The Cambridge natural history, Vol. 04 (of 10)

by Geoffrey Smith; D'Arcy Wentworth Thompson; Cecil Warburton; Walter Frank Raphael Weldon; Henry Woods

TRILOBITA

BY

HENRY WOODS, M.A.

St. John’s College, Cambridge, University Lecturer in Palaeozoology

CHAPTER VIII
TRILOBITA

Among the many interesting groups of fossils found in the Palaeozoic deposits there is none which has attracted more attention than the Trilobites. As early as 1698, Edward Lhwyd, Curator of the Ashmolean Museum in Oxford, recorded in the Philosophical Transactions the discovery of Trilobites in the neighbourhood of Llandeilo in South Wales; and of one of his specimens he remarked that “it must be the Sceleton of a flat Fish.” In the following year the same writer gave in his Lithophylacii Britannici Ichnographia descriptions and figures of two Trilobites which are evidently examples of the species now known as Ogygia buchi and Trinucleus fimbriatus.

Although Trilobites differ so much from living Arthropods that it was difficult to determine even whether they belonged to the Crustacea or the Arachnida, yet one of the earliest writers, Dr. Cromwell Mortimer, Secretary of the Royal Society (1753), recognised their resemblance to Apus (see pp. [19]–36). This view of their affinities was adopted by Linnaeus, and has been supported by many later writers. Another early author, Emanuel Mendez da Costa, thought that the Trilobites were related to the Isopods, an opinion which has been held by some few zoologists of more recent times.

The Trilobites form the only known Order of the Crustacea which has no living representatives. They are found in the oldest known fossiliferous deposits—the Lower Cambrian or Olenellus beds, where they are represented by 19 genera belonging to the families Agnostidae, Paradoxidae, Olenidae, and Conocephalidae. From the variety of forms found and the state of development which they have reached, it is evident that even at that remote period the group must have been of considerable antiquity; but of its pre-Cambrian ancestors nothing is yet known; consequently there is no direct evidence of the origin of the group.

Trilobites form an important part of all the faunas of the Cambrian system; they attain their greatest development in the Ordovician period, after which they become less numerous; their decline is very marked in the Devonian, in which nearly all the genera are but survivals from the Silurian period; in the Carboniferous, evidence of approaching extinction is seen in the small number of genera represented, all of which belong to one family—the Proëtidae, in the relatively few species in each genus and in the small size of the individuals of those species. In Europe no representatives of the group appear to have survived the Carboniferous period, but in America one form has been recorded from deposits of Permian age.

Trilobites seem to have been exclusively marine, since they are found only in association with the remains of marine animals. Their range in depth was evidently considerable, for they occur in many different kinds of sediment, and were apparently able to live regardless of the nature of the sea-floor—whether muddy, sandy, calcareous, or rocky. In some cases they occur in deposits containing reef-building corals and other shallow water animals; in others they are associated with organisms which lived at greater depths. The group appears to have had a world-wide distribution, for the remains of Trilobites are found in the Palaeozoic rocks of all countries. Their range in size is considerable; for whilst a large proportion of the species are about two or three inches in length, some, like Agnostus, are only a quarter of an inch long, others are from ten to twenty inches long, the largest forms including species of Paradoxides, Asaphus, Megalaspis, Lichas, and Homalonotus.

The feature in a Trilobite which first attracts attention is the marked division of the dorso-ventrally flattened body into a median or axial part, and a lateral or pleural part on each side. It was this character that led Walch, in 1771, to give the name by which the group is now known. The axial part of the body contained the alimentary canal, as is shown by the position of the mouth and anus, as well as by casts in mud of the canal which are found in some specimens. The trilobation of the body is quite distinct in the majority of Trilobites, but in a few genera belonging to the Asaphidae and Calymenidae (Fig. [136]) it becomes more or less completely obsolete.

Fig. [136].—Homalonotus delphinocephalus, Green, × 1. Silurian. (After Zittel.)

In most cases the only part of the Trilobite which is preserved is the exoskeleton which covered the dorsal surface of the body. That skeleton consists largely of calcareous material, and shows in sections a finely perforated structure. Generally it is arched above, but in some cases is only slightly convex; in outline it is more or less oval. Three regions can always be distinguished in the body of a Trilobite—the head, the thorax, and the abdomen or pygidium.

The carapace which covers the head is known as the cephalic shield (Fig. [137], A, 1), and is commonly more or less semicircular in outline, but varies considerably in different genera. Only in a few cases, as in some species of Agnostus (Fig. [146]), is its length greater than its breadth. The axial part of the cephalic shield, called the “glabella” (Fig. [137], A, a), is usually more convex than the lateral parts (“cheeks” or “genae”), and is separated from them by longitudinal or axial furrows (b). The shape of the glabella varies greatly; it may be oblong, circular, semi-cylindrical, pyriform, spherical, etc. Its relative size likewise varies; thus in Phacops cephalotes it expands in front and forms the larger part of the head, whilst in Arethusina (Fig. [151], B) it is narrow and short, being only about one-half of the length of the head.

Fig. [137].—Calymene tuberculata, Brünn. × 1. Silurian, Dudley. A, Dorsal surface: 1, head; 2, thorax; 3, pygidium or abdomen. a, Glabella; b, axial furrow; c, glabella-furrow; d, neck-furrow; e, fixed cheek; f, free cheek; g, facial suture; h, eye; i, genal angle; k, axis of thorax; l, pleura. B, Ventral surface of head (after Barrande): a, hypostome; b, doublure; c, c′, facial sutures; d, rostral suture; e, rostral plate. C, One segment of the thorax: a, ring of axis; b, groove; c, articular portion; d, axial furrow; d-f, pleura; d-e, internal part of pleura; e-f, external part of pleura; e, fulcrum; g, groove. D, Coiled specimen: a, glabella; b, eye; c, facial suture; d, pygidium; e, rostral suture; f, continuation of facial suture.

The segmentation of the head is indicated by transverse furrows on the glabella (Fig. [137], A, c, d). In some cases these furrows extend quite across the glabella (Fig. [147]), but commonly they are found on the sides only and divide the glabella into lateral lobes. Only the posterior or “neck-furrow” (Fig. [137], A, d) is continued on to the cheeks, and the segment which it limits anteriorly on the glabella[^[180]] is known as the occipital or neck-ring. In front of the neck-furrow there may be three other furrows, so that altogether five cephalic segments are indicated by the furrows of the glabella. Commonly all the furrows are distinct in the primitive types; but in the more modified forms some, especially the anterior, become either reduced in size or obsolete. The actual number of furrows present consequently varies in different genera, and may even differ in different species of the same genus. In a few genera all the furrows are either indistinct or absent, as for example in Ellipsocephalus (Fig. [150], B). In some cases four furrows are present in addition to the neck-furrow; this is due to the division of the anterior lobe of the glabella by fulcra which are developed for the attachment of muscles.

When the glabella reaches the front border of the head the two cheeks are separated (Fig. [150], I); but in other cases they unite in front of the glabella (Fig. [150], C). The outer posterior angle of the cheeks or genae (“genal angle,” Fig. [137], A, i) may be rounded, pointed, or produced into backwardly directed spines (Fig. [140]). The marginal part of the cephalic shield is often flattened or concave; this border may be quite a narrow rim as in Calymene (Fig. [137], A), but in some genera (e.g. Trinucleus, Fig. [140], B; Harpes, Fig. [150], A; Asaphus) it attains a great development. Each cheek is usually divided by a suture—the “facial suture” (Fig. [137], A, g)—into an inner and an outer part; the former is the “fixed cheek” (e), and the latter the “free cheek” (f). The course of the facial suture varies in different genera: on the posterior part of the head it begins either at the posterior margin (Fig. [150], C) or at the posterior part of the lateral margin (Fig. [151], C, D); at first it is directed inwards, and then bends forward, forming an angle. In front it may (a) end at the front margin (Fig. [147]), or (b) be united beneath the front margin by a rostral suture (Fig. [137], B, d, D, e), or (c) unite with the other suture on the dorsal surface in front of the glabella (Fig. [151], C). In the last case the free cheeks also unite in front of the glabella.

The facial suture is one of the distinguishing features of the Trilobites, and may have been of some use in ecdysis. In only a few forms is it absent, as for example in Agnostus (Fig. [146]) and Microdiscus. In the former, however, Beecher states that a suture is really present, but, unlike that of most other Trilobites, it is situated at the margin of the cephalic shield, and consequently the free cheek, if present, must be on the ventral surface. Lindström and Holm, after a re-examination of well-preserved specimens, deny the existence of a suture in Agnostus. By most authors Olenellus is said to be without a suture, but Beecher maintains that although the fixed and free cheeks have coalesced, yet a raised line passing from the eye-lobe to the posterior margin marks the position of the suture; this view is not accepted by Lindström.

The existence of a facial suture in Trinucleus has likewise been disputed. But Emmerich, Salter, and M‘Coy[^[181]] have maintained that a suture is present in a normal position on the dorsal surface, extending from the posterior margin just within the genal angle to the eye (when present), and from thence bending forward and ending on the front margin near the glabella. It must be admitted that no indications of the suture are seen in the majority of specimens, perhaps owing to the fact that most examples of Trinucleus are in the form of internal casts; perhaps also to the more or less complete coalescence of the fixed and free cheeks, since in no specimen has the free cheek been found separated from the rest of the head, as occurs not uncommonly in many other Trilobites. The probability of the existence of a suture receives some support from the fact that one is found in the allied genera Orometopus and Ampyx (Fig. [140]). Barrande and Oehlert deny its existence in Trinucleus. There is, however, in that genus a suture running close to the margin of the cephalic border,[^[182]] and joining the genal angle so as to cut off the genal spine. Lovén and Oehlert claim that this suture represents the facial suture, but in an abnormal position; this view, however, is not accepted by Beyrich. In this connection it should be noted that in Acidaspis, whilst the majority of the species possess a facial suture, there are two in which it has disappeared owing to the fusion of the fixed and free cheeks. Such being the case, it seems not improbable that the curved line passing backwards from the eye in Harpes may mark the position of the suture; but it is stated that the only suture present in that form runs at the margin of the cephalic border, and is similar to that of Trinucleus. This matter will be referred to again when discussing the nature of the eyes in Trinucleus and Harpes.

The relative sizes of the fixed and free cheeks obviously depend on the position of the facial suture; when this starts on the lateral margin of the cephalic shield and passes forward to the outer part of the front margin, the free cheek will be a narrow strip; when, on the other hand, the suture starts from the posterior margin and runs close to the glabella, the free cheek will be relatively large and the fixed cheek narrow. The fixed cheek is small in Phacops, Cheirurus, and Illaenus; relatively large in Remopleurides, Phillipsia, and Stygina. It was suggested by M‘Coy[^[183]] that the free cheek represents the pleura of an anterior segment which has not become fused with the other cephalic pleurae. The fixed cheek appears to be formed of the coalesced pleurae of the other cephalic segments, but of those pleurae the only indication seen in adult specimens is in the neck-ring; in young specimens of Olenellus, however, the presence of other pleurae is indicated by furrows on the cheeks in front of the neck-furrow.

Fig. [138].—Phacops latifrons, Bronn, × 1. Devonian. Showing large compound eye. (After Zittel.)

A pair of compound eyes are present in the majority of Trilobites. Each eye is situated on the free cheek, at that part of its inner margin where the facial suture bends to form an angle (Figs. [137], A, h, 138). The position of the eye is consequently determined by the position of the facial suture; it may be near the glabella or near the lateral margin of the head, and either as far forward as the first segment of the glabella or nearly as far back as the neck-furrow. In many Trilobites the eye is more or less conical, with its summit truncated or rounded, but in some genera it is ovoid, or crescentic. In Aeglina (Fig. [150], H) the eye is flattened and scarcely raised above the general level of the cheek. The eye of a Trilobite is oriented so that its longer axis is parallel or nearly parallel to the axis of the body (Fig. [150], G); but in one case (Encrinurus intercostatus) it is placed at right angles to this axis. The size of the eye varies considerably; it is largest in Aeglina, in which it covers nearly the whole of the free cheek; it is small in Acidaspis and Encrinurus.

Though the eye is always entirely on the free cheek, the adjoining part of the fixed cheek is raised to form a buttress on which the eye rests; this buttress, which is known as the “palpebral lobe,” is seen clearly when the fixed cheek is removed. The eyes of Trilobites are always sessile; for although in some species, such as Asaphus cornigerus, A. kowalewskii, and Encrinurus punctatus, they are on the summits of prominent stalks, yet those stalks are immovable.

Three types of compound eye have been recognised in Trilobites[^[184]]—holochroal, prismatic, and schizochroal.

Fig. [139].—Eyes of Trilobites. (After Lindström.) A, B, Sphaerophthalmus alatus, Ang. Upper Cambrian. Vertical and horizontal sections, × 100. C, Asaphus fallax, Dalm. Horizontal section, × 60. D, Nileus armadillo, Dalm. Vertical section, × 60, a, prismatic lenses; b, cuticle; c, part of free cheek. E, Dalmanites vulgaris, Salt. Part of eye, × 30. F, Dalmanites imbricatulus, Ang. Vertical section of eye, with a part of the free cheek on the left, × 60. G, H, Harpes vittatus, Barr. G, The two lenses of one eye, × 8; H, vertical section of the same, × 60.

1. In the holochroal eye (Fig. [139], A, B) the lenses are globular or biconvex and close together, so that the cornea is continuous over the entire eye. Examples of this are seen in Bronteus and Sphaerophthalmus.

2. In the prismatic type (Fig. [139], C, D) the lenses are prismatic and plano-convex, and the entire surface of the eye is covered by a smooth cuticle. The lenses are close together and usually hexagonal, but occasionally rhombic or square. Near the margin of the eye the lenses may become irregular, giving rise to a border in which the prismatic structure is more or less indistinct. The prismatic type of eye is found in the genera Asaphus, Nileus, Illaenus, etc.

3. The schizochroal eye occurs only in the family Phacopidae (Fig. [139], E, F). The lenses are biconvex and are separated by portions of the cephalic shield, so that each lens appears to rest in a separate socket, and the cornea is not continuous for the entire eye. The lenses are circular in outline, but owing to the upward and inward growth of the interstitial test they may appear, on the surface, to be hexagonal. The diameter of a lens may be as much as 0·5 mm. The crystalline cones have not been preserved. In specimens of Phacops rana, in which the inner face of the lens is more convex than the outer, J. M. Clarke[^[185]] has obtained evidence of a posterior spheroidal cavity in addition to the anterior corneal cavity. The complete separation of the lenses in this type of eye has led to the suggestion that the schizochroal eye is an aggregate rather than a compound eye. But the difference between this and the holochroal eye is probably less than appears at first sight if the statement made by Clarke is confirmed, namely, that in young specimens of Calymene senaria the lenses are relatively large and similar to those of Phacops, whereas in the adult the eye is holochroal.

These three types of eye, according to Lindström, have appeared successively in chronological order: the prismatic occurring first in the Olenus beds (Upper Cambrian), the holochroal first in the Ceratopyge Limestone (Uppermost Cambrian), and the schizochroal first in the Ordovician. The number of lenses in the eye varies greatly. For example, in Trimerocephalus volborthi there are 14 only, whilst in Remopleurides radians there are as many as 15,000. Even in different species of the same genus there may be considerable differences. Thus Bronteus brongniarti possesses 1000, B. palifer 4000, lenses in each eye. The number increases from the young up to the adult, but decreases in old age. The lenses are usually arranged in alternating rows. In Trilobites with a conical eye the outer segment of the cone bears the visual surface. It has been stated that the eyes of Trilobites resemble those of Isopods,[^[186]] but close comparison is difficult to make, since in Trilobites no part of the eye beneath the lenses is preserved. In some genera a threadlike ridge, called the “eye-line,” passes from the glabella, generally from the front segment, to the eye, where it often ends in the palpebral lobe; this eye-line is found in nearly all genera which are confined to the Cambrian period, and persists in a few of later date, as for example in Triarthrus, Euloma, and some species of Calymene from the Ordovician; in Arethusina and Acidaspis from the Silurian; and in Harpes from the Devonian (Fig. [150], A).

Fig. [140].—Trinucleidae. A, Orometopus elatifrons, Ang. × 5. Restoration based on specimens from the Upper Cambrian (Tremadoc) of Shineton, Shropshire. B, Trinucleus bucklandi, Barr. Ordovician, Bohemia. A complete but not fully-grown individual showing eyes. Natural size. (After Barrande.) C, Ampyx rouaulti, Barr. × 3. Ordovician, Bohemia. (After Barrande.)

In Harpes and in some species of Trinucleus eyes are present, but have been stated to be of a different type. They are described as simple eyes, and have been compared with ocelli; they are never found in Trilobites which possess the compound eyes described above. In Harpes (Fig. [150], A) the eye is near the middle of the cheek, in the position where compound eyes occur in other genera; it appears to consist of two or three granules or tubercles which are really lenses, and is connected with the front of the glabella by an eye-line. No facial suture can be seen, consequently the whole of the cheek is stated to be the fixed cheek.[^[187]] In Trinucleus (Fig. [140], B) a single tubercle is found on the middle of the cheek in the young of some species, and is sometimes connected with the glabella by an eye-line; the latter disappears before the adult state is reached, and in some species the tubercle also disappears, but in others (such as T. seticornis, T. bucklandi) it persists in the adult individuals.

From the lateral position of these eyes they can hardly be compared with the median simple eye of other Crustacea. In Harpes it is more probable that, as suggested by J. M. Clarke, they are schizochroal eyes imperfectly developed. Their structure (Fig. [139], G, H) is somewhat similar to that of schizochroal eyes, and moreover, in one species, H. macrocephalus,[^[188]] there are, in addition to the three main tubercles, other smaller tubercles in regular rows. Further, the eye-line occupies the same position as in other Trilobites which have undoubted compound eyes. The absence of a facial suture cannot be taken as evidence against these eyes being of the ordinary type, since in some species of Acidaspis (e.g. A. verneuili, A. vesiculosa) which possess compound eyes there is, in consequence of the coalescence of the fixed and free cheeks, no suture.

In some species of Trinucleus (Fig. [140], B) the simple eye is found in the same position as the eye in Harpes, and if, as some writers have maintained, there is evidence of the existence of a suture in that genus, then there is no reason for regarding the eye as other than a degenerate form of compound eye. The probability of its being such is supported by the existence of a compound eye in a similar position in the allied form Orometopus (Fig. [140], A) which possesses a facial suture.

In some species of Trinucleus (Fig. [140], B) and Ampyx there is a small median tubercle on the front part of the glabella, which from its position may be a simple unpaired eye, but its structure appears to be unknown.

Some Trilobites possess no eyes. Well-known examples of such are Agnostus, Microdiscus, Ampyx, Conocoryphe, and some species of Illaenus and Trinucleus; such blind Trilobites are almost confined to the Cambrian and Ordovician periods. All the forms of later periods, with the exception of a species of Ampyx, and possibly one or two other species, possess eyes. In addition to those undoubtedly blind forms Lindström considers that most of the Olenidae and Paradoxidae were without eyes. Many of the members of these families possess a lobe closely resembling a palpebral lobe, and a corresponding excavation in the free cheek; such forms have been generally regarded as possessing eyes; and the absence of any indication of lenses in those cases, on which Lindström lays stress, has been explained by the comparatively imperfect preservation of these early Trilobites. The development of the supposed eye-lobe in some of the Paradoxidae and Olenidae differs from that of the eyes in other families of Trilobites. In the latter the eye appears first at the margin of the head and always in connexion with the facial suture. But in Olenellus, in which there is said to be no facial suture, development shows that the crescentic eye-like lobe (Fig. [145], E) is really of the nature of a pleura coming off from the base of the first segment of the glabella. In Paradoxides, which resembles Olenellus in many respects, a facial suture is present and forms the outer boundary of the eye-like lobe, but it is developed subsequently to the appearance of the latter, which seems to be similar to that of Olenellus. In some genera of the Olenidae the eye-line, which comes off from the first segment of the glabella, ends in some cases in a swelling or knob which has hitherto been regarded as a palpebral lobe, but according to Lindström’s view no trace of an eye has been found in connexion with that lobe, nor is there any space between the lobe and the free cheek in which the eye could have occurred. If this view is correct it follows that the majority of the Cambrian Trilobites were blind. The earliest genus with eyes would then be Eurycare found in the Olenus beds of the Upper Cambrian. Sphaerophthalmus and Ctenopyge, found in the higher beds of the Cambrian, also possessed eyes, but Olenus and Parabolina were probably blind.

On the ventral surface of the head there is a flat rim around the margin (Fig. [137], B, b); this rim or “doublure” is the reflexed border of the cephalic shield. In many Trilobites its median part in front is cut off by sutures so as to form a separate plate (e); such is the case when the two facial sutures (c, c′) cut the anterior margin of the cephalic shield and are continued across the doublure, where they are joined by a transverse or rostral suture (d) just below the margin. When, however, as in Phacops and Remopleurides, the two facial sutures unite on the dorsal surface, in front of the glabella, the median part of the doublure is not separated from the lateral parts, or from the dorsal part of the cephalic shield.

Fig. [141].—A, Hypostome of Bronteus polyactin, Ang. showing maculae, × 4. B, Left macula of Bronteus irradians, Lindst. × 12. (After Lindström.)

The “labrum” or “hypostome” is attached to the doublure in front (Fig. [137], B, a); it is commonly an oval or shield-shaped plate, but is occasionally nearly square. Its surface is sometimes divided into two or three areas by shallow transverse grooves (Fig. [141], A), Just behind the middle of the hypostome, or when transverse grooves are present either in or near the anterior groove, there are often found a pair of small patches or “maculae” which are more or less oval or elliptical in outline (Fig. [141]). The maculae may be (1) surrounded by a raised border, or (2) in the form of pits, or (3) raised like tubercles. In some cases the entire surface of a macula is smooth and glossy; in others either the whole or a part is covered with granules, and in the latter case the granules may be limited to the internal third (Fig. [141], B) or to the central portion. Sections of a macula show that the granules are really globular lenses similar to those of the compound eyes on the dorsal surface of the head. Some of the maculae which are without lenses show no structure, but in others there is a spongy or irregularly polyhedric structure with prisms, resembling the marginal zone of the prismatic eyes of some genera. There seems no doubt that the maculae with lenses are visual organs, and those without are degenerate eyes. They occur in some genera which, according to Lindström, are without eyes on the dorsal surface. Maculae do not appear to be present in other Crustacea, but they have been compared with a median organ, found just in front of the hypostome in Branchipus.[^[189]] Maculae, have so far been found in 136 species of Trilobites belonging to 39 genera ranging from Lower Cambrian to Carboniferous.

A “metastoma” or lower lip plate (Fig. [142], Ep) is found just behind the hypostome in Triarthrus, but has not been noticed in any other genus. Between the hypostome and the metastoma lies the mouth.

The segments of the thorax are free, and their number varies from two in Agnostus (Fig. [146]) to twenty-six in Harpes (Fig. [150], A). In the Trilobites confined to the Cambrian period the number (except in the Agnostidae) is usually larger than in the genera found in the Ordovician and later periods. Owing to the free thoracic segments many Trilobites were able to curl up somewhat after the manner of a Wood-louse (Figs. [137], D, 138). The axial part of each thoracic segment is more or less considerably arched. Usually it consists of three parts: (i.) the largest part (Fig. [137], C, a), called the ring, is band-like in form, and is always visible whether the Trilobite is extended or coiled up; (ii.) in front of the ring is a depressed, groove-like part (Fig. [137], C, b) separating it from (iii.) the articular portion (c) which is convex in front and extends beneath the ring of the preceding segment; this part is only visible when the Trilobite is coiled up or when the segments are separated. In some few genera the axial part consists of a simple arched band without either a groove or a specially modified articular portion. The pleurae (Fig. [137], A, l, C, d-f) are fixed firmly to the axis, and have the form of narrow bands with the ends rounded, obtuse, pointed, or spinose. In a few cases the pleurae have a plain surface; but usually they possess either a ridge or a groove (Fig. [137], C, g); the former is generally parallel to the margins of the pleura, the latter is generally oblique, being inclined backwards from the axis. Sometimes in front of the ridge there is a small groove. On the ventral surface each pleura shows, at its outer extremity, a reflexed margin or doublure. At some distance from the axis the pleurae are bent downwards and backwards. The point where this bend occurs is called the “fulcrum” (e); it divides the pleura into an internal and an external part: the internal part (d-e) is flat or slightly convex, and just touches the front and back margins of the adjacent pleurae; the external part (e-f) may be (i.) narrower than the internal part, so that it is separated from the previous and succeeding pleurae; such occurs principally in pleurae with ridges, as in Cheirurus and Bronteus; or (ii.) it may be in the form of a long cylindrical process, as in many species of Acidaspis; or (iii.) the external part may be of the same width, either throughout or in part, as the internal part, and may overlap the next pleura behind; this type is found principally in pleurae with a groove such as in Phacops, Calymene, Sao, Asaphus, Ellipsocephalus.

In some Trilobites there is beyond the fulcrum a smooth, flat, triangular part at the front margin of the pleura; this part is known as the “facet,” and forms a surface articulating with the preceding segment which overlaps it.

In the remarkable form Deiphon (Fig. [151], E) the pleurae are separate throughout their entire length.

In some Trilobites broad and narrow forms of the same species occur—the difference being seen especially in the axis. The former are regarded as females, the latter as males.[^[190]]

The segments of the abdomen or pygidium (Fig. [137], A, 3) are similar to those of the thorax, except that they are fused together. In a few forms, such as Illaenus (Fig. [150], F) and Bumastus, the fusion is so complete that no trace of segmentation can be seen on the dorsal surface. Usually, however, the segments are easily distinguishable; the number seen on the axis is commonly greater than on the lateral parts of the pygidium; this difference is particularly well shown in Encrinurus. In Trilobites which have grooved pleurae the conspicuous grooves seen on the lateral parts of the pygidium are the grooves of the pleurae, the sutures between the pleurae being less distinct. The shape of the pygidium may be semicircular, a segment of a circle, trapezoidal, triangular, semi-parabolic, etc.; its size varies considerably; in the Cambrian forms it is usually small, but in the Trilobites of later periods it becomes relatively larger. The number of segments in the pygidium varies from two to twenty-eight. The axis of the pygidium tapers more rapidly than that of the thorax; sometimes it reaches quite to the posterior end of the body, but is commonly shorter than the pygidium; in Bronteus it is extremely short, and the grooves on the lateral parts of the pygidium radiate from it in a fan-like manner. Occasionally, as in Bumastus, the axis cannot be distinguished from the lateral parts. In a few early Trilobites (Olenellus, Holmia, Fig. [148], Paradoxides, Fig. [147]) the lateral parts of the pygidium are very small. In some genera, such as Asaphus, the marginal part of the pygidium forms a flattened or concave border. The margin may be entire or produced into spines, and sometimes (Fig. [151], C) a caudal spine comes off from the end of the axis. On the ventral surface of the pygidium there is a marginal rim similar to the doublure of the cephalic shield. The anus is on the ventral surface of the last segment of the pygidium.

Although Trilobites are often found in abundance and in an excellent state of preservation, it is only in very rare cases that anything is seen of the ventral surface except the hypostome and the reflexed borders of the cephalic shield, of the thoracic segments, and of the pygidium. The usual absence of appendages is probably due to their tenuity. Billings, in 1870, first obtained clear evidence of the presence of pairs of appendages, in Asaphus platycephalus. Soon afterwards Walcott[^[191]] showed their existence in American specimens of Asaphus megistos, Calymene senaria, and Cheirurus pleurexacanthus. In the two latter species the appendages were found by cutting sections of curled-up specimens obtained from the Trenton Limestone; 2200 examples were sliced, of which 270 showed evidence of the existence of appendages. They were seen to be present on the head, thorax, and pygidium; a ventral uncalcified cuticle with transverse arches was also found. By means of sections of curled-up specimens it was difficult to determine satisfactorily the form and position of the appendages. Subsequently extended specimens of Triarthrus (Fig. [142]) and Trinucleus, showing the ventral surface and appendages clearly, were discovered in the Utica Slate (Ordovician) near Rome, New York. A full account of the appendages in those specimens has been given by Beecher.[^[192]]

Fig. [142].—Triarthrus becki, Green, × 2½. Utica Slate (Ordovician), near Rome, New York. A, Ventral surface with appendages; Ep, metastome; Hy, hypostome. B, second thoracic appendage; en, endopodite; ex, exopodite, × 12. (After Beecher.)

In Triarthrus each segment, except the anal, bears a pair of appendages, all of which, except the first, are biramous. There are five pairs of cephalic appendages; the first pair are attached at each side of the hypostome, and have the structure of antennae, each consisting of a single flagellum formed of short conical joints. The other cephalic appendages increase in size successively. At present the second and third pairs are not satisfactorily known, but appear to have been similar to the fourth and fifth pairs. The second pair is attached at the level of the posterior end of the hypostome. The fourth and fifth pairs have large, triangular coxopodites which served as gnathobases, their inner edges being denticulate; the endopodites consist of stout joints; the exopodites are slender, and bear setae which are often arranged in a fan-like manner.

The first pair of appendages appear to be antennules, whilst the second pair probably represent the antennae, the third pair the mandibles, and the fourth and fifth pairs the maxillae of other Crustacea. The appendages of the thorax and pygidium do not differ essentially from the two posterior cephalic appendages. Those on the anterior part of the thorax are the longest; the others gradually decrease in size in passing posteriorly. Each thoracic leg (Fig. [142], B) consists of a short coxopodite with an inward cylindrical prolongation forming a gnathobase which is best developed on the anterior legs; the endopodite and exopodite are long and nearly equal; the former consists of six joints tapering gradually to the end; the latter, of a long proximal joint with a denticulate edge and a distal part of ten or more joints, and it bears a row of setae along the whole of the posterior edge.

The anterior appendages of the pygidium differ but little from the posterior thoracic legs; but the phyllopodous character, which appears in the latter, becomes more distinct in the appendages of the pygidium, especially those near its posterior end, and is due to the broad, flat, laminar joints of the endopodite.

The more striking features of the appendages of Triarthrus are the small amount of differentiation which they show in different parts of the body, especially the want of specialisation in the cephalic region; the distinctly biramous character of all except the first pair; and the presence of one pair of functional antennae only, and the occurrence of thoracic gnathobases.

In Trinucleus the appendages are not so well known, but they are considerably shorter than in Triarthrus.

In the Palaeozoic rocks of Bohemia, where Trilobites are very perfectly preserved, Barrande[^[193]] discovered the larval forms of several species, and in some cases was able to trace out the development very completely, but in others the earliest stages were not found. In the strata in which Trilobites occur Barrande found minute spheroidal bodies, usually of a black colour, and only a little smaller than the youngest larval stages; those bodies are probably the eggs of Trilobites. Since the publication of Barrande’s work the development of some species found in North America has been studied by Ford, Matthew, Walcott, and Beecher. But even now the development is known in only a very small proportion of the total number of genera of Trilobites. The early larval form (Fig. [143], A) is similar in general character in the various species in which it has been found. It is circular or ovoid in outline, with a length of from 0·4 to 1 mm., and consists of a large cephalic and a small pygidial portion; the axis is distinct and usually shows more or less clear indications of five cephalic segments; the eyes, when present, are found at or near the front margin, and the free cheeks, if visible at all on the dorsal surface, are narrow. For this early larval form Beecher has proposed the name “protaspis”; he regards it as the representative of the Nauplius of other Crustacea, but that view is not accepted by Professor J. S. Kingsley.[^[194]]

The general changes which occur in the course of development are: modifications in the shape and relative size of the glabella, and of the number and depth of the glabella-furrows; the growth of the free cheeks and the consequent inward movement of the facial sutures and eyes; the introduction of and gradual increase in number of the thoracic segments, and the relative decrease in size of the head.

Fig. [143].—Development of Sao hirsuta, Barr. Cambrian. A, Protaspis; B-F, later stages; G, adult. The small outlines below each figure show the actual size of each specimen. (After Barrande.)

Sao hirsuta is a species found in the Cambrian, the development of which was fully described by Barrande. Its earliest protaspis stage (Fig. [143], A) is circular in outline; the glabella expands in front and reaches the anterior margin; the pygidial region is not distinctly separated from the cephalic region; segmentation is indicated in the former, and the neck-ring is present in the latter; the eye-line is seen on each side of the glabella near the anterior margin. In a later stage (Fig. [143], C) the segmentation of the glabella becomes more distinct, indicating the existence of five cephalic segments, and the facial suture appears near the margin limiting a very narrow free cheek. Subsequently (Fig. [143], D-F) the thoracic segments develop, and increase in number until the adult stage (G) is reached; also the eyes appear at the margin of the cephalic shield, and gradually move inwards, and the glabella becomes narrower and rounded in front, and ceases to reach the anterior margin. In this species the eye-line is present in the adult.

Fig. [144].—Triarthrus becki, Green. Ordovician. A, B, Two successive stages of the protaspis, × 45. (After Beecher.)

In the protaspis of Triarthrus (Fig. [144]), found in the Ordovician, the glabella does not reach the front margin nor expand in front as it does in Sao; an eye-line is present, but disappears before the adult stage is reached.

Fig. [145].—Larval stages of Trilobites. A-D, Dalmanites socialis, Barr. Ordovician, Bohemia. The small figures below show the natural size of each specimen. (After Barrande.) E, Mesonacis asaphoides, Emmons, × 10. Lower Cambrian, North America. (After Walcott.) F, Acidaspis tuberculata, Conrad, × 20. Lower Helderberg Group (Lower Devonian or Upper Silurian), Albany County. (After Beecher.)

Dalmanites (Fig. [151], C) is a more advanced type than Sao and Triarthrus, and is found in later deposits. In the earliest stage (Fig. [145], A) the head and pygidium are quite distinct, and there is no eye-line present at this or any stage in development, but large ovoid eyes are found on the front margin, and have their long axes placed transversely to the axis of the body; the glabella is strongly segmented and is rounded in front. In later stages (C, D) the pygidium increases in size relatively, and the thoracic segments are successively introduced; the facial sutures and free cheeks appear on the dorsal surface, and as the free cheeks grow the eyes move inwards and backwards, and gradually swing round until their long axes become parallel with the axis of the body.

The larval form of Acidaspis (Fig. [145], F) is of interest since even in the earliest stage it shows the spiny character which forms such a striking feature of the adult (Fig. [151], F).

Before the discovery of the ventral surface of Trilobites it was thought by some zoologists that their affinities were with the Xiphosura rather than with the Crustacea. But the presence of antennae, and of five pairs of cephalic appendages; the biramous thoracic and pygidial appendages, the hypostome, and the character of the larval form, as well as the absence of a genital operculum, separate the Trilobites from the Xiphosura and connect them with the Crustacea.

The position of the Trilobites in the Crustacea is, however, difficult to determine. Already in the Cambrian period, at least five main groups of the Crustacea were clearly differentiated, namely, the Phyllopoda, Ostracoda, Cirripedia, Trilobita, and Leptostraca (Phyllocarida), and probably also the Copepoda, but of the last no remains have been preserved as fossils. Palaeontology, therefore, furnishes no connecting links between any two of these orders.

The Crustacea to which the Trilobites show some resemblance are the families Apodidae and Branchipodidae of the Order Phyllopoda (see pp. [19]–36). The Trilobita agree with those families in having a large but variable number of trunk-segments, in the possession of a large labrum (hypostome), and in the occurrence of gnathobases on the thoracic appendages; also the foliation of some of the trunk-appendages is somewhat similar. The points of difference, however, are considerable; thus the cephalic appendages are much more specialised in the Apodidae and Branchipodidae than in the Trilobita; in the latter all, with the exception of the antennae, are distinctly biramous, and whilst the basal joints were masticatory the distal parts appear to have been locomotor organs. The appendages of the trunk also differ considerably; in the Trilobita all are clearly biramous, those of the thorax having a schizopodal form. In the possession of a single pair of antennae the Trilobita differ from other Crustacea; but in some forms of Apus the second pair of antennae may be rudimentary or even absent.

There are still other features which characterise the Trilobita: thus the eyes are borne on free cheeks, and differ in structure from those of Phyllopods. The broad pygidium formed of fused segments and without terminal fulcra is quite unlike the slender-jointed abdomen of Apus and Branchipus; and whilst in the Trilobites all the segments bear appendages, in the Phyllopods some, at any rate, of the posterior segments are devoid of appendages. The distinct division of the body into an axial and pleural region is not seen in Phyllopods, and is probably a character of some importance, since it occurs in the great majority of Trilobites, including all the early forms.

The existence of some relationship between the Trilobita and the Leptostraca (Phyllocarida) has been maintained by Professor G. H. Carpenter.[^[195]] He points out that some of the earliest Trilobites, such as Holmia kjerulfi (Fig. [148]), possess nearly the same number of segments as Nebalia (Fig. [76], p. 111), and that in the latter genus the cephalic appendages, especially the mandibles and maxillae, are less specialised than in Apus, and consequently differ less from those of Trilobites than do the appendages of the Apodidae. Further, in another genus of the Leptostraca, Paranebalia, the biramous thoracic legs, in which both endopodite and exopodite are elongate, approach those of Trilobites more nearly than do the thoracic legs of Apus.

The view[^[196]] that some connexion may exist between the Isopoda and the Trilobita seems to have been based on the similar dorso-ventral flattening of the body, its division into three regions—head, thorax, and abdomen—and the presence of sessile eyes. Beyond this it is difficult to find any resemblance; whilst the differences, such as the variable number of thoracic segments and their biramous appendages in Trilobites, are important.

At present, then, we can only conclude that the Trilobita are more primitive than any other Crustacea, and that their resemblance to some of the Phyllopoda is sufficient to make it probable that they had some ancestral connexion;[^[197]] the possibility of such a relationship receives some support from the presence in the Lower Cambrian rocks of Protocaris, a genus of the Phyllopoda which resembles Apus.[^[198]] The primitive characters of Trilobites are the variable and often large number of segments in the thorax and pygidium; the presence of a pair of appendages on every segment except the anal; the biramous form of all except the first pair of appendages; and the lack of specialisation shown by the appendages, especially those of the head.

The classification of Trilobites is due largely to the work of Barrande and Salter, and the families defined by those authors have been, in the main, generally adopted. But the phylogenetic relationship of the families has still, to a large extent, to be established. Salter[^[199]] arranged the families in four groups, but did not claim that that classification was entirely natural. His groups with the families included in each are:—

1. Agnostini. Without eyes or facial suture. Agnostidae.

2. Ampycini. Facial sutures obscure, or submarginal, or absent. Eyes often absent. Trinucleidae.

3. Asaphini. Facial sutures ending on the posterior margin. Acidaspidae, Lichadidae, Harpedidae, Calymenidae, Paradoxidae, Conocephalidae, Olenidae, Asaphidae, Bronteidae, and Proëtidae.

4. Phacopini. Facial sutures ending on the lateral margins. Eyes well developed. Phacopidae, Cheiruridae, and Encrinuridae.

A modification of Salter’s classification has been brought forward by Beecher[^[200]] who divides the Trilobita into three main groups:—

1. Hypoparia. Facial sutures at or near the margin, or ventral. Compound eyes absent. This is equivalent to Salter’s Agnostini and Ampycini with the addition of the Harpedidae.

2. Opisthoparia. Facial sutures extending from the posterior margin to the front margin, but occasionally uniting in front of the glabella. Eyes holochroal or prismatic, but sometimes absent. This comprises the same families as Salter’s Asaphini with the exclusion of the Harpedidae and Calymenidae.

3. Proparia. Facial sutures extending from the lateral margins, and either cutting the anterior margin or uniting in front of the glabella. Eyes holochroal or schizochroal; occasionally absent. This is equivalent to Salter’s Phacopini with the addition of the Calymenidae.

In each of the groups proposed Beecher regards as the more primitive forms those which possess characters similar to those of the early larval stages, such as narrow free cheeks, the absence of compound eyes, and a glabella which is broad in front and reaches the anterior margin of the head.

The modifications introduced by Beecher can scarcely be regarded as making Salter’s classification more natural. For instance, the Agnostidae differ so much from all other families that, at present, there is no evidence to show that they have any close phylogenetic relationship with the Trinucleidae and Harpedidae. Further, the Calymenidae, which Salter recognised as related to the Olenidae, have been shown by the careful work of Professor Pompeckj[^[201]] to have descended from the latter family, and to have no genetic connexion with the Phacopidae with which they are grouped by Beecher. Then in the Trinucleidae the earliest genus, Orometopus[^[202]] (Fig. [140], A), possesses compound eyes and facial sutures which begin at the posterior margin and unite in front of the glabella; so that, according to Beecher’s classification, that genus would belong to the Opisthoparia, whereas the later genera (Trinucleus, etc.) of the same family would be placed in the Hypoparia. At present, therefore, the only classification of Trilobites which can be adopted is a division into families, of which a short account is given below.

Fig. [146].—Agnostus integer, Beyr., × 8. Cambrian. (After Barrande.)

Fam. 1. Agnostidae (Fig. [146]).—Small Trilobites, in which the head and pygidium are of nearly the same size and shape. The thorax is shorter than the head or pygidium, and consists of from two to four segments with grooved pleurae. The length and width of the head are commonly nearly equal, but sometimes the length is greater. Eyes are absent. Facial sutures appear to be absent, but are stated by Beecher to be at the margin of the cephalic shield. From the absence of eyes, the probable absence of facial sutures, the few or indistinct furrows on the glabella, and the smaller number of thoracic segments, the Agnostidae appear to be degenerate forms. Microdiscus is apparently less modified than Agnostus, on account of the larger number of thoracic segments, the more distinct segmentation of the pygidium, and, in some species, the larger number of furrows on the glabella. Cambrian and Ordovician. Genera: Agnostus, Microdiscus.

Fam. 2. Shumardiidae.—The body is very small and oval. The cephalic shield is nearly semicircular and very convex, with a broad glabella which expands in front, and in which the furrows, except the neck-furrow, are indistinct. The facial suture is marginal and eyes are absent. There are six thoracic segments with ridged pleurae; the axis is broader than the pleurae. The pygidium is large, and is formed of about four segments similar to those of the thorax. Upper Cambrian and Ordovician. Genus: Shumardia.

Fam. 3. Trinucleidae (Fig. [140]).—The head is large and has a flat border (except in Ampyx), and long genal spines. In the earliest genus (Orometopus) the facial sutures start from the posterior margin (near the genal angle) and pass obliquely inwards to the compound eye, from whence they continue forward and unite in front of the glabella. In Ampyx the suture starts from just within the genal angle and passes to the front border, cutting off a narrow free cheek; eyes are absent. In most specimens of Trinucleus no sutures are seen, but some examples show indications of what may be a facial suture (see p. [226]), and a suture is sometimes found at the margin of the cephalic border; eyes may occur (see p. [230]). The thorax consists of from five to eight segments, with grooved pleurae. The pygidium is triangular. Principally Ordovician. Genera: Orometopus (Upper Cambrian), Ampyx, Trinucleus, Dionide.

Fam. 4. Harpedidae (Figs. [139], G, H; 150, A).—The head is large and has a broad, flat border which is finely punctate, and extends backwards on each side in the form of a horn-like projection nearly as far as the posterior end of the thorax. The glabella is convex and does not reach the front margin. The cheeks are less convex than the glabella, and bear eyes which usually consist of two or three lenses. An eye-line connects the eye with the anterior part of the glabella. A suture is stated to occur at the external margin of the flat border. The thorax consists of from twenty-five to twenty-nine segments; its axis is narrow, and the pleurae are long and grooved. The pygidium is very small, and consists of three or four segments. Ordovician to Devonian. Genus: Harpes.

Fig. [147].—Paradoxides bohemicus, Barr. × ½. Middle Cambrian. (After Zittel.)

Fig. [148].—Holmia kjerulfi, Linnars. × 1. Lower Cambrian. (After Holm.)

Fam. 5. Paradoxidae (Figs. [147], 148, 149).—The cephalic shield is large, and bears long genal spines. The glabella is more or less swollen in front. The facial sutures appear to be absent in some genera, and when present extend from the posterior to the anterior margin. The palpebral lobes are long, and often more or less semicircular or kidney-shaped. The thorax is long, and consists of from sixteen to twenty segments with their pleurae produced into spines. The pygidium is very small, and plate-like, or sometimes in the form of a long spine. Cambrian. Genera: Olenellus, Holmia, Mesonacis, Olenelloides, Paradoxides, Zacanthoides, Centropleura (Anopolenus). Remopleurides (Fig. [150], D) from the Ordovician is usually included in the Paradoxidae, but probably belongs to a separate family.

Fig. [149].—Clenelloides armatus, Peach. Lower Cambrian, × 3. (After Peach.)

Fam. 6. Conocephalidae (Conocoryphidae) (Fig. [150], E).—The cephalic shield is semicircular, and larger than the pygidium. The glabella narrows in front. The facial suture passes from near the genal angle on the posterior border to the antero-lateral margin, and limits a large fixed cheek and a narrow free cheek. Eyes are absent or rudimentary, but an eye-line is usually present. The thorax consists of from fourteen to seventeen segments with grooved pleurae, which may be pointed, but are not usually produced into spines. The pygidium is small, and formed of few segments. Cambrian. Genera: Conocoryphe, Atops, Ctenocephalus, Bathynotus.

Fam. 7. Olenidae (Figs. [142], 143; 150, B, C).—The cephalic shield is larger than the pygidium. The glabella is either rectangular or parabolic. The facial suture passes from the posterior to the anterior margin. The palpebral lobes are of moderate or rather large size, and are connected by an eye-line with the front part of the glabella. The thorax includes from eleven (occasionally fewer) to eighteen segments with grooved pleurae. The pygidium is usually small, with from two to eight segments. Principally Cambrian. Genera: Ptychoparia, Angelina, Solenopleura, Sao, Agraulos (Arionellus), Ellipsocephalus, Protolenus, Olenus, Peltura, Acerocare, Eurycare, Ctenopyge, Leptoplastus, Triarthrus, Parabolina, Sphaerophthalmus, Parabolinella, Ceratopyge (position doubtful). Dikelocephalus is usually placed in the Olenidae, but perhaps belongs to a distinct family.

Fam. 8. Calymenidae (Figs. [136], 137).—The glabella is broadest behind. The facial suture starts at or near the genal angle—sometimes on the posterior border just inside the angle, sometimes on the lateral border just in front of the angle; the suture may be continuous with the other suture in front of the glabella, or may cut the anterior margin, beneath which it is connected with the other suture by means of a transverse suture (Fig. [137], B, D). The eyes are rather small. The thorax consists of thirteen segments with grooved pleurae; the pygidium of from six to fourteen segments. Ordovician to Devonian. Genera: Calymene, Synhomalonotus, Homalonotus.

Fig. [150].—A, Harpes ungula, Sternb., Ordovician. B, Ellipsocephalus hoffi, Scloth., Cambrian. C, Olenus truncatus, Brünn., Cambrian. (After Angelin.) D, Remopleurides radians, Barr., Ordovician. E, Conocoryphe sulzeri, Barr., Cambrian. F, Illaenus dalmanni, Volb., Ordovician. G, Proëtus bohemicus, Corda, Silurian, × 1½. H, Aeglina prisca, Barr., Ordovician, × 3. I, Phacops sternbergi, Barr., Devonian. (A, D, E, G, H, I, after Barrande; B, F, from Zittel; natural size except G, H.)

Fam. 9. Asaphidae (Fig. [150], F).—The body is oval and commonly rather large. The cephalic shield is large, with its glabella often indistinctly limited and the glabella-furrows often obscure. The facial suture starts from the posterior margin and usually cuts the anterior margin, but is sometimes continued in front of the glabella. The relative size of the fixed and free cheeks varies greatly. The eyes are of variable size. The thorax consists of eight or ten (sometimes fewer) segments; the pleurae are generally grooved, but sometimes plane. The pygidium is large, often being similar in form and size to the head; it consists of numerous segments which, however, may be indistinctly shown; the axis in some forms is obsolete. Upper Cambrian (Tremadoc) to Silurian; common in the Ordovician. Genera: Asaphus (sub-genera, Megalaspis, Asaphellus, Symphysurus, etc.), Ogygia, Barrandia, Niobe, Nileus, Illaenus, Bumastus, Stygina. Aeglina (Fig. [150], H) is usually placed in this family, but its systematic position is doubtful.

Fam. 10. Bronteidae.—The general form is similar to that of the Asaphidae. The glabella broadens rapidly in front, and is marked with furrows on each side, which are usually short, and may be indistinct. The facial suture passes from the posterior margin to the crescentic eye which is situated rather near the posterior border, and from thence to the anterior margin. There are ten thoracic segments with ridged pleurae. The pygidium is longer than the head, and has a very short axis, from which the furrows on the pleural part radiate. Ordovician to Devonian. Genus: Bronteus.

Fam. 11. Phacopidae (Figs. [138]; 150, I; 151, C).—The head and pygidium are of about the same size. The glabella is distinctly limited, and wider in front than behind, with a neck-furrow and three other furrows, of which some of the anterior may be indistinct or obsolete. The eyes are schizochroal and usually large. The facial suture begins at the lateral margin and unites with the suture of the other side in front of the glabella. There are eleven thoracic segments with grooved pleurae. The pygidium is usually large, with a distinct axis and many segments. Ordovician to Devonian. Genera: Phacops, Trimerocephalus, Acaste, Pterygometopus, Chasmops, Dalmanites, Cryphaeus.

Fig. [151].—A, Phillipsia gemmulifera, Phill., Carboniferous. B, Arethusina konincki, Barr., Ordovician. C, Dalmanites limulurus, Green, Silurian. (After Hall.) D, Cheirurus insignis, Beyr., Silurian. E, Deiphon forbesi, Barr., Silurian. F, Acidaspis dufrenoyi, Barr., Silurian. (A, B, from Zittel; D, E, F, after Barrande; natural size.)

Fam. 12. Cheiruridae (Fig. [151], D, E).—The glabella is convex or inflated, and distinctly defined. The facial suture passes from the lateral to the front margin. The free cheeks are small, and the eyes usually rather small. There are from nine to eighteen (usually eleven) thoracic segments; the pleurae have ridges or grooves and free ends. The pygidium is small, consisting of from three to five segments often produced into spines. Upper Cambrian to Devonian. Genera: Cheirurus, Deiphon, Placoparia, Sphaerexochus, Amphion, Staurocephalus.

Fam. 13. Proëtidae (Figs. [150], G; 151, A, B).—The body is rather small, and the head forms about a third of its entire length. The glabella is sharply defined, and its furrows are sometimes indistinct; the posterior furrow curves backward to the neck-furrow, thus limiting a basal lobe on each side of the glabella. The eyes are often large (Fig. [150], G); but in Arethusina (Fig. [151], B), in which an eye-line is present, they are small. The facial sutures pass from the posterior to the anterior margin. The free cheeks are large. There are from eight to twenty-two thoracic segments with grooved pleurae. The pygidium is usually formed of numerous segments, and its margin is usually entire. Ordovician to Permian. Genera: Proëtus, Arethusina, Cyphaspis, Phillipsia, Griffithides, Brachymetopus, Dechenella.[^[203]]

Fam. 14. Encrinuridae.—The cephalic shield is ornamented with tubercles. The free cheeks are narrow, and the eyes very small. The facial suture extends from the lateral margin (or from the genal angle) to the anterior margin. There are from ten to twelve thoracic segments with ridged pleurae. On the axis of the pygidium numerous segments are seen, but usually fewer are indicated on the lateral parts. Ordovician and Silurian. Genera: Encrinurus, Cybele, Dindymene.

Fam. 15. Acidaspidae (Fig. [151], F).—The cephalic shield is broad, with a spinose margin, genal spines, and sometimes spines on the neck-ring. The glabella has a longitudinal furrow on each side, due to the backward bending of the lateral furrows. The facial suture passes from the posterior border (near the genal angle) to the anterior border. The free cheeks are large; the eyes small. There are from eight to ten thoracic segments with ridged pleurae, which are produced into long backwardly directed spines. The pygidium is short, and is formed of two or three segments with long spines at the margin. Ordovician to Devonian. Genus: Acidaspis.

Fam. 16. Lichadidae.—The body is broad, with a granular dorsal surface. The cephalic shield is small and short, with spinose genal angles. The glabella is broad, and its anterior furrows are directed backwards, limiting a convex median lobe and some lateral lobes. The facial suture extends from the posterior to the anterior margin. There are nine or ten thoracic segments with grooved pleurae, which have pointed ends. The pygidium is large and triangular, with a short axis and a toothed margin. Ordovician to Devonian. Genus: Lichas (sub-genera, Arges, Dicranogmus, Conolichas, Ceratolichas).

INTRODUCTION TO ARACHNIDA,
AND
XIPHOSURA

BY

A. E. SHIPLEY, M.A., F.R.S.

Fellow of Christ’s College, Cambridge, and Reader in Zoology in the University

CHAPTER IX
ARACHNIDA—INTRODUCTION

The Arachnida, together with the Crustacea, Insecta, Myriapoda, and Peripatus, make up the great phylum Arthropoda, a phylum which, from the point of view of numbers of species and of individuals, is the dominant one on this planet, and from the point of view of intelligence and power of co-operating in the formation of social communities is surpassed but by the Vertebrata. The Arachnida form a more diverse class than the Insecta; they differ, perhaps, as much inter se as do the Crustacea, and in structure as in size and habit they cover a wide range.

Lankester in his article upon the Arthropoda, in the tenth edition of the Encyclopaedia Britannica, dwells upon the fact that whereas the adult Peripatus has but one persisting segment in front of the head, and its mouth is between the second persisting appendages, in Arachnids the mouth has receded and lies between the bases of the appendages (pedipalpi) of the third persisting segment, while two of the persisting segments, those of the eyes and chelicerae, have passed in front of the mouth. This process has continued in the Crustacea and in the Insecta; in both of these classes there are three embryonic segments in front of the adult mouth, which lies between the appendages of the fourth segment.

In the larger and more complex Arachnida the number of segments is fixed and constant, and though possibly no adult member of the group, owing to the suppression of one or more segments during the ontogeny, ever shows the full number at any one time, the body can be analysed into twenty-one segments. It is interesting to note that the same number of segments occurs in Insecta and in the higher Crustacea.[^[204]] The significance of this fact is not perhaps apparent, but it seems to indicate “a sort of general oneness, if I may be allowed to use so strong an expression,” as Mr. Curdle said when discussing the unities of the drama with Nicholas Nickleby.

These segments are arranged in higher categories or “tagmata,” of which we can recognise three: (i.) the prosoma, (ii.) the mesosoma, and (iii.) the metasoma. The prosoma, sometimes termed the “cephalothorax,” includes all the segments in front of the genital pore. According to this definition the prosoma includes the segment which bears the chilaria in Limulus (the King-crab) and the pregenital but evanescent segment in Scorpions. The mesosoma begins with the segment bearing the genital pore, and ends with the last segment which bears free appendages, six segments in all. The metasoma also consists of six segments which have no appendages; together with the mesosoma it forms the abdomen of some writers. The anus lies posteriorly on the last segment, and behind it comes in the higher forms a post-anal “telson,” taking in Scorpions the form of the sting, in King-crabs that of the spine.

As we have seen, it is only in the more typical and perhaps higher forms that we can find our twenty-one segments, and then they are never present all at once. In many groups of Arachnids the number is reduced at the hinder end, and obscured by the fusion of neighbouring segments. Also segments are dropped as a stitch is dropped when knitting; for instance, in the rostral segment which has a neuromere, and in the Spider Trochosa vestigial antennae, or in Scorpions the pregenital segment.

Primitive Arachnids appear to have lived in the sea and to have breathed by gill-books borne on appendages; when their descendants took to living on land and to breathing air instead of water, the gill-books sank into the body and became lung-books, to which the air was admitted by slit-like stigmata. In other terrestrial forms the lung-books are replaced by tracheae which in their structure and arrangement resemble those of Peripatus rather than those of the Insecta. The circulation, as is usual in Arthropods, is largely lacunar, but in Scorpions and Limulus the arteries form definite channels, and are in fact better developed than in any other Arthropod.

As a rule the alimentary canal in Arachnids is no longer than the distance between the mouth and the anus; but in the King-crab, where the mouth is pushed back almost to the centre of the body, there is a flexure in the median vertical plane. Paired glands, usually called the liver, open into the mesenteron; food passes into the lumen of these glands, and is probably digested there. In many Arachnids these glands extend into the limbs. In those members of the group that have become terrestrial the nitrogenous excreta are separated out by Malpighian tubules which open into the proctodaeum; but coxal glands, homologous with the green gland and shell-glands of Crustacea, may coexist, and in the aquatic Limulus these alone are found. They usually open on the base of one or more pairs of walking legs.

The endosternite, or internal skeletal plate to which muscles are attached, reaches a higher development in the Arachnida than in the Crustacea. In Scorpions it forms a kind of diaphragm incompletely separating the cavities of the pro- and meso-soma.

The supra-oesophageal ganglion supplies the two existing segments which have slipped before the mouth, i.e. those of the eyes and of the chelicerae. The post-oral ganglia in the Acarina, the Pedipalpi, the Solifugae, and the Araneae have fused into a central nerve-mass, from which nerves radiate; but in Limulus the prosomatic appendages are all supplied from the nerve-ring. The chief sense-organs are eyes of the characteristic Arthropod type, and sensory hairs of a great variety of complexity. Scorpions and Spiders have stridulating organs, and we may assume that they have also some auditory apparatus; perhaps some of the hairs just mentioned act as hearing organs.

Arachnids are male and female; they do not reproduce asexually, and there is no satisfactory proof that they ever reproduce parthenogenetically. As a rule there is little external difference between the two sexes, except in Spiders, where the male is as a rule smaller than the female, and when adult has the pedipalpi modified for use in depositing the spermatophores. The ovaries and testes are annular, and with their ducts encircle the alimentary canal in Mites and Phalangids; in Scorpions and King-crabs they have become retiform. Mites, Scorpions, and Pedipalps are viviparous, their eggs developing in the ovary or in a uterus. Other Arachnids lay eggs, and many Spiders and Pseudoscorpions carry their eggs about with them. As a rule the young is but a miniature of the parent, and the Arachnid, unlike the Crustacean or Insect, undergoes little or no metamorphosis.

A certain number of Mites are parasitic in plants and in animals, and a few, together with a few Spiders, have resumed the aquatic life of their remote ancestors. The members of some Orders, such as the Solifugae and Opiliones, are nocturnal, and many are provided with silk-glands and weave webs which reach their highest pitch of perfection amongst the Spiders. At times—especially is this the case with the Mites—enormous numbers of individuals live together, but they never show the least adaptation to communal life, and no individuals are ever specialised to perform certain functions, as is the rule in communities of social Insects.

With the two exceptions that we regard the Trilobites as more nearly allied to the Crustacea, and have therefore considered them apart, and have treated the Pycnogonida independently of the Arachnida, we have followed Lankester in his classification, though not always in his nomenclature:—

Sub-class 1. Delobranchiata[^[205]] (Merostomata).

Order (i.) Xiphosura.

Order (ii.) Eurypterida (= Gigantostraca, Extinct).

Sub-class 2. Embolobranchiata.

Order (i.) Scorpionidea.

Order (ii.) Pedipalpi.

Order (iii.) Araneae.

Order (iv.) Palpigradi.

Order (v.) Solifugae.

Order (vi.) Chernetidea (= Pseudoscorpiones).

Order (vii.) Podogona.

Order (viii.) Phalangidea (= Opiliones).

Order (ix.) Acarina.

Appendices

(i.) Tardigrada.

(ii.) Pentastomida.

CHAPTER X
ARACHNIDA (CONTINUED)—DELOBRANCHIATA = MEROSTOMATA—XIPHOSURA

SUB-CLASS I.—DELOBRANCHIATA = MEROSTOMATA.

Order I. Xiphosura.[^[206]]

In his recent classification of the Arachnida, Lankester[^[207]] has grouped the Xiphosura or King-crabs with the extinct Eurypterids or Gigantostraca under the name of Delobranchiata, better known under the name Merostomata[^[208]] of Dana. The chief character of this group, and one which differentiates it from all the animals placed together by Lankester in the group Embolobranchiata, is that they have gills patent and exposed. The Xiphosura are, in fact, with the exception of a few marine Mites, the only Arachnids which now live in the sea as did their allies the Eurypterids in Palaeozoic times. With a few fresh-water exceptions, all other Arachnids have taken to life on land, and with a change from water-breathing to air-breathing came a change in the respiratory system, the gills becoming “lung-books,” or possibly tracheae, or disappearing altogether.

A few years ago Pocock re-classified the Xiphosura, and his classification will be found on pp. 276, 277. It will be noticed that in his classification the generic name Limulus has disappeared. I have, however, retained it in this article, partly because I regard the name as so well established that every one knows what it denotes, and partly because in a group which contains confessedly very few species, differing inter se comparatively slightly, it seems unnecessary to complicate matters with sub-families and new names.

Looked at from above a Limulus presents a horse-shoe-like outline, from the posterior end of which projects a long spine. It is often called in America the Horsefoot-crab, but its common or vulgar name is the King-crab. Across the middle of the body is a joint, and this joint separates the prosoma from the meso- and metasoma which are in King-crabs fused together. The prosoma comprises all the segments up to and including the segment which carries the chilaria;[^[209]] the mesosoma begins with the segment bearing the genital pores, and ends with the last segment which bears appendages; the metasoma comprises all the segments posterior to the last segment which carries appendages. The prosoma corresponds with the “cephalothorax” of some authors, and the meso-plus the metasoma are equivalent to their “abdomen.”

Dorsally, then, the prosoma is a vaulted structure with a smooth, horse-shoe-shaped anterior and lateral margin. Its posterior edge, the line where the meso-plus the metasoma are hinged, is a re-entrant bay with three sides. The meso- and metasoma are in the King-crabs fused together and form a hexagon. Three sides of this hexagonal double region form the hinge, two form the lateral margins and slope inwards; these bear six fused and six-jointed spines which have a segmental value. The sixth or posterior side is indented, and its concavity forms the area to which the large post-anal, unsegmented tail or spine is hinged.

The whole body is covered by a smooth chitinous sheath varying from sage-green to black in colour, and it is kept very clean, probably by some excretion which hinders various sessile animals attaching themselves to it as they do, for instance, on many Copepods. Burrowing animals like Limulus are usually free from these messmates. King-crabs have a self-respecting, well-groomed appearance. On the rounded dorsal surface the chitinous covering is produced into a certain number of spines arranged in a median and two lateral rows. The anterior median spine overhangs the median eyes, and the anterior lateral spine on each side overshadows the large lateral eyes.

Fig. [152].—Dorsal view of the King-crab, Limulus polyphemus, × ½. From Shipley and MacBride. 1, Carapace covering prosoma; 2, meso- and metasoma; 3, telson; 4, median eye; 5, lateral eye.

The vaulted carapace is turned in on the under side, where there is a flat rim which widens anteriorly, and on the inner edge this rim borders a sunken area, into the concavity of which the numerous appendages project. Thus, although when viewed from above a Limulus looks as though it had a solid body shaped something like half a pear, when viewed from below, especially if the appendages be removed, it will be found that the body is thin and hollowed, and almost leaf-like, as if most of the edible part of the half-pear had been scooped out. Within the hollow thus formed the appendages lie, and here they move about, seldom or never protruding beyond the edge of the carapace,—in fact, all except the pedipalps and ambulatory legs are too short to project beyond this limit.

Fig. [153].—Ventral view of the King-crab, Limulus polyphemus, × ½. From Shipley and MacBride. 1, Carapace covering prosoma; 2, meso- and metasoma; 3, telson; 4, chelicera; 5, pedipalp; 6, 7, 8, 9, 3rd to 6th appendages, ambulatory limbs; 10, genital operculum turned forward to show the genital apertures; 11, 12, 13, 14, 15, appendages bearing gill-books; 16, anus; 17, mouth; 18, chilaria.

The body of a King-crab can be analysed into twenty-one segments, but these do not all persist to the adult stage. They are grouped together in higher aggregates, or “tagmata” as Lankester calls them, and most of the segments bear paired appendages.

The segments with their respective appendages and their grouping into tagmata are shown in the following scheme:—

Appendages.
I.	Segment	Median eyes	Preoral	Prosoma
II.	„	Rostrum	„
III.	„	Chelicerae	„
IV.	„	Pedipalpi	Lateral to mouth
V.	„	1st Walking Legs	Postoral
VI.	„	2nd Walking Legs	„
VII.	„	3rd Walking Legs	„
VIII.	„	4th Walking Legs	„
IX.	„	Chilaria	„
X.	„	Genital operculum	„	Mesosoma
XI.	„	1st Gill-books	„
XII.	„	2nd Gill-books	„
XIII.	„	3rd Gill-books	„
XIV.	„	4th Gill-books	„
XV.	„	5th Gill-books	„
XVI.	„	No appendages	„	Metasoma
XVII.	„	„	„
XVIII.	„	„	„
XIX.	„	„	„
XX.	„	„	„
XXI.	„	„	„

We have followed Carpenter[^[210]] in inserting the rostral segment. This corresponds with the segment that in Insects and Crustacea bears the antennae or first antennae respectively, the absence of these organs being one of the characteristic but negative features of all Arachnids. The evidence for the existence of this evanescent segment rests partly upon the observation of von Jaworowski[^[211]] on the vestigial feelers in an embryo Spider, Trochosa, and perhaps more securely on the fact that, according to Korschelt and Heider, there is a distinct neuromere for this segment, between the proto-cerebral neuromere which supplies the eyes and the trito-cerebral neuromere which supplies the chelicerae. According to Brauer[^[212]] the chelicerae of Scorpions are also supplied by the third neuromere.

The bases of the chelicerae do not limit the mouth, but between and behind them is a ridge or tubercle which has the same relationship to the mouth of Limulus that the labrum has in Insects and some Crustacea. Posteriorly the mouth is bounded by the “promesosternite,” a large median plate which lies between the bases of the ambulatory limbs. The pedipalps and all the ambulatory limbs have their bases directed towards the mouth, their gnathobases or sterno-coxal processes are cushion-like structures covered with spines—all pointing inwards—and with crushing teeth. They form a very efficient manducatory apparatus. The boundary of the mouth is finally completed by the chilaria.

Certain of the appendages which persist will be described with the functions they subserve, the eyes with the sense-organs, the genital operculum with the generative organs, the gill-books with the respiratory system, but the chelicerae, pedipalpi, and walking limbs, which have retained the functions of prehension and locomotion usual to limbs, merit a little attention.[^[213]] The chelicerae are short and composed of but three joints. They are, like the succeeding segments, chelate, and the chelae of all are fine and delicate like a pair of forceps rather than like a Lobster’s claw. In the female L. polyphemus the pedipalp is remarkably like the three ambulatory legs which succeed it, and all four are chelate, but in the adult male the penultimate joint of the pedipalp is not prolonged to form one limb of the chela, which is therefore absent, and the appendage is thicker and heavier than in the other sex. In L. longispina and L. moluccanus the first walking leg, as well as the pedipalp, ends in a claw and not in a chela; the immature males resemble the females. The first three walking legs in both sexes of L. polyphemus resemble the pedipalpi of the female, and like them have six joints. The fourth and last pair of ambulatory appendages is not chelate, but its distal joints carry a number of somewhat flattened structures, which are capable of being alternately divaricated and approximated or bunched together. This enables them to act as organs for clearing away sand or mud from beneath the carapace as the creature lies prone on the bottom of the sea. To quote Mr. Lloyd,[^[214]] the “two limbs are, sometimes alternately and sometimes simultaneously, thrust backward below the carapace, quite beyond the hinder edge of the shell; and in the act of thrusting, the lobes or plates on each leg encounter the sand, the resistance or pressure of which causes them to open and fill with sand, a load of which at every thrusting operation is pushed away from under the king-crab, and deposited outside the carapace. The four plates then close and are withdrawn closed, previous to being opened and charged with another load of sand; and at the deposit of every load the whole animal sinks deeper into its bed, till it is hidden all except the eyes.” There seems little doubt that the action of these appendages in removing the sand from under the carapace is reinforced by the fanning action of the respiratory appendages, which set up a current that helps to wash the particles away. But the posterior walking legs are not the only organs used in burrowing. The Rev. Dr. Lockwood,[^[215]] who observed the habits of L. polyphemus off the New Jersey coast, says, “The king-crab delights in moderately deep water, say from two to six fathoms. It is emphatically a burrowing animal, living literally in the mud, into which it scoops or gouges its way with great facility. In the burrowing operation the forward edge of the anterior shield is pressed downward and shoved forward, the two shields being inflected, and the sharp point of the tail presenting the fulcrum as it pierces the mud, whilst underneath the feet are incessantly active scratching up and pushing out the earth on both sides. There is a singular economy of force in this excavating action; for the doubling up or inflecting and straightening out of the two carapaces, with the pushing purchase exerted by the tail, accomplish both digging and subterranean progression.”

At night-time Limulus is apt to leave the sand and progress by a series of short swimming hops, the respiratory appendages giving the necessary impetus, whilst between each two short flights the animal balances itself for a moment on the tip of its tail. During this method of progressing the carapace is slanting, forming an angle of about 45° with the ground. The unsegmented tail is also used when a King-crab falls on its back. “The spine is then bent, i.e. its point is planted in the sand so that it makes an acute angle with the carapace, which is then so far raised that some of the feet are enabled to grasp a projecting surface, either longitudinal or vertical, or at some combination of the two; and the crab then turns over.”

Fig. [154].—A sagittal section of Limulus, seen from the right side, somewhat smaller than natural size. After Patten and Redenbaugh.
All the prosomatic appendages, except the chelicera (4) and chilarium (33) of the right side, are omitted. The genital operculum (32) and the five gills (28) are represented.
The muscles are omitted except the fibres running from the occipital ring to the posterior side of the oesophagus, the chilarial muscles, the sphincter ani (27), and the levator ani (24).
The endosternite (34), with the occipital ring and the capsuliginous bar, is seen from the side, and the positions of the abdominal endochondrites (31) are indicated.
The mouth (1) leads into the oesophagus, which passes through the brain to the proventriculus (12). A constriction, which marks the position of the pyloric valve, separates the proventriculus from the intestine (23) which passes posteriorly to the anus (26). A pair of hepatic ducts (15) enter the intestine opposite the endocranium.
The heart (16) surrounded by the pericardial sinus lies above the intestine. The pericardium is shown between the heart and the intestine. The ostia (17) of the heart and the origins of the four lateral arteries (19) are indicated; the frontal artery (13) and the aortic arches (14) curving down to the brain, arise from the anterior end of the heart; the superior abdominal artery and the opening of the collateral artery into it are shown.
The brain surrounding the oesophagus is seen in side view upon the neural side of the endosternite (34). The ventral cord (35) passes through the occipital ring into the abdominal region. The anterior commissure (3), with the three rostral nerves (2) innervating the rostrum, or labrum, and four of the post-oral commissures, are represented.
The cheliceral nerve with the small external pedal branch is shown entire, but the next five neural nerves are cut off. The chilarial nerve, the opercular nerve, and the five branchial nerves, enter their respective appendages, the two former passing through the occipital ring.
From the fore-brain the three olfactory nerves (5) pass anteriorly to the olfactory organ; the median eye-nerve (10) passes to the right of the proventriculus (12) to the median eyes (11); the lateral eye-nerve (7) passes forward and is represented as cut off opposite the proventriculus. The lateral nerve (9) or first haemal nerve is also cut off just beyond the point where it fuses with the second haemal nerve (8). The stomodaeal nerve (6) ramifies over the oesophagus and proventriculus.
The second haemal nerve (8) passes to the anterior extremity of the carapace; its haemal branch is cut off opposite the proventriculus. An intestinal branch arises from near its base and disappears behind the anterior cornu of the endosternite.
The next three haemal nerves (36) are cut off close to the brain, and the following nine haemal nerves are cut off beyond the cardiac branches. The fifteenth haemal nerve (29) is cut off beyond its branch to the telson muscles. Both branches of the haemal nerve are represented extending into the telson (25).
The intestinal nerves are shown arising from the haemal nerves and entering the intestine. Those from the sixth and seventh neuromeres pass through foramina in the endosternite, and communicate with a plexus in the longitudinal abdominal muscles before entering the intestine. The eighth passes just posterior to the endosternite and joins the same plexus. Those from the first four branchial neuromeres arise very near the abdominal ganglia, and are double in their origins, the anterior branches joining the above-mentioned plexus, and the posterior branches entering the intestine. The fifteenth extends far back towards the rectum and anastomoses with the sixteenth, which arises from the caudal branch of the sixteenth haemal nerve, and innervates the rectum and anal muscles.
The segmental cardiac nerves (18) arise from the haemal nerves of the sixth to the thirteenth neuromeres respectively. The most anterior one passes to the inter-tergal muscles and the epidermis in the median line, but the connections with the cardiac plexus have not been made out. The next two (18) fuse to form a large nerve, which passes to the inter-tergal muscles and epidermis, but has not been observed to connect directly with the cardiac plexus. It, however, sends posteriorly a branch, the pericardial nerve (20), which in turn gives a branch to each of the cardiac nerves of the branchial neuromeres, and then continues onward to the posterior margin of the abdomen. This nerve lies in the epidermis. The median and lateral cardiac nerves (22 and 21) are seen upon the walls of the heart. The five cardiac nerves from the branchial neuromeres pass, in the epidermis, to the median line, and dip down to the median nerve (22) of the heart opposite the last five pairs of ostia (17). They communicate with the pericardial nerve (20) and also with the lateral sympathetic nerve (30).
Two post-cardiac nerves pass from the first and second post-branchial nerves to the epidermis posterior to the heart.
The last cardiac nerve and the two post-cardiac nerves give off branches which anastomose with each other and innervate the extensors of the telson.
The lateral sympathetic nerve (30) receives branches from all the neuromeres from the eighth to the fourteenth, either through the cardiac nerves or the haemal nerves, and innervates the branchio-thoracic muscles, extending with these far into the cephalothorax.
1, Mouth; 2, rostral nerve in labrum; 3, anterior commissure; 4, chelicera; 5, olfactory nerves; 6, stomodaeal nerve; 7, lateral eye-nerve; 8, 2nd haemal nerve; 9, lateral nerve; 10, median eye-nerve; 11, median eye; 12, proventriculus; 13, frontal artery; 14, aortic arch; 15, anterior hepatic duct of liver; 16, heart; 17, 2nd ostium; 18, 7th and 8th segmental cardiac nerves; 19, one of the lateral arteries; 20, pericardial nerve; 21, lateral cardiac nerve; 22, median cardiac nerve; 23, intestine; 24, levator ani muscle; 25, telson; 26, anus; 27, sphincter ani muscle; 28, last branchial appendage; 29, 15th haemal nerve; 30, lateral sympathetic nerve; 31, 8th abdominal endochondrite; 32, genital operculum; 33, chilarium; 34, endosternite; 35, ventral nerve-cord; 36, 6th haemal nerve; 37, origin of 6th neural nerve.

Limulus feeds partly on bivalves, but mainly on worms, especially Nereids, which it catches with its chelate limbs as it burrows through the sand. The food is held immediately under the mouth by the chelicerae, aided at times by the succeeding appendages; it is thus brought within range of the gnathobases of the walking legs, and these by an alternate motion “card” the food into fragments, which when sufficiently comminuted pass into the mouth. At times its appendages are caught between the valves of Venus mercenaria, a burrowing bivalve known in America as the “quahog” or “round clam.” The Limulus has seized with its chelate claws the protruding siphon of this mollusc, which, being rapidly drawn in, drags with it the limb of the king-crab, and the valves of the clam are swiftly snapped to.

As a rule in Arachnids the alimentary canal is no longer than the body, and runs straight from mouth to anus, but in Limulus, the mouth being pushed far backward, there is a median loop, and the narrow oesophagus which leads from the mouth, having traversed the nerve-ring, passes forward towards the anterior end of the carapace. Here it enters into a somewhat ⸧ shaped and spacious proventriculus; posteriorly the proventriculus opens by a funnel-shaped valve into the anterior end of the narrow intestine. All these structures are derived from the stomodaeum, are lined with chitin and are provided with very muscular walls whose internal surface is thrown into longitudinal ridges. The intestine runs straight backward, diminishing in its diameter, and ends in a short, chitin-lined, and muscular rectum which is derived from the proctodaeum; the anus is a longitudinal slit. A large gland, usually called the liver, consisting of innumerable tubules, pours its secretions into the broader anterior end of the intestine by two ducts upon each side; it extends into the meso- and metasoma, and, together with the reproductive organs, forms a “packing” in which the other organs are embedded. The contents of the alimentary canal are described as “pulpy and scanty,” and probably much of the actual digestion goes on inside the lumen of the above-mentioned gland.

The vascular system of Limulus, like that of the Scorpions, is more completely developed than is usually the case in Arthropods. For the most part the blood runs in definite arteries, and when it passes as it does into venous lacunae these are more definite in position and in their retaining walls than in other members of the phylum.

The heart lies in a pericardial space with which it communicates by eight[^[216]] pairs of ostia. Eight paired bands of connective tissue, the “alary muscles” of authors, sling the heart to the pericardial membrane. Posteriorly the pericardial chamber receives five paired veins on each side coming from the gills and returning the purified blood to the heart.

Eleven arteries arise from the heart. These are (i.) a median frontal artery which, passing forward, divides into a right and left marginal artery. These run round the edge of the carapace to its posterior angle, where each receives a branch of the collateral artery mentioned below. (ii.) and (iii.) are the aortic arches (Fig. [154]), paired vessels running round and supplying the proventriculus and oesophagus. These unite ventrally in a vascular ring which encloses the nerve-ring, and is continued along the ventral nerve-cord as the ventral artery and along some of the chief nerves. This vascular ring supplies the lateral eyes and all the appendages mentioned on p. 263 up to and including the genital operculum. The ventral artery supplies the respiratory appendages, and gives branches to the rectum, caudal spine, etc. Two of its branches encircle the rectum, and uniting open into the superior abdominal artery. iv.–xi. are paired lateral arteries which leave the heart beneath the anterior four ostia, and soon enter a longitudinal pair of collateral arteries which unite behind in the just mentioned superior abdominal artery; they also give off branches to the muscles and to the intestine, and a stout branch mentioned above which passes into the marginal artery posteriorly. The venous system is lacunar, and the blood is collected from the irregular spaces between the various organs into a pair of longitudinal sinuses, whence it passes into the operculum and the five pairs of gills. A large branchio-cardiac canal returns the blood from each gill to the cavity of the pericardium, and so through the ostia to the heart. Eight veno-pericardiac muscles run from the under surface of the pericardium to be inserted into the upper surface of the longitudinal sinus; they occur opposite the ostia, and play an important part in the mechanism of the circulation. The blood is coloured blue by haemocyanin; amoeboid corpuscles float in the plasma.

The respiratory organs are external gills borne on the posterior face of the exopodite of the lamella-like posterior five mesosomatic limbs. Each gill consists of a series of leaves like the leaves of a book, and some 150–200 in number. Within the substance of each leaf the blood flows, while without the oxygen-carrying water circulates between the leaves. These gillbearing appendages can be flapped to and fro, and they seem to be at times held apart by the flabellum, a spatulate process which Patten and Redenbaugh regard as a development of the median sensory knob on the outer side of the coxopodite of the last pair of walking limbs.

Fig. [155].—Diagram of the first gill of Limulus, from the posterior side, showing the distribution of the gill-nerve to the gill-book (about natural size). After Patten and Redenbaugh. 1, Inner lobe of the appendage; 2, outer lobe of appendage; 3, median lobe of appendage; 4, gill-book; 5, neural nerve of the ninth neuromere; 6, internal branchial nerve; 7, gill-nerve; 8, median branchial nerve; 9, external branchial nerve.

Limulus has no trace of Malpighian tubules, structures which seem often to develop only when animals cease to live in water and come to live in air. The Xiphosura have retained as organs of nitrogenous excretion the more primitive nephridia, or coxal glands as they are called, in the Arachnida. They are redbrick in colour, and consist of a longitudinal portion on each side of the body, which gives off a lobe opposite the base of the pedipalps and each of the first three walking legs—in the embryo also of the chelicerae and last walking legs, but these latter disappear during development. A duct leads from the interior of the gland and opens upon the posterior face of the last pair of walking legs but one.

The nervous system has been very fully described by Patten and Redenbaugh, and its complex nature plays a large part in the ingenious speculations of Dr. Gaskell as to the origin of Vertebrates. It consists of a stout ring surrounding the oesophagus and a ventral nerve-cord, composed—if we omit the so-called fore-brain—of sixteen neuromeres. The fore-brain supplies the median and the lateral eyes, and gives off a median nerve which runs to an organ, described as olfactory by Patten, situated in front of the chelicerae on the ventral face of the carapace. Patten distinguishes behind the fore-brain a mid-brain, which consists solely of the cheliceral neuromere, a hind-brain which supplies the pedipalps and four pair of walking legs, and an accessory brain which supplies the chilaria and the genital operculum. This is continued backward into a ventral nerve-cord which bears five paired ganglia supplying the five pairs of gills and three pairs of post-branchial ganglia; the latter are ill-defined and closely fused together. As was mentioned above, the whole of the central nervous system is bathed in the blood of the ventral sinus.

The sense-organs consist of the olfactory organ of Patten, the median and lateral eyes, and possibly of certain gustatory hairs upon the gnathobases. The lateral eyes in their histology are not so differentiated as the median eyes, but both fall well within the limits of Arachnid eye-structure, and their minute anatomy has been advanced as one piece of evidence amongst many which tend to demonstrate that Limulus is an Arachnid.

Both ovaries and testes take the form of a tubular network which is almost inextricably entangled with the liver. From each side a duct collects the reproductive cells which are formed from cells lining the walls of the tubes, and discharges them by a pore one on each side of the hinder surface of the genital operculum. As is frequently the case in Arachnids the males are smaller than the females, and after their last ecdysis the pedipalps and first two pairs of walking legs, or some of these appendages, end in slightly bent claws and not in chelae. Off the New Jersey coast the king-crabs (L. polyphemus) spawn during the months of May, June, and July, Lockwood states at the periods of highest tides, but Kingsley[^[217]] was never “able to notice any connexion between the hours when they frequent the shore and the state of the tide.” “When first seen they come from the deeper water, the male, which is almost always the smaller, grasping the hinder half of the carapace of the female with the modified pincer of the second pair of feet. Thus fastened together the male rides to shallow water. The couples will stop at intervals and then move on. Usually a nest of eggs can be found at each of the stopping-places, and as each nest is usually buried from one to two inches beneath the surface of the sand, it appears probable that the female thrusts the genital plate into the sand, while at the same time the male discharges the milt into the water. I have not been able to watch the process more closely because the animals lie so close to the sand, and all the appendages are concealed beneath the carapace. If touched during the oviposition, they cease the operation and wander to another spot or separate and return to deep water. I have never seen the couples come entirely out of the water, although they frequently come so close to the shore that portions of the carapace are uncovered.”[^[218]]

Fig. [156].—A view of the nervous system of Limulus from below. (About natural size.) After Patten and Redenbaugh.
The carapace is represented as transparent. The appendages have been removed, but the outlines of the left entocoxites (6) have been sketched in. The positions of the abdominal appendages are indicated by the external branchial muscles (17), the branchial cartilages (19), the tendinous stigmata (18), and the abdominal endochondrites (21). In the cephalothorax (1) all the tergo-coxal and plastro-coxal muscles have been dissected away, leaving the endosternite (11) with the occipital ring exposed. One of the left tergo-proplastral muscles (4) and the left branchio-thoracic muscles (16) are represented. The longitudinal abdominal muscles are also seen. All the muscles of the right side have been omitted except the haemo-neural muscles (23), of which the last two are represented upon the left side also. At the base of the telson the flexors (29) and extensors (27) of the caudal spine are represented as cut off near their insertions. The sphincter ani (26), levator ani, and occludor ani (25), and their relations to the anus (28), are shown.
The oesophagus runs forward to the proventriculus (3). From this the intestine (20) passes posteriorly.
The brain lies upon the neural side of the endosternite, and the ventral cord (22) passes back through the occipital ring. The neural nerves are cut off, but the left haemal nerves and those from the fore-brain (12) are represented entire.
The first pair of neural nerves go to the chelicerae. The second to sixth pairs go to the next five cephalothoracic appendages, which are represented by the entocoxites (6). The seventh pair of neural nerves go to the chilaria, and the eighth pair to the operculum. The neural nerves from the ninth to the thirteenth arise from the abdominal ganglia and innervate the five pairs of gills.
From the fore-brain a median olfactory nerve (9) and two lateral ones (8) pass forward to the olfactory organ; a median eye-nerve (2) passes anteriorly and haemally upon the right of the proventriculus (3) to the median eyes; and a pair of lateral eye-nerves pass to the lateral eyes (15).
The first haemal nerve, or lateral nerve, follows the general course of the lateral eye-nerve, but continues posteriorly far back on to the neural side of the abdomen.
The haemal nerves of the hind-brain radiate from the brain to the margins of the carapace, and each one passes anterior to the appendage of its own metamere. The integumentary portions divide into haemal and neural branches, of which the haemal branches (5) are cut off. Each haemal branch gives off a small nerve which turns back toward the median line upon the haemal side of the body.
The haemal nerves of the accessory brain pass through the occipital ring to the sides of the body between the operculum and the sixth cephalothoracic appendage. The seventh innervates the posterior angles of the cephalothorax, the eighth the opercular portion of the abdomen. The next five haemal nerves arise from the five branchial neuromeres, pass out anterior to the gills to the sides of the abdominal carapace, and innervate the first five spines upon the sides of the abdomen.
The first post-branchial nerve innervates the last abdominal spine; the second post-branchial nerve and one branch of the third post-branchial innervate the posterior angles of the abdomen and the muscles of the telson; and the caudal branch of the third post-branchial nerve innervates the telson.
Intestinal branches arise from all the haemal nerves from the sixth to the sixteenth, and pass to the longitudinal abdominal muscles and to the intestine.
Cardiac nerves arise from all the haemal nerves from the sixth to the thirteenth. Six of the cardiac nerves communicate with the lateral sympathetic nerve (24), which innervates the branchio-thoracic muscles (16).
Two post-cardiac nerves arise from the first two post-branchial nerves, and passing to the haemal side anastomose with a branch from the last cardiac nerve, and innervate the extensors (27) of the telson and the epidermis behind the heart.
1, Cephalothorax; 2, median eye-nerve; 3, proventriculus; 4, tergo-proplastral muscles; 5, haemal branch of integumentary nerve; 6, entocoxites; 7, 2nd haemal nerve; 8, right olfactory nerve; 9, median olfactory nerve; 10, intestine; 11, endosternite; 12, fore-brain; 13, origin of 4th neural nerve; 14, lateral nerve; 15, lateral eye; 16, branchio-thoracic muscles; 17, external branchial muscles; 18, tendinous stigmata; 19, branchial cartilages; 20, intestine; 21, abdominal endochondrites; 22, ventral cord; 23, haemo-neural muscles; 24, lateral sympathetic nerve; 25, occludor ani; 26, sphincter ani; 27, extensors of telson; 28, anus; 29, flexors of telson; 30, lateral projections of abdomen; 31, nerves of spines; 32, external branchial muscles.

Fig. [157].—The markings on the sand made by the female Limulus when depositing eggs. Towards the lower end the round “nests” cease to be apparent, the king-crab being apparently exhausted. (From Kishinouye.) About natural size.

The developing ova and young larvae are very hardy, and in a little sea-water, or still better packed in sea-weed, will survive long journeys. In this way they have been transported from the Atlantic to the Pacific coasts of the United States, and for a time at any rate flourished in the western waters. Three barrels full of them consigned from Woods Holl to Sir E. Ray Lankester arrived in England with a large proportion of larvae alive and apparently well.

According to Kishinouye, L. longispina spawns chiefly in August and between tide-marks. “The female excavates a hole about 15 cm. deep, and deposits eggs in it while the male fertilises them. The female afterwards buries them, and begins to excavate the next hole.”[^[219]] A line of nests (Fig. [157]) is thus established which is always at right angles to the shore-line. After a certain number of nests have been formed the female tires, and the heaped up sand is not so prominent. In each “nest” there are about a thousand eggs, placed first to the left side of the nest and then to the right, from which Kishinouye concludes that the left ovary deposits its ova first and then the right. Limulus rotundicauda and L. moluccanus do not bury their eggs, but carry them about attached to their swimmerets.

The egg is covered by a leathery egg-shell which bursts after a certain time, and leaves the larva surrounded only by the blastodermic cuticle; when ripe it emerges in the condition known as the “Trilobite larva” (Fig. [158]), so-called from a superficial and misleading resemblance to a Trilobite. They are active little larvae, burrowing in the sand like their parents, and swimming vigorously about by aid of their leaf-like posterior limbs. Sometimes they are taken in tow-nets. After the first moult the segments of the meso- and metasoma, which at first had been free, showing affinities with Prestwichia and Belinurus of Palaeozoic times, become more solidified, while the post-anal tail-spine—absent in the Trilobite larva—makes its first appearance. This increases in size with successive moults. We have already noted the late appearance of the external sexual characters, the chelate walking appendages only being replaced by hooks at the last moult.

Fig. [158].—Dorsal and ventral view of the last larval stage (the so-called Trilobite stage) of Limulus polyphemus before the appearance of the telson. 1, Liver; 2, median eye; 3, lateral eye; 4, last walking leg; 5, chilaria. (From Kingsley and Takano.)

Limulus casts its cuticle several times during the first year—Lockwood estimates five or six times between hatching out in June and the onset of the cold weather. The cuticle splits along a “thin narrow rim” which “runs round the under side of the anterior portion of the cephalic shield.”[^[220]] This extends until it reaches that level where the animal is widest. Through this slit the body of the king-crab emerges, coming out, not as that of a beetle anteriorly and dorsally, but anteriorly and ventrally, in such a way as to induce the unobservant to exclaim “it is spewing itself out of its mouth.” In one nearly full-sized animal the increase in the shorter diameter of the cephalic shield after a moult was from 8 inches to 9½ inches, which is an indication of very rapid growth. If after their first year they moult annually Lockwood estimates it would take them eight years to attain their full size.

The only economic use I know to which Limulus is put is that of feeding both poultry and pigs. The females are preferred on account of the eggs, of which half-a-pint may be crowded into the cephalic shield. The king-crab is opened by running a knife round the thin line mentioned on p. 275. There is a belief in New Jersey that this diet makes the poultry lay; undoubtedly it fattens both fowls and pigs, but it gives a “shocking” flavour to the flesh of both.

CLASSIFICATION.

But five species of existing King-crabs are known, and these are grouped by Pocock into two sub-families: (i.) the Xiphosurinae, and (ii.) the Tachypleinae. These together make up the single family Xiphosuridae which is co-extensive with the Order. The following is Pocock’s classification.[^[221]] The names used in this article are printed in italic capitals.

Order Xiphosura.

Family 1. Xiphosuridae.

Sub-Fam. 1. Xiphosurinae.

This includes the single species Xiphosura polyphemus (Linn.) (= Limulus polyphemus, Latreille), “which is said to range from the coast of Maine to Yucatan.”

Sub-Fam. 2. Tachypleinae.

Genus A. Tachypleus includes three species: (i.) T. gigas, Müll. (= Limulus gigas, Müll., and L. moluccanus, Latreille), widely distributed in Malaysia; (ii.) T. tridentatus, Leach (= L. tridentatus, Leach, and L. longispina, Van der Hoeven), extending from British North Borneo to China and Southern Japan; and (iii.) T. hoeveni, Pocock (= L. moluccanus, Van der Hoeven), found in the Moluccas.

Genus B. Carcinoscorpius with one species, C. rotundicauda (Latreille) (= L. rotundicauda, Latreille). It occupies a more westerly area than T. gigas or than T. tridentatus, having been recorded from India and Bengal, the Gulf of Siam, Penang, the Moluccas, and the Philippines.

With regard to the affinities of the group it is now almost universally accepted that they are Arachnids. The chief features in which they differ from other Arachnids are the presence of gills and the absence of Malpighian tubules, both being features associated with aquatic life. As long ago as 1829 Straus-Dürckheim emphasised the points of resemblance between the two groups, and although the view was during the middle of the last century by no means universally accepted, towards the end of that epoch the painstaking researches of Lankester and his pupils, who compared the King-crab and the Scorpion, segment with segment, organ with organ, tissue with tissue, almost cell with cell, established the connexion beyond doubt. Lankester would put the Trilobites in the same phylum, but in this we do not follow him. With regard to the brilliant but, to our mind, unconvincing speculations as to the connexion of some Limulus-like ancestor with the Vertebrates, we must refer the reader to the ingenious writings of Dr. Gaskell,[^[222]] recently summarised in his volume on “The Origin of Vertebrates,” and to those of Dr. Patten in his article “On the Origin of Vertebrates from Arachnids.”[^[223]]

Fossil Xiphosura.[^[224]]

Fig. [159].—A., Hemiaspis limuloides, Woodw., Upper Silurian, Leintwardine, Shropshire. Natural size. (After Woodward.) B., Prestwichia (Euroöps) danae (Meek), Carboniferous, Illinois, × ⅔. (After Packard.)

Limulus is an example of a persistent type. It appears first in deposits of Triassic age, and is found again in the Jurassic, Cretaceous, and Oligocene. In the lithographic limestone of Solenhofen in Bavaria, which is of Upper Jurassic age, Limulus is common and is represented by several species. One species is known from the Chalk of Lebanon, and another occurs in the Oligocene of Saxony. No other genus of the Xiphosura appears to be represented in the Mesozoic and Tertiary deposits, but in the Palaeozoic formations (principally in the Upper Silurian, the Old Red Sandstone, and the Coal Measures) several genera have been found, most of which differ from Limulus in having some or all of the segments of the abdomen free; in this respect they resemble the Eurypterida, but differ from them in the number of segments. In Hemiaspis (Fig. [159], A), from the Silurian, the segments of the abdomen are divisible into two groups (mesosoma and metasoma) in the same way that they are in Eurypterids; the first six segments have broad, short terga, the lateral margins of the sixth being divided into two lobes, probably indicating the presence of two fused segments; the last three segments are narrower and longer than the preceding, and at the end is a pointed tail-spine. In Belinurus (Fig. [160]) from the Carboniferous, the two regions of the abdomen are much less distinct; there are eight segments, the last three of which are fused together, and a long tail-spine. In Neolimulus, from the Silurian, there seems to be no division of the abdomen into two regions, and apparently all the segments were free. On the other hand, in Prestwichia (Carboniferous), all the segments of the abdomen, of which there appear to be seven only, were fused together (Fig. [159], B).

Fig. [160].—Belinurus reginae, Baily, Coal Measures, Queen’s Co., Ireland, × 1. (After Woodward).

In the Palaeozoic genera the median or axial part of the dorsal surface is raised and distinctly limited on each side, so presenting a trilobed appearance similar to that of Trilobites. In Neolimulus, Belinurus, and Prestwichia, lateral eyes are present on the sides of the axial parts of the carapace, and near its front margin median eyes have been found in the two last-named genera.

In nearly all the specimens of Palaeozoic Xiphosura[^[225]] which have been found nothing is seen but the dorsal surface of the body; in only a very few cases have any traces of the appendages been seen,[^[226]] but, so far as known, they appear to have the same general character as in Limulus.

Aglaspis, found in the Upper Cambrian of Wisconsin, has been regarded as a Xiphosuran. If that view of its position is correct, then Aglaspis will be the earliest representative of the group at present known. Other genera of Palaeozoic Xiphosura are Bunodes, Bunodella, and Pseudoniscus in the Silurian; Protolimulus in the Upper Devonian; and Prolimulus in the Permian.