agreed as to whether they are aberrant Coleoptera or a distinct Order. The newly-hatched larva is a minute triungulin (Fig. 154), somewhat like that of Meloe; it fixes itself to the skin of the larva of a Hymenopterous Insect, penetrates into the interior, and there undergoes its metamorphoses, the male emerging to enjoy a brief period of an abnormally active, indeed agitated, existence, while the female never moves. It is important to note that these Strepsiptera do not, like most other internal parasites, produce the death of their hosts; these complete their metamorphosis, and the development of the parasite goes on simultaneously with that of the host, so that the imago of the Strepsipteron is found only in the imago of the host.[^[158]] After the young Stylops has entered its host it feeds for a week or so on the fat-body (apparently by a process of suction), then moults and assumes the condition of a footless maggot, in which state it remains till growth is completed. At the latter part of this period the history diverges according to sex; the female undergoes only a slight metamorphic development of certain parts, accompanied apparently by actual degradation of other parts; while the male goes on to pupation, as is normal in Insects. (We may remark that the great features of the development of the sexes are parallel with those of Coccidae in Hemiptera.) When the Hymenopterous larva changes to a pupa, the larva of the Strepsipteron pushes one extremity of its body between two of the abdominal rings of its host, so that this extremity becomes external, and in this position it completes its metamorphosis, the male emerging very soon after the host has become an active winged Insect, while the female undergoes no further change of position, but becomes a sac, in the interior of which young develop in enormous numbers, finally emerging from the mother-sac in the form of the little triungulins we have already mentioned. This is all that can be given at present as a general account; many points of the natural history are still obscure, others have been merely guessed; while some appear to differ greatly in the different forms. A few brief remarks as to these points must suffice.

Bees carrying, or that have carried, Strepsiptera, are said to be stylopised (it being a species of the genus Stylops that chiefly infests bees); the term is also used with a wider application, all Insects that carry a Strepsipterous parasite being termed stylopised, though it may be a Strepsipteron of a genus very different from Stylops that attacks them. The development of one or more Strepsiptera in an Insect usually causes some deformity in the abdomen of its host, and effects considerable changes in the condition of its internal organs, and also in some of the external characters. Great difference of opinion prevails as to what these changes are; it is clear, however, that they vary much according to the species, and also according to the extent of the stylopisation. Usually only one Stylops is developed in a bee; but two, three, and even four have been observed:[^[159]] and in the case of the wasp, Polistes, Hubbard has observed that a single individual may bear eight or ten individuals of its Strepsipteron (Xenos, n. sp.?).

Fig. 154—Young larva of Stylops on a bee's-hair. Greatly magnified. (After Newport.)

There is no exact information as to how the young triungulins find their way to the bee-larvae they live in. Here again the discrepancy of opinion that prevails is probably due to great difference really existing as to the method. When a Stylops carried by an Insect (a Hymenopteron, be it noted, for we have no information whatever as to Hemiptera) produces young, they cover the body of the host as if it were powdered, being excessively minute and their numbers very great; many hundreds, if not thousands, of young being produced by a single Stylops. The species of the wasp genus Polistes are specially subject to the attacks of Stylops; they are social Insects, and a stylopised specimen being sickly does not as a rule leave the nest; in this case the Stylops larva may therefore have but little difficulty in finding its way to a Hymenopterous larva, for even though it may have to live for months before it has the chance of attaching itself to a nest-building female, yet it is clearly in the right neighbourhood. The bee genus Andrena has, however, quite different habits; normally a single female makes her nest underground; but in the case of a stylopised female it is certain that no nest is built, and no larvae produced by a stylopised example, so that the young triungulins must leave the body of the bee in order to come near their prey. They can be active, and have great powers of leaping, so that it is perhaps in this way possible for them to attach themselves to a healthy female bee.

Fig. 155.—Portion of early stages of Xenos rossii. (After von Siebold.) A, Small male larva; B, small female larva; C, full-grown male larva; D, full-grown female larva; E, the so-called "cephalothorax" and adjacent segment of adult female. (The newly-hatched larva is very much like that of Stylops shown in Fig. 154.)

We have still only very imperfect knowledge as to the structure and development of Strepsiptera. Indeed but little information has been obtained since 1843.[^[160]] Before that time the mature female was supposed to be a larva, and the triungulins found in it to be parasites. Although the erroneous character of these views has been made clear, the problems that have been suggested present great difficulties. Apparently the change from the triungulin condition (Fig. 154) to the parasitic larvae (Fig. 155, A, B) is extremely great and abrupt, and it appears also that during the larval growth considerable sexual differentiation occurs (Fig. 155, C, D); details are, however, wanting, and there exists but little information as to the later stages. Hence it is scarcely a matter for surprise that authorities differ as to which is the head and which the anal extremity of the adult female. Von Siebold apparently entertained no doubt as to the part of the female that is extruded being the anterior extremity; indeed he called it a cephalothorax. Supposing this view to be correct, we are met by the extraordinary facts that the female extrudes the head for copulatory purposes, that the genital orifice is placed thereon, and that the young escape by it. Meinert[^[161]] contends that the so-called cephalothorax of the adult is the anal extremity, and that fertilisation and the escape of the young are effected by the natural passages, the anterior parts of the body being affected by a complete degeneration. Nassonoff, in controversion of Meinert, has recently pointed out that the "cephalothorax" of the young is shown by the nervous system to be the anterior extremity. It still remains, however, to be shewn that the "cephalothorax" of the adult female corresponds with that of the young, and we shall not be surprised if Meinert prove to be correct. The internal anatomy and the processes of oogenesis appear to be of a very unusual character, but their details are far from clear. Brandt has given some particulars as to the nervous system; though he does not say whether taken from the male or female, we may presume it to be from the former; there is a supra-oesophageal ganglion, and near it a large mass which consists of two parts, the anterior representing the sub-oesophageal and the first thoracic ganglia, while the posterior represents two of the thoracic and most of the abdominal ganglia of other Insects; at the posterior extremity, connected with the other ganglia by a very long and slender commissure, there is another abdominal ganglion.[^[162]]

It is a matter of great difficulty to procure material for the prosecution of this study; the fact that the instars to be observed exist only in the interior of a few Hymenopterous larvae, which in the case of the bee, Andrena, are concealed under ground; and in the case of the wasps, Polistes, placed in cells in a nest of wasps, adds greatly to the difficulty. It is therefore of interest to know that Strepsiptera occur in Insects with incomplete metamorphosis. They have been observed in several species of Homoptera; and the writer has a large Pentatomid bug of the genus Callidea, which bears a female Strepsipteron apparently of large size. This bug[^[163]] is abundant and widely distributed in Eastern Asia, and it may prove comparatively easy to keep stylopised examples under observation. Both v. Siebold and Nassonoff think parthenogenesis occurs in Strepsiptera, but there appear to be no facts to warrant this supposition. Von Siebold speaks of the phenomena of Strepsipterous reproduction as paedogenesis, or pseudo-paedogenesis, but we must agree with Meinert that they cannot be so classed.

The males of Strepsiptera live for only a very short time, and are very difficult of observation. According to Hubbard the males of Xenos dash about so rapidly that the eye cannot see them, and they create great agitation amongst the wasps in the colonies of which they are bred. Apparently they are produced in great numbers, and their life consists of only fifteen or twenty minutes of fiery energy. The males of Stylops are not exposed to such dangers as those of Xenos, and apparently live somewhat longer—a day or two, and even three days are on record. The individuals of Andrena parasitised by Stylops are apparently greatly affected in their economy and appear earlier in the season than other individuals; this perhaps may be a reason, coupled with their short lives, for their being comparatively rarely met with by entomologists.

Fig. 156.—Abdomen of a wasp (Polistes hebraeus) with a Strepsipteron (Xenos ♀) in position, one of the dorsal plates of the wasp's abdomen being removed. a, Projection of part of the parasite; b, line indicating the position of the removed dorsal plate.

It is not possible at present to form a valid opinion as to whether Stylopidae are a division of Coleoptera or a separate Order. Von Siebold considered them a distinct Order, and Nassonoff, who has recently discussed the question, is also of that opinion.

CHAPTER VI

LEPIDOPTERA—OR BUTTERFLIES AND MOTHS

Order VI. Lepidoptera.

Wings four; body and wings covered with scales usually variegate in colour, and on the body frequently more or less like hair: nervures moderate in number, at the periphery of one wing not exceeding fifteen, but little irregular; cross-nervules not more than four, there being usually only one or two closed cells on each wing, occasionally none. Imago with mouth incapable of biting, usually forming a long coiled proboscis capable of protrusion. Metamorphosis great and abrupt; the wings developed inside the body; the larva with large or moderate head and strong mandibles. Pupa with the appendages usually adpressed and cemented to the body so that it presents a more or less even, horny exterior, occasionally varied by projections that are not the appendages and that may make the form very irregular: in many of the smaller forms the appendages are only imperfectly cemented to the body.

Lepidoptera, or butterflies and moths, are so far as ornament is concerned the highest of the Insect world. In respect of intelligence the Order is inferior to the Hymenoptera, in the mechanical adaptation of the parts of the body it is inferior to Coleoptera, and in perfection of metamorphosis it is second to Diptera. The mouth of Lepidoptera is quite peculiar; the proboscis—the part of the apparatus for the prehension of food—is anatomically very different from the proboscis of the other Insects that suck, and finds its nearest analogue in the extreme elongation of the maxillae of certain Coleoptera, e.g. Nemognatha. The female has no gonapophyses, though in certain exceptional forms of Tineidae, there are modifications of structure connected with the terminal segments, that have as yet been only imperfectly investigated. As a rule, the egg is simply deposited on some living vegetable and fastened thereto. Lepidoptera are the most exclusively vegetarian of all the Orders of Insects; a certain number of their larvae prey on Insects that are themselves filled with vegetable juices (Coccidae, Aphidae) and a very small number (Tinea, etc.) eat animal matter. In general the nutriment appears to be drawn exclusively from the fluids of the vegetables, the solid matter passing from the alimentary canal in large quantity in the form of little pellets usually dry, and called frass. Hence the quantity of food ingested is large, and when the individuals unduly increase in number, forest trees over large areas are sometimes completely defoliated by the caterpillars.

Fig. 157.—Metamorphosis of a Lepidopteron (Rhegmatophila alpina, Notodontidae). (After Poujade, Ann. Soc. ent. France, 1891.) Europe. A, Egg; B, young larva, about to moult; C, adult larva; D, head and first body-segment of adult larva, magnified; E, pupa, × 2⁄1; F, male moth in repose; G, female moth in repose.

Lepidoptera pass a larger portion of their lives in the pupal stage than most other Insects do; frequently during nine months of the year the Lepidopteron may be a pupa. In other Orders of Insects it would appear that the tendency of the higher forms is to shorten the pupal period, and when much time has to be passed between the end of the feeding up of the larva and the appearance of the imago, to pass this time as much as possible in the form of a resting-larva, and as little as may be in the form of a pupa; in Lepidoptera the reverse is the case; the resting-larva period being usually reduced to a day or two. Hence we can understand the importance of a hard skin to the pupa. There are, however, numerous Lepidopterous pupae where the skin does not attain the condition of hardness that is secured for the higher forms by the chitinous exudation we have mentioned; and there are also cases where there is a prolonged resting-larva period: for instance Galleria mellonella spins a cocoon in the autumn and remains in it as a resting larva all the winter, becoming a pupa only in the spring. In many of these cases the resting-larva is protected by a cocoon. It is probable that the chief advantage of the perfect chitinous exudation of the Lepidopterous pupa is to prevent the tiny, complex organisation from the effects of undue transpiration. Bataillon has suggested that the relation of the fluid contents of the pupa to air and moisture are of great importance in the physiology of metamorphosis.

The duration of life is very different in various forms of Lepidoptera. It is known that certain species (Ephestia kuehniella, e.g.) may go through at least five generations a year. On the other hand, certain species that feed on wood or roots may take three years to complete their life-history; and it is probable that some of the forms of Hepialidae are even longer lived than this.

Lepidoptera have always been a favourite Order with entomologists, but no good list of the species has ever been made, and it would be a difficult matter to say how many species are at present known, but it can scarcely be less than 50,000. In Britain we have about 2000 species.

The close affinity of the Order with Trichoptera has long been recognised: Réaumur considered the latter to be practically Lepidoptera with aquatic habits, and Speyer pointed out the existence of very numerous points of similarity between the two. Brauer emphasised the existence of mandibles in the nymph of Trichoptera as an important distinction: the pupa of Micropteryx (Fig. 211) has however been recently shown to be similar to that of Trichoptera, so that unless it should be decided to transfer Micropteryx to Trichoptera, and then define Lepidoptera and Trichoptera as distinguished by the condition of the pupa, it would appear to be very difficult to retain the two groups as distinct.

Structure of Imago.—The head of a Lepidopteron is in large part made up of the compound eyes; in addition to these it frequently bears at the top a pair of small, simple eyes so much concealed by the scales as to cause us to wonder if seeing be carried on by them. The larger part of the front of the head is formed by the clypeus, which is separated by a well-marked line from the epicranium, the antennae being inserted on the latter near its point of junction with the former. There is sometimes (Saturnia, Castnia) on each side of the clypeus a deep pocket projecting into the head-cavity. The other parts of the head are but small. The occipital foramen is very large.[^[164]]

Fig. 158—External structure of a female butterfly, Anosia plexippus. (After Scudder.) a, Base of antenna; b, pronotum; b², scutum of mesothorax; c, clypeus; cx, coxa; d, scutellum; d¹, scutellum of metathorax; e, post-scutellum (= base of phragma); em, epimeron; ep, episternum; f, scutum of metathorax; m, basal part of proboscis (= maxilla); o, eye; p, labial palp; r, mesosternum; s, prothoracic spiracle; t, tegula; tr, trochanter; 1-9, dorsal plates of abdomen.

The antennae are always conspicuous, and are very various in form; they are composed of numerous segments, and in the males of many species attain a very complex structure, especially in Bombyces and Psychidae; they doubtless function in such cases as sense-organs for the discovery of the female.

The largest and most important of the mouth-parts are the maxillae and the labial palpi, the other parts being so small as to render their detection difficult. The labrum is a very short, comparatively broad piece, visible on the front edge of the clypeus; its lateral part usually forms a prominence which has often been mistaken for a mandible; Kellogg has applied the term "pilifer" to this part. In the middle of the labrum a small angular or tongue-like projection is seen just over the middle of the base of the proboscis; this little piece is considered by several authorities to be an epipharynx.

Fig. 159—Mouth of Lepidoptera. Tiger-moth, Arctia caja. A, Seen from front; B, from front and below, a, Clypeus; b, labrum; c, epipharynx; d, mandibular area; d′, prominence beneath mandibular area; e, one side of haustellum or proboscis; f, maxillary palp; g, labial palp.

Mandibles.—Savigny, Westwood, and others considered the parts of the labrum recently designated pilifers by Kellogg to be the rudimentary mandibles, but Walter has shown that this is not the case.[^[165]] The mandibles are usually indistinguishable, though they, or some prominence possibly connected with them,[^[166]] may frequently be detected in the neighbourhood of the pilifers; they are, according to Walter, largest and most perfectly developed in Eriocephala, a genus that was not distinguished by him from Micropteryx and was therefore termed "niedere Micropteryginen," i.e. lower Micropteryges. The opinion entertained by Walter that Micropteryx proper (his "höhere Micropteryginen") also possesses rudimentary mandibles is considered by Dr. Chapman, no doubt with reason, to be erroneous.[^[167]] The mandibles, however, in the vast majority of Lepidoptera can scarcely be said to exist at all in the imago; there being only an obtuse projection—without trace of articulation—on each side of the labrum; and even this projection is usually absent. Meinert recognised these projections as mandibles in Smerinthus populi, and Kellogg in Protoparce carolina, another large Sphinx moth. They appear to be unusually well developed in that group. In Castnia they are even more definite than they are in Sphingidae.

The Maxillae are chiefly devoted to the formation of the proboscis. Their basal portions are anatomically very indefinite, though they exist very intimately connected with the labium. Each usually bears a small tubercle or a segmented process, the representative of the maxillary palpus. The proboscis itself consists of the terminal, or outer, parts of the two maxillae, which parts are closely and beautifully coadapted to form the spirally coiled organ, that is sometimes, though incorrectly, called the tongue. The exact morphology of the Lepidopterous proboscis has not been established. The condition existing in the curious family Prodoxidae (see p. [432]), where a proboscis coexists with another structure called a maxillary tentacle, suggests a correspondence between the latter and the galea of a typical maxilla; and between the proboscis and the lacinia or inner lobe of a maxilla: but J. B. Smith is of opinion that the tentacle in question is a prolongation of the stipes. The condition of the parts in this anomalous family (Prodoxidae) has not, however, been thoroughly investigated, and Packard takes a different view of the proboscis; he considers that "it is the two galeae which become elongated, united and highly specialised to form the so-called tongue or glossa of all Lepidoptera above the Eriocephalidae."[^[168]] The proboscis in some cases becomes very remarkable, and in certain Sphingidae is said to attain, when unrolled, a length of ten inches. In some cases the maxillary lobes do not form a proboscis, but exist as delicate structures, pendulous from the mouth, without coadaptation (Zeuzera aesculi, the Wood-leopard moth). In other forms they are absent altogether (Cossus, e.g.), and in Hepialus we have failed to detect any evidence of the existence of the maxillae. On the other hand, in Micropteryx the maxillae are much more like those of a mandibulate Insect; and various other Microlepidoptera approach more or less a similar condition. In the genus last mentioned the maxillary palpi are largely developed, flexible and slender. According to Walter various forms of palpus intermediate between that of Micropteryx and the condition of rudimentary tubercle may be found amongst the Microlepidoptera.[^[169]]

Labium.—The labial palpi are usually largely developed, though but little flexible; they form conspicuous processes densely covered with scales or hairs, and curve forwards or upwards, rarely downwards, from the under side of the head, somewhat in the fashion of tusks. The other parts of the labium are frequently represented merely by a membranous structure, united with the maxillae and obstructing the cavity of the pharynx. Where the proboscis is absent it is difficult to find any orifice leading to the alimentary canal, such opening as may exist being concealed by the overhanging clypeus and labium. In some forms, Saturnia, e.g., there appears to be no buccal orifice whatever. In Hepialus the labium is in a very unusual condition; it projects externally in the position usually occupied by the labial palpi, these organs being themselves extremely short. It is very difficult to form an opinion as to the structure of the labium and other mouth-parts when the maxillae are not developed, as in these cases the parts are of a delicate membranous nature, and shrivel after death. This is the explanation of the fact that in descriptive works we find vague terms in use such as "mouth aborted" or "tongue absent."

The mouth of the Lepidopterous imago is a paradoxical structure; it differs very greatly from that of the larva, the changes during metamorphosis being extreme. We should thus be led to infer that it is of great importance to the creatures; but, on the other hand, the various structures that make up the mouth, as we have remarked, are frequently absent or reduced to insignificant proportions; and even in forms where the apparatus is highly developed the individuals seem to be able to accomplish oviposition without taking food, or after taking only very minute quantities. It is therefore difficult to understand why so great a change should occur during the metamorphosis of the Insects of this Order. It has been ascertained that in some forms where the mouth is atrophied the stomach is in a correlative condition; but we are not aware that any investigations have been made as to whether this correspondence is general or exceptional.

The exact mode in which the proboscis acts is in several respects still obscure, the views of Burmeister and Newport being in some points erroneous. Towards the tip of the proboscis there are some minute but complex structures considered by Fritz Müller to be sense-organs, and by Breitenbach to be mechanical instruments for irritating or lacerating the delicate tissues of blossoms. It is probable that Müller's view will prove to be correct. Nevertheless the proboscis has considerable power of penetration; there being a moth, "Ophideres fullonica" that causes considerable damage to crops of oranges by inserting its trunk through the peel so as to suck the juices.[^[170]] The canal formed by each maxilla opens into a cavity inside the front part of the head. This cavity, according to Burgess,[^[171]] is a sort of sac connected with five muscles, and by the aid of this apparatus the act of suction is performed: the diverticulum of the alimentary canal, usually called a sucking-stomach, not really possessing the function formerly attributed to it.

The Prothorax is very small, being reduced to a collar, between the head and the alitrunk, of just sufficient size to bear the front pair of legs. Its most remarkable feature is a pair of processes, frequently existing on the upper surface, called "patagia." These in many cases (especially in Noctuidae) are lobes capable of considerable movement, being attached only by a narrow base. In Hepialus, on the contrary, they are not free, but are merely indicated by curved marks on the dorsum. The patagia are styled by many writers "tegulae." They are of some interest in connection with the question of wing-like appendages on the prothorax of Palaeozoic insects, and they have been considered by some writers[^[172]] to be the equivalents of true wings. The Mesothorax is very large, especially its upper face, the notum, which is more or less convex, and in the higher forms attains a great extension from before backwards. The notum consists in greater part of a large anterior piece, the meso-scutum, and a smaller part, the meso-scutellum behind. In front of the scutum there is a piece termed prae-scutum by Burgess. It is usually small and concealed by the front part of the scutum; but in Hepialus it is large and horizontal in position. It is of importance as being the chief point of articulation with the prothorax. The scutellum is more or less irregularly rhomboidal in form; its hinder margin usually looks as if it were a lobe or fold placed in front of the base of the abdomen or metathorax, according to whether the latter is concealed or visible. In some of the higher forms this meso-scutellar lobe is prominent, and there may be seen under its projection a piece that has been called the post-scutellum, and is really the base of the great mesophragma, a chitinous piece that descends far down into the interior of the body. In addition to the front pair of wings the mesothorax bears on its upper surface another pair of appendages, the tegulae: in the higher forms they are of large size; they are fastened on the front of the mesothorax, and extend backwards over the joint of the wing with the body, being densely covered with scales so that they are but little conspicuous. These appendages are frequently erroneously called patagia, but have also been called scapulae, pterygodes, paraptera, and shoulder-tufts, or shoulder-lappets. The lower surface of the mesothorax is much concealed by the large and prominent coxae, but the sternum and the two pleural pieces on each side, episternum and epimeron, are easily detected. The area for attachment of the anterior wing on each side is considerable, and appears to be of rather complex structure; its anatomy has been, however, but little studied.

The Metathorax is small in comparison with the preceding segment, to which it is intimately co-adapted, though the two are really connected only by delicate membrane, and can consequently be separated with ease by dissection. The metanotum consists of (1) the scutum, which usually appears externally as an anterior piece on each side; (2) the scutellum, forming a median piece placed behind the scutum, which it tends to separate into two parts by its own extension forwards. In order to understand the structure of the metathorax it is desirable to dissect it off from the larger anterior segment, and it will then be found that its appearance when undissected is deceptive, owing to its being greatly arched, or folded in the antero-posterior direction. A broad, but short phragma descends from the hind margin of the metascutellum into the interior of the body. It should be noted that though the metanotum is forced, as it were, backwards by the great extension of the mesonotum in the middle line of the body, yet at the sides the metanotum creeps forward so as to keep the points of attachment of the hind wings near to those of the front wings. In many forms of Hesperiidae, Sphingidae, Noctuidae, etc. the true structure of the metanotum is further concealed by the back of the mesoscutellum reposing on, and covering it.

Difference of opinion exists as to the thoracic Spiracles; there is one conspicuous enough in the membrane behind the pronotum, and it is thought by some writers that no other exists. Westwood and Scudder, however, speak of a mesothoracic spiracle, and Dr. Chapman considers that one exists. Minot describes[^[173]] a structure behind the anterior wing, and thinks it may be an imperfect spiracle, and we have found a similar stigma in Saturnia pavonia. At the back of the thorax there is on each side in some Lepidoptera (Noctuidae, Arctia, etc.), a curious large cavity formed by a projection backwards from the sides of the metasternum, and a corresponding development of the pleura of the first abdominal segment. Minot and others have suggested that this may be an organ of hearing.

The Abdomen differs according to the sex. In the female seven segments are conspicuous dorsally, but only six ventrally, because the first segment is entirely membranous beneath, and is concealed between the second abdominal ventral plate and the posterior coxae. Besides these segments there are at the hind end two others smaller, more or less completely withdrawn into the body, and in certain cases forming an ovipositor. These nine segments are usually considered to constitute the abdomen; but according to Peytoureau,[^[174]] a tenth dorsal plate is represented on either side of the anal orifice, though there is no trace of a corresponding ventral plate. In the male the segments, externally conspicuous, are one more than in the female. According to the authority quoted,[^[175]] this sex has also truly ten abdominal segments, the ninth segment being withdrawn to a greater or less extent to the inside of the body, and modified to form part of a copulatory apparatus; its dorsal portion bears a process called the "uncus"; the anal orifice opens on the inner face of this process, and below it there is another process—developed to a greater or less extent—called the "scaphium." The ventral portion of the ninth segment bears a lobe, the "saccus" (Peytoureau, l.c.). On each side of the ninth abdominal segment there is a process called the "valve," the internal wall of which bears some hook-like or other processes called "harpes"; it is continued as a membrane surrounding the "oedeagus," or penis, and—bearing more or less distinct prominences—connects with the scaphium. In many forms the parts alluded to, other than the valves, are concealed by the latter, which come together when closed, and may be covered externally with scales like the rest of the abdomen. Peytoureau considers that the uncus is really the dorsal plate of a tenth segment, and that the scaphium is the tenth ventral plate. Thus, according to this view, the ninth segment is extensive and complex, being very highly modified in all its parts: while the tenth segment is greatly reduced. The structure of the male organs is simpler in Lepidoptera, and less varied than it is in the other great Orders of Insects. There are seven pairs of abdominal spiracles on the upper parts of the membranous pleurae.

Fig. 160—Acherontia atropos. The termination of ♂ body, one side removed. IX, Ninth dorsal plate; IX’, ninth ventral; s, lobe, saccus, of ninth ventral plate; X, tenth dorsal plate, or uncus; sc, scaphium, or tenth ventral plate; a, position of anus; b, chitinised band of scaphium; V, valve or clasper; c, hooks, or harpes, of clasper; p, penis (or oedeagus). (After Peytoureau.)

Legs.—The legs are long, slender, covered with scales, and chiefly remarkable from the fact that the tibiae sometimes bear articulated spurs on their middle as well as at the tip. The front tibia usually possesses on its inner aspect a peculiar mobile pad; this seems to be in some cases a combing organ; it also often acts as a cover to peculiar scales. The tarsi are five-jointed, with two small claws and a small apparatus, the functional importance of which is unknown, between the claws.

Wings.—The wings are the most remarkable feature of this Order; it is to them that butterflies owe their beauty, the surfaces of the wings being frequently adorned with colours and patterns of the most charming and effective nature. These effects are due to minute scales that are implanted in the wing-membrane in an overlapping manner, somewhat similar to the arrangement of slates on the roof of a house. The scales are very readily displaced, and have the appearance of a silky dust. We shall describe their structure and allude to their development subsequently. The wings are usually of large size in comparison with the Insect's body: in the genus Morpho, the most gorgeous of the butterflies, they are enormous, though the body is small; so that when deprived of these floats the Insect is insignificant. The great expanse of wing is not correlative with great powers of flight, though it is perhaps indicative of flying with little exertion; for the small-winged Lepidoptera, Sphingidae, etc., have much greater powers of aërial evolution than the large-winged forms. The area of the wing is increased somewhat by the fact that the scales on the outer margin, and on a part or on the whole of the inner margin, project beyond the edges of the membrane that bears them: these projecting marginal scales are called fringes. In many of the very small moths the actual size of the wing-membranes is much reduced, but in such cases the fringes may be very long, so as to form the larger part of the surface, especially of that of the hind wings. Frequently the hind wings are of remarkable shape, being prolonged into processes or tails, some of which are almost as remarkable as those of Nemoptera in the Order Neuroptera.

The wings are very rarely absent in Lepidoptera; this occurs only in the female sex, no male Lepidopterous imago destitute of wings having been discovered. Although but little is known of the physiology of flight of Lepidoptera, yet it is clearly important that the two wings of the same side should be perfectly coadapted or correlated. This is effected largely by the front wing overlapping the hind one to a considerable extent, and by the two contiguous surfaces being pressed, as it were, together. This is the system found in butterflies and in some of the large moths, such as Lasiocampidae and Saturniidae; in these cases the hind wing always has a large shoulder, or area, anterior to its point of insertion. In most moths this shoulder is absent, but in its place there are one or more stiff bristles projecting forwards and outwards, and passing under a little membranous flap, or a tuft of thick scales on the under face of the front wing; the bristle is called the "frenulum," the structure that retains it a "retinaculum." In Castnia (Fig. 162) and in some Sphingidae there is the unusual condition of a highly-developed shoulder (s) coexisting with a perfect frenulum (f) and retinaculum (r). The frenulum and retinaculum usually differ in structure, and the retinaculum in position, in the two sexes of the same moth; the male, which in moths has superior powers of flight, having the better retaining organs. Hampson says "the form of the frenulum is of great use in determining sex, as in the males of all the forms that possess it, it consists of hairs firmly soldered together so as to form a single bristle, whilst in nearly all females it consists of three or more bristles which are shorter than that of the male; in one female Cossid I have found as many as nine. Also in the large majority of moths the retinaculum descends from the costal nervure in the male, while in the female it ascends from the median nervure."[^[176]] This sexual difference in a structure for the discharge of a function common to the two sexes is a very remarkable fact. There are a few—very few—moths in which the bases of the hind wings are not well coadapted with the front wings, and do not possess a frenulum, and these species possess a small more or less free lobe at the base of the front wing that droops towards the hind wing, and may thus help to keep up an imperfect connexion between the pair; this lobe has been named a jugum by Professor Comstock. Occasionally there is a jugum on the hind as well as on the front wing. There is usually a very great difference between the front and the hind wings; for whereas in the front wing the anterior portion is doubtless of great importance in the act of flight and is provided with numerous veins, in the hind wing, on the other hand, the corresponding part has not a similar function, being covered by the front wing; hence the hind wing is provided with fewer nervures in the anterior region, the divisions of the subcostal being less numerous than they are in the front wing. In the moths possessing a jugum the two wings differ but little from one another, and it is probable that they function almost as four separate wings instead of as two pairs.