In Astacus, where the structure is perhaps more distinct, the margins of the approximating somites are seen to be compressed together, the anterior margin of one with the posterior of the next, and to thin out and ultimately combine together into a thin wall or plate of partition, separating the several sets of muscles connected with appendages belonging to one somite from those belonging to adjoining ones. Independently of being walls of separation they are points of attachment on which some of the muscles are securely fixed. Not only do they exist near the lateral margin, but continue inwards and extend forwards until they reach the corresponding processes on the opposite side of the percion, and also anteriorly until they unite with a similar system of osseous plates in the adjoining somite. Each plate appears to form a basis on which a strong muscle may take root on either side, thus forming a fulcrum for muscular power and a means of separating one set of muscles from another. In Palinurus these plates, when they approximate the median line, turn over and lie horizontally with the longitudinal axis of the animal. These plates thus displayed form a perforated floor on which the larger and more important internal viscera rest. This osseous system continues from the postmandibular somite persistently to the penultimate somite of the percion, where it is united with the floor of the pereion by a central and lateral point of contact.

The anterior margins of the two halves of the first somite of the percion meet together in the centre and form an oblique and prominent bridge that supports the posterior portion of the stomachic viscera, while the internal processes of the apodema, as they are termed by M. Milne-Edwards, that spring from the posterior two somites of the cephalon, are closely attached to, and at their extremities are perfectly ossified with, the lateral and central parts of the apodema of the anterior somite of the pereion, a point of union that the structure of the animal requires to be of considerable strength, as the enormous processes of the internal movable mandibular plates occupy so large a space that their points of attachment necessitate a structure of greater resistance and strength than the impoverished character and condition of the two posterior somites of the cephalon are capable of securing to them, without the additional support which they receive from a union of a more or less perfect character with the anterior somite of the pereion.

The apodema that support the internal viscera are perforated by a series of foramina that, while they correspond in form on each side of the central line, yet differ in size and shape according to the relative proportions of the organism that are connected with them. The dimensions of the foramina, through which the muscles move the large and more important appendages, are larger and more conspicuous than they that relate to those that move the less efficient and smaller organs of the body. Thus we find that, generally, the largest and most conspicuous foramina correspond with the third somite of the pereion in Palinurus, Astacus, &c., whereas in those genera where the great prehensile hand is produced by the increased growth and proportions of any other pair of appendages, the foramina in the apodemal plate correspond with the increase of their dimensions.

In the Anomura, of which we will take Lithodes as the type, the internal and apodemal plates do not project so as to reach the corresponding processes on the opposite side. There are only six somites fused together on the ventral surface, or, I should rather say, contributing to the formation of the sternal plastron; the seventh somite exists as a separate and distinct ring, both dorsally and ventrally free from ossified union with the anterior somites of the pereion.

In this genus the sternal plate, as an anatomical part of the animal, is wanting, or represented only in a theoretical character by the median line of fusion.

The coxæ are existent without fusion with each other for some extent, visible on the ventral surface before their close contact reaches ossification so perfect that their line of union is represented by marks of depression only on the external surface, and corresponding crests or ridges on the internal surface. Dorsally this appears to be similarly repeated, and the lines of contact are imperfect in their fusion until the plates have thinned out into a membrane. Laterally the walls of the cox of the several pairs of appendages are so closely compressed that their lines of union are with difficulty determined not to be fused together. That they exist for some distance as thin plates in close contact is certain; but they ultimately reach a point where the distinction is lost in perfect ossification. The internal plates approach the corresponding ones on the opposite side in the first two somites only, which form a bridge that supports the posterior extremity of the stomachic region; behind this the ventral surface rapidly widens, but the apodema or internal plates abruptly terminate, leaving a largo expansion for the internal viscera to occupy.

In the Brachyura the central fusion of the sternal plates is still more perfect, and the ventral portion of the somites appears to be covered entirely; this exists in a vertical plate that appears to be formed by being compressed between the coxæ of the corresponding pairs of appendages, the external surface of which may be traced to a sinus (Pl. II. fig. 6 a) that opens in the median line between the third and fourth somites. The segments of the pereion in this order of Crustacea, as may be seen in the genus Cancer, are very closely compressed, and apparently overlap each other dorsally, while ventrally the several appendages, from their proportionate dimensions, preclude the possibility of too close a contact. The consequence is that the general arrangement of the entire muscular system that moves the appendages or the pereion, together with the osseous structure that supports them, is arranged in a circular form, the superior or extensor muscles forming the upper or dorsal arc, and the inferior or flexor muscles forming the lower or ventral arc. The plate, therefore, that is produced internally in the median line is in continuation with the anterior portion of the ventral floor of the pereion, and is the homologue of the sternal plate. This tendency of the muscles to form round a common osseous centre appears to give a similar relation of the several somites to one another. Thus we find that the apodema narrows the dorsal extremity corresponding to each somite to such a degree that a deep notch or fold takes place over the fourth pair of appendages, at which point the curvature is greatest (fig. 5). It is this circular portion of the muscles that facilitates that peculiar arrangement by which the posterior two pairs of legs in Dromia, Doripe, &c. appear to be attached to the dorsal surface of the animal, which enables them to adhere to floating pieces of wood or weed, or securely attach themselves to univalve shells by means of these appendages.

The pleon, or that portion of the animal to which the appendages are attached which, in their most perfect condition, are adapted for swimming, undergoes a great variety of forms. It is perhaps most perfectly developed, in accordance with the value and usefulness of its parts, in the Macrurous division of Crustacea.

In the Edriophthalmia it is perhaps more simple in character; but it is in the Anisopoda, or that intermediate stage that unites the Isopoda and the

Amphipoda, that we are enabled to determine the true homological relation of one part to the other.

In all Crustacea above the Entomostracous forms the several somites are distinguished by a dorsal and a ventral arc. The dorsal is invariably a hard, strong, and osseous plate. The ventral arc is mostly represented by an osseous band that reaches across the animal, and is united anteriorly and posteriorly to the contiguous somites by large and flexible membranous tissues. The dorsal arc is wide, and dips under the adjoining one anteriorly in all except the second somite in the Macrura, which overrides the plates of the adjoining somites both anteriorly and posteriorly. This arrangement does not exist in the Edriophthalmia, because, there being no dorsal carapace protecting the pereion, all the somites have a separate and distinct dorsal arc. The consequence is that each somite posteriorly overlaps the anterior margin of the next succeeding ring, except the first or anterior somite of the pereion, which overlaps anteriorly the posterior margin of the cephalon and posteriorly the anterior margin of the second somite of the pereion. In each of these orders of Crustacea we find that the greatest power of flexion is given to the animal at these points.

In all the distinguishable somites of the Edriophthalmia, from one extremity of the animal to the other, each separate one is observed to support laterally a large plate. These, in the percion, are firmly attached to their respective somites, but not ossified to them; in the pleon they are so united by ossific matter that one part is not capable of being separated anatomically or distinguished in structure from the other. It is these parts in this particular division of Crustacea that originated the idea of the theory of the Crustacean somite as enunciated in 1830 by Prof. Milne-Edwards. The fact that the supposed side-plates, or epimera, were merely the first joint of the normal legs or appendages has been satisfactorily demonstrated in the Edriophthalmia, as far as relates to the somites of the pereion; but hitherto the relation of the side-plates of the pleon to the normal condition of the mobile appendages had not been demonstrated until the structure of the dermal anatomy of the genus Apseudes had been made out*; that "one interesting and, as far as we know, unique feature in these Crustacea yet remains to be noticed. The segments of the pleon have the lateral walls (long known as the epimera of Milne-Edwards, called also the pleura by many authors) existing as articulated appendages, demonstrating two important features in the homologies of these parts: 1st, that they are all really portions of the appendages, being the first joint or coxæ of the pleopod . . . and 2nd, that, since the peduncle consists of three joints, the second branch in the appendages of the pleon, as in other parts, is shown to take place invariably at the extremity of the third joint." In the Macrura and higher Stomapods the coxal joint of the several appendages is united to the dorsal arc in a very perfect and complete state of ossification, with the exception of the first somite, where there are no appendages, and the sixth, where the coxa is free and articulates, with small lateral motions, with the dorsal arc of the respective somite. The seventh somite (telson) is reduced in character and altered in form; it universally covers and holds the terminal exit of the alimentary canal, the inferior arc of which is represented by a membranous tissue. In the Amphipodous order of Crustacea the fifth and sixth somites carry their appendages with free cox, and the terminal somite exists only in the form of a scale very liable to vary in shape, or separated into two of minute * Hist. Brit. Sessile-eyed Crust. vol. ii. p. 146 (Apscudes).

1876.

G

dimensions. In the Isopoda the sixth somite only has the coxæ free, and the appendages attached to them bear no very distant analogy to the homologous pair as they exist in the Macrura. In numerous genera of Isopods the sixth somite is developed to a very large size, and either absorbs or displaces the terminal somite or telson altogether, which in some genera is represented by a notch or cavity only, while in many others it is produced to a point or terminates in a smooth and even margin; with the exception of some of the Anisopod genera, the telson probably is absent throughout the order of Isopods.

The form of the pleon in the Brachyura bears as close a resemblance to that of the Isopoda belonging to the tribe Liberatica as that of the Macrura resembles Parasitica in the same order.

The coxæ or side-pieces, as they have been very commonly supposed to be, are, in the Brachyura, very densely ossified with the dorsal arc, and this to such an extent in the male animals that it is very difficult to determine their presence. In the female, where the lateral development assumes a greater extent, the line of union is capable of being determined by a marked depression that defines the limit of the somites and the altered position of the appen→ dages; but that they are homologically present in both sexes there can be no reasonable cause of doubt. This, I think, may be generally depended onthat the more the coxa departs from the normal type of the joint, as we see in the Macrurous Crustacea, and becomes associated with the dorsal arc of the theoretical somite, the more the character of the appendage becomes simplified or depreciated; but, on the other hand, the more intimately it becomes associated with the ventral arc, the more it becomes developed in its connexion with the requirements of the animal, and any variation of form is dependant on the value of its position and the habits and necessities of the creature. Thus we find that all the appendages of the cephalon and pereion are associated with the ventral arc in the Brachyura and Macrura, but in the Edriophthalmia those of the pereion are associated with the dorsal arc; whereas the appendages of the pleon are, in all divisions of Crustacea, so intimately associated with the dorsal arc that in most cases the coxa is incorporated with the somite, and generally the remainder of the appendages disappear or are reduced to merely a rudimentary condition, useful in some females for the attachment of ova; while in the males they disappear more or less completely, or in the general conditions of life become variated so as to fulfil special requirements or peculiar functions.

Thus the 21 somites of which the typical Crustacean consists each supports in its most simple condition a single pair of appendages; and if we were to suppose every segment of the animal to be reduced to its most simple character, and the appendages attached to each segment reduced to the most simple form of articulated limbs, and all of them uniform in size, the animal would bear a close analogy to a segmented annelid.

This we must take as the archetype of a crustaceous animal, and assume that the appendages are attached to the spaces that exist between the dorsal and the ventral arcs of each somite. Thus when we observe any extreme variation of form, we must consider the earliest and most simple condition of the appendage in the archetype; and it is not at variance with our idea of progression to assume that any great departure from the most simple type that appears to be common to the entire or a large portion of the subkingdom of Crustacea had its origin at an earlier period in the history of its evolution.

The organs of vision are common to all the Crustacea; and in those species

that are blind in their adult condition, the eyes are generally well developed in the younger stages.

The eyes are, independent of their value as organs of vision, of great importance in the study of the natural arrangement of the various forms of animals in the subkingdom. They vary in form and character from the most incipient ophthalmic spot to the compound eye erected on pedestals; but whether single or compound, solitary or in pairs, their form and composition is generally so persistent with certain forms and characteristics of the life and habits of the animals that the few exceptions to the general rule do not preclude them from being an important and valuable means of arranging Crustacea.

This was first appreciated by Leach, in 1815, in his Classification of the two great divisions of these animals. He arranged them under the two great heads of PODOPHTHALMIA and EDRIOPHTHALMIA-or those Crustacea that in their adult stage have the eyes elevated on peduncles or footstalks, and those which have them sessile or without any footstalk. To this general observation the exceptions are very few. Among some genera that inhabit subterranean passages and live in the dark, the footstalks are so reduced in size that they can only be said to exist theoretically, inasmuch as we find them well exhibited in their young and early stage. We must therefore assume that they have depreciated from their normal condition through adverse circumstances. On the other hand, among the Edriophthalmia we have the genus Tanais with its compound eyes elevated on their own pedestals, differing from the pedunculated form only in being rigid and incapable of movement.

In the Podophthalmia the eyes are implanted at the extremities of appendages that are supported upon a separate and distinct somite.

In 1837 Prof. Milne-Edwards demonstrated this to be the case in the genus Squilla; in 1854 he states, in his "Observations sur le Squelette tégumentaire des Crustacés décapodes," Ann. des Sciences Nat. p. 254, which I have since confirmed (fig. 7), that in the genus Palinurus (the Langouste)" l'anneau ophthalmique est parfaitement distinct, et se présente sous la forme d'une pièce sclérodermique impaire, courte et large, située en avant du bord frontal de la carapace, et au-dessus de l'anneau antennulaire. Les appendices ophthalmiques, ou tiges oculaires, naissent des deux extrémités de ce segment, et se composent chacun de deux articles: une pièce que j'appellerai basophthalmite, et une seconde, qui porte à son extrémité la cornée transparente, et qu'on peut nommer podophthalmite."

Milne-Edwards in the same manner shows how in several species of Palinurus the antero-median portions of the carapace project more or less completely over the ophthalmic ring, and so (l. c. p. 255) "par conséquent, ouvert à ses deux extrémités latérales pour le passage des tiges oculaires, et l'espèce de cadre ainsi constitué autour de la base de ces tiges forme la portion fondamentale de l'orbite ou trou orbitaire."

Thus the orbit in Crustacca is formed by the third or second antennal somite reaching over and coming into contact more or less perfectly with the first antennal somite. The greater or less in degree the separation between the second and third somite above the ophthalmic somite the more or less complete is the orbit in which the eye is protected. This varies in different genera, and is very complete in the genus Cancer (Pl. II. fig. 9), where the ophthalmic somite is enclosed entirely by the union without fusion of the anterodorsal projection of the posterior antennal somite with the anterior antennal somite; but, according to Milne-Edwards, in the genus Palinurus this perfection of the orbit varies. In P. vulgaris (fig. 8) the ophthalmic somite is naked,