PYCNOGONIDA

BY

D’ARCY W. THOMPSON, C.B., M.A. Trinity College

Professor of Natural History in University College, Dundee

CHAPTER XXI
PYCNOGONIDA[[393]]

Remote, so far as we at present see, from all other Arthropods, while yet manifesting the most patent features of the Arthropod type, the Pycnogons constitute a little group, easily recognised and characterised, abundant and omnipresent in the sea. The student of the foreshore finds few species and seldom many individuals, but the dredger in deep waters meets at times with prodigious numbers, lending a character to the fauna over great areas.

Fig. [262].—Pycnogonum littorale, Ström, × 2.

The commonest of our native species, or that at least which we find the oftenest, is Pycnogonum littorale (Phalangium littorale, Ström, 1762). We find it under stones near low water, or often clinging louse-like to a large Anemone. The squat segmented trunk carries, on four pairs of strong lateral processes, as many legs, long, robust, eight-jointed, furnished each with a sharp terminal claw. In front the trunk bears a long, stout, tubular proboscis, at the apex of which is the mouth, suctorial, devoid of jaws; the body terminates in a narrow, limbless, unsegmented process, the so-called “abdomen,” at the end of which is the anal orifice. The body-ring to which is attached the first pair of legs, bears a tubercle carrying four eye-spots; and below, it carries, in the male sex, a pair of small limbs, whose function is to grasp and hold the eggs, of which the male animal assumes the burden, carrying them beneath his body in a flattened coherent mass. In either sex a pair of sexual apertures open on the second joints of the last pair of legs. The integument of body and limbs is very strongly chitinised, brown in colour, and raised into strong bosses or tubercles along the middle line of the back, over the lateral processes, and from joint to joint of the limbs. The whole animal has a singular likeness to the Whale-louse, Cyamus mysticeti (well described by Fr. Martins in 1675), that clings to the skin of the Greenland Whale as does Pycnogonum to the Anemone, a resemblance close enough to mislead some of the older naturalists, and so close that Linnaeus, though in no way misled thereby, named it Phalangium balaenarum. The substance of the above account, and the perplexity attending the classification of the animal, are all included in Linnaeus’s short description:[[394]] “Simillimus Onisco Ceti, sed pedes omnes pluribus articulis, omnes perfecti, nec plures quam octo. Dorsum rubrum, pluribus segmentis; singulis tribus mucronibus. Cauda cylindrica, brevissima, truncata. Rostrum membranaceum, subsubulatum, longitudine pedum. Genus dubium, facie Onisci ceti; rostro a reliquis diversum. Cum solo rostro absque maxillis sit forte aptius Acaris aut proprio generi subjiciendum.... Habitat in mari norvegico sub lapidibus.”[[395]]

Fig. [263].—Dorsal view of Nymphon brevirostre, Hodge, × 6. Britain.

The common Pycnogonum is, by reason of the suppression of certain limbs, rather an outlying member than a typical representative of the Order, whose common characters are more strikingly and more perfectly shown in species, for instance, of Nymphon. Of this multiform genus we have many British species, some of the smaller being common below tide-marks, creeping among weeds or clinging like Caprellae with skeleton limbs to the branches of Zoophytes, where their slender forms are not easily seen. In contrast to the stouter body and limbs of Pycnogonum, the whole fabric of Nymphon tends to elongation; the body is drawn out so that the successive lateral processes stand far apart, and a slender neck intervenes between the oculiferous tubercle and the proboscis; the legs are produced to an amazing length and an extreme degree of attenuation: “mirum tam parvum corpus regere tam magnos pedes,” says Linnaeus. Above the base of the proboscis are a pair of three-jointed appendages, the two terminal joints of which compose a forcipate claw; below and behind these come a pair of delicate, palp-like limbs of five joints; and lastly, on the ventral side, some little way behind these, we find the ovigerous legs that we have already seen in the male Pycnogonum, but which are present in both sexes in the case of Nymphon. At the base of the claw which terminates each of the eight long ambulatory legs stands a pair of smaller accessory or “auxiliary” claws. The generative orifices are on the second joint of the legs as in Pycnogonum, but as a rule they are present on all the eight legs in the female sex, and on the two hindmost pairs in the male. One of the Antarctic Nymphonidae (Pentanymphon) and one other Antarctic genus less closely related (Decolopoda) have an extra pair of legs. No other Pycnogon, save these, exhibits a greater number of appendages than Nymphon nor a less number than Pycnogonum, nor are any other conspicuous organs to be discovered in other genera that are not represented in these two: within so narrow limits lie the varying characters of the group.

Fig. [264].—Nymphon brevirostre, Hodge. Head, from below, showing chelophores, palps, and ovigerous leg.

In framing a terminology for the parts and members of the body, we encounter an initial difficulty due to the ease with which terms seem applicable, that are used of more or less analogous parts in the Insect or the Crustacean, without warrant of homology. Thus the first two pairs of appendages in Nymphon have been commonly called, since Latreille’s time, the mandibles and the palps (Linnaeus had called them the palps and the antennae), though the comparison that Latreille intended to denote is long abandoned; or, by those who leaned, with Kröyer and Milne-Edwards, to the Crustacean analogy, mandibles and maxillae. Dohrn eludes the difficulty by denominating the appendages by simple numbers, I., II., III., ... VII., and this method has its own advantages; but it is better to frame, as Sars has done, a new nomenclature. With him we shall speak of the Pycnogon’s body as constituted of a trunk, whose first (composite) segment is the cephalic segment or head, better perhaps the cephalothorax, and which terminates in a caudal segment or abdomen; the “head” bears the proboscis, the first appendages or “chelophores,” the second or “palps,” the third, the false or “ovigerous” legs, and the first of the four pairs of “ambulatory” legs. The chelophores bear their chela, or “hand,” on a stalk or scape; the ambulatory legs are constituted of three coxal joints, a femur, two tibial joints, a tarsus, and a propodus, with its claws, and with or without auxiliary claws.

The Body.—The trunk with its lateral processes may be still more compact than in Pycnogonum, still more attenuated than in Nymphon.

In a few forms (e.g. Pallene, Ammothea, Tanystylum, Colossendeis) the last two, or even more, segments of the trunk are more or less coalescent. In Rhynchothorax the cephalic segment is produced into a sharp-pointed rostrum that juts forward over the base of the proboscis. The whole body and limbs may be smooth, tuberculated, furnished with scattered hairs, or sometimes densely hispid.

Fig. [265].—A, Colossendeis proboscidea, Sabine, Britain; B, Ammothea echinata, Hodge, Britain; C, Phoxichilus spinosus, Mont., Arctic Ocean. (The legs omitted.)

The proboscis varies much in shape and size. It may be much longer or much shorter than the body, cylindrical or tumid, blunt or pointed, straight or (e.g. Decolopoda) decurved; usually firmly affixed to the head and pointing straight forwards; sometimes (Eurycide, Ascorhynchus) articulated on a mobile stalk and borne deflexed beneath the body.

Chelophores.—The first pair of appendages or chelophores are wanting in the adult Pycnogonum, Phoxichilus, Rhynchothorax, and Colossendeis.[[396]]

In Ammothea and its allies they are extremely rudimentary in the adult, being reduced to tiny knobs in Tanystylum and Trygaeus, and present as small two-jointed appendages in Ammothea; in this last, if not in the others also, they are present in complete chelate form in the later larval stages.

Fig. [266].—A, B, Chelophores of Ascorhynchus abyssi, G.O.S. A, Young; B, adult. (After Sars.) C, Anterior portion of Ammothea hispida, Hodge, Jersey: late larval stage (= Achelia longipes, Hodge), showing complete chelae. D, Chela of Eurycide hispida, Kr.

In Eurycide, Ascorhynchus, and Barana they are usually less atrophied, but yet comparatively small and with imperfect chelae, while in some Ascorhynchi (A. minutus, Hoek) they are reduced to stumps.

Fig. [267].—Chelae of species of Nymphonidae: A, Nymphon brevirostre, Hodge; B, Boreonymphon robustum, Bell; C, Chaetonymphon macronyx, G.O.S.; D, Nymphon elegans, Hansen.

Fig. [268].—Proboscis and chelophores of Cordylochele longicollis, G.O.S. (After Sars.)

In Pallenopsis the scape of the chelophore consists of two joints, as also in Decolopoda and some Ascorhynchus: in Nymphon, Phoxichilidium, Pallene, and Cordylochele of one only; in all these the terminal portion or “hand” forms a forcipate “chela,” of which the ultimate joint forms the “movable finger.” In some species of Nymphon the chela is greatly produced and attenuated, and armed with formidable serrate teeth on its opposing edges; in others it is shortened, with blunter teeth; in Boreonymphon robustum the claws are greatly curved, with a wide gap between. In this last, and in Phoxichilidium, the opposing edges are smooth and toothless. In Cordylochele the hand is almost globular, the movable finger being shortened down, and half enclosed by the other.

Fig. [269].—Eurycide hispida, Kr., showing stalked proboscis and zigzag palps.

Palpi.—The second pair of appendages, or palps, are absent, or all but absent, in the adult Pycnogonum, Phoxichilus, Phoxichilidium, Pallene, and their allies. In certain of these cases, e.g. Phoxichilidium, a knob remains to mark their place; in others, e.g. Pallenopsis, a single joint remains; in a few Pallenidae a sexual difference is manifested, reduction of the appendage being carried further in the female than in the male. The composition of the palps varies in the genera that possess them. In Nymphon there are five joints, and their relative lengths (especially of the terminal ones) are much used by Sars in defining the many species of the genus. The recently described Paranymphon, Caullery, has palps of six or seven joints. In the Ammotheidae the number of joints ranges from five or six in Tanystylum to nine (as a rule) in Ammothea and Oorhynchus, or ten, according to Dohrn, in certain species of Ammothea. Colossendeis and the Eurycididae have a ten-jointed palp, which in this last family is very long and bent in zigzag fashion, as it is, by the way, also in Ammothea. The terminal joints of the palp are in all cases more or less setose, and their function is conjecturally tactile.

Ovigerous Legs.—Custom sanctions for these organs an inappropriate name, inasmuch as it is only in the males that they perform the function which the name connotes.[[397]] They probably also take some part, as Hodgson suggests, in the act of feeding.

Fig. [270].—Ovigerous legs of A, Phoxichilus spinosus, Mont.; B, Phoxichilidium femoratum, Rathke; C, Anoplodactylus petiolatus, Kr.; D, Colossendeis proboscideus, Sab.

Fig. [271].—Terminal joints of ovigerous leg of Rhynchothorax mediterraneus, Costa.

Fig. [272].—Nymphon brevirostre, Hodge. Terminal joints of ovigerous leg, with magnified “tooth.”

In Pycnogonum, Phoxichilus, Phoxichilidium, and their immediate allies they are absent in the female; in all the rest they are alike present in both sexes, though often somewhat smaller in the female than in the male. They are always turned towards the lower side of the body, and in many cases even their point of origin is wholly ventral. The number of joints varies: in Phoxichilidium five, Anoplodactylus six, Phoxichilus seven; in Paranymphon eight; in Pycnogonum nine, with, in addition, a terminal claw; in the Ammotheidae from seven (Trygaeus) to ten, without a claw; in Pallenidae ten, with or without a claw; in Rhynchothorax, Colossendeis, Eurycide, Ascorhynchus, Nymphon, ten and a claw. The appendage, especially when long, is apt to be wound towards its extremity into a spiral, and its last four joints usually possess a peculiar armature. In Rhynchothorax this takes the form of a stout toothed tubercle on each joint; in Colossendeis of several rows of small imbricated denticles; in Nymphon and Pallene of a single row of curious serrate and pointed spines, each set in a little membranous socket.

Fig. [273].—Nymphon strömii, Kr. Male carrying egg-masses on his ovigerous legs.

Fig. [274].—Terminal joints (tarsus and propodus) of legs. 1, Chaetonymphon hirtum, Fabr.; 2, N. strömii, Kr.; 3, Nymphon brevirostre, Hodge; 4, Ammothea echinata, Hodge; 5, Ascorhynchus abyssi, G.O.S. (All after Sars.)

Legs.—The four pairs of ambulatory legs are composed, in all cases without exception, of eight joints if we exclude, or nine if we include, the terminal claw. They vary from a length about equal to that of the body (Pycnogonum, Rhynchothorax, Ammothea) to six or seven times as much, perhaps more, in Nymphon and Colossendeis, the fourth, fifth, and sixth joints being those that suffer the greatest elongation. The seventh joint, or tarsus, is usually short, but in some Nymphonidae is much elongated; the eighth, or propodus, is usually somewhat curved, and usually possesses a special armature of simple or serrate spines. The auxiliary claws, sometimes large, sometimes small, lie at the base of the terminal claw in Ammotheidae, Phoxichilidae, in Phoxichilidium, in most Pallenidae, in nearly all Nymphonidae. Their presence or absence is often used as a generic character, helping to separate, e.g., Pallene from Pseudopallene and Pallenopsis, and Phoxichilidium from Anoplodactylus; nevertheless they may often be detected in a rudimentary state when apparently absent. The legs are smooth or hirsute as the body may happen to be.

Fig. [275].—Legs of A, Pallene brevirostris, Johnston; B, Anoplodactylus petiolatus, Kr.; C, Phoxichilus spinosus, Mont.; D, Colossendeis proboscidea, Sabine; E, Ammothea echinata, Hodge, ♂.

Fig. [276].—Boreonymphon robustum, Bell. Male with young, slightly enlarged. Faeroe Channel.

Glands.—In some or all of the appendages of the Pycnogonida may be found special glands with varying and sometimes obscure functions. The glands of the chelophores (Fig. [280], p. 522) are present in the larval stages only. They consist of a number of flask-shaped cells[[398]] lying within the basal joint of the appendage, and generally opening at the extremity of a long, conspicuous, often mobile, spine (e.g. Ammothea (Dohrn), Pallene, Tanystylum (Morgan), Nymphon brevicollum and N. gracile (Hoek)). They secrete a sticky thread, by means of which the larvae attach themselves to one another and to the ovigerous legs of the male parent. In Nymphon hamatum, Hoek, the several filaments secreted by the separate sacculi of the gland issue separately. In Pycnogonum the spine on which the gland opens is itself prolonged into a long fine filament, and here, according to Hoek, the gland is in all probability functionless and rudimentary. Hoek has failed to find the gland in Ascorhynchus, and also in certain Nymphonidae (e.g. Boreonymphon robustum, Bell), in which the young are more than usually advanced at the time of hatching. The gland has also been described by Lendenfeld and others in Phoxichilidium, whose larvae do not cling together but live a parasitic life; in this genus the long spine or tubercle is absent on which the orifice is usually situated, and, according to Lendenfeld, the secretion issues from many small orifices set along the opposing edges of the chela. Of the two species described by Dohrn as Barana castelli and B. arenicola, the former has the spine of inordinate length, more than twice as long as the whole body, chelophore and all; while in the latter (which species rather resembles Ascorhynchus) the spine is altogether absent.

In the palps and ovigerous legs of the adult are found glandular bodies of a hollow vesicular form with a simple lining of cells, the vesicle being divided within by a septum with a central orifice, the outer and smaller half opening to the exterior. These glands are probably of general occurrence, but they have been but little investigated. They lie usually in the fourth and fifth joints of the palp, and the third and fourth joints of the ovigerous leg. Hoek describes them in Discoarachne (Tanystylum) as lying within the elongated third joint of the palp, and opening by a sieve-plate at the end of the second joint. In Ammothea (Dohrn) and Ascorhynchus (Hoek) they open on a small tubercle situated on the fifth joint of the palp. In Nymphon, Hoek describes them as opening by a small pore on the fourth joint of the ovigerous leg. Dohrn failed to find them in Pycnogonum, but in Phoxichilus, Phoxichilidium and Pallene he discovered the glands appertaining to the palps, though the palps themselves have disappeared in those genera; he has found the glands also in Ammothea, in larvae that have not yet attained their full complement of legs.

The males in nearly all cases are known to possess glands in the fourth joints or thighs of all the ambulatory legs, and these glands without doubt act as cement-glands, emitting, like the chelophoral glands of the larvae, a sticky thread or threads by which the eggs and young are anchored to the ovigerous legs. In some species of Nymphon and of Colossendeis Hoek could not find these, and he conjectures them to be conspicuous only in the breeding season. While in most cases these glands open by a single orifice or by a few pores grouped closely together, in Barana, according to Dohrn, and especially in B. arenicola, the pores are distributed over a wide area of the femoral joint.[[399]] In Discoarachne (Loman) and Trygaeus they open into a wide chitinised sac with tubular orifice. While the function of these last glands and of the larval glands seems plain enough, that of those which occur in the palps and ovigerous legs of both sexes remains doubtful.

In their morphological nature the two groups of glands are likewise in contrast, the former being unicellular glands, such as occur in various parts of the integument of the body and limbs of many Crustacea; while the latter are segmentally arranged and doubtless mesoblastic in origin, like the many other segmental excretory organs (or coelomoducts) of various Arthropods.

By adding colouring matters (acid-fuchsin, etc.) to the water in which the animals were living, Kowalevsky demonstrated the presence of what he believed to be excretory organs in Phoxichilus, Ammothea, and Pallene. These are small groups of cells, lying symmetrically near the posterior borders of the first three body-segments, and also near the bases of the first joints of the legs, dorsal to the alimentary canal.[[400]]

Fig. [277].—Longitudinal section through one “antimere” of the proboscis in Phoxichilus charybdaeus. G, g′, Principal and secondary ganglia; h, sieve-hairs; L, lip; mt, oral tooth; N, N′, inner and outer nerve-cords; t, proboscis-teeth. (After Dohrn.)

Alimentary System.—The proboscis is a very complicated organ, and has been elaborately described by Dohrn.[[401]] It is a prolongation of the oral cavity, containing a highly developed stomodaeum, but showing no sign of being built up of limbs or gnathites. The mouth, situated at its apex, is a three-sided orifice, formed by a dorsal[[402]] and two lateral lobes; and hence the proboscis has been assumed by some, on no competent evidence, to be constituted of a degenerate pair of appendages and a labrum or upper lip. Each of the three lobes which bounds the mouth shows the following structures: firstly, a lappet of external chitinised integument, overlapping, as the finger-nail overlaps the finger, a cushion-like lip, ridged after the fashion of a fine-cut file in some species, hairy in others, on the inner surface where the three lips meet to close the orifice of the mouth. Below this again is a prominent tooth (Fig. [277], mt), supported, as are the lips, by a system of chitinous rods, which are but little developed in the genus here figured, though conspicuous and complicated in others. Transverse ridges run across the angles where adjacent lips meet, and the whole mechanism constitutes an efficient valve, preventing the escape of swallowed food. The greater portion of the proboscis is occupied by a masticating or triturating apparatus, the oesophageal cavity expanding somewhat and having its walls densely covered, in three bands corresponding to the antimeres, with innumerable minute spines (h) or needles, sometimes supplemented by large teeth (t) that point forwards somewhat obliquely to the axis of the proboscis.[[403]]

In the curious East Indian genus Pipetta (Loman) the sucking and sifting mechanism is low down in the proboscis, and the organ is prolonged into a very fine tube, the lips growing together till they leave an aperture of only ·007 mm. for the absorption of liquids.

In some cases, where the proboscis itself is short, as in Pallene, this mechanism is carried backwards into the fore-part of the body; and, in the latter genus, the narrow oesophagus which succeeds the masticatory apparatus is likewise provided with extrinsic muscles.

Fig. [278].—Transverse sections through the proboscis of Ph. charybdaeus. A, Anterior, through the principal ganglionic mass (G); B, posterior, at the level of the sieve-hairs (h). Coec, Intestinal caeca; Dil. M, dilator muscles; N, inner nerve-ganglion, with circular commissure; N′, outer nerve; or, chitinous lining of oral cavity; R M, Ret.M, retractor muscles. (After Dohrn.)

Fig. [279].—Transverse section through the basal joint of the third leg in Phoxichilus charybdaeus, ♀. Cut, Cuticle; Hyp, hypodermis; Int, intestinal caecum; N, nerve-cord; Ov, ovary; Sept, septum. (After Dohrn.)

The oesophagus is followed by a long gastric cavity, which sends forth caecal diverticula into the chelophores (when these are present), and four immensely long ones into the ambulatory legs. The caeca are attached to the walls of the limb cavities, especially at their extremities in the tarsi, by suspensory threads of connective tissue, and the whole gut, central and diverticular, is further supported by a horizontal septal membrane, running through body and legs, which separates the dorsal blood-vessel and sinus from the gut, the nervous system and the ventral sinus, giving support also to the reproductive glands. A short and simple rectum follows the gastric cavity.

In Phoxichilus, which lacks the three anterior appendages in the female and the two anterior in the male, two pairs of caeca run from the gut into the cavity of the proboscis (Fig. [278], B, coec.).[[404]]

Circulatory System.—The heart has been especially studied by Dohrn in Phoxichilus. It consists of a median vessel running from the level of the eyes to the abdomen, furnished with two pairs of lateral valvular openings, and sometimes, though not always, with an unpaired one at the posterior end. The walls are muscular, but with this peculiarity that the muscular walls do not extend around the heart dorsally, in which region its lumen is only covered by the hypodermis and cuticle of the back. The blood-spaces of the body are separated into dorsal and ventral halves by the septal membrane already referred to, which is perforated in the region of the lateral processes by slits placing the two cavities in communication; this septal membrane runs through the limbs to their tips, and far into the proboscis, where it is attached to the edge of the superior antimere. The blood is a colourless plasma with several kinds of corpuscles, of which the most remarkable are amoeboid, actively mobile, often coalescing into plasmodia. The course of the circulation is on the whole outwards in the inferior or ventral sinus, inwards towards the heart in the superior, save in the proboscis, where the systole of the heart drives the blood forwards in the dorsal channel. The beat is rapid, two or three times in a second, according to Loman, in Phoxichilidium. Especially in the species with small body and exaggerated legs, the movement of the circulatory fluid is actuated more by the movements of the limbs and the contractions of the intestinal caeca than by the direct impulse of the heart.

Nervous System.—The nerve-chain consists of a fused pair of supra-oesophageal ganglia, which innervate (at least in the adult) the chelophores, and of ventral ganglia, whence proceed the nerves to the other limbs. The ganglia of the second and third appendages are fused with one another, sometimes also with the ganglia of the first ambulatory legs; the ganglia of the three posterior pairs of legs are always independent (though the development of their longitudinal commissures varies with the body-form), and they are succeeded by one or two pairs of ganglia, much reduced in size, situated in the abdomen, of which the posterior one innervates the muscles of the abdomen and of the anal orifice. Each lateral nerve divides into two main branches, which supply the parts above and below the septal membrane. The nerve-supply of the proboscis is very complicated. Its upper antimere is supplied from the pre-oral, its two lateral antimeres from the first post-oral, ganglion, and each of these three nerves divides into two branches, of which the inner bears six to eight or more small ganglia, which annular commissures passing round the pharynx connect one to another. Of these ganglia and commissures the anterior are the largest, and with these the outer lateral nerve-branches of the proboscis merge. The immediate origin of the nerves to the chelophores is from the median nerve that springs from the under side of the supra-oesophageal ganglion to run forward into the proboscis, but it is noteworthy that the chelophores receive twigs also from the lateral nerves of the proboscis which arise from the post-oral ganglia.

Eyes.—Eyes are the only organs of special sense known in the Pycnogons. The deep-water Pycnogons, in general those inhabiting depths below four or five hundred fathoms, have in most cases imperfect organs, destitute of lens and of pigment, so imperfect in many cases as to be described as wanting. It is rare for the eyes to be lacking in shallow-water species, as they are, for instance, in Ascorhynchus minutus, Hoek, dredged by the Challenger in 38 fathoms, but, on the other hand, it is no small minority of deep-water species that possess them of normal character and size, even to depths of about 2000 fathoms.

In all cases where eyes are present, they are simple or “monomeniscous” eyes, four in number, and are situated in two pairs on an “oculiferous tubercle,” sometimes blunt and low, sometimes high and pointed, placed on the so-called cephalothorax, or first, compound, segment of the body. The anterior pair are frequently a little larger, sometimes, as in Phoxichilidium mollissimum, Hoek, very much larger, than the posterior. The minute structure of the eye has been investigated by Dohrn, Grenacher, Hoek, and Morgan. The following account is drawn in the first instance from Morgan’s descriptions.[[405]]

The eye of a Pycnogon (Phoxichilidium) is composed of three layers, an outer layer of specialised ectoderm cells (hypodermis) that secrete the cuticular lens, a middle layer of visual or retinal elements, and an inner layer of pigment-cells. The elements of the middle layer consist of much elongated cells, whose branching outer ends are connected with nerve-fibrils and interwoven in a protoplasmic syncytium, whose middle parts are occupied by the nuclei and whose inwardly directed ends form the retinal rods or bacilli. The pigment-cells of the inner layer are of various forms, those towards the middle of the eye being small and flattened, those at the sides being, for the most part, long and attenuated, so seeming, as Morgan remarks, to approximate in character to the retinal elements. The pigment-layer is easily dispersed and reveals beneath it a median vertical raphe, caused by the convergence of the cells of the middle layer from either side, and along the line of this raphe the optic nerve joins the eye, though its subsequent course to its connection with the retinal elements is obscure. It is at least clear that the retina is an “inverted” retina, with the nerve-connected bases of its cells lying outwards and their bacillar extremities directed inwards.

In a longitudinal vertical section of the eye of a larva (Tanystylum), at a stage when three pairs of walking legs are present, Morgan shows us the pigment-layer apparently continuous with the hypodermis just below the eye, and in close connection with the middle layer at the upper part of the eye. From this we are permitted to infer a development by invagination, in which the long invaginated sac is bent and pushed upwards till it comes into secondary contact with the hypoderm, so giving us the three layers of the developed eye. This manner of formation is precisely akin to that described by Parker, Patten, Locy, and others for the median eyes of Scorpions and of Spiders, and the organ is structurally comparable to the Nauplius- or median eye of Crustacea. But neither in these cases nor in that of the Pycnogon is the whole process clear, in consequence chiefly of the obscurity that attends the course of the optic nerve in both embryo and adult. For various discussions and accounts, frequently contradictory, of these phenomena, the reader is referred to the authors quoted, or to Korschelt and Heider’s judicious summary.[[406]]

There seems to be a small structure, of some sort or other, between the ocelli on either side. Dohrn thought it might be auditory, Loman that it might be secretory, but its use is unknown.

Integument.—The chitinised integument is perforated by many little cavities, some of them conical and tapering to a minute external pore, the others more regularly tubular. Sometimes, but according to Hoek rarely, the tubular pore-canals communicate with, or arise from, the conical cavities. The pore-canals transmit a nerve for the supply of sensory hairs, often forked, which arise from the orifice of the canal in little groups of two or more, sometimes in rosettes of eight or nine. These setae are small or rudimentary in Ascorhynchus and totally wanting in Colossendeis; they appear to be extremely large and stellate in Paranymphon. The conical cavities contain proliferated epithelial cells, blood corpuscles, and cells of more doubtful nature that are perhaps glandular. According to Dohrn, glands exist in connection with both kinds of integumentary perforations, and he suspects that they secrete a poisonous fluid in response to stimuli affecting the sensory hairs; Hoek, on the other hand, is inclined to ascribe a respiratory function to the cavities; but indeed, as yet, we must confess that their use is undetermined.

Reproductive Organs.—In each sex the generative organs consist of a pair of ovaries or testes lying above the gut on either side of the heart; in the adult they are fused together posteriorly at the base of the abdomen, and send long diverticula into the ambulatory legs. In the female Phoxichilidium, at least, as Loman has lately shown, the fusion is complete, and the ovary forms a thin broad plate, spreading through the body and giving off its lateral diverticula. The diverticula of the testes reach to the third joint of the legs, those of the ovaries to the fourth, or sometimes farther. The ova ripen within the lateral diverticula, chiefly, and sometimes (Pallene) exclusively, in the femora or fourth joints of the legs,[[407]] which, in many forms, are greatly swollen to accommodate them; the spermatozoa, on the other hand, are said to develop both within the legs and within the thoracic portions of the testis. The genital diverticula may end blindly within the leg, or communicate through a duct with the exterior by a valvular aperture placed on the second coxal joint. Such apertures occur, as a rule, on all the legs in the females, in Rhynchothorax and Pycnogonum on the last only. In the males an aperture is present on all the legs in Decolopoda and Phoxichilidium; on the last three in Nymphon and Phoxichilus; in most genera on the last two; in Pycnogonum and Rhynchothorax on the last only.

Very commonly the female individuals are somewhat larger than the males, and in some species (Ammothea, Trygaeus) the latter are distinguished by a greater development of spines or tubercles on the body and basal joints of the legs (Dohrn).

The act of fecundation has been observed by Cole[[408]] in Anoplodactylus. The animal reproduces towards the end of August. Consorting on their Eudendrium (Hydroid) colony, the male climbs upon the female and crawls over her head to lie beneath her, head to tail; and then, fertilisation taking place the while, the hooked ovigerous legs of the male fasten into the extruding egg-masses and tear them away. The whole process is over in five minutes. The fresh egg-masses are more or less irregular in shape, and white in colour like little tufts of cotton.

Each ball of eggs that the male carries represents the entire brood of one female, and in Phoxichilidium Loman has seen a male carrying as many as fourteen balls. Fertilisation is external, taking place while the eggs are being laid. The spermatozoa have small rounded heads and long tails, and are thus unlike the spermatozoa of most Crustacea.

Development.—Until the hatching of the embryo, the eggs of the Pycnogons are carried about, agglutinated by cement-substance into coherent packets, on the ovigerous legs of the males. They are larger or smaller according to the amount of yolk-substance present, very small in Phoxichilidium and Tanystylum (Morgan), where they measure only ·05 mm. in diameter; larger in Pallene (·25 mm.); larger still (·5–·7 mm.) in Nymphon. In Pallene each egg-mass commonly contains only two eggs; in the other genera they are much more numerous, rising to a hundred or more in Ammothea (Dohrn). The egg-masses may be one or more on each ovigerous leg, sometimes (Phoxichilidium angulatum, Dohrn) a single egg-mass is held by both legs; they are extremely numerous in Phoxichilus, and in Pycnogonum they coalesce to form a broad pad beneath the body. The fact that it is the male and not the female that carries the eggs was only announced in 1877 by Cavanna;[[409]] before, and by some even after his time, the two sexes were constantly confused.[[410]]

Fig. [280].—Young larva (nat. size ·1 mm.) of Ammothea fibulifera, Dohrn. C.G, Brain; gl, gld, gland and duct of chelophore; pr, proboscis; I, II, III, IV, appendages. (After Dohrn.)

Segmentation is complete, symmetrical in the forms with smaller eggs, unequal in those burdened with a preponderance of yolk (Morgan). In Pallene, as in the Spider’s egg, what is described as at first a total segmentation passes into a superficial or centrolecithal one by the migration outwards of the nuclei and the breaking down of the inner ends of the wedge-shaped segmentation-cells. The blastoderm so formed becomes concentrated at the germinal pole of the egg. A thickened portion of the blastoderm (which Morgan compares to the “cumulus primitivus” of the Spider’s egg) forms an apparently blastoporal invagination (though Morgan calls it the stomodaeum), and from its sides are budded off the mesodermal bands. Meisenheimer has recently given a minute account of the early development of Ammothea, a form with small yolkless eggs. Here certain cells of the uniform and almost solid blastosphere grow inwards till their nuclei arrange themselves in an inner layer of what (so far as they are concerned) is a typical gastrula, but without any central cavity. The inner layer subsequently, but slowly, differentiates into the mid-gut, and into dorsal and lateral offshoots, the sources of the heart and of the muscles and connective tissues respectively. The further development of the egg takes place, as is usual in Arthropods, by the appearance, in a longitudinal strip or germ-band which enwraps the yolk, of paired thickenings which represent the cerebral and post-oral ganglia, and of others from which arise the limbs. Of these latter, the chelophores are the first to appear, on either side of the mouth; in Pallene the fourth pair appears next in order, followed by the fifth and sixth, and by the third and seventh just before the hatching out of the embryo; the second is lacking in this particular genus. Thus in Pallene (Dohrn, Morgan), and in some others, e.g. Nymphon brevicollum (Hoek), the free larva is from the first provided with its full complement of limbs. Certain other species of Nymphon hatch out in possession of four or five pairs of limbs, but in the great majority of cases studied the larval Pycnogon is at first provided with three pairs only, the three anterior pairs of the typical adult.[[411]] Numerical coincidence, and that alone, has often led this “Protonymphon” larva to be compared with the Crustacean Nauplius. In the annexed figure of a young larval Ammothea (Achelia), we see the unsegmented body, the already chelate chelophores (furnished with the provisional cement-glands already described), the other two pairs of appendages each with a curious spine at its base, the gut beginning to send out diverticula (of which the first pair approach the chelophores) but still destitute of the anus (which is only to be formed after the development of the abdomen), the proboscis, and one pair of eyes situated close over the pre-oral ganglia. The subsequent changes are in this genus extremely protracted, and terminate with the loss of the chelae, a process which occurs so late in life that the chelate individuals were long looked upon as belonging to a separate genus, the original Ammothea of Hodge, until Hoek proved their identity with the clawless Achelia.

The developmental history of Phoxichilidium and Anoplodactylus is peculiar. The young larvae have the claws of the second and third appendages hypertrophied to form enormous stiff tendril-like organs, with which they affix themselves to the bodies of Hydroid Zoophytes (Coryne, Eudendrium, Tubularia, Hydractinia, etc.), feeding as the adults do: afterwards losing these elongated tendrils in a moult, they pass into the gastral cavity of the Hydroid; in our native species the larva issues from the Hydroid and begins its independent life at a stage when three pairs of ambulatory legs are present and the fourth is in bud.[[412]] The Phoxichilidium larvae were first noticed by Gegenbaur in Eudendrium,[[413]] again by Allman in Coryne eximia.[[414]] George Hodge made detailed and important observations,[[415]] and showed, in opposition to Gegenbaur, that it was the larva which entered the Hydroid and not the egg that was laid therein.[[416]]

Fig. [281].—Larva of Phoxichilidium sp., showing tendril-like appendages of the larval palps and ovigerous legs. (After Dohrn.)

Moseley has the following interesting note in his Challenger Report:[[417]] “The most interesting parasite observed was a form found in the gastric cavities of the gastrozoids of Pliobothrus symmetricus (West Indies, 450 f.), contained in small capsules. These capsules were badly preserved, but there seemed little doubt that they contained the remains of larvae of a Pycnogonid, so that the deep-sea Pycnogonids, which are so abundant, very possibly pass through their early stages in deep-sea Stylasteridae.... The gastrozoids containing the larvae were partly aborted.”

A Pycnogon larva, doubtfully ascribed to Nymphon, has been found living in abundance ectoparasitically on Tethys in the Bay of Naples.[[418]]

Habits.—Of the intimate habits of the Pycnogons we can say little. Pycnogonum we often find clinging, as has been said, close appressed to some large Anemone (Tealia, Bolocera, etc.), whose living juices it very probably imbibes. The more slender species we find climbing over sea-weeds and Zoophytes, where sometimes similarity of colour as well as delicacy of form helps to conceal them; thus Phoxichilidium femoratum (Orithyia coccinea, Johnston) is red like the Corallines among which we often find it, P. virescens green like the filamentous Ulvae, the Nymphons yellowish like the Hydrallmania and other Zoophytes which they affect. On the New England coast, according to Cole, the dark purple Anoplodactylus lentus, Wilson (Phoxichilidium maxillare, Stimpson), is especially abundant on colonies of Eudendrium, whose colour matches its own, the yellowish Tanystylum orbiculare frequents a certain yellowish Hydroid, and of these two species neither is ever found on the Hydroid affected by the other; while, on the other hand, Pallene brevirostris, whose whitish, almost transparent body is difficult to see, is more generally distributed.[[419]] The deep-sea Pycnogons (Colossendeis, Nymphon) are generally (if not universally) of a deep orange-scarlet colour, a common dress of many deep-sea Crustacea.

The movements of the Pycnogons are singularly slow and deliberate; they are manifestly not adapted to capture or to kill a living prey. Linnaeus accepted from J. C. König the singular statement that they enter and feed upon bivalve shells, “Mytilorum testes penetrat et exhaurit”; but the statement has never been reaffirmed.[[420]]

Loman describes Phoxichilidium as feeding greedily on Tubularia larynx, and especially on the gonophores. It grasps them with its claws, sucks them in bit by bit till the proboscis is filled as far as the sieve, whereupon that part of the proboscis squeezes and kneads the mass, letting only juices and fine particles pass through into the alimentary canal. The lateral caeca and the rectum are separated by sphincter muscles from the stomach; the former are in turn filled with food and again emptied; the contents of the alimentary canal are in constant rolling movement, and the faeces are eliminated by the action of a pair of levatores ani, in round pellets.

The Pycnogons, or some of them, can swim by “treading water,” and Pallene is said by Cole to swim especially well; they more often progress half by swimming, half by kicking on the bottom. They move promptly towards the light, unless they have Hydroids to cling to, and Cole points out that when they crawl with all their legs on the bottom they move forwards towards the light,[[421]] but backwards when they swim in part or whole. The legs move mostly in a vertical plane, horizontal movements taking place chiefly between the first and second joints. Tanystylum is uncommonly sluggish and inert; it sinks to the bottom, draws its legs over its back and remains quiet, while Pallene, by vigorous kicks, remains suspended.

The long legs of the Pycnogons are easily injured or lost, and easily repaired or regenerated. This observation, often repeated, is as old as Fabricius: “Mutilatur etiam in libertate sua, redintegrandum tamen; vidi enim in quo pedes brevissimi juxta longiores enascentes, velut in asteriis cancris aliisque redintegratis.” In such cases of redintegration of a leg, the reproductive organ, the genital orifice, and the cement-gland are not restored until the next moult.[[422]]

Systematic Position.—To bring this little group into closer accord with one or other of the greater groups of Arthropods is a problem seemingly simple but really full of difficulty.

The larval Pycnogon, with its three pairs of appendages, resembles the Crustacean Nauplius in no single feature save this unimportant numerical coincidence; nor is there any significance in the apparent outward resemblance to isolated forms (e.g. Cyamus) that induced some of the older writers, from Fabricius downwards and including Kröyer and the elder Milne-Edwards, to connect the Pycnogons with the Crustacea. To refer them, or to approximate them to the Arachnids, has been a stronger and a more lasting tendency.[[423]] Linnaeus (1767) included the two species of which he was cognisant in the genus Phalangium, together with P. opilio. Lamarck, who first formulated the group Arachnida (1802), let it embrace the Pycnogons; and Latreille (1804, 1810), who immediately followed him, defined more clearly the Pycnogonida as a subdivision of the greater group, side by side with the subdivision that corresponds to our modern Arachnida (“Arachnides acères”), and together with a medley of lower Crustacea, Myriapoda, Thysanura, and Parasitic Insects; he was so cautious as to add “j’observerai seulement, que je ne connais pas encore bien la place naturelle des Pycnogonides et des Parasites,” and Cuvier, setting them in a similar position, adds a similar qualification.[[424]]

Leach (1814), whose great service it was to dissociate the Edriophthalmata and the Myriapoda from the Latreillian medley, left the group Arachnida as we still have it (save for the inclusion of the Dipterous Insect Nycteribia), and divided the group (with the same exception) into four Orders of which the Podosomata, i.e. the Pycnogonida, are one. Savigny (1816), less philosophical in this case than was his wont, assumed the Crustacean type to pass to the Arachnidan by a loss of several anterior pairs of appendages, and appears to set the Pycnogons in an intermediate grade, marking the pathway of the change. He considered the seven pairs of limbs of the Pycnogons to represent thoracic limbs of a Malacostracan, and, like so many of his contemporaries, was much biased by the apparent resemblance of Cyamus to Pycnogonum. The reader may find in Dohrn’s Monograph a guide to many other opinions and judgments, some of them of no small morphological interest and historical value[[425]]; but it behoves us to pass them by, and to inspect, in brief, the case as it stands at present. The obvious features in which a Pycnogon resembles a Spider or other typical Arachnid, are the possession of four pairs of walking legs, and the pre-oral position and chelate form of the first pair of appendages; we may perhaps also add, as a more general feature of resemblance, the imperfect subservience of limbs to the mouth as compared with any of the Crustacea. The resemblance would still be striking, in spite of the presence of an additional pair of legs in a few Pycnogons, were it not for the presence of the third pair of appendages or ovigerous legs of the Pycnogon, whose intercalation spoils the apparent harmony. We are neither at liberty to suppose, with Claus, that these members, so important in the larva, have been interpolated, as it were, anew in the Pycnogon; nor that they have arisen by subdivision of the second pair, as Schimkewitsch is inclined to suppose; nor that they have dropped out of the series in the Arachnid, whose body presents no trace of them in embryo or adult. In a word, their presence precludes us from assuming a direct homology between the apparently similar limbs of the two groups,[[426]] and at best leaves it only open to us to compare the last legs of the Pycnogon with the first abdominal, or genital, appendages of the Scorpion and the Spider. On the other hand, if we admit the seventh (as we must admit the occasional eighth) pair of appendages of Pycnogons to be unrepresented in the prosoma of the Arachnids, then, in the cephalothorax of the former, with its four pairs of appendages, we may find the homologue of the more or less free and separate part of the cephalothorax in Koenenia, Galeodes, and the Tartaridae. There is a resemblance between the two groups in the presence of intestinal diverticula that run towards or into the limbs, as in Spiders and some Mites, and there are certain histological and embryological resemblances that have been in part referred to above; but these, such as they are, are not adequate guides to morphological classification. We must bear in mind that such resemblances as the Pycnogons seem to show are not with the lower Arachnids but with the higher; they are either degenerates from very advanced and specialised Arachnida, or they are lower than the lowest. Confronted with such an issue, we cannot but conclude to let the Pycnogons stand apart, an independent group of Arthropods[[427]]; and I am inclined to think that they conserve primitive features in the usual presence of generative apertures on several pairs of limbs, and probably also in the non-development of any special respiratory organs. But inasmuch as the weight of evidence goes to show that subservience of limbs to mouth is a primitive Arthropodan character, the fact that the basal elements of the anterior appendages have here (as in Koenenia) no such relation to the mouth must be taken as evidence, not of antiquity, but of specialisation. In like manner the suctorial proboscis cannot be deemed a primitive character, and the much reduced abdomen also is obviously secondary and not primitive.

Classification.—No single genus more than another shows signs of affinity with other groups, and no single organ gives us, within the group, a clear picture of advancing stages of complexity. On the contrary, the differences between one genus and another depend very much on degrees of degeneration of the anterior appendages, and we have no reason to suppose that these stages of degeneration form a single continuous series, but have rather reason to believe that degeneration has set in independently in various ways and at various points in the series. But while we are unable at present to form a natural classification[[428]] of the Pycnogons, yet at the same time a purely arbitrary or artificial classification, conveniently based on the presence or absence of certain limbs, would run counter to such natural relationships as we can already discern.

The classification here adopted is a compromise between a natural system, so far as we can detect it, and an artificial one.

Two forms, separated from one another by many differences, show a minimum of degeneration, namely Decolopoda on the one hand, and the Nymphonidae on the other. The former genus has five pairs of legs, and this peculiarity is shared by Pentanymphon. In both groups the three anterior limbs are all present and well formed, save only that the ovigerous legs, which have ten joints in Decolopoda, are reduced to five joints in the Nymphons, and their denticulate spines, of which several rows are present in the former, are reduced to one row in the latter; on the other hand, a greater or a less degeneration of these limbs marks each and all of the other families.

Decolopoda is very probably the most primitive form known, though it has characters which seem to be the reverse of primitive in the dwarfish size of its chelophores and the crowded coalescent segmentation of the trunk. Colossendeis, in spite of its vanished chelophores, is probably closely allied: the shape and segmentation of the body and the several rows of smooth denticles on the ovigerous legs are points in common. The Eurycydidae are closely allied to Colossendeidae; they agree with Decolopoda in the two-jointed scape of the chelophore, and with Ammotheidae in the deflexed mobile proboscis. The true position of Rhynchothorax is very doubtful.

The Nymphonidae and Pallenidae are closely allied, and the Phoxichilidiidae have points of resemblance, especially with the latter. Nymphon compares with Decolopoda in the completeness of its parts, and is more typical in its long well-segmented body, and in its highly-developed chelae; but it already shows reduction in the scape of the chelophore, in the palps, and in the armature of the ovigerous legs.

The Phoxichilidae and Pycnogonidae (Agnathonia, Leach; Achelata, Sars), though differing greatly in aspect, are not improbably allied to one another; and whether this be so or not, the complete absence of chelophores and of palps affords an arbitrary character by which they are conveniently separated from all the rest.

The following table epitomises the chief characters of the several families:—

Pycnogonida. Proboscis. Chelophores. Palps. Ovigerous legs. Teeth on do. Legs. Trunk-segments. Genital Openings.
♂ ♀
(Cryptochelata, Sars)—
 Decolopodidae Fixed, decurved Complete, small, scape 2–jointed 10 joints 10 joints ♂, ♀ Four rows, simple 5 Condensed, coalescent 1, 2, 3, 4, 5, 1, 2, 3, 4, 5
 Colossendeidae Somewhat mobile, sometimes decurved 0 10 10 ♂, ♀ Many rows, simple 4 Coalescent 1, 2, 3, 4 1, 2, 3, 4
 Eurycididae Mobile, stalked, deflexed Scape 2–jointed, chelae rudimentary 10 10 ♂, ♀ More than one row, serrate 4 Well segmented 3, 4 1, 2, 3, 4
 Hannonia „ Rudimentary 0 10 ♂, ♀ Scattered spines 4 „ „ „
 Ammotheidae Mobile, deflexed „ 4–9 10 (or less) ♂, ♀ Few, scattered, serrate or smooth 4 Condensed, segmented „ „
 ? Rhynchothoracidae Large, fixed, aberrant 0 8 (5) 10 ♂, ♀ Toothed tubercles 4 „ 4 4
(Euchelata, Sars)—
 Nymphonidae Large, fixed Large, scape 1–jointed 5 (7) 8–10 ♂, ♀ One row, serrate 4–5 Well segmented 2, 3, 4 (5) 1, 2, 3, 4 (5)
 Pallenidae „ „ 0 or rudimentary 10 ♂, ♀ „ 4 „ (1, 2), 3, 4 „
 Phoxichilidiidae „ „ 0 5–6 ♂ One row, simple 4 „ 1, 2, 3, 4 „
(Achelata, Sars)—
 Phoxichilidae Large, fixed 0 0 7 ♂ Scattered, simple 4 Well segmented 2, 3, 4 1, 2, 3, 4
 Pycnogonidae „ 0 0 9 ♂ Small, irregular 4 Segmented, condensed 4 4

CLASS PYCNOGONIDA.[[429]]

Marine Arthropoda, with typically seven (and very exceptionally eight) pairs of appendages, of which none have their basal joints subservient to mastication, the first three are subject to suppression, the first (when present) are chelate, the second palpiform, the third ovigerous, and the rest form ambulatory limbs, usually very slender and long; with a suctorial proboscis, a limbless, unsegmented abdomen, and no manifest respiratory organs.

Fig. [282].—Decolopoda australis, Eights. A, × 1: from a specimen obtained at the South Shetlands by the Scotia Expedition. B, First appendage, or chelophore. (A, original; B, after Hodgson.)

Fam. 1. Decolopodidae.—Appendage I. dwarfed, but complete and chelate, scape with two joints; II. 9–10–jointed; III. well developed in both sexes, 10–jointed, the terminal joints with about four rows of teeth; five pairs of legs, destitute of accessory claws; genital apertures on all the legs (Bouvier).

Decolopoda australis, Eights[[430]] (1834), a remarkable form from the South Shetlands, recently re-discovered by the Scotia expedition. The animal is large, seven inches or more in total span, in colour scarlet; it was found in abundance in shallow water and cast upon the shore. The body is greatly condensed, the proboscis is “clavate, arcuated downwards,” and beset with small spines. A second Antarctic species, D. antarctica, has been described by Bouvier. The presence of a fifth pair of legs distinguishes Decolopoda from all known Pycnogons, except Pentanymphon. Stebbing would ally Decolopoda with, or even include it in, the Nymphonidae; but the presence of a second joint in the chelophoral scape, the number of joints in, and the armature on, the ovigerous legs, and the deflexed proboscis, are all characters either agreeing with or tending towards those of the Eurycididae; while the Colossendeidae would be very like Decolopoda were it not for the complete suppression of the chelophores. It seems convenient to constitute a new family for this remarkable form.

Fam. 2. Colossendeidae (Pasithoidae, Sars).—Appendage I. absent in adult; appendage II. very long, 10–jointed; appendage III. 10–jointed, clawed, with many rows of teeth; auxiliary claws absent; segments of trunk fused; proboscis very large, somewhat mobile; genital apertures, in at least some cases, on all the legs.

Pasithoe, Goodsir (1842), which Sars assumes as the type of the family, is here relegated to Ammothea.[[431]] Colossendeis, Jarszynsky (1870) (Anomorhynchus, Miers (1881), Rhopalorhynchus, Wood-Mason (1873)), remains as the only genus commonly accepted: large, more or less slender short-necked forms; world-wide, principally Arctic, Antarctic, and deep-sea; about twenty-five species.[[432]] The largest species, C. gigas, Hoek, from great depths in the Southern Ocean, has a span of about two feet. The North Atlantic C. proboscidea and Antarctic C. australis are very closely related to one another. Carpenter would retain the genus Rhopalorhynchus for R. kröyeri, W.-M. (Andamans), R. clavipes, Carp. (Torres Straits), and R. tenuissimus, Haswell (Australia), all more or less shallow-water species, excessively attenuated, with the second and third body-segments elongated, the caudal segment excessively reduced, the club-shaped proboscis on a slender stalk, and other common characters. Pipetta weberi, Loman (1904), is a large and remarkable form from the Banda Sea, apparently referable, in spite of certain abnormal features, to this family; the proboscis is extraordinarily long and slender; the palps have eight joints, the ovigerous legs eleven.

Fam. 3. Eurycididae (Ascorhynchidae, Meinert).—Appendage I. more or less reduced; appendage II. 10–jointed (absent in Hannonia); appendage III. 10–jointed, clawed, with more than one tow of serrated teeth; proboscis movably articulated and more or less bent under the body; auxiliary claws absent.

Fig. [283].—Eurycide hispida, Kr.; side view.

Eurycide, Schiödte (1857) (Zetes, Kröyer, 1845): Appendage I. with two-jointed scape, without chelae in adult; one species (E. hispida, (Kr.)), from the North Atlantic and Arctic, and two others from the East Indies, recently described by Loman. Barana arenicola, Dohrn (1881), is nearly allied. Ascorhynchus, G. O. Sars (1876) (Gnamptorhynchus, Böhm, 1879; Scaeorhynchus, Wilson, 1881), very similar to Eurycide, with which, according to Schimkewitsch, it should be merged, includes large, smooth, elongated forms, with long neck and expanded frontal region, and a long proboscis lacking the long scape that supports the proboscis in Eurycide; about twelve species, world-wide, mostly deep-water. Barana castelli, Dohrn, from Naples is akin to the foregoing genera, but seems to deserve generic separation from B. arenicola. Ammothea longicollis, Haswell, from Australia, is, as Schimkewitsch has already remarked, almost certainly a Eurycide, as is also, probably, Parazetes auchenicus, Slater, from Japan.

Hannonia typica, Hoek (1880), from Cape Town, is a remarkable form, lately redescribed by Loman. The chelophores are much reduced, the palps are absent; the ovigerous legs are 10–jointed, and clawed; the terminal joints of the latter bear long straight spines, scattered over their whole surface; the proboscis is borne on a narrow stalk, and sharply deflexed. The eggs form a single flattened mass, as in Pycnogonum. While the lack of palps would set this genus among the Pallenidae, the remarkable proboscis seems to be better evidence of affinity with Ascorhynchus and Eurycide.[[433]]

Nymphopsis, Haswell (1881), is a genus of doubtful affinities, placed here by Schimkewitsch. The first appendage is well-developed and chelate; the palps are 9–jointed, the ovigerous legs are 7–jointed, none of the joints being provided with the compound spines seen in Nymphon and Pallene. It is perhaps an immature form. Schimkewitsch has described another species, N. korotnevi, and Loman a third, N. muscosus, both from the East Indies.

Fam. 4. Ammotheidae.—Akin to Eurycididae in having the proboscis more or less movably jointed to the cephalic segment, and appendage I. reduced, non-chelate in the adult; the body is compact and more or less imperfectly segmented; appendage II. 4–9–jointed; appendage III. clawless, and the number of joints sometimes diminished, with a sparse row of serrated spines; auxiliary claws usually present.

Ammothea, Leach (1815) (including Achelia, Hodge (1864) = the old non-chelate individuals): appendage I. very small, 2–jointed; appendage II. 8–9–jointed; caudal segment fused with last body-segment; about eighteen species, four from the South Seas, two or three from the East Indies, the rest mostly Mediterranean and North Atlantic, in need of revision. Ammothea longipes, Hodge, is the young of Achelia hispida, Hodge; and Ammothea magnirostris, Dohrn, is apparently the same species. A. fibulifera, Dohrn, seems identical with Achelia echinata, Hodge (of which A. brevipes, Hodge, is the young), and so probably is A. achelioides, Wilson; Endeis didactyla, Philippi (1843), is very probably the same species. A. uniunguiculata, Dohrn (? Pariboea spinipalpis, Philippi (1843)), has no auxiliary claws. Leionymphon, Möbius (1902), contains nine Antarctic forms, allied to Ammothea (including A. grandis, Pfeffer, and Colossendeis gibbosa, Möb., which two are probably identical), with characteristic transverse ridges on the body, a large proboscis, a 9–jointed palp, and somewhat peculiar ovigerous legs. Cilunculus, Fragilia, and Scipiolus are new genera more or less allied to Leionymphon, described by Loman (1908) from the Siboga Expedition.[[434]] Tanystylum, Miers (1879) (including Clotenia, Dohrn (1881), and Discoarachne, Hoek (1880)), has appendage I. reduced to a single joint or a small tubercle, and appendage II. 4–6–jointed; world-wide; about eight species. Austrodecus glacialis and Austroraptus polaris are two allied Antarctic species, described by Hodgson (1907), the former a curious little form with a pointed, weevil-like proboscis, no chelophores, and 6–jointed palp. Trygaeus communis, Dohrn (1881), from Naples, has a 7–jointed, and Oorhynchus aucklandiae, Hoek (1881), a 9–jointed palp; the former has only seven joints in the ovigerous leg. Lecythorhynchus armatus, Böhm (1879), with rudimentary 2–jointed chelophores, and L. (Corniger) hilgendorfi, Böhm, with small tubercles in their place, both from Japan, have also 9–jointed palps: the former, at least, is apparently an Ammothea. Several insufficiently described genera, Phanodemus, Costa (1836), Platychelus, Costa (1861), Oiceobathes, Hesse (1867), and Böhmia, Hoek (1880), seem to be referable to this group; all have chelate mandibles, and may possibly be based on immature forms.

Goodsir’s Pasithoe vesiculosa[[435]] is, in my opinion, undoubtedly Ammothea hispida, Hodge, and so also, I believe, is his Pephredo hirsuta; P. umbonata, Gould[[436]] (Long Island Sound), is, with as little doubt, Tanystylum orbiculare, Wilson.

Fam. 5. Rhynchothoracidae.—The animal identified by Dohrn as Rhynchothorax mediterraneus, Costa (1861), is a minute and very remarkable form, without chelophores, with large 8–jointed palps, reduced by fusion to five joints, and 10–jointed, clawed ovigerous legs, which last are provided on the last five joints with peculiar toothed tubercles. The general aspect of the body is somewhat like that of an Ammothea, which genus it resembles in the ventral insertion of the ovigerous legs and the somewhat imperfect segmentation of the body. It differs from Ammotheidae in the possession of a claw on appendage III. It is highly peculiar in the structure of the mouth, in having a long forward extension of the oculiferous tubercle jutting out over the proboscis, in the extreme shortness of the intestinal caeca and ovaries which scarcely extend into the legs, and in the absence of cement-glands from the fourth joint of the legs; these last are present only in the third joint of the penultimate legs. A single pair of generative orifices are found on the last legs. A second species, R. australis, Hodgson, comes from the Antarctic.

Fig. [284].—Rhynchothorax mediterraneus, Costa. A, Body and bases of legs; B, terminal joints of palp. (After Dohrn.)

Fam. 6. Nymphonidae.—Appendage I. well-developed, chelate; II. well-developed, usually 5–jointed; III. well-developed in both sexes, usually 10–jointed, the terminal joints with one row of denticulated spines.

Nymphon, Fabr. (1794), about forty-five recognised species, of which some are but narrowly defined. Closely allied are Chaetonymphon, G. O. Sars (1888), including thick-set, hairy species, about eight in number, from the North Atlantic, Arctic, and Antarctic; and Boreonymphon, G. O. Sars (1888), with one species (B. robustum, Bell, Fig. [276]), also northern, in which the auxiliary claws are almost absent. Nymphon brevicaudatum, Miers (= N. horridum, Böhm), an extraordinary hispid form from Kerguelen,[[437]] is also peculiar. Pentanymphon, Hodgson (1904), from the Antarctic (circumpolar), differs in no respect save in the presence of a fifth pair of legs; one species.

The only other genus is Paranymphon, Caullery (1896) (one species, Gulf of Gascony, West of Ireland, Greenland), in which the palp is (6–)7–jointed, the ovigerous leg 8–jointed, and the auxiliary claws are absent.

Fam. 7. Pallenidae.—As in Nymphon, but appendage II. absent or rudimentary.

Fig. [285].—Pallene brevirostris, Johnston, ♀, Plymouth.

Pallene, Johnston (1837): about ten species (Mediterranean, North Atlantic, Arctic, Australia). P. languida, Hoek, Australia, lacks auxiliary claws, and is otherwise distinct; but P. novaezealandiae, G. M. Thomson, is typical. Pseudopallene, Wilson (1878):[[438]] appendage III. clawed; auxiliary claws absent; four (or more) species (North Atlantic, Arctic, Antarctic). P. (Phoxichilus) pygmaea, Costa (1836), and P. spinosa, Quatref., seem to belong to this genus or to Pallene. Cordylochele, G. O. Sars (1888): closely allied, but with front of cephalic segment much expanded and chelae remarkably swollen, includes three very smooth, elongated, northern species, to which Bouvier has added one from the Antarctic; Pallene laevis, Hoek, from Bass’s Straits, is somewhat similar. Neopallene, Dohrn (1881): as in Pallene, but with a rudimentary second appendage in the female, and no generative aperture on the last leg in the male (one species, Mediterranean). Parapallene, Carpenter (1892): as in Pallene, but without auxiliary claws, and with the two last segments of the trunk (which in Pallene are coalesced) independent (about ten species, East Indies and Australia); Pallene grubii, Hoek (Phoxichilidium sp., Grube, 1869), is probably congeneric. Pallenopsis, Wilson (1881): appendage I. 2–jointed; appendage II. rudimentary, 1–jointed; appendage III. clawless; auxiliary claws present; slender forms, including some formerly referred to Phoxichilidium; about fifteen species, world-wide. Pallene dimorpha, Hoek, from Kerguelen, with 4–jointed palps, deserves a new generic appellation. P. longiceps, Böhm, from Japan, with rudimentary 2–jointed palps in the male, is also peculiar.

Fig. [286].—Phoxichilidium femoratum, Rathke, Britain. A, The animal with its legs removed; B, leg and chela.

Fam. 8. Phoxichilidiidae.—Appendage I. well-developed; II. absent; III. present only in the male, having a few simple spines in a single row. The last character is conveniently diagnostic, but nevertheless the Phoxichilidiidae come very near to the Pallenidae, with which, according to Schimkewitsch and others, they should be merged; the two families resemble one another in the single row of spines on the ovigerous legs and in the extension of the cephalic segment over the base of the proboscis.

Phoxichilidium, M.-E. (1840): appendage III. 5–jointed; five or six species (Mediterranean, North Atlantic, Arctic, Australia, Japan). Anoplodactylus, Wilson (1878): appendage III. 6–jointed; auxiliary claws absent or very rudimentary; about twelve species, cosmopolitan, of which many were first referred to Phoxichilidium. A. neglectus, Hoek, comes from 1600 fathoms off the Crozets. Oomerus stigmatophorus, Hesse (1874), from Brest, seems to belong to one or other genus, but is unrecognisable. Anaphia, Say (1821), is in all probability identical with Anoplodactylus, and if so the name should have priority. Halosoma, Cole (1904), is an allied genus from California.

Fig. [287].—Anoplodactylus petiolatus, Kr., Britain. A, Dorsal view; B, side view.

Fam. 9. Phoxichilidae.[[439]]—Appendage I. and II. absent; appendage III. present only in the males, 7–jointed, with minute scattered spines; auxiliary claws well-developed; body and legs slender. The only genus is Phoxichilus (auctt., non Latreille, Chilophoxus, Stebbing, 1902); the type is P. spinosus, Mont. (non Quatrefages), from the N. Atlantic, and P. vulgaris, Dohrn, P. charybdaeus, Dohrn, and P. laevis, Grube, are all very similar. Endeis gracilis, Philippi (1843), is probably identical with P. spinosus, or one of its close allies. There are also known P. meridionalis, Böhm, P. mollis, Carp., and P. procerus, Loman, from the East Indies; P. australis, Hodgson, from the Antarctic; P. böhmii, Schimk., of unknown locality; and forms ascribed to P. charybdaeus by Haswell and by Schimkewitsch from Australia and Brazil.

Fam. 10. Pycnogonidae.—Appendages I. and II. absent; appendage III. present only in the male, 9–jointed, with small, simple spines; auxiliary claws absent or rudimentary; body and legs short, thick-set.

The only genus is Pycnogonum, Brünnich (1764) (Polygonopus, Pallas, 1766); the type is P. littorale, Ström, of the N. Atlantic (0–430 fathoms), to which species have also been ascribed forms from various remote localities, e.g. Japan, Chile, and Kerguelen. P. crassirostre, G. O. Sars, a northern and more or less deep-sea form, is distinct, and so also are P. nodulosum and P. pusillum, Dohrn, from Naples. P. stearnsi, Ives, from California, is like P. littorale, except for the rostrum, which resembles that of P. crassirostre. P. magellanicum, Hoek, P. magnirostre, Möbius, both from the Southern Ocean; P. microps, Loman, from Natal, and four others described by Loman from the East Indies, are the other authenticated species. Of P. philippinense, Semper, I know only the bare record; and P. australe, Grube, is described only from a larval form with three pairs of legs. P. orientale, Dana (first described as Astridium, n.g.), is also described from an immature specimen, and more resembles a Phoxichilus.

The British Pycnogons.

Dr. George Johnston,[[440]] the naturalist-physician of Berwick-on-Tweed, Harry Goodsir,[[441]] brother of the great anatomist, who perished with Sir John Franklin, and George Hodge[[442]] of Seaham Harbour, a young naturalist of singular promise, dead ere his prime, were in former days the chief students of the British Pycnogons. Of late, Carpenter[[443]] has studied the Irish species; and the cruises of the Porcupine, Triton, and Knight Errant have given us a number of deep-water species from the verge of the British area.

In compiling the following list, I have had the indispensable advantage of access to Canon Norman’s collection, and the still greater benefit of his own stores of endless information.[[444]]

Pseudopallene circularis, Goodsir: Firth of Forth.

Phoxichilidium femoratum, Rathke (P. globosum, Goodsir; Orithyia coccinea, Johnston) (Figs. [270], B; 286): East and West coasts, Shetland, Ireland.

Anoplodactylus virescens, Hodge (? Phoxichilidium olivaceum, Gosse): South coast.

A. petiolatus, Kr. (Figs. [270], C; 275, B; 287) (Pallene attenuata and pygmaea, Hodge; Phoxichilidium exiguum and longicolle, Dohrn): Plymouth, Firth of Forth, Cumbrae, Irish coasts.

Ammothea (Achelia) echinata, Hodge (Fig. [265], B; 274, 4; 275, E): Plymouth, Channel Islands, Isle of Man, Cumbrae, Durham (Hodge), West of Ireland. We have not found it on the East of Scotland. A. brevipes, Hodge, is presumed to be the young. Two of Dohrn’s Neapolitan species, A. fibulifera and A. franciscana, are in my opinion not to be distinguished from one another, nor from the present species.

A. hispida, Hodge (Fig. [266], C) (A. longipes, Hodge (juv); A. magnirostris, Dohrn;? Pasithoe vesiculosa, Goodsir;? Pephredo hirsuta, Goodsir): Cornwall and Devon (Hodge and Norman), Jersey. The form common on the East of Scotland would seem to be this species. The Mediterranean A. magnirostris, Dohrn, appears to be identical.

A. laevis, Hodge: Cornwall (Hodge), Devon (Norman), Jersey (Sinel).

Tanystylum orbiculare, Wilson (Clotenia conirostre, Dohrn): Donegal (Carpenter).

Phoxichilus spinosus, Mont. (Fig. [265], C; 270, A; 275, C): South Coast, Moray Firth, Firth of Clyde, Ireland. A smaller and less spiny form occurs, which Carpenter records as P. laevis, Grube, but Norman unites the two under the name of Endeis spinosus (Mont.).

Pycnogonum littorale, Ström (Fig. [262]): on all coasts, and to considerable depths (150 fathoms, West of Ireland).

Nymphon brevirostre, Hodge (N. gracile, Sars) (Figs. [263], 264, 267, A; 272, 274, 3): common on the East Coast; Herm (Hodge), Dublin, Queenstown (Carpenter). Our smallest species of Nymphon.

N. rubrum, Hodge (N. gracile, Johnston; N. rubrum, G. O. Sars): common on the East Coast; Oban (Norman), Ireland (Carpenter).

N. grossipes, O. Fabr., Johnston (N. johnstoni, Goodsir): Northumberland, East of Scotland, Orkney, etc., not uncommon.

N. gracile, Leach (N. gallicum, Hoek; ♂ N. femoratum, Leach): South of England, West of Scotland, and Ireland.

N. strömii, Kr. (N. giganteum, Goodsir) (Figs. [273], 274, 2): East Coast, from Holy Island to Shetland.

Chaetonymphon hirtum, Fabr. (Fig. [274], 1): Northumberland (Hodge), Margate (Hoek), East of Scotland, and Ireland, not uncommon. There seems to be no doubt that British specimens agree with this species as figured and identified by Sars. N. spinosum, Goodsir (East of Scotland, Goodsir; Belfast, W. Thompson), is, according to Norman, the same species. Sars’ Norwegian specimens figured under the latter name are not identical, and have been renamed by Norman C. spinosissimum, but are said by Meinert and Möbius to be identical with C. hirtipes, Bell.

Hodge (1864) records Nymphon mixtum, Kr., and N. longitarse, Kr., from the Durham coast. His full list of the recorded species of other authors also includes the following doubtful or unrecognised species: N. pellucidum, N. simile, and N. minutum, all of Goodsir.

Pallene brevirostris, Johnston (P. empusa, Wilson;? P. emaciata, Dohrn) (Figs. [275], A; 285): all coasts. Examples differ considerably in size and proportions, as do Dohrn’s Neapolitan species one from another. We have specimens from the Sound of Mull that come very near, and perhaps agree with, Sars’ P. producta, a species that scarcely differs from P. brevirostris, save in its greater attenuation; the same species has also been recorded from Millport and from Port Erin.

P. spectrum, Dohrn: Plymouth (A. H. Norman).

Besides the above, all of which are littoral or more or less shallow-water species, we have another series of forms, or, to speak more correctly, we have two other series of forms, from the deep Atlantic waters within the British area. In the cold area of the Faeroe Channel we have Boreonymphon robustum, Bell; Nymphon elegans, Hansen; N. sluiteri, Hoek; N. stenocheir, Norman; Colossendeis proboscidea, Sabine; C. angusta, Sars. In the warm waters south and west of the Wyville-Thomson ridge we have Chaetonymphon spinosissimum, Norman; Nymphon gracilipes, Heller (non Fabr.); N. hirtipes, Bell; N. longitarse, Kr.; N. macrum, Wilson; Pallenopsis tritonis, Hoek (= P. holti, Carpenter); Anoplodactylus oculatus, Carpenter, and A. typhlops, G. O. Sars; and to the list under this section Canon Norman has lately made the very interesting addition of Paranymphon spinosum, Caullery, from the Porcupine Station XVII., S.S.E. of Rockall, in 1230 fathoms. Lastly, and less clearly related to temperature, we have Chaetonymphon tenellum, Sars; N. gracilipes, Fabr.; N. leptocheles, Sars; N. macronyx, Sars; N. serratum, Sars; and Cordylochele malleolata, Sars.

Of the species recorded in the above list as a whole, Anoplodactylus virescens, Nymphon gracile, and Pallene spectrum reach their northern limit in the southern parts of our own area; Ammothea echinata, Anoplodactylus petiolatus, Pallene brevirostris, and Phoxichilus spinosus (or very closely related forms) range from the Mediterranean to Norway, the last three also to the other side of the Atlantic; Nymphon brevirostre and N. rubrum range from Britain, where they are in the main East Coast species, to Norway. Of the Atlantic species, other than the Arctic ones, the majority are known to extend to the New England coast.