THE
CAMBRIDGE NATURAL HISTORY

CRUSTACEA

By Geoffrey Smith, M.A. (Oxon.), Fellow of New College, Oxford; and the late W. F. R. Weldon, M.A. (D.Sc., Oxon.), formerly Fellow of St. John’s College, Cambridge, and Linacre Professor of Human and Comparative Anatomy, Oxford

TRILOBITES

By Henry Woods, M.A., St. John’s College, Cambridge; University Lecturer in Palaeozoology

INTRODUCTION TO ARACHNIDA, AND KING-CRABS

By A. E. Shipley, M.A., F.R.S., Fellow and Tutor of Christ’s College, Cambridge; Reader in Zoology

EURYPTERIDA

By Henry Woods, M.A., St. John’s College, Cambridge; University Lecturer in Palaeozoology

SCORPIONS, SPIDERS, MITES, TICKS, ETC.

By Cecil Warburton, M.A., Christ’s College, Cambridge; Zoologist to the Royal Agricultural Society

TARDIGRADA (WATER-BEARS)

By A. E. Shipley, M.A., F.R.S., Fellow and Tutor of Christ’s College, Cambridge; Reader in Zoology

PENTASTOMIDA

By A. E. Shipley, M.A., F.R.S., Fellow and Tutor of Christ’s College, Cambridge; Reader in Zoology

PYCNOGONIDA

By D’Arcy W. Thompson, C.B., M.A., Trinity College, Cambridge; Professor of Natural History in University College, Dundee

All the ingenious men, and all the scientific men, and all the fanciful men, in the world, with all the old German bogypainters into the bargain, could never invent ... anything so curious, and so ridiculous, as a lobster.

The Editors feel that they owe an apology and some explanation to the readers of The Cambridge Natural History for the delay which has occurred in the issue of this, the fourth in proper order, but the last to appear of the ten volumes which compose the work. The delay has been due principally to the untimely death of Professor W. F. E. Weldon, who had undertaken to write the Section on the Crustacea. The Chapter on the Branchiopoda is all he actually left ready for publication, but it gives an indication of the thorough way in which he had intended to treat his subject. He had, however, superintended the preparation of a number of beautiful illustrations, which show that he had determined to use, in the main, first-hand knowledge. Many of these figures have been incorporated in the article by Mr. Geoffrey Smith, to whom the Editors wish to express their thanks for taking up, almost at a moment’s notice, the task which had dropped from his teacher’s hand.

A further apology is due to the other contributors to this volume. Their contributions have been in type for many years, and owing to the inevitable delays indicated above they have been called upon to make old articles new, ever an ungrateful labour.

The appearance of this volume completes the work the Editors embarked on some sixteen years ago. It coincides with the cessation of an almost daily intercourse since the time when they “came up” to Cambridge as freshmen in 1880.

March 1909.

The Crustacea are almost exclusively aquatic animals, and they play a part in the waters of the world closely parallel to that which insects play on land. The majority are free-living, and gain their sustenance either as vegetable-feeders or by preying upon other animals, but a great number are scavengers, picking clean the carcasses and refuse that litter the ocean, just as maggots and other insects rid the land of its dead cumber. Similar to insects also is the great abundance of individuals which represent many of the species, especially in the colder seas, and the naturalist in the Arctic or Antarctic oceans has learnt to hang the carcasses of bears and seals over the side of the boat for a few days in order to have them picked absolutely clean by shoals of small Amphipods. It is said that these creatures, when crowded sufficiently, will even attack living fishes, and by sheer press of numbers impede their escape and devour them alive. Equally surprising are the shoals of minute Copepods which may discolour the ocean for many miles, an appearance well known to fishermen, who take profitable toll of the fishes that follow in their wake. Despite this massing together we look in vain for any elaborate social economy, or for the development of complex instincts among Crustacea, such as excite our admiration in many insects, and though many a crab or lobster is sufficiently uncanny in appearance to suggest unearthly wisdom, he keeps his intelligence rigidly to himself, encased in the impenetrable reserve of his armour and vindicated by the most powerful of pincers. It is chiefly in the variety of structure and in the multifarious phases of life-history that 4the interest of the Crustacea lies. Before entering into an examination of these matters, it will be well to take a general survey of Crustacean organisation, to consider the plan on which these animals are built, and the probable relation of this plan to others met with in the animal kingdom.

The Crustacea, to begin with, are a Class of the enormous Phylum Arthropoda, animals with metamerically segmented bodies and usually with externally jointed limbs. Their bodies are thus composed of a series of repeated segments, which are on the whole similar to one another, though particular segments may be differentiated in various respects for the performance of different functions. This segmentation is apparent externally, the surface of a Crustacean being divided typically into a number of hard chitinous rings, some of which may be fused rigidly together, as in the carapace of the crabs, or else articulated loosely.

Each segment bears typically a pair of jointed limbs, and though they vary greatly in accordance with the special functions for which they are employed, and may even be absent from certain segments, they may yet be reduced to a common plan and were, no doubt, originally present on all the segments.

Passing from the exterior to the interior of the body we find, generally speaking, that the chief system of organs which exhibits a similar repetition, or metameric segmentation, is the nervous system. This system is composed ideally of a nervous ganglion situated in each segment and giving off peripheral nerves, the several ganglia being connected together by a longitudinal cord. This ideal arrangement, though apparent during the embryonic development, becomes obscured to some extent in the adult owing to the concentration or fusion of ganglia in various parts of the body. The other internal organs do not show any clear signs of segmentation, either in the embryo or in the adult; the alimentary canal and its various diverticula lie in an unsegmented body-cavity, and are bathed in the blood which courses through a system of narrow canals and irregular spaces which surround all the organs of the body. A single pair, or at most two pairs of kidneys are present.

The type of segmentation exhibited by the Crustacea is thus of a limited character, concerning merely the external skin with its appendages, and the nervous system, and not touching any 5of the other internal organs.^[1] In this respect the Crustacea agree with all the other Arthropods, in the adults of which the segmentation is confined to the exterior and to the nervous system, and does not extend to the body-cavity and its contained organs; and for the same reason they differ essentially from all other metamerically segmented animals, e.g. Annelids, in which the segmentation not only affects the exterior and the nervous system, but especially applies to the body-cavity, the musculature, the renal, and often the generative organs. The Crustacea also resemble the other Arthropoda in the fact that the body-cavity contains blood, and is therefore a “haemocoel,” while in the Annelids and Vertebrates the segmented body-cavity is distinct from the vascular system, and constitutes a true “coelom.” To this important distinction, and to its especial application to the Crustacea, we will return, but first we may consider more narrowly the segmentation of the Crustacea and its main types of variation within the group. In order to determine the number of segments which compose any particular Crustacean we have clearly two criteria: first, the rings or somites of which the body is composed, and to each of which a pair of limbs must be originally ascribed; and, second, the nervous ganglia.

Around and behind the region of the mouth there is very little difficulty in determining the segments of the body, if we allow embryology to assist anatomy, but in front of the mouth the matter is not so easy.

In the Crustacea the moot point is whether we consider the paired eyes and first pair of antennae as true appendages belonging to two true segments, or whether they are structures sui generis, not homologous to the other limbs. With regard to the first antennae we are probably safe in assigning them to a true body-segment, since in some of the Entomostraca, e.g. Apus, the nerves which supply them spring, not from the brain as in more highly specialised forms, but from the commissures which pass round the oesophagus to connect the dorsally lying brain to the ventral nerve-cord. The paired eyes are always innervated from the brain, but the brain, or at least part of it, is very 6probably formed of paired trunk-ganglia which have fused into a common cerebral mass; and the fact that under certain circumstances the stalked eye of Decapods when excised with its peripheral ganglion^[2] can regenerate in the form of an antenna, is perhaps evidence that the lateral eyes are borne on what were once a pair of true appendages.

Now, with regard to the segmentation of the body, the Crustacea fall into three categories: the Entomostraca, in which the number of segments is indefinite; the Malacostraca, in which we may count nineteen segments, exclusive of the terminal piece or telson and omitting the lateral eyes; and the Leptostraca, including the single recent genus Nebalia, in which the segmentation of head and thorax agrees exactly with that of the Malacostraca, but in the abdomen there are two additional segments.

It has been usually held that the indefinite number of segments characteristic of the Entomostraca, and especially the indefinitely large number of segments characteristic of such Phyllopods as Apus, preserves the ancestral condition from which the definite number found in the Malacostraca has been derived; but recently it has been clearly pointed out by Professor Carpenter^[3] that the number of segments found in the Malacostraca and Leptostraca corresponds with extraordinary exactitude to the number determined as typical in all the other orders of Arthropoda. This remarkable correspondence (it can hardly be coincidence) seems to point to a common Arthropodan plan of segmentation, lying at the very root of the phyletic tree; and if this is so, we are forced to the conclusion that the Malacostraca have retained the primitive type of segmentation in far greater perfection than the Entomostraca, in some of which many segments have been added, e.g. Phyllopoda, while in others segments have been suppressed, e.g. Cladocera, Ostracoda. It may be objected to this view of the primitive condition of segmentation in the Crustacea that the Trilobites, which for various reasons are regarded as related to the ancestral Crustaceans, exhibit an indefinite and often very high number of segments; but, as Professor Carpenter has pointed out, the oldest and most primitive of Trilobites, such as Olenellus, possessed 7few segments which increase as we pass from Cambrian to Carboniferous genera.

The following table shows the segmentation of the body in the Malacostraca, as compared with that of Limulus (cf. p. 263), Insecta, the primitive Myriapod Scolopendrella, and Peripatus. It will be seen that the correspondence, though not exact, is very close, especially in the first four columns, the number of segments in Peripatus being very variable in the different species.

The appendages of the Crustacea exhibit a wonderful variety 8of structure, but these variations can be reduced to at most two, and possibly to one fundamental plan. In a typical Crustacean, besides the paired eyes, which may be borne on stalks, possibly homologous to highly modified limbs, there are present, first, two pairs of rod-like or filamentous antennae, which in the adult are usually specialised for sensory purposes, but frequently retain their primitive function as locomotory limbs even in the adult, e.g. Ostracoda; while in the Nauplius larva, found in almost all the chief subdivisions of the Crustacea, the two pairs of antennae invariably aid in locomotion, and the base of the second antennae is usually furnished with sharp biting spines which assist mastication. Following the antennae is a pair of mandibles which are fashioned for biting the food or for piercing the prey, and posterior to these are two pairs of maxillae, biting organs more slightly built than the mandibles, whose function it is to lacerate the food and prepare it for the more drastic action of the mandibles. So far, with comparatively few exceptions, the order of specialisation is invariable; but behind the maxillae the trunk-appendages vary greatly both in structure and function in the different groups.

As a general rule, the first or first few thoracic limbs are turned forwards toward the mouth, and are subsidiary to mastication; they are then called maxillipedes; this happens usually in the Malacostraca, but to a much less extent in the Entomostraca; and in any case these appendages immediately behind the maxillae never depart to any great extent from a limb-like structure, and they may graduate insensibly into the ordinary trunk-appendages. The latter show great diversity in the different Crustacean groups, according as the animals lead a natatory, creeping, or parasitic method of life; they may be foliaceous, as in the Branchiopoda, or biramous, as in the swimming thoracic and abdominal appendages of the Mysidae, or simply uniramous, as in the walking legs of the higher Decapoda, and the clinging legs of various parasitic forms.

Without going into detailed deviations of structure, many of which will be described under the headings of special groups, it is clear from the foregoing description and from Fig. 1 (p. 10), that three main types of appendage can be distinguished: first, the foliaceous or multiramous; second, the biramous; and, third, the uniramous.

9We may dismiss the uniramous type with a few words: it is obviously secondarily derived from the biramous type; this can be proved in detail in nearly every case. Thus, the uniramous second antennae of some adult forms are during the Nauplius stage invariably biramous, a condition which is retained in the adult Cladocera. Similarly the uniramous walking legs of many Decapoda pass through a biramous stage during development, the outer branches or exopodites of the limbs being suppressed subsequently, while the primitively biramous condition of the thoracic limbs is retained in the adults of the Schizopoda, which doubtless own a common ancestry with the Decapoda. The only Crustacean limb which appears to be constantly uniramous both in larval and adult life is the first pair of antennae.

We are reduced, therefore, to two types—the foliaceous and biramous. Sir E. Ray Lankester,^[6] in one of his most incisive morphological essays, has explained how these two types are really fundamentally the same. He compares, for instance, the foliaceous first maxillipede (Fig. 1, A), or the second maxilla (Fig. 1, B) of a Decapod, e.g. Astacus, with the foliaceous thoracic limb of Branchipus (Fig. 1, D), and with the typically biramous first maxillipede of a Schizopod (Fig. 1, F).

In each case there is present, on the outer edge of the limb, one or more projections or epipodites which are generally specialised for respiratory purposes, and may carry the gills. The 6th and 5th “endites” in the foliaceous limb (Fig. 1, D) are compared with the exopodite and endopodite respectively of the biramous limb, while the endites 4–1 of the foliaceous limb are found in the basal joints of the biramous limb. Lankester presumes that the biramous type of limb throughout has been derived from the foliaceous type by the suppression of the endites 1–4, as discrete rami, and the exaggerated development of the endites 5 and 6, as above indicated.

10

Fig. 1.—Appendages of Crustacea (A-G) and Trilobita (H). A, First maxillipede of Astacus; B, second maxilla of Astacus; C, second walking leg of Astacus; D, thoracic limb of Branchipus; E, first maxillipede of Mysis; F, first maxillipede of Gnathophausia; G, thoracic limb of Nebalia; H, thoracic limb of Triarthrus. bp, basipodite; br, bract; cp, carpopodite; cxp, coxopodite; cx.s, coxopoditic setae; dp, dactylopodite; end, endopodite; ep, epipodite; ex, exopodite; ip, ischiopodite; mp, meropodite; pp, propodite; 1–6, the six endites.

The essential fact that the two types of limb are built on the same plan may be considered as established; but it may be urged that the biramous type represents this common plan more nearly than the foliaceous. It is, at any rate, certain that in the maxillipedes of the Decapoda we witness the conversion of the biramous type into the foliaceous by the expansion of the basal joints concomitantly with the assumption by the maxillipedes of masticatory functions. Thus in the Decapoda the first maxillipede is decidedly foliaceous owing to the expanded “gnathobases” (Fig. 1, A, bp, cxp), and the second maxillipedes are flattened, with their basal joints somewhat expanded and furnished with biting hairs; but in the “Schizopoda” 11(e.g. Mysis) the first maxillipede is a typical biramous limb, though the expanded gnathobases in some forms are beginning to project (Fig. 1, E), while the limb following, which corresponds to the second maxillipede of Decapods, is simply a biramous swimming leg. Besides this obvious conversion of a biramous into a foliaceous limb, further evidence of the fundamental character of the biramous type is found, first, in its invariable occurrence in the Nauplius stage, which does not necessarily mean that the ancestors of the Crustacea possessed this type of limb in the adult, but which does imply that this type of limb was possessed at some period of life by the common ancestral Crustacean; and, second, the limbs of the Trilobita, a group which probably stands near the origin of the Crustacea, have been shown by Beecher to conform to the biramous type (Fig. 1, H). Furthermore, the thoracic limbs of Nebalia, an animal which combines many of the characteristics of Entomostraca and Malacostraca, and is therefore considered as a primitive type, despite their flattened character, are really built upon a biramous plan (Fig. 1, G).

In conclusion, we may point out that this view of the Crustacean limb, as essentially a biramous structure, agrees with the conclusion derived from our consideration of the segmentation of the body, and points less to the Branchiopoda as primitive Crustacea and more to some generalised Malacostracan type.

So far we have shortly dealt with those systems of organs which are clearly affected by the metameric segmentation of the body; we must now expose the condition of the body-cavity to a similar scrutiny. If we remove the external integument of a Crustacean, we find that the internal organs do not lie in a spacious and discrete body-cavity, as is the case in the Annelids and Vertebrates, but that they are packed together in an irregular system of spaces (“haemocoel”) in communication with the vascular system and containing blood. In the Entomostraca and smaller forms generally, a definite vascular system hardly exists, though a central heart and artery may serve to propel the blood through the irregular lacunae of the body-cavity; but in the larger Malacostraca a complicated system of arteries may be present which pour the blood into fairly definitely arranged spaces surrounding the chief organs. These spaces return the 12blood to the pericardium, and so to the heart again through the apertures or ostia which pierce its walls.

This condition of the body-cavity or haemocoel is reproduced in the adults of all Arthropods, but in some of them by following the development we can trace the steps by which the true coelom is replaced by the haemocoel. In the embryos of all Arthropods except the Crustacea, a true closed metamerically segmented coelom is formed as a split in the mesodermal embryonic layer of cells, distinct from the vascular system. During the course of development the segmented coelomic spaces and their walls give rise to the reproductive organs and to certain renal organs in Peripatus, Myriapoda, and Arachnida (nephridia and coxal glands), but the general body-cavity is formed as an extension of the vascular system, which is laid down outside the coelom by a canaliculisation of the extra-coelomic mesoderm. In the embryos of the Crustacea, however, there is never at any time a closed segmented coelom, and in this respect the Crustacea differ from all other Arthropods. The only clear instance in which metamerically repeated mesodermal cavities have been seen in the embryo Crustacean is that of Astacus; here Reichenbach^[7] states that in the abdomen segmental cavities are formed which subsequently break down; but even in this instance no connexion has been shown to subsist between these embryonic cavities and the reproductive and excretory organs of the adult.

Since the connexion between the coelom and the excretory organs is always a very close one throughout the animal kingdom, interest naturally centres upon the renal organs in Crustacea, and it has been suggested that these organs in Crustacea represent the sole remains, with the possible exception of the gonads, of the coelom. Since, at any rate, a part of the kidneys appears to be developed as a closed sac in the mesoderm, and since they possess a possible segmental value, this suggestion is plausible; but, on the other hand, since there are never more than two pairs of kidneys, and since they are totally unconnected with the gonads or with any other indication of a segmented coelom, the suggestion remains purely hypothetical.

The renal organs of the Crustacea, excluding the Malpighian tubes present in some Amphipods which open into the alimentary canal, and resemble the Malpighian tubes of Insects, consist of 13two pairs—the antennary gland, opening at the base of the second antenna, and the maxillary gland, opening on the second maxilla. These two pairs of glands rarely subsist together in the adult condition, though this is said to be the case in Nebalia and possibly Mysis; the antennary glands are characteristic of adult Malacostraca^[8] and the larvae of the Entomostraca, while the maxillary glands (“shell-glands”) are present in adult Entomostraca and larval Malacostraca, that is to say, the one pair replaces the other in the two great subdivisions of the Crustacea. The shell-gland of the Entomostraca is a simple structure consisting of a coiled tube opening to the exterior on the external branch of the second maxilla, and ending blindly in a dilated vesicle, the end-sac. The antennary gland of the Malacostraca is usually more complicated: these complications have been studied especially by Weldon,^[9] Allen, and Marchal^[10] in the Decapoda. In a number of forms we have a tube opening to the exterior at the base of the second antenna, and expanding within to form a spacious bladder into which the coiled tubular part of the kidney opens, while at the extremity of this coiled portion is the vesicle called the end-sac. This arrangement may be modified; thus in Palaemon Weldon described the two glands as fusing together above and below the oesophagus, the dorsal commissure expanding into a huge sac stretching dorsally down the length of the body. This closed sac with excretory functions thus comes to resemble a coelomic cavity, and the view that it is really coelomic has indeed been upheld.

A modified form of this view is that of Vejdovský, who describes a funnel-apparatus leading from the coiled tube into the end-sac of the antennary gland of Amphipods; he regards the end-sac alone as representing the coelom, while the funnel and coiled tube represent the kidney opening into it.

Not very much is known of the development of these various structures. Some authors have considered that both antennary and maxillary glands are developed in the embryo from ectodermal inpushings, but the more recent observations of Waite^[11] on Homarus americanus indicate that the antennary gland at 14any rate is a composite structure, formed by an ectodermal ingrowth which meets a mesodermal strand, and from the latter are produced the end-sac and perhaps the tubular excretory portions of the gland with their derivatives.

With regard to the possible metameric repetition of the renal organs, it is of interest to note that by feeding Mysis and Nebalia on carmine, excretory glands of a simple character were observed by Metschnikoff situated at the bases of the thoracic limbs.

The alimentary canal of the Crustacea is a straight tube composed of three parts—a mid-gut derived from the endoderm of the embryo, and a fore- and hind-gut formed by ectodermal invaginations in the embryo which push into and fuse with the endodermal canal. The regions of the fore- and hind-gut can be recognised in the adult by the fact of their being lined with the chitinous investment which is continued over the external surface of the body forming the hard exoskeleton, while the mid-gut is naked. The chitinous lining of fore- and hind-gut is shed whenever the animal moults. In the Malacostraca, in which a complicated “gastric mill” may be present, the chitinous lining of this part of the gut is thrown into ridges bearing teeth, and this stomach in the crabs and lobsters reaches a high degree of complication and materially assists the mastication of the food. The gut is furnished with a number of secretory and metabolic glands; the so-called liver, which is probably a hepatopancreas, opening into the anterior end of the mid-gut, is directed forwards in most Entomostraca and backwards in the Malacostraca, in the Decapoda developing into a complicated branching organ which fills a large part of the thorax. In the Decapoda peculiar vermiform caeca of doubtful function are present, a pair of which open into the gut anteriorly where fore-passes into mid-gut, and a single asymmetrically placed caecum opens posteriorly into the alimentary tract where mid- passes into hind-gut.

The disposition of these caeca, marking as they do the morphological position of fore-, mid-, and hind-gut, is of peculiar interest owing to the variations exhibited. From some unpublished drawings of Mr. E. H. Schuster, which he kindly lent me, it appears that in certain Decapods, e.g. Callianassa subterranea, the length of the mid-gut between the anterior and posterior caeca is very long; in Carcinus maenas it is considerable; 15in Maia squinado it is greatly reduced, the caeca being closely approximated; while in Galathea strigosa the caeca are greatly reduced, and the mid-gut as a separate entity has almost disappeared. The relation of these variations to the habits of the different crabs and to their modes of development is unknown.

The reproductive organs usually make their appearance as a small paired group of mesodermal cells in the thorax comparatively late in life; and neither in their early development nor in the adult condition do they show any clear signs of segmentation or any connexion with a coelomic cavity. The sexes are usually separate, but hermaphroditism occurs sporadically in many forms, and as a normal condition in some parasitic groups (see pp. 105–107). The adult gonads are generally simple paired tubes, from the walls of which the germ-cells are produced, and as these grow and come to maturity they fill up the cavities of the tubes; special nutrient cells are rarely differentiated, though in some cases (e.g. Cladocera) a few ova nourish themselves by devouring their sister-cells (see p. 44). The oviducts and vasa deferentia are formed as simple outgrowths from the gonadial tubes, which acquire an opening to the exterior; they are usually poorly supplied with accessory glands, the epithelium of the canals often supplying albuminous secretions for cementing the eggs together, while the lining of the vasa deferentia may be instrumental in the formation of spermatophores for transferring large packets of spermatozoa to the female. In the vast majority of Crustacea copulation takes place, the male passing spermatophores or free spermatozoa into special receptacles (spermathecae), or into the oviducts of the female. The spermatophores are hollow chitinous structures in which the spermatozoa are packed; they are often very large and assume characteristic shapes, especially in the Decapoda.

The spermatozoa show a great variety of structure, but they conform to two chief types—the filiform, which are provided with a long whip-like flagellum; and the amoeboid, which are furnished with radiating pseudopodia, and are much slower in their movements. The amoeboid spermatozoa of some of the Decapoda contain in the cell-body a peculiar chitinous capsule, and Koltzoff^[12] has observed that when the spermatozoon has 16settled upon the surface of the egg the chitinous capsule becomes suddenly exceedingly hygroscopic, swells up, and explodes, driving the head of the spermatozoon into the egg. We cannot enter here into a description of the embryological changes by which the egg is converted into the adult form. Crustacean eggs as a whole contain a large quantity of yolk, but in some forms total segmentation occurs in the early stages, which is converted later into the pyramidal type, i.e. the blastomeres are arranged round the edge, and the yolk in the centre is only partly segmented to correspond with them. The eggs during the early stages of development are in almost all cases (except Branchiura, p. 77, and Anaspides, p. 116) carried about by the female either in a brood-pouch (Branchiopoda, Ostracoda, Cirripedia, Phyllocarida, Peracarida), or agglutinated to the hind legs or some other part of the body (Copepoda, Eucarida), or in a chamber formed from the maxillipedes (Stomatopoda). Development may be direct, without a complicated metamorphosis, or indirect, the larva hatching out in a form totally different to the adult state, and attaining the latter by a series of transformations and moults. The various larval forms will be described under the headings of the several orders.

The respiratory organs are typically branchiae, i.e. branched filamentous or foliaceous processes of the body-surface through which the blood circulates, and is brought into close relation with the oxygen dissolved in the water. In most of the smaller Entomostraca no special branchiae are present, the interchange of gases taking place over the whole body-surface; but in the Malacostraca the gills may reach a high degree of specialisation. They are usually attached to the bases of the thoracic limbs (“podobranchiae”), to the body-wall at the bases of these limbs, often in two series (“arthrobranchiae”), and to the body-wall some way above the limb-articulations (“pleurobranchiae”). In an ideal scheme each thoracic appendage beginning with the first maxillipede would possess a podobranch, two arthrobranchs, and a pleurobranch, but the full complement of gills is never present, various members of the series being suppressed in the various orders, and thus giving rise to “branchial formulae” typical of the different groups.

After this brief survey of Crustacean organisation we 17may be able to form an opinion upon the position of the Crustacea relative to other Arthropoda, and upon the question debated some time ago in the pages of Natural Science^[13] whether the Arthropoda constitute a natural group. The Crustacea plainly agree with all the other Arthropoda in the possession of a rigid exoskeleton segmented into a number of somites, in the possession of jointed appendages metamerically repeated, some of which are modified to act as jaws; they further agree in the general correspondence of the number of segments of which the body is primitively composed; the condition of the body-cavity or haemocoel is also similar in the adult state. An apparently fundamental difference is found in the entire absence during development of a segmented coelom, but since this organ breaks down and is much reduced in all adult Arthropods, it is not difficult to believe that its actual formation in the embryo as a distinct structure might have been secondarily suppressed in Crustacea.

The method of breathing by gills is paralleled by the respiratory structures found in Limulus and Scorpions; the transition, if it occurred, from branchiae to tracheae cannot, it is true, be traced, but the separation of Arthropods into phyletically distinct groups of Tracheata and Branchiata on this single characteristic is inadmissible. On the whole the Crustacea may be considered as Arthropods whose progenitors are to be sought for among the Trilobita, from whose near relations also probably sprang Limulus and the Arachnids.

SUB-CLASS I.—ENTOMOSTRACA.

The Entomostraca are mostly small Crustacea in which the segmentation of the body behind the head is very variable, both in regard to the number of segments and the kind of differentiation exhibited by those segments and their appendages. An unpaired simple eye, known as the Nauplius eye from its universal presence in that larval form, often persists in the adult, and though lateral compound eyes may be present they are rarely borne on movable stalks. In the adult the excretory gland (“shell-gland”) opens on the second maxillary segment, but in the larval state or early stages of development a second antennary gland may also be present, which disappears in the adult. The liver usually points forwards, and is simple and saccular in structure, and the stomach is not complicated by the formation of a gastric mill. With the exception of most Cladocera and Ostracoda the young hatch out in the Nauplius state.

Order I. Branchiopoda.^[14]

The Branchiopods are of small or moderate size, with flattened and lobate post-cephalic limbs, and with functional gnathobases. Median and lateral eyes are nearly always present. The labrum is large, and the second maxillae are small or absent in the adult.

Branchiopods are found in every part of the world; a few are marine, but the great majority are confined to inland lakes and ponds, or to slowly-moving streams. The fresh waters, from the 19smallest pools to the largest lakes, often swarm with them, as do those streams which flow so slowly that the creatures can obtain occasional shelter among vegetation along the sides and bottom without being swept away, while even rivers of considerable swiftness contain some Cladocera. Several Branchiopods are found in the brackish waters of estuaries, and some occur in lakes and pools so salt that no other Crustacea, and few other animals of any kind, can live in them. The great majority swim about with the back downwards, collecting food in the ventral groove between their post-oral limbs, and driving it forwards, towards the mouth, by movements of the gnathobases (p. 10). The food collected in this way consists largely of suspended organic mud, together with Diatoms and other Algae, and Infusoria; the larger kinds, however, are capable of gnawing objects of considerable size, Apus being said to nibble the softer insect larvae, and even tadpoles. Many Cladocera (e.g. Daphnia, Simocephalus) may be seen to sink to the bottom of an aquarium, with the ventral surface downwards, and to collect mud, or even to devour the dead bodies of their fellows, while Leptodora is said to feed upon living Copepods, which it catches by means of its antennae.

The Branchiopoda fall naturally into two Sub-orders, the Phyllopoda including a series of long-bodied forms, with at least ten pairs of post-cephalic limbs, and the Cladocera with shorter bodies and not more than six pairs of post-cephalic limbs.

Sub-Order 1. Phyllopoda.

The Phyllopoda include a series of genera which differ greatly in appearance, owing to differences in the development of the carapace, which are curiously correlated with differences in the position of the eyes. Except in these points, the three families which the sub-order contains are so much alike that they may conveniently be described together.

In the Branchipodidae the carapace is practically absent, being represented only by the slight backward projection on each side of the head which contains the kidney (Fig. 2); the paired eyes are supported on mobile stalks, and project freely, one on either side of the head.

In the Apodidae^[15] the head is broad and depressed, the ventral 20side being nearly flat, the dorsal surface convex; the hinder margin of the head is indicated dorsally by a transverse cervical ridge, bounded by two grooves, behind which the carapace projects backwards as a great shield, covering at least half the body, but attached only to the back of the head. In Lepidurus productus the head and carapace together form an oval expansion, deeply emarginate at the hinder, narrower end, the sides of the emargination being toothed. The carapace has a strong median keel. The kidneys project into the space between the folds of skin which form the carapace, and their coils can be seen on each side, the terminal part of each kidney-tube entering the head to open at the base of the second maxilla. In all Branchiopoda with a well-developed carapace the kidney is enclosed in it in this way, whence the older anatomists speak of it as the “shell-gland.”

Fig. 2.—Chirocephalus diaphanus, female, × 5, Sussex. D.O, Dorsal organ; H, heart; Ov, ovary; U, uterus; V, external generative opening.

Associated with the development of the carapace, in this and in the next family, is a remarkable condition of the lateral eyes, which are sessile on the dorsal surface of the head, and near the middle line, the median eye being slightly in front of them. During embryonic life a fold of skin grows over all three eyes, so that a chamber is formed over them, which communicates with the exterior by a small pore in front.

In the Limnadiidae the body is laterally compressed, and the carapace is so large that at least the post-cephalic part of the body, and generally the head also, can be enclosed within it.

Fig. 3.—Limnetis brachyura, × 15. (After G. O. Sars.)

In Limnetis (Fig. 3) the dorsal surface of the head is bent downwards and is much compressed, the carapace being attached 21to it only for a short distance near the dorsal middle line. The sides of the carapace are bent downwards, and their margins can be pulled together by a transverse adductor muscle, so that the whole structure forms an ovoid or spheroidal case, from which the head projects in front, while the rest of the body is entirely contained within it. When the adductor muscle is relaxed the edges of the carapace gape slightly, like the valves of a Lamellibranch shell, and food-particles are drawn through the opening thus formed into the ventral groove by the movements of the thoracic feet, locomotion being chiefly effected by the rowing action of the second antennae, as in the Cladocera, to which all the Limnadiidae present strong resemblances in their method of locomotion, in the condition of the carapace, and in the form of the telson.

In Limnadia and Estheria the carapace projects not only backwards from the point of attachment to the head, but also forwards, so that the head can be enclosed by it, together with the rest of the body.

In all these genera the carapace is flexible along the middle dorsal line; in Estheria especially the softening of the dorsal cuticle goes so far that a definite hinge-line is formed, and this, together with the deposition of the lateral cuticle in lines concentrically arranged round a projecting umbo, gives the carapace a strong superficial likeness to a Lamellibranch shell, for which it is said to be frequently mistaken by collectors.

The eyes of the Limnadiidae are enclosed in a chamber formed by a growth of skin over them, as in Apodidae, but the pore by which this chamber communicates with the exterior is even more minute than in Apus. The paired eyes are so close together that they may touch (Limnadia, Estheria) or fuse (Limnetis); they are farther back than in the Apodidae, while the ventral curvature of the head causes the median eye to lie below them. In all 22these points the eyes of the Limnadiidae are intermediate between those of Apus and those of the Cladocera.

Dorsal Organ.—A structure very characteristic of adult Phyllopods is the “dorsal organ” (Figs. 2, 5, D.O), whose function is in many cases obscure. It is always a patch of modified cephalic ectoderm, supplied by a nerve from the anterior ventral lobe of the brain on each side; but its characters, and apparent function, differ in different forms. In the Branchipodidae the dorsal organ is a circular patch, far forward on the surface of the head (Figs. 2, 5, D.O). Its cells are arranged in groups, which remind one of the retinulae in a compound eye; each cell contains a solid concretion, and the concretions of a group may be so placed as to look like a badly-formed rhabdom. Claus,^[16] who first called attention to this structure in the Branchipodidae, regarded it as a sense-organ. In Apodidae the dorsal organ is an oval patch of columnar ectoderm, immediately behind the eyes; it is slightly raised above the surrounding skin, and is covered by a very delicate cuticle (with an opening to the exterior?), and below it is a mass of connective tissue permeated by blood; Bernard has suggested that it is an excretory organ.

Most Limnadiidae resemble the Cladocera in the possession of a “dorsal organ” quite distinct from the above; in Limnetis and Estheria it has the form of a small pit, lined by an apparently glandular ectoderm, and this is its condition in many Cladocera; in Limnadia lenticularis it is a patch of glandular epithelium on a raised papilla. Limnadia has been observed to anchor itself to foreign objects by pressing its dorsal organ against them, and many Cladocera do the same thing; Sida crystallina, for example, will remain for hours attached by its dorsal organ to a waterweed or to the side of an aquarium. Structures resembling a dorsal organ occur in the larvae of many other Crustacea, but the presence of this organ in the adult is confined to Branchiopods, and indeed in many Cladocera it disappears before maturity. It is certain that the sensory and adhesive types of dorsal organ are not homologous, especially as rudimentary sense-organs may exist on the head of Cladocera together with the adhesive organ.

The telson differs considerably in the different genera. In the Branchipodidae^[17] the anus opens directly backwards; and 23the telson carries two flattened backwardly directed plates, one on each side of the anus, the margins of each plate being fringed with plumose setae. In Artemia the anal plates are rarely as large as in Branchipus, and never have their margins completely fringed with setae; in A. salina from Western Europe, and in A. fertilis (Fig. 4, A) from the Great Salt Lake of Utah, there is a variable number of setae round the apical half of each lobe, but in specimens of A. salina from Western Siberia the number of setae may be very small, or they may be absent; in the closely allied A. urmiana from Persia the anal lobes are well developed in the male, each lobe bearing a single terminal hair, but they are altogether absent in the female. Schmankewitch and Bateson have shown that there is a certain relation between the salinity of the water in which Artemia salina occurs and the condition of the anal lobes, specimens from denser waters having on the whole fewer setae; the relation is, however, evidently very complex, and further evidence is wanted before any more definite statements can be made.

Fig. 4.—A, Ventral view of the anal region in Artemia fertilis, from the Great Salt Lake; B, ventral view of the telson and neighbouring parts of Lepidurus productus; C, side view of the telson and left anal lobe of Estheria (sp.?).

In the Apodidae the anal lobes have the form of two-jointed cirri, often of considerable length; in Apus the anus is terminal, but in Lepidurus (Fig. 4, B) the dorsal part of the telson is prolonged backwards, so as to form a plate, on the ventral face of which the anus opens, much as in the Malacostraca.

In the Limnadiidae (Fig. 4, C) the telson is laterally compressed 24and produced, on each side of the anus, into a flattened, upwardly curved process, sharply pointed posteriorly, and often serrate; the anal lobes are represented by two stout curved spines, while in place of the dorsal prolongation of Lepidurus we find two long plumose setae above the anus. In the characters of the telson and anal lobes, as in those of the head, the Limnadiidae approximate to the Cladocera. In Limnetis brachyura the ventral face of the telson is produced into a plate projecting backwards below the anus, in a manner which has no exact parallel among other Crustacea.

The appendages of the Phyllopoda are fairly uniform in character, except those affected by the sexual dimorphism, which is usually great.

Fig. 5.—Chirocephalus diaphanus, male. Side view of head, showing the large second antenna, A₂, with its appendage Ap, above which is seen the filiform first antenna; D.O, dorsal organ; E₁, median eye.

25

Fig. 6.—Chirocephalus diaphanus. Second antenna of male, uncoiled.

Of the cephalic appendages, the first antennae are generally small, and are never biramous; in Branchipus and its allies they are simple unjointed rods, in some species of Artemia they are three-jointed, in Apus they are feebly divided into two joints, while in Estheria they are many-jointed. The second antennae are the principal organs of locomotion in the Limnadiidae, where they are large and biramous; in all other Phyllopoda they are uniramous in the female, being either unjointed triangular plates as in Chirocephalus (Fig. 2), or minute vestigial filaments as in Apus, in which genus Zaddach, Huxley, and Claus have all failed to find any trace of a second antenna in some females. In the male Branchipodidae the second antennae are modified to form claspers, by which the female is seized, the various degrees of complication which these claspers exhibit affording convenient generic characters. In Branchinecta each second antenna is a thick, three-jointed rod, the last joint forming a claw, while the second joint is serrate on its inner margin; in Branchipus the base is much thickened, and bears on its inner side a large filament (perhaps represented by the proximal tubercle of Branchinecta and Artemia), which looks like an extra antenna. In Streptocephalus the terminal joint of the antenna is bifid, and there is a basal filament like that of Branchipus; in Chirocephalus diaphanus (Figs. 5, 6) the main branch of the antenna consists of two large joints, the terminal joint being a strong claw with a serrated process at its base, while the proximal joint bears two appendages on its inner side; one of these is a small, subconical tubercle, the second is more complicated, consisting of a main stem and five outgrowths. The main stem is many-jointed and flexible, its basal joint being longer than the others, and bearing on its outer side a large, triangular, membranous appendage, and four soft cylindrical appendages, the main stem and its appendages being beset with curious tubercles, ending in short spines, whose structure is not understood. Except during the act of copulation this remarkable apparatus is coiled on the inner side of the antennary claw, the jointed stem being so coiled that it is often compared to the 26coiled proboscis of a butterfly, and the triangular membrane folded like a fan beside it, so that much of the organ is concealed, and the general appearance of the head is that shown in Fig. 5. During copulation, the whole structure is widely extended.

Fig. 7.—Artemia fertilis. Front view of the head of a male, showing the large second antennae, A.2; A.1, first antennae.

The males of Artemia (Fig. 7) have the second antenna two-jointed, the basal joint bearing an inner tubercle, the terminal joint being flattened and bluntly pointed, its outer margin provided with a membranous outgrowth. In A. fertilis the breadth of the second joint varies greatly, the narrower forms presenting a certain remote resemblance to Branchinecta. In the males of Polyartemia the second antennae have a remarkable branched form not easily comparable with that found in other Branchipodidae.

The cephalic jaws are fairly uniform throughout the order. The mandibles have an undivided molar surface, and no palp; the first maxilla is very generally a triangular plate, with a setose biting edge; mandibles and maxillae are covered by the labrum. The second maxilla generally lies outside the chamber formed by the labrum, and is a simple oval plate, with or without a special process for the duct of the kidney.

The thoracic limbs, in front of the genital segments, are not as a rule differentiated into anterior maxillipedes and posterior locomotive appendages, as in higher forms; we have seen, however, that all these limbs take part in the prehension of food, and except in the Limnadiidae they all assist in locomotion. One of the middle thoracic legs of Artemia (Fig. 8, A) has a flattened stem, with seven processes on its inner, and two on its outer margin. The gnathobase (gn) is large, and fringed with long plumose setae, each of which is jointed; this is followed by four smaller “endites” (or processes on the median side), and then by two larger ones, the terminal endite (the sixth, 27excluding the gnathobase) being very mobile and attached to the main stem by a definite joint. On the outer side are two processes; a proximal “bract,” a flat plate with crenate edges, partly divided by a constriction into two, and a distal process, cylindrical and vascular, called by Sars and others the “epipodite.” In other Branchipodidae we have essentially the same condition, except that the fifth endite often becomes much larger than in Artemia, throwing the terminal endite well over to the outer edge of the limb; such a shift as this, continued farther, might well lead to the condition found in the Limnadiidae, or Apodidae, where the lobe which seems to represent the terminal endite of Artemia is entirely on the outer border of the limb, forming what most writers have called the exopodite (Lankester’s “flabellum”).^[18] In the two last-named families the basal exite or bract of the Branchipodidae does not appear to be represented.

Fig. 8.—A, Thoracic limb of Chirocephalus diaphanus; B, prehensile thoracic limb of male Estheria. gn, Gnathobase; 1–6, the more distal endites.

The limbs of the Apodidae are remarkable in two ways; those in front of the genital opening (very constantly ten pairs) 28are not so nearly alike as in most genera of the sub-order, the first two pairs especially having the axis definitely jointed, while the endites are elongated and antenniform; further, while the first eleven segments bear each a single pair of limbs, as is usual among Crustacea, many of the post-genital segments bear several pairs; thus in Apus cancriformis there are thirty-two post-cephalic segments in front of the telson, the first eleven having each one pair of limbs, while the next seventeen have fifty-two pairs between them, the last four segments having none.

In all the Phyllopoda some of the post-cephalic limbs are modified for reproductive purposes; in the Branchipodidae the last two pairs (the 12th and 13th generally, the 20th and 21st in Polyartemia) are so modified in both sexes. In the female these appendages fuse at an early period of larval life, and surround the median opening of the generative duct (Fig. 2); in the male the two pairs also fuse, but traces of the limbs are left as eversible processes round the paired openings of the vasa deferentia.

In the other families, one or more limbs of the female are adapted for carrying or supporting the eggs. In the Apodidae the appendages of the eleventh segment have the exopodite in the form of a rounded, watchglass-shaped plate, fitting over a similarly shaped process of the axis of the limb, so that a lens-shaped box is formed, into which the eggs pass from the oviduct. In Limnadiidae the eggs are carried in masses between the body and the carapace, and are kept in position by special elongations of the exopodites of two or three legs, either those near the middle of the thorax (Estheria, Limnadia), or at its posterior end (Limnetis). In female Limnetis the last thoracic segments bear two remarkable lateral plates, which apparently also help to support the eggs. In the male Limnadiidae, the first (Limnetis) or the first two thoracic feet (Limnadia, Estheria) are prehensile (Fig. 8, B).

Alimentary Canal.—The mouth of the Phyllopoda is overhung by the large labrum, so that a kind of atrium is formed, outside the mouth itself, in which mastication is performed; numerous unicellular glands, opening on the oral face of the labrum, pour their secretion into the atrial chamber, and may be called salivary, though the nature of their secretion is not known. The mouth has commonly two swollen and setose 29lips, running longitudinally forwards from the bases of the first maxillae, and often wrapping round the blades of the mandibles. It leads into a vertical oesophagus, which opens into a small globular stomach, lying entirely within the head; the terminal part of the oesophagus is slightly invaginated into the stomach, so that a valvular ring is formed at the junction of the two. The stomach opens widely behind into a straight intestine, which runs backwards to about the level of the telson, where it joins a short rectum, leading to the terminal or ventral anus. The stomach and intestine are lined by a columnar epithelium, and covered by a thin network of circularly arranged muscle-fibres; the rectum has a flatter epithelium, and radial muscles pass from it to the body-wall, so that it can be dilated. The only special digestive glands are two branched glandular tubes, situated entirely within the head, which open into the stomach by large ducts, one on each side. In Chirocephalus the gastric glands are fairly small and simple; in the Apodidae their branches are more complex and form a considerable mass, filling all that portion of the head which is not occupied by the nervous system and the muscles. Backwardly directed gastric glands, like those of the higher Crustacea, are not found in Branchiopods; both forms occur together in the genus Nebalia, but with this exception the forwardly-directed glands are peculiar to Branchiopods.

Heart.—In Branchipus and its allies, and in Artemia, the heart extends from the first thoracic segment to the penultimate segment of the body, and is provided with eighteen pairs of lateral openings, one pair in every segment through which it passes except the last; it is widely open at its hinder end, and is prolonged in front for a short distance as a cephalic aorta, the rest of the blood-spaces being lacunar.

In most, at least, of the other Branchiopods, the heart is closed behind and is shortened; in Apus and Lepidurus it only extends through the first eleven post-cephalic segments, while in the Limnadiidae it is shorter still, the heart of Limnetis passing through four segments only. In all cases there is a pair of lateral openings in every segment traversed by the heart.

The blood of the Branchipodidae and Apodidae contains dissolved haemoglobin, the quantity present being so small as to give but a faint colour to the blood in Branchipus, while 30Artemia has rather more, and the blood of Apus is very red. The only other Crustacea in which the blood contains haemoglobin are the Copepods of the genus Lernanthropus,^[19] so that the appearance of this substance is as irregular and inexplicable in Crustacea as in Chaetopods and Molluscs.

The nervous system of Branchipus may be described as an illustration of the condition prevailing in the group. The brain consists of two closely united ganglia, in each of which three main regions may be distinguished; a ventral anterior lobe, a dorsal anterior lobe, and a posterior lobe. The ventral anterior lobes give off nerves to the median eye, to the dorsal organ, and to a pair of curious sense-organs, comparable with the larval sense-knobs of many higher forms, situated one on each side of the median eye; in late larvae Claus describes the terminal apparatus of each frontal sense-organ as a single large hypodermic cell; W. K. Spencer^[20] has lately described several terminal cells, containing peculiar chitinous bodies, in the adult. The homologous sense-organs of Limnetis are apparently olfactory. The dorsal anterior lobes give off the large nerves to the lateral eyes, while the posterior lobes supply the first antennae. The oesophageal connectives have a coating of ganglion-cells, and some of these form the ganglion of the second antenna, the nerve to this appendage leaving the connective just behind the brain. The post-oral nerve-cords are widely separate, each of them dilating into a ganglion opposite every appendage, the two ganglia being connected by two transverse commissures. The ganglia of the three cephalic jaws, so often fused in the higher Crustacea, are here perfectly distinct. Closely connected with each thoracic ganglion is a remarkable unicellular gland, opening to the exterior near the middle ventral line; it is conceivable that these cells may be properly compared with the larval nephridia of a Chaetopod,^[21] but no evidence in support of such a comparison has yet been adduced.

Behind the genital segments, where there are no limbs, the nerve-cords run backwards without dilating into segmental ganglia, except in the anterior two abdominal segments where 31small ganglionic enlargements occur. In Apodidae, on the other hand, those segments which carry more than one pair of appendages have as many pairs of ganglia, united by transverse commissures, as they have limbs.

A stomatogastric nervous system exists in Apus, where a nerve arises on each side from the first post-oral commissure, and runs forward to join its fellow of the opposite side on the anterior wall of the oesophagus. From the loop so formed a larger median and a series of smaller lateral nerves pass to the wall of the alimentary canal. A second nerve to the oesophagus is given off from the mandibular ganglion of each side.

Reproductive Organs.—In Chirocephalus the ovaries (Fig. 2, Ov) are hollow epithelial tubes, lying one on each side of the alimentary canal, and extending from the sixth abdominal segment forwards to the level of the genital opening; at this point the two ovaries are continuous with ducts, which bend sharply downwards and open into the single uterus contained within the projecting egg-pouch and opening to the exterior at the apex of that organ. Short diverticula of the walls of the uterus receive the ducts of groups of unicellular glands, the bodies of which contain a peculiar opaque secretion, said to form the eggshells. In Apodidae the ovaries are similar in structure, but they are much larger and branch in a complex manner, while each ovary opens to the exterior independently of the other in the eleventh post-cephalic segment; nothing like the median uterus of the Branchipodidae being formed. The epithelium of the ovarian tubes proliferates, and groups of cells are formed; one becoming an ovum, the others being nutrient cells like those which will be more fully described in the Cladocera.

In Chirocephalus the testes are tubes similar in shape and position to the ovaries, each communicating in front with a short vas deferens, which dilates into a vesicula seminalis on its way to the eversible penis; an essentially similar arrangement is found in all Branchipodidae, but in Apodidae and Limnadiidae there is no penis.

All the Branchiopoda are dioecious,^[22] and many are parthenogenetic. Among Branchipodidae Artemia is the only genus known to be parthenogenetic, but parthenogenesis is common in 32all Apodidae, while the males of several species of Limnadia are still unknown, although the females are sometimes exceedingly common. In Artemia, generations in which the males are about as numerous as the females seem to alternate fairly quickly with others which contain only parthenogenetic females; in Apus males are rarely abundant, and often absent for long periods; during five consecutive years von Siebold failed to discover a male in a locality in Bavaria, though he examined many thousands of individuals; near Breslau he found on one occasion about 11 per cent of males (114 in 1026), but in a subsequent year he found less than 1 per cent; the greatest recorded percentage of males is that observed by Lubbock in 1863, when he found 33 males among 72 individuals taken near Rouen.

The eggs of most genera can resist prolonged periods of desiccation, and indeed it seems necessary for the development of many species that the eggs should be first dried and afterwards placed in water. Many eggs (e.g. of Chirocephalus diaphanus and Branchipus stagnalis) float when placed in water after desiccation, the development taking place at the surface of the water.

Habitat.—All the Phyllopoda, except Artemia, are confined to stagnant shallow waters, especially to such ponds as are formed during spring rains, and dry up during the summer. In waters of this kind the species of Branchipus, Apus, etc., develop rapidly, and produce great numbers of eggs, which are left in the dried mud at the bottom after evaporation of the water, where they remain quiescent until a fresh rainy season. The mud from the beds of such temporary pools often contains large numbers of eggs, which may be carried by wind, on the legs of birds, and by other means, to considerable distances. Many exotic species have been made known to European naturalists by their power of hatching out when mud brought home by travellers is placed in water. The water of stagnant pools quickly dissolves a certain quantity of solid matter from the soil, and often receives dissolved solids through surface drainage from the neighbouring land; such salts may remain as the water evaporates, so that the water which remains after evaporation has proceeded for some time may be very sensibly denser than that in which the Branchiopods were hatched; these creatures must therefore be able to endure a considerable increase in the salinity of the surrounding waters during 33the course of their lives. My friend Mr. W. W. Fisher points out that the plants present in such a pond would often precipitate the carbonate of lime, so that this might be removed as evaporation went on, but that chlorides would probably remain in solution; from analyses which Mr. Fisher has been kind enough to make for me, it is seen that this happened in a small aquarium in my laboratory, in which Chirocephalus diaphanus lived for four months. In April, mud from the dry bed of a pond, known to contain eggs of Chirocephalus, was placed in this aquarium in Oxford, and water was added from the tap. Oxford tap-water contains about 0·3 grm. salts per litre, the chlorine being equivalent to 0·023 grm. NaCl. Water was added from time to time during May and June, but in July evaporation was allowed to proceed unchecked. At the end of July there was about half the original volume of water, the Chirocephalus being still active; the residue contained 0·96 grm. dissolved solids per litre, with chlorine equal to 0·19 grm. NaCl, so that the percentage of chlorides was about eight times the initial percentage, but there were only three and a fifth times the original amount of total solid matter in solution, the carbonate of lime having precipitated as a visible film.

Some species of Branchipus (e.g. B. spinosus, M. Edw.) and of Estheria (E. macgillivrayi, Baird, E. gubernator, Klutzinger) occur in salt pools, but Artemia flourishes in waters beside whose salinity that endured by any other Branchiopod is insignificant. In the South of Europe, Artemia salina may be found in swarms, as it used to be found in Dorsetshire, in the shallow brine-pans from which salt is commercially prepared; Rathke quotes an analysis showing that a pool in the Crimea contained living Artemia when the salts in solution were 271 grms. per litre, and the water was said to have the colour and consistency of beer.

The behaviour of the animals in the water differs a little; in normal feeding all the species swim with the back downwards, as has already been said; the Branchipodidae rarely settle on the ground, or on foreign objects, but the Apodidae occasionally wriggle along the bottom on their ventral surface, and Estheria burrows in mud.

The greater number of species are found in pools in flat, low-lying regions, and many appear to be especially abundant near 34the sea; Apus cancriformis has, however, been found in Armenia at 10,000 feet above sea level.

Wells and underground waters do not generally contain Phyllopods; but a species of Branchipus and one of Limnetis, both blind, have been described from the caves of Carniola.

One of the many puzzles presented by these creatures is the erratic way in which they are scattered through the regions they inhabit; a single small pond, a few yards or less in diameter, may be the only place within many miles in which a given species can be found; in this pond it may, however, appear regularly season after season for some time, and then suddenly vanish.

Geographically, the Phyllopoda are cosmopolitan, representatives of every family and of some genera (e.g. Streptocephalus, Lepidurus, Estheria) being found in every one of the great zoological regions, though a few aberrant genera are of limited range, thus Polyartemia is known only from the northern Palaearctic and Nearctic regions, Thamnocephalus only from the Central United States. The genus Artemia is not at present known in Australia.^[23] The only recorded British species are Chirocephalus diaphanus, Artemia salina, and Apus cancriformis,^[24] but other continental islands, for example the West Indian group, are better supplied. The distribution of the species is very imperfectly known, but on the whole every main zoological region seems to have its own peculiar species, which do not pass beyond its boundaries. Branchinecta paludosa and Lepidurus glacialis are circumpolar, both occurring in Norway, in Lapland, in Greenland, and in Arctic North America; but with these exceptions the Palaearctic and Nearctic species seem to be distinct. The European species Apus cancriformis occurs in Algiers, but the relations between the species of Northern Africa as a whole and those of Southern Europe on the one hand, or of Central and Southern Africa on the other, have yet to be worked out.

The soft-bodied Branchipodidae are not known in the fossil condition;^[25] an Apus, closely related to the modern A. cancriformis, has been found in the Trias, but the most numerous remains have been left, as might be expected, by the hard-shelled Limnadiidae; 35carapaces, closely resembling those of the modern Estheria, are known in beds of all ages from the Devonian period to recent times; these carapaces are in several cases associated with fossils of an apparently marine type. None of the fossil species differ in any important characters from those now living, so that the Phyllopoda have existed in practically their present form for an enormously long period; this fact, and the evidence that species of existing genera were at one time marine, explain the wide distribution of animals at present restricted to a remarkably limited range of environmental conditions.

Summary of the Characters of the Genera.

Sub-Order Phyllopoda.—Branchiopoda with an elongated body, provided with at least ten pairs of post-cephalic limbs, the heart extending through four or more thoracic segments, and having at least four pairs of ostia.

Fam. 1. Branchipodidae.^[26]—Carapace rudimentary, eyes stalked; the second antennae flat and unjointed in the female, jointed and prehensile in the male; female generative opening single; telson not laterally compressed, bearing two flattened lobes, or none. The heart extending through the thorax and the greater part of the abdomen.

A. Eleven pairs of praegenital ambulatory limbs.

a. Abdomen of six well-formed segments and a telson; anal lobes well formed, their margins setose.

Branchinecta, Verrill—Second antennae of ♂ without lateral appendages; ovisac of ♀ elongated. B. paludosa, O. F. Müll.—Circumpolar.

Branchiopodopsis, G. O. Sars^[27]—Second antennae of ♂ as in Branchinecta; ovisac of ♀ short. B. hodgsoni, G. O. Sars—Cape of Good Hope.

Branchipus, Schaeffer—Second antennae of ♂ with simple internal filamentous appendage. B. stagnalis, Linn.—Central Europe.

Streptocephalus, Baird—Second antennae of ♂ 3–jointed, the last joint bifid; an external filamentous appendage. S. torvicornis, Wagn., Poland.

Chirocephalus, Prévost—Second antennae of ♂ 3–jointed, with a jointed internal appendage, which bears secondary processes, four cylindrical and one lamellar. C. diaphanus, Prévost (Fig. 2, p. 20).—Britain, Central Europe.

b. Abdominal segments five or fewer, and a telson. Anal lobes small or 0, sparsely or not at all setose.

Artemia, Leach—Second antennae of ♂ without filamentous 36appendage, 2–jointed, the second joint lamellar. A. salina, Linn.—Brine pools of the Palaearctic region.

c. Hinder abdominal segments united with telson to form a fin; anal lobes absent.

Thamnocephalus, Packard—Head with a branched median process of unknown nature. Only species T. platyurus, Packard—Kansas, U.S.A.

B. Nineteen pairs of praegenital ambulatory limbs.

Polyartemia, Fischer—Second antennae of ♂ forcipate; ovisac of ♀ very short. Only species P. forcipata, Fisch.

Fam. 2. Apodidae.^[28]—Carapace well developed as a depressed shield, covering at least half the body. Eyes sessile, covered; no male clasping organs; anal lobes long, jointed cirri.

Apus, Scopoli—Telson not produced backwards over the anus; endites of first thoracic limb very long. A. cancriformis, Schaeffer—Britain, Europe, Algiers, Tunis. A. australiensis, Central Australia.

Lepidurus, Leach—Telson produced backwards to form a plate above the anus; endites of first thoracic limb short. L. productus, Bosc.—Central Europe. L. viridis, Southern Australia, New Zealand, L. patagonicus, Bergh, Argentines.

Fam. 3. Limnadiidae.—Body compressed; carapace in the form of a bivalve shell, the two halves capable of adduction by means of a strong transverse muscle; second antennae biramous, alike in both sexes; in the male, the first or the first and second thoracic limbs prehensile; telson laterally compressed.

A. Only the first thoracic limbs prehensile in the male; the carapace spheroidal, without lines of growth; head not included within the carapace-chamber.

Limnetis, Lovén—Compound eyes fused; anal spines absent; ambulatory limbs 10–12. L. brachyura, O. F. Müll (Fig. 3, p. 21).—Norway, Central Europe.

B. The first and second thoracic limbs prehensile in the male; carapace distinctly bivalve, enclosing the head, with concentric lines of growth round a more or less prominent umbo.

Eulimnadia, Packard—Carapace narrowly ovate, with few (4–5) lines of growth. E. mauritani, Guérin—Mauritius. E. texana, Packard—Texas, Kansas.

Limnadia, Brongniart—Carapace broadly ovate, with numerous lines of growth, without distinct umbones; L. lenticularis, Linn.—Northern and Central Europe.

Estheria, Rüppell—Carapace with well-marked umbones and numerous lines of growth, oval; E. tetraceros, Kryneki—Central Europe.

Leptestheria,^[29] G. O. Sars—Carapace compressed, oblong. Rostrum 37with a movable spine; thoracic limbs with accessory lappet on the exopodite. L. siliqua, G. O. Sars—Cape Town.

Cyclestheria,^[30] G. O. Sars. C. hislopi, Baird—Queensland, India, East Africa, Brazil.

Sub-Order 2. Cladocera.

The Cladocera are short-bodied Branchiopods, with not more than six pairs of thoracic limbs. The second antennae are important organs of locomotion, and are nearly always biramous; the first antennae are small, at least in the female; the second maxillae are absent in the adult. The carapace may extend backwards so as to enclose the whole post-cephalic portion of the body, or may be reduced to a small dorsal brood-pouch, leaving the body uncovered.

The Cladocera or “Water-fleas” are never of great size; Leptodora hyalina, the largest, is only about 15 mm. long, while many Lynceidae are not more than 0·1 or 0·2 mm. in length.

The head is bent downwards in all the Cladocera, so that parts which are morphologically anterior, such as the median eye and the first antennae, lie ventral to or even behind the compound eyes and the second antennae (cf. Fig. 10).

The compound lateral eyes fuse at an early period of embryonic life, so that they form a single median mass in the adult, over which a fold of ectoderm grows, to make a chamber over the eye, like that found in the Limnadiidae, except that it is completely closed. The fused eyes are generally large and conspicuous; in some deep-water forms the retinular elements of the dorsal portion are larger than those of the ventral (e.g. Bythotrephes, Fig. 13). In one or two species which live at very great depths, or in caves, the eyes are altogether absent.

The appendages of the head are fairly uniform, the most variable being the first antennae. In the females of many genera the first antennae are short and immovable, consisting of a single joint, with a terminal bunch of sensory hairs, and often a long lateral hair, as in Simocephalus (Figs. 9, 10), Daphnia, etc. In the female Moina (Fig. 16) they are movable, as they are in Ceriodaphnia and some others; in Bosmina (Fig. 22) and many Lyncodaphniidae they are elongated and imperfectly divided 38into joints by rings of spines, while in Macrothrix they are flattened plates. In the males the first antennae are elongated and mobile (cf. Figs. 11, 19).

Fig. 9.—Simocephalus vetulus, female. Ventral view, without the carapace; A₁, A₂, first and second antennae; For, head; Md, mandible; Te, telson; I-IV, first to fourth thoracic appendages.

The second antennae, the chief organs of locomotion, are biramous in all genera except Holopedium; the number of joints in each ramus, and the number of the long plumose hairs with which they are provided, are remarkably constant in whole series of genera, and are therefore useful for purposes of classification. The creatures row themselves by quick strokes of these appendages, the movement being slow and irregular in the rounder forms, such as Simocephalus or Daphnia, rapid and well directed in such elongated lacustrine forms as Bythotrephes or Leptodora.

The mandibles have no palp; the first maxillae are very small, and the second maxillae are absent (Fig. 9).

The carapace varies very much. In most genera (the Calyptomera of Sars) it is a large, backwardly-projecting fold of skin, bent downwards at the sides so as to form a bivalve shell, enclosing the whole post-cephalic portion of the body, as in Simocephalus (Fig. 10). The eggs are laid into the space between the carapace and the dorsal part of the thorax, both the carapace and the thorax itself being often modified for their protection and nutrition. In a few forms, the Gymnomera of Sars, the carapace serves only as a brood-pouch, which is distended when eggs are laid, but collapses to an inconspicuous appendage at the back of the head when it is empty (e.g. Leptodora, Fig. 24, Bythotrephes, Fig. 13). In the Calyptomera the surface of the carapace is frequently provided with a series of ridges, which may be parallel, rarely branching, as in Simocephalus; or in two sets which cross nearly at right angles, as in Daphnia; or so arranged as to form a hexagonal pattern, as 39in Ceriodaphnia. In a few forms the whole surface is irregularly covered with spines or scales. The hinder edge of the carapace is often produced into a median dorsal spine (Daphnia, Fig. 19), or more rarely there are two spines, one at each ventro-lateral corner (Scapholeberis, Fig. 20).

Fig. 10.—Simocephalus vetulus, × 30. Side view of female, showing the arrangement of the principal organs. A.2, Second antenna; C.S, cervical suture; E, fused compound eyes; H, heart; L, forwardly-directed gastric caeca; N, dorsal organ.

The cuticle of the carapace is often separated from that of the head by a cervical suture, as in Simocephalus (Fig. 10, C.S.) and near the line of demarcation many forms exhibit patches of glandular ectoderm which seem to be homologous with the dorsal adhesive organs of the Limnadiidae. The commonest condition is that of a median dorsal pit (Fig. 10, N.) by means of which the animal can fix itself to foreign objects. Certain forms may remain for long periods of time attached by the dorsal organ to plants, or to the sides of an aquarium, the only movement being a slow vibration of the feet, by which a current of water, sufficiently rapid for respiratory purposes, is established round it.^[31] In Sida crystallina (Fig. 11) the dorsal organ is represented by three structures; in front there is a median raised 40patch (N.m) of columnar ectoderm, containing concretions like those described in the Branchipodidae, and behind this is a pair of cup-shaped organs (N.e), with raised margins.

Fig. 11.—Sida crystallina, male, × 27. Oxford. A.1, Elongated first antenna; N.e, paired element of dorsal organ; N.m, median element of dorsal organ; Te, testes; ♂, opening of vas deferens.

The fold of skin which forms the carapace contains the coils of the single pair of kidneys, and it forms an important organ of respiration, partly from the great size of the blood-vessels it contains, and partly from the presence of red, blue, or brown respiratory pigments in the tissue of the skin itself.

In most Cladocera the cuticle of the carapace is cast at every ecdysis, with that of other parts of the body; but in Iliocryptus and a few others it remains after each moult, giving the carapace an appearance of “lines of growth,” like that seen in many Limnadiidae.

The segmentation of the body behind the head is obscure, but we can generally recognise (1) a thorax, of as many segments as there are pairs of limbs; (2) an abdomen of three segments; and (3) a telson.

The thoracic limbs of the Calyptomera are flattened, and resemble those of the Phyllopoda; as a type we may examine the third thoracic limb of Simocephalus (Fig. 12, C), in which the axis bears a large setose gnathobase (Gn) on its inner edge, followed by two small endites; the terminal process, or exopodite (Ex) is a large flattened plate, with six long plumose hairs on its edge. The outer margin of the axis bears a bract (Br) and an epipodite.

In Simocephalus, as in the other Daphniidae, there are five pairs of thoracic limbs, of which the third and fourth are alike; in the female each limb of the first pair consists of a jointed axis, with strong biting hairs on the inner border, and a rudimentary epipodite (Fig. 12, A), the second limb being more like the third, but with a more prominent gnathobase and a narrower exopodite (B), while the limbs of the fifth pair have the gnathobase and the exopodite filamentous (D).

41In the Sididae there are six pairs of thoracic limbs, which are nearly alike in the female; in the Bosminidae there are six pairs, the first two modified for prehension, the last much reduced.

Fig. 12.—Thoracic limbs of female Simocephalus vetulus. A, The first; B, the second; C, the third; D, the fifth. Br, Bract; Ep, epipodite; Ex, exopodite; Gn, gnathobase.

In the male, the first thoracic limb is usually provided with a long sensory process and a prehensible hook (Figs. 11, 19).

In the Gymnomera the limbs are cylindrical, jointed rods, 42with a gnathobase on the inner side in the Polyphemidae, but not in Leptodora. The number varies from four to six pairs.

The abdomen bears no appendages. The telson is compressed in the Calyptomera, and is produced into two flattened plates, one on each side of the anal opening. The backwardly directed margins of these plates are commonly serrated, and the lower corner of each is produced into a curved spine, which carries secondary teeth. The number and arrangement of these teeth, though often extremely variable in the same species, are used extensively as specific characters. Above the anus the telson commonly bears two long plumose hairs, which are directed backwards.

Fig. 13.—Bythotrephes cederströmii, female, × 20, North Wales, from a specimen found by A. D. Darbishire. Car, carapace.

In the Gymnomera the telson is not bilaterally compressed, and it may be produced into a long spine, dorsal to the anus (e.g. Bythotrephes, Fig. 13).

The alimentary canal is extremely simple. The labrum is large, and forms a chamber above the mouth, into which food is driven by the limbs, as in the Phyllopoda, food being taken while the animal swims or lies on its back. The oesophagus runs vertically to join a small stomach, which bends sharply backwards and passes gradually into an intestine. In the last segment of the abdomen the intestine joins a short, thin-walled rectum, provided with radial muscles, by means of which it can be dilated. The dilatation of the rectum leads to an inhalation of water through the anus, which may possibly serve as a means of respiration. In the Daphniidae and Bosminidae there are two forwardly-directed digestive glands which open into the stomach, and in Eurycercus there is a large caecum at the junction of the rectum with the intestine. The 43intestine is usually straight, but in Lynceidae and in some Lyncodaphniidae it is coiled (e.g. Peracantha, Fig. 14).

In Leptodora the alimentary canal is altogether remarkable; the oesophagus is a long and very narrow tube, which runs back through the whole length of the thorax and joins the mid-gut in the third abdominal segment. The mid-gut is not differentiated into stomach and intestine; it has no diverticula of any kind, and runs straight backwards to join the short rectum a little in front of the anus.

Fig. 14.—Peracantha truncata, female, × 100. Oxford.

The heart is always short, and never has more than a single pair of lateral openings; it is longest in the Sididae, which show some approximation to the Phyllopods in this, as in the slight degree of difference between their anterior and posterior thoracic limbs. The pericardium lies in the one or two anterior thoracic segments, dorsal to the gut. From the heart the blood runs forwards to the dorsal part of the head, and passes backwards by three main channels, one entering each side of the carapace, while the third runs down the body, beneath the alimentary canal to dilate into a large sinus round the rectum. This ventral blood-channel gives a branch to each limb, which forms a considerable dilatation in the epipodite, the blood from the limb returning to the pericardium by a lateral sinus. From the rectum a large sinus runs forwards to the pericardium along the dorsal wall of the body. The blood which enters each half of the carapace is collected in a median vessel and returned through this to the pericardium.

Those spaces between the viscera which are not filled with blood are occupied by a peculiar connective tissue, consisting of rounded or polyhedral cells, charged with drops of a fatty material which is often brightly coloured.

The reproductive organs are interesting because of the peculiar phenomena connected with the nutrition of the two kinds of eggs. The ovaries or testes are epithelial sacs, one on 44each side of the body, each continuous with a duct which opens to the exterior behind the last thoracic limb. In the female, the opening is dorsal (Fig. 10), in the male it is ventral (Fig. 11). The external opening is usually simple; but in the male there is sometimes a penis-like process, on which the vas deferens opens (Daphnella).

The eggs are of two kinds, the so-called “summer-eggs,” with relatively little yolk, which develop rapidly without fertilisation, and the so-called “winter-eggs,” containing much yolk, which require to be fertilised and then develop slowly.

At one end of the ovary, generally that nearest to the oviduct, there is a mass of protoplasm, containing nuclei which actively divide; this is the germarium (Fig. 15, A, B, C). As a result of proliferation in the germarium, nucleated masses are thrown off into the cavity of the ovary; each such mass contains four nuclei, and its protoplasm soon becomes divided into four portions, one round each nucleus, so that four cells are produced. In the simpler ovaries, such as that of Leptodora (Fig. 15, A), these sets of four cells are arranged in a linear series within the tube of ovarian epithelium; in other cases, as in Daphnia, the arrangement is more irregular. In the normal development of parthenogenetic eggs, one cell out of each set of four becomes an ovum, the other three feeding it with yolk and then dying. Weismann^[32] has shown that the ovum is always formed from the third cell of each set, counting from the germarial end, so that in the ovary of Leptodora drawn in Fig. 15, A, the ova will be formed from the cells marked E₁, E₂, E₃. At certain times, one or two sets of germinal cells fail to produce ova; the epithelial wall of the ovary thickens round these cells, so that they become incompletely separated from the rest in a so-called “nutrient chamber” (Fig. 15, B, N.C). Germ-cells enclosed in a nutrient chamber degenerate and are ultimately devoured by the ovarian epithelium. The significance of these nutrient chambers is unknown.

45

Fig. 15.—A, Ovary of a parthenogenetic Leptodora hyalina; B, base of another ovary of the same species, showing a so-called “nutrient chamber”; C, ovary of a female Daphnia, showing the formation of a winter-egg. E, E₁–E₃, Parthenogenetic egg; Ep, ovarian epithelium; G, germarium; N.C, nutrient chamber; O.D, oviduct; W, winter-egg; 1, 2, 4, the other three cells of the same group; II, III, two other groups of germ-cells.

The production of a winter-egg is a more complicated process. The epithelium of the ovarian tube swells up, so that the lumen is nearly obliterated, and several sets of four germ-cells pass from the germarium to lie among the swollen epithelial cells. All these groups of germ-cells, except one, disintegrate and are 46devoured by the ovarian epithelium, one cell of the remaining group enlarging to form a winter-egg, fed during its growth not only by the three cells of its own set but also by the epithelial cells of the ovarian tube, which have devoured the germ-cells of other sets. An ovary never contains more than a single winter-egg at the same time, the number of germ-cells which are devoured during its formation varying in the different species; the Daphnia drawn in Fig. 15, C, has produced three groups of germ-cells, of which two (II, III), will die, while the cell W from the remaining group will develop into an ovum; in Moina, Weismann finds that as many as a dozen cell-groups may be thrown into the ovary before the production of a winter-egg, so that only one out of forty-eight germ-cells survives as an ovum.

Fig. 16.—Sketch of a parthenogenetic Moina rectirostris, × 45, the brood-pouch being emptied and the side of the carapace removed, showing the dome of thickened epithelium on the thorax, by which nutrient material is thrown into the brood-pouch, and the ridge which fits against the carapace in the natural condition so as to close the brood-pouch.

Fig. 17.—Moina rectirostris, ♀, × 40, showing the ephippial thickening of the carapace which precedes the laying of a winter-egg.

The summer-eggs are always carried until they are hatched by the parthenogenetic female which produces them. The brood-pouch is the space between the dorsal wall of the thorax and the carapace. This space is always more or less perfectly closed at the sides by the pressure of the carapace against the body, and behind by vascular processes from the abdominal segments (Figs. 10, 16, etc.). The presence of a large blood-sinus 47beneath the dorsal wall of the thorax and in the middle line of the carapace suggests the possibility that some special nutrient substances may pass from the body of the parent into the brood-chamber, and in some species the thoracic ectoderm is specially modified as a placenta. In Moina (Fig. 16) the dorsal wall of the thorax is produced into a dome, covered by a columnar ectoderm, which contains a dilatation of the dorsal blood-sinus; and in this form it has been shown that the fluid in the brood-pouch contains dissolved proteids. Associated with the apparatus for supplying the brood-pouch with nutriment is a special apparatus for closing it, in the form of a raised ridge, which projects from the back and sides of the thorax and fits into a groove of the carapace.

A somewhat similar nutrient apparatus exists in the Polyphemidae, where the edges of the small carapace are fused with the thorax, so that the brood-pouch is completely closed, and the young can only escape when the parent casts her cuticle. In some genera of this family (e.g. Evadne) the young remain in the parental brood-pouch until they are themselves mature, so that when they are set free they may already bear parthenogenetic embryos in their own brood-pouches.

Fig. 18.—Newly-cast ephippium of Daphnia, containing two winter-eggs.

The winter-eggs are fertilised in the same part of the carapace of the female in which the parthenogenetic eggs develop, but after fertilisation they are thrown off from the body of the mother, either with or without a protective envelope formed from the cuticle of the carapace. The eggs of Sida are surrounded by a thin layer of a sticky substance, and when cast out of the maternal carapace they adhere to foreign objects, such as water-weeds; those of Polyphemus have a thick, gelatinous coat; in Leptodora and Bythotrephes the egg secretes a two-layered chitinous shell. In these forms the cuticle of the 48parent is not used as a protection for the winter-eggs, although it is generally, if not invariably, thrown off when the eggs are laid. In the Lynceidae the cuticle is moulted in such a way that the winter-eggs remain within it, at least for a time; the cuticle is occasionally modified before it is thrown off; thus in Camptocercus macrurus the cuticle of the carapace, in the region of the brood-pouch, becomes thickened and darkly coloured, forming a fairly strong case round the eggs. The modification of the cuticle round the brood-pouch is much more pronounced in the Daphniidae, where it leads to the formation of a saddle-shaped cuticular box, the “ephippium,” in which the winter-eggs are enclosed. The ripening of a winter-egg in the ovary of a Daphnia is accompanied by a great thickening of the cuticle of the carapace (cf. Fig. 18), so that a strong case is formed in the position of the brood-pouch. The winter-eggs are laid between the two valves of this case, and shortly afterwards the parent moults. The eggs are retained within the ephippium, from which the rest of the cuticle breaks away (Fig. 18). After separation, the ephippium, which contains a single egg (Moina rectirostris) or usually two (Daphnia, etc.), either sinks to the bottom, as in Moina, or floats.

The winter-eggs usually go through the early stages of segmentation within a short time after they are laid, but after this a longer or shorter period of quiescence occurs, during which the eggs may be dried or frozen without injury. The sides and floor of a dried-up pond are often crowded with ephippia, containing winter-eggs which develop quickly when replaced in water; and the resting-stage of winter-eggs produced in aquaria can often be materially shortened by drying the ephippia which contain them, though such desiccation does not appear to be necessary for development. Under normal conditions large numbers of winter-eggs remain quiescent through the winter and hatch in the following spring.

The individual developed from a sexually fertilised winter-egg 49is invariably a parthenogenetic female: the characters of the succeeding generations differ in different cases.

In a few forms, of which Moina is the best known, the parthenogenetic female, produced from a winter-egg, may give rise to males, to sexual females, and to parthenogenetic females, so that the cycle of forms which intervene between one winter-egg and the next is short. A sexual female produces one or two winter-eggs, and if these are fertilised they are enclosed in an ephippium and cast off; if, however, the eggs when ripe are not fertilised, they atrophy, and the female produces parthenogenetic eggs, being thenceforward incapable of forming sexual “winter” eggs. An accidental absence of males may thus lead to the occurrence of parthenogenesis in the whole of the second generation. The regular production of sexual individuals in the second generation from the winter-egg appears to depend on a variety of circumstances not yet understood. Mr. G. H. Grosvenor tells me that Moina from the neighbourhood of Oxford may give rise to several successive generations of parthenogenetic individuals, when grown in small aquaria.

In the greater number of Daphniidae, the parthenogenetic female, produced from a winter-egg, gives rise only to parthenogenetic forms, and it is not until after half a dozen parthenogenetic generations have been produced that a few sexual forms appear, mixed with the others. Such sexual forms are fairly common in April or May in this country; they produce “winter” eggs and then die, the generations which succeed them through the summer being entirely parthenogenetic. In late autumn sexual individuals are again produced, giving rise to a plentiful crop of winter-eggs, but many parthenogenetic females are still found, and some of these appear to live and to reproduce through the winter.

In Sida, in the Polyphemidae and Leptodoridae, and in most of the Lynceidae, sexual individuals are produced only once in every year, while in a few forms which inhabit great lakes the sexual condition occurs so rarely that it is still unknown.

Weismann^[33] has pointed out that the sexual forms, with their property of producing eggs which can endure desiccation, recur most frequently in species such as Moina, which inhabit small pools liable to be dried up at frequent intervals, while the 50species which produce sexual forms only once a year are all inhabitants either of great lakes which are never dry, or of the sea. Many suggestions have been made as to the environmental stimulus which induces the production of sexual individuals, but nothing is definitely known upon the subject.

We have said that even in those generations which contain sexual males and females there are always some parthenogenetic individuals; there is therefore nothing in the behaviour of Daphniidae, either under natural conditions or when observed in aquaria, to suggest that there is any natural or necessary limit to the number of generations which may be parthenogenetically produced.

The parthenogenetic Daphniidae are extremely sensitive to changes in their surroundings; small variations in the character and amount of substances dissolved in the water are often followed by changes in the length of the posterior spine, in the shape and size of crests on the head, and in other characters affecting the appearance of the creatures, so that the determination of species is often a matter of great difficulty. It is remarkable that the green light which has passed through the leaves of water-plants appears to have a prejudicial effect upon some species. Warren has shown that Daphnia magna reproduces more slowly when exposed to green light, and that individuals grown in this way are more readily susceptible to injury from the presence of small quantities of salt (sodium chloride) in the water than individuals which have been exposed to white light.

The majority of the Cladocera belong to the floating fauna of the fresh waters and seas; a few are littoral in their habits, clinging to water-weeds near the shore, a very few live near the bottom at considerable depths, but the majority belong to that floating fauna to which Haeckel gave the name of “plankton.” The Crustacea are an important element in the plankton, whether in fresh waters or in the sea, the two great groups which contribute most largely to it being the Cladocera and the Copepoda. For this reason it will be more convenient to discuss the habits and distribution of individual Cladocera and Copepoda together in a chapter specially devoted to the characters of pelagic faunas (cf. Chap. VII.). We will only add to the present chapter a table of the families with a diagnosis of the British genera.

Tribe I. Calyptomera, Sars.—The post-cephalic portion of the body enveloped in a free fold or carapace.

A. Six pairs of thoracic feet, the first pair not prehensile (Ctenopoda).

Fam. 1. Sididae: second antennae biramous in both sexes. Sida, Straus (Fig. 11): second antenna with three joints in the dorsal ramus, two in the ventral; the rostrum large, the teeth on the telson many. Latona, Straus: second antenna with two joints in the dorsal ramus, three in the ventral, the proximal joint of the dorsal ramus provided with a setose appendage. Daphnella, Baird: second antenna with the joints as in Latona, but with no setose appendage.

Fam. 2. Holopediidae: second antennae not biramous in the female; a rudimentary second ramus in the male. Holopedium, Zaddach.

B. Four to five or six pairs of thoracic feet, the anterior pair prehensile (Anomopoda).

A. Ventral ramus of second antenna with three joints, the dorsal ramus with four.

Fam. 3. Daphniidae: five pairs of thoracic feet, with a gap between the fourth and fifth pairs. The stomach with two forwardly-directed diverticula.

Fig. 19.—Daphnia obtusa, male, × about 50. Oxford. A.1, First antenna; Th.1, first thoracic appendage.

i. First antennae of female short.

α A median dorsal spine on posterior margin of carapace. Daphnia, O. F. Müller (Fig. 19): first antennae of female not mobile. The head separated from the thorax only by a slight constriction or not at all. Cuticle with a quadrate rhomboid pattern. Ceriodaphnia, Dana: first antennae of 52female mobile. The head separated by a deep depression from the thorax. Cuticle with a polygonal pattern.

β A pair of ventral spines on posterior margin of carapace. Scapholeberis, Schoedler (Fig. 20).

Fig. 20.—Scapholeberis mucronata, female, × 25. Oxford.

γ No spine on posterior margin of carapace. Simocephalus, Schoedler (Fig. 10, p. 39): the cuticle with a pattern of parallel branching ridges.

Fig. 21.—Moina rectirostris, female, × 24. Oxford.

ii. First antennae of female long, mobile. Moina, Baird (Figs. 16, 17, 21): median eye absent. Posterior margin of carapace without a spine.

Fig. 22.—Bosmina sp., female, × about 80. Lake Constance.

Fig. 23.—Acroperus leucocephalus, × about 35. Oxford.

53Fam. 4. Bosminidae: feet equidistant, five or six pairs; the first antennae of the female immobile, with sense-hairs arranged in rings, not forming an apical tuft. The intestine uncoiled; no caeca. Bosmina, Baird (Fig. 22).

Fam. 5. Lyncodaphniidae: four, five, or six pairs of equidistant thoracic limbs; the first two pairs prehensile. First antennae of female mobile, with apical sense-hairs. Intestine coiled or straight.

i. Four pairs of thoracic limbs. Lathonura, Lilljeborg.

ii. Five pairs of thoracic limbs.

a. The four-jointed ramus of the second antenna with four swimming hairs. Macrothrix, Baird: the first antennae of the female flattened, curved. The intestine simple, straight. Streblocerus, Sars: first antennae of the female very little flattened, curved backwards and outwards. The intestine coiled, the stomach with two forwardly-directed caeca.

b. The four-jointed ramus of the second antenna with only three swimming hairs. Drepanothrix, Sars.

iii. Six pairs of thoracic limbs; the labrum provided with an appendage. Acantholeberis, Lilljeborg: appendage of labrum long, pointed, and setose. Intestine without caecum. Ilyocryptus, Sars: appendage of the labrum short, truncated. Intestine with a caecum.

B. Both rami of second antenna three-jointed.

Fam. 6. Lynceidae^[35]: five or six equidistant pairs of thoracic feet. Intestine coiled.

i. Six pairs of thoracic limbs. Head and thorax separated by a deep depression. Intestine with one caecum, stomach with two. Female carries many summer-eggs. Eurycercus, Baird.

ii. Five pairs of thoracic limbs. Head and thorax separated by a slight groove or not at all. Anterior digestive caeca absent. Female carries only one or two summer-eggs.

A. Body elongate, oval.

a. Head carinate, the eye far from the anterior cephalic margin. Camptocercus, Baird: body laterally compressed. Second antennae with seven swimming hairs. Telson more than half as long as the shell. Acroperus, Baird (Fig. 23): body compressed. Second antennae with eight swimming hairs, of which one is very small. Telson less than half as long as the shell.

b. Head not carinate, the eye near the anterior cephalic margin. Alonopsis, Sars: terminal claws of telson with three accessory teeth. Alona, Baird: terminal claws of telson with one accessory tooth (includes sub-genera Leydigia, Alona, Harporhynchus, Graptoleberis). Peracantha, Baird (Fig. 14): terminal 54claws of telson with two accessory teeth (includes sub-genera Alonella, Pleuroxus, Peracantha).

B. Body small, spheroidal; the head depressed. Chydorus, Leach: compound eye present. Monopsilus, Sars: compound eye absent.

Tribe II. Gymnomera, Sars.—The carapace forms a closed brood-pouch, which does not cover the body; all the thoracic limbs prehensile.

Fam. 7. Polyphemidae: four pairs of thoracic limbs, provided with a gnathobase.

Fresh-water genera.—Polyphemus, Müller, with no rudimentary exites on first three thoracic limbs. Bythotrephes, Leydig (Fig. 13), with no trace of processes on the outer sides of the limbs.

Marine genera.—Evadne, Lovén, the head not separated by a constriction from the thorax. Podon, Lovén, with deep cervical constriction.

Fig. 24.—Leptodora hyalina, × 6. Lake Bassenthwaite. A.1, First antenna; Car, carapace; I, VI, first and sixth thoracic appendages.

Fam. 8. Leptodoridae: six pairs of thoracic limbs, with no gnathobase. Only genus, Leptodora, Lilljeborg (Fig. 24), from fresh water.

Note.—For extra-European Cladocera consult Daday, “Microskopische Süsswassertiere aus Patagonien und Chili,” Termés Füzetek, xxv., 1902, p. 201; for Paraguay, Bibliotheca Zoologica, Heft 44; for Ceylon, Termés Füzetek, xxi., 1898; and for Australia, Sars, Christiania Vidensk. Forhand. 1885, No. 8, and 1888, No. 7; and Arch. f. Math. og Naturvid. xviii., 1896, No. 3, and xix., 1897, No. 1.—G. W. S.

Order II. Copepoda.

The Copepods are small Crustacea, composed typically of about sixteen segments, in which the biramous type of limb predominates. They are devoid of a carapace. Development proceeds gradually by the addition posteriorly of segments to a Nauplius larval form. Paired compound eyes are absent, except in Branchiura, the adult retaining the simple eye of the Nauplius.

In a typical Copepod, such as Calanus hyperboreus (Fig. 25), we can distinguish the following segments with their appendages: a cephalothorax, carrying a pair of uniramous first antennae (1^st Ant.); a pair of biramous second antennae (2^nd Ant.); mandibles (Md.) with biting gnathobases and a palp, and a pair of foliaceous first maxillae (Mx.¹). Two pairs of appendages follow, which were looked upon as the two branches of the second maxillae, but it is now certain that they represent two pairs of appendages, which may be called second maxillae (Mx.²) and maxillipedes (Mxp.) respectively. Behind these are five pairs of biramous swimming feet, the first pair (Th.¹) attached to the cephalothorax, the succeeding four pairs to four distinct thoracic somites. Behind the thorax is a clearly delimited abdomen composed of five segments, the first of which (Abd.¹) carries the genital opening, and the last a caudal furca.

The Copepods exhibit a great variety of structure, and their classification is attended with great difficulties. Claus^[36] based his attempt at a natural classification on the character of 56the mouth and its appendages, dividing the free-living and semi-parasitic forms as Gnathostomata from the true parasites or Siphonostomata. This division, although convenient, breaks down in many places, and it is clear that the parasitic mode of life has been acquired more than once in the history of Copepod evolution, while the free-living groups do not constitute a natural assemblage.

Fig. 25.—Calanus hyperboreus, × 30. Abd¹, First abdominal segment; 1st Ant, 2nd Ant, 1st and 2nd antennae; Md, mandible; Mx¹, Mx², 1st and 2nd maxillae; Mxp, maxillipede; Th¹, 1st thoracic appendage. (After Giesbrecht.)

57Giesbrecht has more recently^[37] founded a classification of the free-living pelagic Copepods upon the segmentation of the body and certain secondary sexual characters, and he has hinted^[38] that this scheme of classification applies to the semi-parasitic and parasitic forms. Although much detail remains to be worked out and the position of some families is doubtful, Giesbrecht’s scheme is the most satisfactory that has hitherto been suggested, and will be adopted in this chapter.

The peculiarity in structure of the Argulidae, a small group of ectoparasites on fresh water fish, necessitates their separation from the rest of the Copepods (Eucopepoda) as a separate Branch, Branchiura.

BRANCH I. EUCOPEPODA.

Sub-Order 1. Gymnoplea.

The division between the front and hind part of the body falls immediately in front of the genital openings and behind the fifth thoracic feet. The latter in the male are modified into an asymmetrical copulatory organ.

TRIBE I. AMPHASCANDRIA.

The first antennae of the male are symmetrical, with highly-developed sensory hairs.

Fam. Calanidae.—The Calanidae are exclusively marine Crustacea, and form a common feature of the pelagic plankton in all parts of the world. Some species of the genus Calanus often occur in vast shoals, making the sea appear blood-red, and they furnish a most important article of fish food. These swarms appear to consist chiefly of females, the males being taken rarely, and only at certain seasons of the year. Some of the Calanidae are animals of delicate and curious form, owing to the development of plumed iridescent hairs from various parts of their body, which may often exhibit a marked asymmetry, as 58in the species figured, Calocalanus plumulosus (Fig. 26), from the Mediterranean.

Fig. 26.—Calocalanus plumulosus, × 15. (After Giesbrecht.)

Sars makes a curious observation^[39] with regard to the distribution of certain Calanidae. He reports that along the whole route of the “Fram,” species such as Calanus hyperboreus and Euchaeta norwegica were taken at the surface, which, in the Norwegian fjords, only occur at depths of over 100 fathoms. He suggests that the Norwegian individuals, instead of migrating northwards as the warmer climate supervened, have sought boreal conditions of temperature by sinking into the deeper waters.

TRIBE II.
HETERARTHRANDRIA.

The first antennae of the male are asymmetrical, one, usually the right, being used as a clasping organ.

The males of the Centropagidae, Candacidae and Pontellidae, besides possessing the asymmetrically modified thoracic limbs of the fifth pair also exhibit a modification of one of the first antennae, which is generally thickened in the middle, and has a peculiar joint in it, or geniculation, which enables it to be flexed and so used as a clasping organ for holding the female.

Fam. 1.—Centropagidae.—These Copepods are very common in the pelagic plankton, and some of the species vie with the 59Calanidae in plumed ornaments, e.g. Augaptilus filigerus, figured by Giesbrecht in his monograph. The use of these ornaments, which are possessed by so many pelagic Copepods, is entirely obscure.^[40] Certain of the Centropagidae live in fresh water. Thus Diaptomus is an exclusively fresh-water genus, and forms a most important constituent of lake-plankton; various species of Heterocope occur in the great continental lakes, and certain Eurytemora go up the estuaries of rivers into brackish water.

An excellent work on the fresh-water Copepods of Germany has been written by Schmeil,^[41] who gives analytical tables for distinguishing various genera and species. The three fresh-water families are the Centropagidae, Cyclopidae, and Harpacticidae (see p. 62). The Centropagidae may be sharply distinguished from the other fresh-water families by the following characters:—The cephalothorax is distinctly separated from the abdomen; the first antennae are long and composed of 24–25 segments, in the male only a single antenna (generally the right) being geniculated and used as a clasping organ. The fifth pair of limbs are not rudimentary; a heart is present, and only one egg-sac is found in the female. The second antennae are distinctly biramous.

Diaptomus.—The furcal processes are short, at most three times as long as broad; endopodite of the first swimming appendage 2–jointed, endopodites of succeeding legs 3–jointed.

Heterocope.—The furcal processes are short, at most twice as long as broad; endopodites of all swimming legs 1–jointed.

Eurytemora.—The furcal processes are long, at least three and a half times as long as broad; the endopodite of the first pair of legs 1–jointed, those of the other pairs 2–jointed.

Fig. 27.—Dorsal view of Anomalocera pattersoni, ♂, × 20. (After Sars.)

It has been known for a long time that some of the marine Copepods are phosphorescent, and, indeed, owing to their numbers in the plankton, contribute very largely to bring about that liquid illumination which will always excite the admiration of seafarers. In northern seas the chief phosphorescent Copepods belong to Metridia, a genus of the Centropagidae; but in the Bay of Naples Giesbrecht^[42] states that the phosphorescent species are the following Centropagids: Pleuromma abdominale and P. gracile, Leuckartia flavicornis and 60Heterochaeta papilligera; Oncaea conifera is also phosphorescent. It is often stated that Sapphirina (p. 69) is phosphorescent, but its wonderful iridescent blue colour is purely due to interference colours, and has nothing to do with phosphorescence. Giesbrecht has observed that the phosphorescence is due to a substance secreted in special skin-glands, which is jerked into the water, and on coming into contact with it emits a phosphorescent glow. This substance can be dried up completely in a desiccated specimen and yet preserve its phosphorescent properties, the essential condition for the actual emission of light being contact with water. Similarly, specimens preserved in glycerine for a long period will phosphoresce when compressed in distilled water. From this last experiment Giesbrecht concludes that the phosphorescence can hardly be due to an oxidation process, but the nature of the chemical reaction remains obscure.

Fam. 2. Candacidae.—This family comprises the single genus Candace, with numerous species distributed in the plankton of all seas. Some species, e.g. C. pectinata, Brady, have a practically world-wide distribution, this species being recorded from the Shetlands and from the Philippines.

Fam. 3. Pontellidae.—This is a larger family also comprising widely distributed species found in the marine plankton. Anomalocera pattersoni (Fig. 27) is one of the commonest elements in the plankton of the North Sea.

The boundary between the fore and hind part of the body falls in front of the fifth thoracic segment. The appendages of the fifth thoracic pair in the male are never modified as copulatory organs.

TRIBE I. AMPHARTHRANDRIA.

The first antennae in the male differ greatly from those in the female, being often geniculated and acting as prehensile organs.

Fig. 28.—Euterpe acutifrons, ♀, × 70. Abd.1, 1st abdominal segment; Th.5, 5th thoracic segment. (After Giesbrecht.)

Fig. 29.—First antenna of Euterpe acutifrons, ♂. (After Giesbrecht.)

Fams. 1–2. Cyclopidae and Harpacticidae, and other allied families, are purely free-living forms; they are not usually pelagic in habit, but prefer creeping among algae in the littoral zone or on the sea-bottom, or especially in tidal pools. Some genera are, nevertheless, pelagic; e.g. Oithona among Cyclopidae; Setella, Clytemnestra, and Aegisthus among Harpacticidae.

The sketch (Fig. 28) of Euterpe acutifrons ♀, a species widely 62distributed in the Mediterranean and northern seas, exhibits the structure of a typical Harpacticid, while Fig. 29 shows the form of the first antenna in the male.

Several fresh-water representatives of these free-living families occur. The genus Cyclops (Cyclopidae) is exclusively fresh-water, while many Harpacticidae go up into brackish waters: for example on the Norfolk Broads, Mr. Robert Gurney has taken Tachidius brevicornis, Müller, and T. littoralis, Poppe; Ophiocamptus brevipes, Sars; Mesochra lilljeborgi, Boeck; Laophonte littorale, T. and A. Scott; L. mohammed, Blanchard and Richard; and Dactylopus tisboides, Claus.

Schmeil^[43] gives the following scheme for identifying the fresh-water Cyclopidae and Harpacticidae (see diagnosis of Centropagidae on p. 59):—

Fam. 1. Cyclopidae.—The cephalothorax is clearly separated from the abdomen. The first antennae of the female when bent back do not stretch beyond the cephalothorax; in the male both of them are clasping organs. The second antennae are without an exopodite. The fifth pair of limbs are rudimentary, there is no heart, and the female carries two egg-sacs.

Cyclops.—Numerous species, split up according to segmentation of rudimentary fifth pair of legs, number of joints in antennae, etc.

Fam. 2. Harpacticidae.—The cephalothorax is not clearly separated from the abdomen. The first antennae are short in both sexes, both being clasping organs in the male. The second antennae have a rudimentary exopodite. The fifth pair of limbs are rudimentary and plate-shaped; a heart is absent, and the egg-sacs of the female may be one or two in number.

1. Ophiocamptus (Moraria).—Body worm-shaped; first antennae of female 7–jointed, rostrum forming a broad plate.

2. Body not worm-shaped; first antennae of female 8–jointed, rostrum short and sharp.

(a) Endopodites of all thoracic limbs 3–jointed. The first antennae in female distinctly bent after the second joint. Nitocra.

(b) Endopodite of at least the fourth limb 2–jointed; first antennae in female not bent. Canthocamptus.

3. Ectinosoma.—Body as in 2, but first antennae are very short, and the maxillipede does not carry a terminal hooked seta as in 1 and 2.

63Fam. 3. Peltiidae.^[44]—This is an interesting family, allied to the Harpacticidae, and includes species with flattened bodies somewhat resembling Isopods, and a similar habit of rolling themselves up into balls. No parasitic forms are known, though Sunaristes paguri on the French and Scottish coasts is said to live commensally with hermit-crabs.

We have now enumerated the chief families of free-living Copepods; the rest are either true parasites or else spend a part of their lives as such. A number of the semi-parasitic and parasitic Copepods can be placed in the tribe Ampharthrandria owing to the characters of their antennae; but it must be remembered that many parasitic forms have given up using the antennae as clasping organs; however, the sexual differences in the antennae, and the fact that many of the species which have lost the prehensile antennae in the male have near relations which preserve it, enable us to proceed with some certainty. The adoption of this classification necessitates our separating many families which superficially may seem to resemble one another, e.g. the semi-parasitic families Lichomolgidae and Ascidicolidae, and the Dichelestiidae from the other fish-parasites; it also necessitates our treating the presence of a sucking mouth as of secondary importance. This characteristic must certainly, however, have been acquired more than once in the history of the Copepods, for instance in the Asterocheridae and in the fish-parasites, while it sometimes happens that genera belonging to a typically Siphonostomatous group possess a gnathostome, or biting mouth, e.g. Ratania among the Asterocheridae. Again, it is impossible even if we use the character of the mouth as a criterion to place together all the true parasites on fishes in one natural group, because the Bomolochidae and Chondracanthidae, which are otherwise closely similar to the rest of the fish-parasites, possess no siphon. It seems plain, therefore, that the parasitic habit has been acquired several times separately by diverging stocks of free-swimming Copepods, and that it has resulted in the formation of convergent structures.

64

Fig. 30.—Haemocera danae, × 40. A, Side view ♀; B, ventral view ♂. Ant.1, 1st antenna; e, eye; ov, ovary; ovd, oviduct; St, stomach; Th.1, 1st thoracic appendage; Th.5, 5th thoracic segment; vd, vas deferens. (After Malaquin.)

Fig. 31.—Free-swimming Nauplius larva of Haemocera danae; Ant.1, Ant.2, 1st and 2nd antennae; e, remains of eye; Md, mandible. (After Malaquin.)

Fam. 4. Monstrillidae.^[45]—These are closely related to the Harpacticidae. The members of this curious family are parasitic during larval life and actively free-swimming when adult. There are three genera, Monstrilla, Haemocera, and Thaumaleus. The best known type is Haemocera danae (often described as Monstrilla danae). In the adult state (Fig. 30) there are no mouth-parts; the mouth is exceedingly small and leads into a very small stomach, which ends blindly, while the whole body contains reserve food-material in the form of brown oil-drops. The sole appendages on the head are the first antennae; but on the thorax biramous feet are present by means of which the animal can swim with great rapidity. This anomalous organisation receives an explanation from the remarkable development through which the larva passes. The larva is liberated from the parent as a Nauplius with the structure shown in Fig. 31; it does not possess an alimentary canal. It makes its way to a specimen of the Serpulid worm, Salmacina dysteri, into the epidermis of which it penetrates by movements of the antennae, hanging on all the time by means of the hooks on the mandibles. From the epidermis it passes through the muscles into the coelom of the worm, and thence into the blood-vessels, usually coming to rest in the ventral blood-vessel. 65As the Nauplius migrates, apparently by amoeboid movements of the whole body, it loses all its appendages, the eye degenerates, and the body is reduced to a minute ovoid mass of cells, representing ectoderm and endo-mesoderm, surrounded by a chitinous membrane (Fig. 32, A). Arrived in the ventral blood-vessel it begins to grow, and the first organ formed is a pair of fleshy outgrowths representing the second antennae (Fig. 32, B), which act as a nutrient organ intermediary between host and parasite. The adult organs now begin to be differentiated, as shown in Fig. 32, C, from the undifferentiated cellular elements of the Nauplius, the future adult organism being enclosed in a spiny coat from which it escapes. At this stage it occupies a large part of its host’s body, lying in the distended ventral blood-vessel, and it escapes to the outside world by rupturing the body-wall of the worm, leaving behind it the second antennae, which have performed their function as a kind of placenta. Malaquin, to whom we owe this account, makes the remarkable statement that if two or three Monstrillid Nauplii develop together in the same host they are always males, if only one it may be either male or female. The only parallel to this extraordinary life-history is found in the Rhizocephala (see pp. 96–99).

Fig. 32.—Later stages in the development of Haemocera danae. Abd, Abdomen; Ant.1, Ant.2, 1st and 2nd antennae; ch, chitinous investment; e, eye; Ect, ectoderm; En, endoderm; Mes, mesoderm; Mes & en, mesoderm and endoderm; R, rostrum; St, mouth and stomach; Th, thoracic appendages. (After Malaquin.)

66

Fig. 33.—Side view of Doropygus pulex, ♀, × 106. Abd.1, 1st abdominal segment; Ant.1, 1st antenna; b.p, brood-pouch; Th.1, 1st thoracic appendage; Th.4, 4th thoracic segment. (After Canu.)

Fam. 5. Ascidicolidae.^[46]—Although the members of this family, which live semiparasitically in the branchial sac or the gut of Ascidians, betray their Ampharthrandrian nature by the sexual differences of their first antennae, only two genera, Notodelphys and Agnathaner, possess true prehensile antennae. According as the parasitism is more or less complete, the buccal appendages either retain their masticatory structure or else become reduced to mere organs of fixation. In Notodelphys both sexes can swim actively and retain normal mouth-parts; they live parasitically, or perhaps commensally, in the branchial cavities of Simple or Compound Ascidians, feeding on the particles swept into the respiratory chamber of the host. They leave their host at will in search of a new home, and are frequently taken in the plankton.

Doropygus (Fig. 33), a genus widely distributed in the North Sea and Mediterranean, also inhabiting the branchial sac of Ascidians, is more completely parasitic, and the female cannot swim actively. Forms still more degraded by a parasitic habit are Ascidicola rosea (especially abundant in the stomach of 67Ascidiella scabra at Concarneau), in which the female has lost its segmentation, the mouth-parts and thoracic legs being purely prehensile, and various species of Enterocola, parasitic in the stomach of Compound Ascidians, in which the female is a mere sac incapable of free motion, while the male preserves its swimming powers and a general Cyclops-form (Fig. 34). We have here the first instance of the remarkable parallelism between the degree of parasitism and the degree of sexual dimorphism, a parallelism which holds with great regularity among the Copepoda, and can be also extended to other classes of parasitic animals.

Fig. 34.—Enterocola fulgens. A, Ventral view of ♀, × 35; B, side view of ♂, × 106. Abd.1, 1st abdominal segment; Ant.1, Ant.2, 1st and 2nd antennae; c.m, gland-cells; n, ventral nerve-cord; og, oviducal gland; ov, ovary; po, vagina; Th.1, 1st thoracic appendage; Th.4, Th.5, 4th and 5th thoracic segments. (After Canu.)

Fig. 35.—Asterocheres violaceus, ♀, with egg-sacs, × 57. (After Giesbrecht.)

Fig. 36.—Diagrammatic transverse section through the distal part of the siphon of Rhynchomyzon purpurocinctum (Asterocheridae). Md, mandible. (After Giesbrecht.)

Fam. 6. Asterocheridae.^[47]—These forms retain the power of swimming actively, and are very little modified in outward appearance by their parasitic mode of life (Fig. 35), though they 68possess a true siphon in which the styliform mandibles work. The siphon is formed by the upper and lower lips, which are produced into a tube with three longitudinal ridges; in the outer grooves are the mandibles, while the inner groove forms the sucking siphon (see transverse section, Fig. 36). In Ratania, however, there is no siphon. The first antennae possess a great number of joints, and may be geniculated in the male (Cancerilla). The members of this family live as ectoparasites on various species of Echinoderms, Sponges, and Ascidians, but they frequently change their hosts, and it appears that one and the same species may indifferently suck the juices of very various animals, and even of Algae. Cancerilla tubulata, however, appears to live only on the Brittle Starfish, Amphiura squamata.

Fam. 7. Dichelestiidae.—The males and females are similarly parasitic, and the body in both is highly deformed, the segmentation being suppressed and the thoracic limbs being produced into formless fleshy lobes; they are placed among the Ampharthrandria owing to sexual differences in the form of the first antennae. There is a well-developed siphon in which the mandibular stylets work, except in Lamproglena, parasitic on the gills of Cyprinoid fishes; the succeeding mouth-parts are prehensile.

The majority of the species are parasitic on the gills of various fish (Dichelestium on the Sturgeon, Lernanthropus^[48] on Labrax lupus, Serranus scriba, etc.), but Steuer^[49] has recently described a Dichelestiid (Mytilicola) from the gut of Mytilus galloprovincialis off Trieste. This animal and Lernanthropus are unique among Crustacea through the possession of a completely closed blood-vascular system which contains a red fluid; the older observers believed this fluid to contain haemoglobin, but Steuer, as the result of careful analysis, denies this. The parasite on the gills of the Lobster, Nicothoe astaci, possibly belongs here.

The inclusion of Nicothoe and the Dichelestiidae among the Ampharthrandria rests on a somewhat slender basis; this basis is afforded by the fact that none of the parasitic Isokerandria have more than seven joints in the first antennae, whereas 69Nicothoe and some of the Dichelestiidae^[50] have more numerous joints. In most of the Dichelestiidae, however, the number of joints is less than seven and practically equal in the two sexes.

TRIBE II. ISOKERANDRIA.

The first antennae are short, similar in the two sexes, and are never used by the male as clasping organs. This function may be subserved by the second maxillae.

Fams. Oncaeidae, Corycaeidae, Lichomolgidae, Ergasilidae, Bomolochidae, Chondracanthidae, Philichthyidae, Nereicolidae, Hersiliidae, Caligidae, Lernaeidae, Lernaeopodidae, Choniostomatidae.

The families Oncaeidae and Corycaeidae contain pelagic forms of flattened shape and great swimming powers, but the structure of the mouth-parts in the Corycaeidae points to a semi-parasitic habit.

Fam. 1. Oncaeidae.—This family, including the genera Oncaea, Pachysoma, etc., does not possess the elaborate eyes of the next family, nor is the sexual dimorphism so marked.

Fam. 2. Corycaeidae.—These are distinguished from the Oncaeidae, not only by their greater beauty, but also by the possession of very elaborate eyes, which are furnished with two lenses, one at each end of a fairly long tube. The females of Sapphirina are occasionally found in the branchial cavity of Salps, and their alimentary canal never contains solid particles, but is filled with a fluid substance perhaps derived by suction from their prey. S. opalina may occur in large shoals, when the wonderful iridescent blue colour of the males makes the water sparkle as it were with a sort of diurnal phosphorescence. The animal, however, despite the opinion of the older observers, is not truly phosphorescent. It may be that the ornamental nature of some of the males is correlated with the presence of the curious visual organs, which are on the whole better developed in the females than in the males. As in so many pelagic Copepods, the body and limbs may bear plumed setae of great elaboration and beautiful colour, e.g. Copilia vitrea (Fig. 37).

We now pass on to the rest of the parasitic Copepods,^[51] which 70probably belong to the tribe Isokerandria, and we meet with the same variety of degrees of parasitism as in the Ampharthrandria, often leading to very similar results.

Fig. 37.—Copilia vitrea (Corycaeidae), ♀, × 20. (After Giesbrecht.)

In the first seven families mentioned below there is no siphon. The Lichomolgidae and Ergasilidae have not much departed from the free-living forms just considered, retaining their segmentation, though in the Ergasilidae the body may be somewhat distorted (Fig. 39). In both families the thoracic swimming feet are of normal constitution.

Fig. 38.—Lichomolgus agilis, × 10. Abd. 1, 1st abdominal segment; cpth, cephalothorax; Th.1, 1st thoracic segment; Th.5, 5th thoracic appendage. (After Canu.)

Fam. 3. Lichomolgidae.^[52]—These are semi-parasitic in a number of animals living on the sea-bottom, such as Actinians, 71Echinoderms, Annelids, Molluscs, and Tunicates. Lichomolgus agilis (Fig. 38) occurs in the North Sea, Atlantic, and Mediterranean, on the gills of large species of the Nudibranch, Doris, while L. albeus is found in the peribranchial cavity and cloaca of various Ascidians. Sabelliphilus may infect the gills of Annelids such as Sabella, and is common at Liverpool.

Fam. 4. Ergasilidae.—Thersites (Fig. 39) is parasitic on the gills of various fishes, e.g. T. gasterostei, which is common on Gasterosteus aculeatus on the French and North Sea coasts, and may even be found on specimens of the fish that have run up the River Forth into fresh water. The animal possesses claw-like second antennae by which it clings to its host.

Fig. 39.—Thersites gasterostei. A, ♀, × 10; B, ♂, × 20. Abd. 1 & 2, Fused 1st and 2nd abdominal segments; Ant.1, Ant.2, 1st and 2nd antennae; e.s, egg-sac; Th, thoracic appendages. (After Gerstaecker.)

Similarly characterised by the absence of a siphon are three other families of fish-parasites, the Bomolochidae, Chondracanthidae, and Philichthyidae.

Fam. 5. Bomolochidae.—Bomolochus (Fig. 40), parasitic on the skin of the Sole (Solea) and in the nostrils of Cod (Gadus), is held to be related to the Ergasilidae. The first thoracic limb is remarkably modified. Were it not for the absence of a siphon, it would be hard to separate this family from the Caligidae.

72

Fig. 40.—Bomolochus, sp. (Bomolochidae), × 8. Abd. 1, 1st abdominal segment; Ant.1, Ant.2, 1st and 2nd antennae; Mx.1, Mx.2, 1st and 2nd maxillae; Mxp, maxillipede; Th.1, 1st thoracic appendage. (After Gerstaecker.)

Fig. 41.—Chondracanthus zei, ♀, × 4.

Fig. 42.—Dwarf male of Lernentoma cornuta (Chondracanthidae), × 10. Ant.1, Ant.2, 1st and 2nd antennae; Th.1, 1st thoracic segment. (After Gerstaecker.)

Fam. 6. Chondracanthidae.—These Copepods infest the gills and even the mouth of various marine fish, such as the Gurnard, Plaice, Skate, Sole, and many others. The sexual dimorphism is very marked, the female being large, indistinctly segmented, and with irregular paired processes protruding from the sides of the body, giving the animal a monstrous form (Fig. 41); while the male (Fig. 42) is very small, has a 73completely segmented thorax, and lives clinging on to the female by its prehensile second antennae—Chondracanthus, Lernentoma.

Fam. 7. Philichthyidae.—These parasites, which are hardly known to occur in British waters,^[53] are mucus-feeders and infest the skin of Teleosts, e.g. the Sole; often taking up a position in the lateral line or in a slime canal. They show a similar sexual dimorphism to the foregoing family, the adult female being extraordinarily drawn out into finger-like processes (e.g. Philichthys)^[54] or else long, slender, and Nematode-like, with much reduced appendages (Lernaeascus), while the male retains a more normal structure. As in all the foregoing forms there is no siphon.

We now return to two semi-parasitic families, Fam. 8, Nereicolidae, and Fam. 9, Hersiliidae, in which there is certainly no well-developed siphon, but the upper and under lips protrude, forming a hollow between them in which the mouth-parts work. Both families are ectoparasites which frequently leave their hosts, and they retain their segmentation and powers of swimming. Perhaps the best-known form is the Hersiliid, Giardella callianassae, which lives in the adult state in the galleries excavated in the sand by Callianassa subterranea, gaining its nourishment as an ectoparasite on the Decapod. The larvae are pelagic, and are said by Thomson^[55] to occur in Liverpool Bay.

List^[56] describes Gastrodelphys, a parasite of doubtful position, from the gills of tubicolous worms, such as Myxicola and Sabella, which possesses a perfectly siphonostomatous mouth.

The remaining families to be dealt with are those containing all the fish-parasites which possess a true siphonostome, as well as the siphonostomatous family Choniostomatidae, which is parasitic on other Crustacea. In all these forms the mouth is prolonged into a tube in which the styliform mandibles work.

Fam. 10. Caligidae.—Ectoparasites on fish, lodging most frequently in the gill-chamber. In most of the genera the segmentation and power of swimming are retained in both sexes, the sexual dimorphism not being very well marked, though the males are smaller than the females, and were in some cases originally described as belonging to a special genus Nogagus. 74The females carry two long egg-sacs; the general structure may be made out from the ventral view of Caligus nanus (Fig. 43).

Some of the Caligidae are distinguished by the terga of the thoracic segments being expanded to form large chitinous elytra, e.g. Cecrops, found parasitic on the gills of the Tunny and on the Sun-fish (Orthagoriscus mola). Caligus rapax is parasitic on the skin and in the gills of Sea-Trout, Pollan, etc.; and C. lacustris is common in fresh-water lakes and streams on Pike and Carp.

Fig. 43.—Caligus nanus, × 10. Abd.1, 1st abdominal segment; Ant.1, Ant.2, 1st and 2nd antennae; Mx.1, Mx.2, 1st and 2nd maxillae; Mxp, maxillipede; s, siphon; Th.1, Th.5 1st and 5th thoracic appendages. (After Gerstaecker.)

Fig. 44.—Lernaea branchialis from the Haddock, ♀, × 1. Ceph, cephalothorax; e.s, egg-sacs. (After Scott.)

Fam. 11. Lernaeidae.—These parasites burrow with their heads deep into the skin, or even into the blood-vessels or body-cavity, of various marine fish. The body of the adult female Lernaea is extraordinarily deformed, consisting of a mere shapeless sac with irregular branched processes on the head, and two egg-sacs attached behind (Fig. 44). Pennella sagitta^[57] bores so deeply into the flesh of its host, Chironectes marmoratus, that only the egg-sacs and some remarkable branchial processes attached to its abdomen protrude outside the host to the 75exterior. Peroderma cylindricum bores similarly into the flesh of the Sardine, and where it is common, inflicts considerable damage. The males of these curious animals are of more normal structure (Fig. 45). Claus^[58] states that fertilisation takes place when both sexes are free-swimming, and of a more or less similar structure, and that subsequently the female becomes fixed to her host and degenerates into the shapeless mass shown in Fig. 44.

Fig. 45.—Lernaea branchialis, ♂, × 10. Ant.1, Ant.2, 1st and 2nd antennae; Br, brain; e, eye; g, stomach; t, testis; vd, vas deferens; ves. sem, vesicula seminalis. (After Claus.)

Fig. 46.—Achtheres percarum. A, ♀, × 4; B, ♂, × 4. Ant.2, 2nd antenna; g, stomach; Mx.2, 2nd maxilla; Mxp, maxillipede; ov, ovary; ovd, oviduct. (After Gerstaecker.)

Fam. 12. Lernaeopodidae.—This family may be illustrated by the common gill-parasite of Perch and Trout, known as Achtheres percarum. The female (Fig. 46), which is much larger than the male, and is not clearly segmented, is attached to the host by means of the maxillipedes, which are fused distally into a pad armed with chitinous hooks. In the male the maxillipedes 76are prehensile, but are not so fused. Besides Achtheres there are other fresh-water forms, e.g. Lernaeopoda salmonea on Salmon, and a number of marine genera. It appears that the larvae fix themselves to their hosts by means of a long glandular thread, which proceeds from the middle of the forehead.^[59]

Fig. 47.—Ventral view of Stenochotheres egregius (Choniostomatidae), ♂. A, A′, 1st and 2nd antennae; M, mouth; Mx, 2nd maxilla; T, 1st thoracic leg. (After Hansen.)

Fam. 13. Choniostomatidae.^[60]—The members of this family are all parasitic on other Crustacea. The majority live parasitically in the marsupial pouches of female Amphipods, Isopods, Mysidae, and Cumacea, e.g. Sphaeronella and Stenochotheres in the marsupia of Gammarids; but Choniostoma occurs in the branchial cavity of Hippolyte, Homoeoscelis is common in the branchial cavity of Diastylis and Iphinoe, and Aspidoecia on the outside of the body of the Mysid Erythrops. The males and females live together in the same marsupium, but the adult males retain the power of roving about, and do not feed so much as the females, though their mouth-parts are similarly constructed (Fig. 47). Representatives occur all over the world, but the majority of species known at present are from the North Sea, the most abundant being Stenochotheres egregius, parasitic on the Gammarid Metopa bruzelii, Goës.

The male bears a median glandular thread on the forehead by which it attaches itself to the females or to the host. Hansen considers that the family is most closely allied to the Lernaeopodidae.

77

Fig. 48.—Argulus foliaceus, young ♂, × 15. a¹, a², First and second antennae; ab, abdomen; E, compound eye; l, liver; m, mandibles and first maxillae; mx, second maxilla (the median eye is seen between the two second maxillae); mxp, maxillipede; s.g, shell-gland; sp, spine; t, testis; 1, 4, first and fourth swimming appendages. (After Claus.)

BRANCH II. BRANCHIURA.

Fam. Argulidae.^[61]—We have yet to mention this group of fish-parasites, related to the Copepoda, but occupying an isolated position. They are ectoparasites upon various species of fish, Argulus foliaceus being common in the fresh waters of Europe, infesting the branchial chamber or the skin of fresh-water fish, but being frequently taken swimming freely in the water. Both males and females can swim with great agility, and they leave their hosts regularly at the breeding season in spring and autumn; fertilisation is internal, and the female deposits the eggs on stones and other objects. After leaving its host, an Argulus, if it cannot find a fish of the same species, can live on almost any other species, and may even attack Frog tadpoles; while the kinds that infest migratory fish can change with their 78hosts from salt to fresh water, or the reverse. America appears to be the home of the Argulidae.^[62]

The structure of an Argulid is exhibited in Fig. 48. In front of the siphon, within which the styliform mandibles and first maxillae work, there is a poison-spine (sp); the appendages which correspond to the second maxillae (mx) are modified into sucking discs, but in the genus Dolops they terminate in normal claws. The next pair of appendages, usually spoken of as maxillipedes (mxp), are clasping organs, and behind follow four pairs of thoracic swimming feet (1–4). The body is foliaceous, and they always apply themselves to their hosts with the long axis pointing forwards and parallel to that of the host, while on various parts of the under surface of the body are spines pointing backward which prevent the parasite being brushed off by the passage of the host through the water. These animals, alone among the Copepoda, possess compound eyes.

A short sketch has now been given of the variations in Copepod organisation, but we cannot leave the subject without pointing out the rich field which still remains for the morphologist, especially in determining the true relationships of the parasitic families.

Order III. Cirripedia.

The Cirripedes are medium-sized Crustacea, with the body consisting of few segments, and enveloped in a mantle formed as a fold of the external integument, which may be strongly protected by calcified plates. The abdomen is greatly reduced. The larva, after hatching out as a Nauplius, and passing through a Cypris stage, when it resembles an Ostracod, fixes itself to a foreign object by means of the first antennae, and becomes a pupa, which after profound changes gives rise to the adult.

All the Cirripedes, when adult, live either a fixed or parasitic existence, and as so frequently happens with animals of this kind, they have departed widely from the ordinary structure of the class to which they belong. Their anomalous appearance and the mystery surrounding their propagation gave rise, probably, to the old legend that the Barnacles (Lepadidae), which live attached to pieces of floating timber, hatched out into Barnacle geese^[63]; and even so late as 1678, in the Royal 80Society’s Transactions, Sir Robert Moray describes what he takes to be little birds enclosed in Barnacle shells, washed ashore on the coast of Scotland: “The little Bill like that of a Goose, the Eyes marked, the Head, Neck, Breast, Wings, Tail, and Feet formed, the Feathers everywhere perfectly shaped and blackish coloured, and the Feet like those of other Water-fowl, to my best remembrance.” Cuvier in his classification of the animal kingdom included them in the Mollusca; and it was not until 1830 that J. V. Thompson described their larval stages, and showed conclusively that they belonged to the Crustacea. Since the work of this naturalist a number of observers have securely founded our knowledge of the group, but we may especially mention the epoch-making works of Darwin,^[64] Hoek,^[65] and latterly of Gruvel.^[66]

The young Cirripede is hatched out from the maternal mantle-cavity as a free-swimming Nauplius, a larval form common to most of the Entomostraca and to some Malacostraca; the Cirripede Nauplius (Fig. 49) is characterised by the presence of well-developed frontal horns, and usually by the long spiny processes which spring from various parts of the body. As an introduction to the study of the group, it will be well to follow the transformations of this larva in Lepas up to the period when it begins its sessile existence. The liberated Nauplii swim freely near the surface of the sea, and remaining in this condition for several days are dispersed widely from their birthplace; they are then transformed by the process of moulting into the second larval stage, known as the Cypris (Fig. 50), from its resemblance to a bivalve Ostracod. The Cypris larva continues to swim about by means of the six pairs of biramous thoracic legs until it finds a suitable place on which to fix; in the case of Lepas fixation usually takes place on loose floating logs; the Cypris fixes itself by means of the first antennae, at the bases of which a large cement-gland secretes an adhesive substance. The biramous swimming legs are cast off, and six pairs of biramous cirri characteristic of the adult take their 81place; at this stage the body has the appearance shown in Fig. 51. The region of the head at the base of the antennae now becomes greatly swollen and elongated to form the peduncle or stalk of the adult; the larval bivalve carapace is cast off and on the external surface of the mantle the calcifications begin which will give rise to the exoskeletal plates of the adult. This region is known as the “capitulum” of the adult, as opposed to the “peduncle.” The young Cirripede is now known as a pupa, and from this stage the adult form is reached by a gradual transition.

Fig. 49.—Nauplius larva of Lepas fascicularis, × 12. A₁, A₂, 1st and 2nd antennae; B, brain; E, eye; H, fronto-lateral horn; M, mandible; S, stomach. (After Groom.)

82

Fig. 50.—Cypris stage in the development of Lepas australis, × 15. A, Peduncle; A.M, adductor muscle; C, caecum of oesophagus; C.g, cement-glands; Cr, cirri (thoracic appendages); E, compound eye; E¹, simple eye; G, ventral ganglia; I, intestine; M, mouth; M.C, mantle-cavity; O, ovary; S, stomach. (After Hoek.)

Fig. 51.—Pupa of Lepas pectinata, × 8. A, Antenna; C, carina; M, adductor muscle; S, scutum; T, tergum. (After Gruvel.)

The body of the adult Lepas is retracted into the mantle, and lies free in the mantle-cavity, but is continuous anteriorly with the tissues of the peduncle, into which the mantle does not extend. The thorax, with its six pairs of legs, can be protruded from the mantle-cavity through the slit-like opening which separates the two valves of the mantle along the ventral middle line; and when the animal is feeding, the thoracic legs are so protruded, and by their concerted waving action they drive the food-particles in the water along the channel between them, until the particles reach the oral cone, where they are masticated by the mandibles and two pairs of maxillae, and so passed into the alimentary canal. When the animal is disturbed it rapidly retracts its limbs, the valves of the mantle are closed by means of a strong adductor muscle in the head, and the animal is protected from all external influences. In the acorn-barnacles (Operculata), 83which live in great numbers attached to rocks and other objects between tide-marks, the body is constructed on a similar plan, save that there is no stalk, and the body is completely enclosed in a hard calcareous box formed from the mantle, which, when the valves are closed, as they always are during low tide, completely protect the animal inside from desiccation or danger of any kind. Besides the cement-glands situated in the peduncle, we can distinguish the generative organs, consisting of a pair of ovaries and testes, the majority of Cirripedes being hermaphrodite. The testes open at the end of an elongated median penis behind the thoracic limbs, while the ovaries, situated in the peduncle, have paired openings into the mantle-cavity on either side of the head. A pair of maxillary glands or kidneys are present, and the alimentary canal is provided with various digestive glands. Special branchial organs are not present in the Pedunculate Cirripedes, but in the Operculate genera two branchiae are formed from the plications of the internal surface of the mantle. There is no contractile heart, and the circulatory system is poorly developed. The Cirripedes are badly furnished with sensory organs; the remains of a simple Nauplius eye may persist, situated on the upper part of the stomach, but the chief sense-organs are the sensory hairs upon the limbs.

Fig. 52.—A, Dwarf male of Scalpellum vulgare, × 27; B, diagram of Stalked Barnacle. a, Peduncle; al, alimentary canal; b, brain; c, carina; e, remains of Nauplius eye; gl, cement-gland; m, mantle-cavity; o, its opening; ov, ovary; p, penis; s, scutum; t, testis; tm, tergum, seen in A as the shaded body above the reference-line of e and to the right of the carina, on the left of the figure.

The recent Cirripedes fall into six clearly defined Sub-orders.

In this division, sometimes combined with the Operculata as Thoracica, owing to the extremely reduced state of the abdomen, the body is borne on a distinct stalk, and the bivalve arrangement of the mantle is clearly retained. The mantle is protected externally by a number of calcareous plates, the arrangement of which is typical of the various genera. It appears that in the most primitive and geologically oldest Cirripedes, the probable ancestors of the Pedunculate and Operculate sub-orders, the arrangement of the plates was somewhat irregular, and they were far more numerous than in the modern forms, so that passing from these older types to modern times we witness a reduction in the number and a greater precision in the arrangement of the skeletal parts.

Fig. 53.—A, Turrilepas wrightianus (Silurian), × 1; B, Archaeolepas redtenbacheri (Jurassic), × 1. C, carina; R, rostrum; S, scutum; T, tergum. (After Zittel.)

One of the most ancient Cirripedes known is Turrilepas, which occurs in the Silurian deposits of England, but it is also known from earlier deposits, while undoubted Cirripedes have been found in the Cambrian of North America. The body of Turrilepas is enclosed in imbricating plates, as shown in Fig. 53, A.

In Archaeolepas of the Upper Jurassic (Lithographic slates of Bavaria) the arrangement of scutes typical of the Lepadidae is foreshadowed, but the whole of the peduncle is protected by rows of plates (Fig. 53, B), as in Turrilepas.

The above-mentioned genera did not survive into the Cretaceous period, their places being taken by the genera Pollicipes and Scalpellum, which first appeared in the Silurian and persist to the present time, the older and more primitive Pollicipes being represented by about half a dozen living species, while the species of Scalpellum are exceedingly numerous.

Fam. 1. Polyaspidae.—This family includes the three genera, Pollicipes, Scalpellum, and Lithotrya.

Fig. 54.—Pollicipes mitella, × 1. (After Darwin.)

Pollicipes is not only very ancient geologically (being found from the Ordovician upward), but it preserves the primitive characteristic 85of numerous skeletal plates, the peduncle being frequently covered with small calcareous pieces, which graduate into the larger more regularly placed scutes on the capitulum (Fig. 54). The species of this genus, many of which are among the largest Cirripedes, are widely distributed in the temperate and tropical seas, living for the most part attached to rocks and often in deep water. P. cornucopia occurs off the English and Scottish coasts.

The members of the genus Scalpellum, which is represented by exceedingly numerous species in the Cretaceous period, also possess a large number of plates on the capitulum, and often on the peduncle as well, but never so many as in Pollicipes. Although the arrangement of the plates varies much in the different species, we may describe a fairly typical case, that of the common Scalpellum vulgare (Fig. 55, B).

The valves of the capitulum are held together by the median dorsal piece called the “carina”; the other unpaired skeletal piece is the “rostrum,” in front, just below the place where the valves gape to allow the protrusion of the limbs. The paired pieces receive the names “scutum,” “tergum,” and “laterals,” and the peduncle is covered with rows of small plates.

The genus Scalpellum is a very large one, and is widely distributed, though at the time at which Darwin wrote only six species were known. The reason for this is to be found in the fact that the great majority of the species live at great depths, so that they remained unknown until the expeditions of the Challenger and other deep-sea expeditions brought them to light. They may affix themselves, generally in considerable numbers together, on branching organisms, such as Corals, Polyzoa, and Hydroids, but often also on empty shells, rocks, and other foreign bodies. The body is colourless or of a pale flesh colour, but a colony of these animals, expanded and drooping in various attitudes from a piece of coral, gives the appearance of some graceful exotic flower.

86

Fig. 55.—A, Complemental male of Scalpellum peronii, × 20; B, hermaphrodite individual of S. vulgare, × 2. a, Complemental males, in situ; b, rostrum. (A, after Gruvel; B, after Darwin.)

Perhaps the most interesting feature of the genus is the remarkable variation in the sexual constitution of some of the species. The great majority of the Pedunculata and all the Operculata are hermaphrodites, which habitually cross-fertilise one another, and this they are well fitted to do, since they all live gregariously and are provided with a long exsertile penis for transferring the spermatozoa from one to the other. In Pollicipes, however, the individuals of which often live solitarily, it appears that self-fertilisation may occur. In Scalpellum three different kinds of sexual constitution may occur: (1) According to Hoek in S. balanoides, taken by the Challenger, the individuals are ordinary cross-fertilising hermaphrodites. (2) In the great majority of species, including the common S. vulgare, as originally described by Darwin, and since confirmed by Hoek and Gruvel,^[67] the individuals are hermaphrodite, but there are present affixed to the adult hermaphrodites, just inside the opening of the valves in a pocket of the mantle, a varying number of exceedingly minute males, called by Darwin “complemental males.” These tiny organisms are really little more 87than bags of spermatozoa, but they possess to varying degrees the ordinary organs of the adult in a reduced condition. The male of S. peronii (Fig. 55, A) retains the shape and skeletal plates of the ordinary form, and differs chiefly in its reduced size; but the more common condition is exhibited by the male of S. vulgare (Fig. 52, A), where the scutes are reduced to vestiges round the mantle-opening, and almost the whole of the body is occupied by the greatly developed generative organs. (3) In a few species, e.g. S. velutinum and S. ornatum, the individuals are purely dioecious, being either females of the ordinary structure resembling the hermaphrodites of the other Lepadidae, or dwarfed males resembling closely the complemental males described above for S. vulgare.

Fig. 56.—Lithotrya dorsalis, x 1. B, Basal calcareous cup; C, carina; R, rostrum; S, scutum; T, tergum. (After Darwin.)

The nature and derivation of these various conditions will be discussed when the parallel cases found in Ibla and among the Rhizocephala have been described.

The remaining genus of the Polyaspidae, also characterised by the presence of numerous skeletal plates on the capitulum, is Lithotrya (Fig. 56), which bores into rocks and shells, and is an inhabitant of the warm and tropical seas.

The peduncle of the full-grown animal is completely imbedded in the rock or shell to which it is attached, and at the basal end of the peduncle is situated a cup composed of large irregular calcified pieces. This cup is, however, not formed until the animal has ceased to burrow. The excavation of the substratum is effected by means of a number of small rasping plates which cover the peduncle, the whole being set in motion by the peduncular muscles.

Fig. 57.—Conchoderma virgata, × 1. C, Carina; S, scutum; T, tergum. (After Darwin.)

Fam. 2. Pentaspidae.—In this family are placed a number of genera, and among them the common Lepas, the species of which possess typically five skeletal plates, viz., a carina and a pair of scuta and of terga, the peduncle being naked. These forms are a later development of Cirripede evolution, and did not come into existence till Tertiary times. Some 88of them, e.g. Oxynaspis, live at considerable depths attached to corals, etc., but large numbers float on the surface of the sea, fixed often on logs and wreckage of various kinds. Dichelaspis is found attached to the shells of large Crustacea.

Conchoderma is an interesting genus, the species of which live affixed to various floating objects, the keels of ships, etc.; the mantle is often brilliantly coloured, as in C. virgata, and the skeletal plates are reduced to the merest vestiges, leaving the greater part of the body fleshy.

Fig. 58.—Ibla cumingii, ♀, × 1. S, Scutum; T, tergum. (After Darwin.)

Fig. 59.—Ibla cumingii, dwarf male, × 32. A, Antennae; B, part of male imbedded in the female, to which the torn membrane M belongs; E, eye; Th, thoracic appendages or cirri. (After Darwin.)

Fam. 3. Tetraspidae.—This family includes the single genus Ibla (Fig. 58), which possesses only four skeletal plates, a pair of terga and of scuta, coloured blue, while the peduncle is covered with brown spines. There are only two very similar species known, I. cumingii, which is found attached to the peduncle of Pollicipes mitella, and I. quadrivalvis, living on masses of the Siphonophore Galeolaria decumbens. These two species are quite different in the partition of the sexes. In I. cumingii the large individuals of normal structure are females, inside the mantle-cavities of which are attached dwarf males of the form shown in Fig. 59.

These organisms have the peduncle buried completely in the substance of the female’s mantle, inside 89which they live; they exhibit a degenerate structure, but still retain two pairs of cirri. The large individuals of I. quadrivalvis, on the other hand, are hermaphrodites, but they harbour within their mantles minute complemental males similar to those of I. cumingii, though they are rather larger.

Fig. 60.—Diagram of the shell of an Operculate Cirripede. a “Ala,” or overlapped portion of a “compartment”; B, basis; C, carina; C.L, carino-lateral; L, lateral; R, rostrum; r, “radius,” or overlapping portion of a compartment; R.L, rostro-lateral. (After Darwin.)

Fam. 4. Anaspidae.—This includes the remaining pedunculate genera, characterised by the fleshy nature of the mantle and peduncle, which are both entirely devoid of calcifications. The species of Alepas live upon Echinoderms and various other animals; Chaetolepas upon Sertularia, and Gymnolepas upon Medusae. Anelasma squalicola is an interesting form, living parasitically upon the Elasmobranch fishes, Selache maxima and Spinax niger in the North Sea. The peduncle is deeply buried in the flesh of the host, so that only a portion of the dark blue capitulum protrudes to the surface. From the whole surface of the peduncle a system of branching processes is given off, which ramify far into the tissues of the fish, and communicate inside the peduncle with the lacunar tissue, which is packed round all the organs of the Cirripede. There can be small doubt that the Anelasma derives its nutriment parasitically through this root-system, since the cirri are mere fleshy lobes unadapted to securing food, and the alimentary canal is always empty. This animal has a suggestive bearing on the Rhizocephala, which, as will be shown, derive their nutriment from a system of roots penetrating the host and growing out from what corresponds morphologically to the peduncle.

Sub-Order 2. Operculata.

Fig. 61.—Balanus tintinnabulum, with the right half of the shell and of the operculum removed, seen from the right side. A, Antennae, the size of which is exaggerated; A.M, adductor muscle; B, basis; C, carina; Cr, cirri or thoracic appendages; D, oviduct; G, ovary; L, lateral compartment; Lb, labrum or upper lip; M, M, depressor muscles of scutum and tergum; M.C, mantle-cavity; O, orifice of excretory organ; O.M, opercular membrane; R, rostrum; S, scutum; St, region of stomach; T, tergum. (After Darwin.)

The “acorn-barnacles” appear later in geological time than the earlier stalked forms. Verruca and Chthamalus are found in the Chalk, and survive down to the present day, but Balanus does not occur until middle Tertiary times. Representatives of the last-named genus are familiar to every one, 90as the hard sharp objects which cover rocks and piles near high-water mark on every sea-coast. If we examine the hard skeleton of one of these animals, we find that, unlike the Pedunculata, they possess no stalk, the capitulum being fused on to the surface of attachment by a broad basal disc. Typically, there may be considered to be eight skeletal pieces forming the outer ring which invests the soft parts of the animal, an unpaired rostrum and carina, and laterally a pair of rostro-lateral, lateral, and carino-lateral “compartments,” as shown in Figs. 60, 63.

The skeletal ring is roofed over by a pair of terga at the carinal end and a pair of scuta at the rostral end; these four plates make up the operculum by which the animal can shut itself completely up in its shell, or between the valves of which it can protrude its limbs for obtaining food.

Fig. 62.—Diagrammatic section of the growing shell of Balanus porcatus. C, Canals; Ct, cuticle; H, hypodermis (= epidermis); H′, part of shell secreted by the hypodermis; Hl, hypodermal lamina; M, part of shell secreted by the mantle. (After Gruvel.)

The relation of the animal to its shell is shown in Fig. 61. The shell in the Operculata is not merely secreted as a dead structure on the external surface of the epidermis, but represents a living calciferous tissue interpenetrated by living laminae 91(Fig. 62, Hl) derived partly from the external hypodermis and partly from the lining of the mantle. The hard parts of the shell usually also contain spaces and canals (C).

Fig. 63.—Diagrams of shells of Operculata. A, Catophragmus (Octomeridae); B, Balanus, Coronula, etc. (Hexameridae); C, Tetraclita (Tetrameridae). C, carina; C.L, carino-lateral; L, lateral; R, rostrum; R.L, rostro-lateral.

The various forms of Acorn-barnacle may be classified according to the number of pieces that go to make up the skeleton; thus starting with the typical number eight (Fig. 63, A), we find that in various degrees a fusion between neighbouring pieces has taken place in the different families.

Fam. 1. Verrucidae.—The ancient genus Verruca, which is still widely distributed in all seas, and is found fixed upon foreign objects on the sea-bottom at various depths, is interesting on account of the asymmetry of its shell, which bears a different aspect according to which side one regards it from. This asymmetry is brought about by the skeletal pieces (carina, rostrum, and paired terga and scuta) shifting their positions after fixation has taken place.

Fam. 2. Octomeridae.—In this family the eight plates composing the shell are separate and unfused (Fig. 63, A). The majority of the species come from the Southern hemisphere, e.g. the members of the genera Catophragmus and Octomeris, but Pachylasma giganteum occurs in deep water in the Mediterranean, where it has been found fixed upon Millepore corals.

Fam. 3. Hexameridae.—This family includes by far the greater number of the Acorn-barnacles, in which only six plates are present, the laterals having fused with the carino-laterals (Fig. 63, B). The very large genus Balanus belongs here, the common B. tintinnabulum of our coasts being found all over the world, and occurring under a number of inconstant varietal forms. Especial interest attaches to certain other genera, from their habit of living parasitically on soft-bodied animals, whose flesh they penetrate.

Coronula diadema and Tubicinella trachealis live embedded 92in the skin of whales, the shell of the first-named being of a highly complicated structure with hollow triangular compartments into which the mantle is drawn out.

Xenobalanus globicipitis lives attached to various Cetacea, and is remarkable for the rudimentary condition of its skeleton, the six plates of which form a mere disc of attachment from which the greatly elongated naked body rises, resembling one of the naked Stalked Barnacles.

Fam. 4 Tetrameridae.—In this family only four skeletal plates are present (Fig. 63, C). This family is chiefly confined to tropical seas or those of the Southern hemisphere. The chief genera are Tetraclita and Pyrgoma, found in British seas.

Sub-Order 3. Acrothoracica.

Gruvel includes in this sub-order four genera (Alcippe, Cryptophialus, Kochlorine, and Lithoglyptes) the species of which live in cavities excavated in the shells of molluscs or in the hard parts of corals.

Fig. 64.—Alcippe lampas. A, ♀, × about 10, seen from the right side, with part of the right half of the animal removed; B, dwarf male, × about 30. A.M, adductor muscle; An, antenna; C, 1st pair of cirri; Cr, posterior thoracic appendages; E, eye; G, testis; M.C, mantle-cavity; O, ovary; P, penis; T, penultimate thoracic segment; V. vesicula seminalis. (After Darwin.)

Darwin discovered and described Cryptophialus minutus, and placed it in a sub-order Abdominalia, believing that it was 93distinguished from all the foregoing Cirripedes by the presence of a well-developed abdomen. Since the discovery of other allied genera, it has been decided that the abdomen is equally reduced in these forms, and that the terminal appendages do not belong to this region, but to the thorax.

The sexes are separate. The body of the female (Fig. 64, A) is enclosed in a chitinous mantle, armed with teeth by which the excavation is effected, and is attached to the cavity in the host by means of a horny disc. Upon this disc the dwarf males (B) are found.

Alcippe lampas inhabits holes on the inner surface of dead Fusus and Buccinum shells; Cryptophialus minutus the shells of Concholepas peruviana; C. striatus ^[68] the plates of Chiton; Kochlorine hamata the shells of Haliotis; and Lithoglyptes varians shells and corals from the Indian Ocean.

Sub-Order 4. Ascothoracica.

These are small hermaphrodite animals completely enveloped in a soft mantle, which live attached to and partly buried in various organisms, such as the branching Black Corals (Gerardia). They retain the thoracic appendages in a modified state, and the body is segmented into a number of somites, the last of which probably represents an abdomen.

Laura gerardiae, described by Lacaze Duthiers,^[69] is parasitic on the stem of the “Black Coral,” Gerardia (vol. i. p. 406); it has the shape of a broad bean, the body being entirely enclosed in a soft mantle, with the orifice in the position corresponding to the hilum of a bean. The body lying in the mantle is composed of eleven segments, and is curved into an S-shape. Its internal anatomy is entirely on the plan of an ordinary Cirripede.

Petrarca bathyactidis, G. H. Fowler,^[70] was found in the mesenteric chambers of the coral Bathyactis, dredged by the Challenger from 4000 metres. The body is nearly spherical, and the mantle-opening forms a long slit on the ventral surface. The mantle is soft, but is furnished on the ventral surface with short spines.

The antennae, which form the organs of fixation, remain 94very much in the state characteristic of the Cypris larvae of other Cirripedes, being furnished with two terminal hooks by which attachment is effected. The thoracic appendages, of which there are the normal number six, are reduced flabellate structures, and the abdomen forms an indefinitely segmented lobe of considerable size.

The animal appears to be in an arrested state of development, and so retains some of the characteristics of the Cypris larvae, but it is very doubtful how far these characters can be considered primitive.

Other forms are Dendrogaster astericola on Echinoderms, and Synagoga mira on the “Black Coral,” Parantipathes larix, at Naples.

Sub-Order 5. Apoda.

Fig. 65.—Proteolepas bivincta, × 26. A, Antennae; a, b, 1st and 2nd abdominal segments; O, ovary; P, penis; T, telson; 1–8, thoracic segments. (After Darwin.)

Darwin described a small hermaphrodite parasite in the mantle chamber of Alepas cornuta from Saint Vincent, West Indies, which he named Proteolepas bivincta.

The body (Fig. 65) is distinctly segmented into eleven somites, the last three of which are supposed to belong to the abdomen; there are no appendages except the antennae by which fixation is effected. The mouth-parts are of normal constitution.

This animal has not been found again since Darwin’s discovery, but Hansen^[71] describes a number of peculiar Nauplius larvae taken in the plankton of various regions, which he argues probably belong to members of this group. A wide field of work is offered in attempting to find the adults into which various larvae grow.

These remarkable animals are Cirripedes which have taken to living parasitically on various kinds of Crustacea; the majority infest species of Decapoda, e.g. Peltogaster on Hermit-Crabs, Sacculina on a number of Brachyura, Sylon on Shrimps, Lernaeodiscus on Galathea; but one genus, Duplorbis, has been found in the marsupium of the Isopod Calathura brachiata from Greenland. Most of the species are solitary, but a few, e.g. Peltogaster sulcatus, are social. In the adult state the body consists of two portions: a soft bag-like structure, external to the host, carrying the reproductive, nervous, and muscular organs, and attached to some part of the host’s abdomen by means of a chitinous ring; and a system of branching roots inside the host’s body, which spring from the ring of attachment and supply the external body with nourishment.

Fig. 66.—Nearly median longitudinal section (diagrammatic) of Peltogaster. gn, Brain; m, mantle; mc, mantle-cavity; mes, mesentery; op, mantle-opening; ov, ovary; ovd, oviduct; ring, ring of attachment; t, testis; vd, vas deferens.

Fig. 67.—Diagrammatic median longitudinal section through a normal Cirripede, gn, Brain; op, mantle-opening; ovd, oviduct; vd, vas deferens.

The structure of the external bag-like portion is very simple, and varies only in details, chiefly of symmetry, in the different genera. In Peltogaster, which preserves the simplest symmetrical arrangement of the organs, a diagrammatic section through the long axis of the body (Fig. 66) shows that it consists of a muscular mantle (m) surrounding a visceral mass, and enclosing a mantle-cavity (mc) or brood-pouch, which stretches everywhere between mantle and visceral mass, except along the surface by which the parasite is attached to its host, where a mesentery (mes) is formed. The ring of attachment is situated in the middle of this mesentery; the mantle-cavity, which is completely 96lined externally and internally with chitin, opens anteriorly by means of a circular aperture (op) guarded by a sphincter muscle. The visceral mass is composed chiefly of the two ovaries (ov), which open on either side of the mesentery by means of a pair of oviducts (ovd); the paired testes (t) are small tubes lying posteriorly in the mesentery, and the nervous ganglion (gn) lies in the mesentery between oviducts and mantle-opening. A comparison with the condition of a normal Cirripede (Fig. 67) shows us that the mesenterial surface of the parasite by which it is fixed corresponds to the dorsal surface of an ordinary Pedunculate Cirripede, and that the ring of attachment corresponds with the stalk or peduncle of a Lepas. The root-system passes out through the ring of attachment into the body of the host, and ramifies round the organs of the crab; the roots are covered externally with a thin chitinous investment, and consist of an epithelium and an internal mass of branching cells continuous with the lacunar tissue in the visceral mass.

97

Fig. 68.—Development of Sacculina neglecta. A, Nauplius stage, × about 70; B, Cypris stage, × about 70. A₁, A₂, 1st and 2nd antennae of Nauplius; Ab, abdomen; Ant, antenna of Cypris; E, undifferentiated cells; F, frontal horn; G, glands of Cypris; H, tendon of Cypris; M, mandible; T, tentacles.

The developmental history of the Rhizocephala is one of the most remarkable that embryology has hitherto revealed. It has been most accurately followed in the case of Sacculina. The young are hatched out in great numbers from the maternal mantle-cavity as small Nauplii (Fig. 68, A) of a typical Cirripede nature, but without any alimentary canal. They swim near the surface of the sea, and become transformed into Cypris larvae of a typical character (Fig. 68, B). The Cypris larva, after a certain period of free existence, seeks out a crab and fixes itself by means of the hooks on its antennae to a hair on any part of the crab’s body. Various races of Sacculina are known which infest about fifty different species of crabs in various seas; the best known are S. carcini parasitic on Carcinus maenas at Plymouth and Roscoff, and S. neglecta on Inachus mauritanicus at Naples. The antenna, by which the Cypris is fixed, penetrates the base of the hair; the appendages are thrown away, and a small mass of undifferentiated cells is passed down the antenna into the body-cavity of the crab. Arrived in the body-cavity it appears that this small mass of cells is carried about in the blood-stream until it reaches the spaces round the intestine in the thorax. Here it becomes applied to the intestine, usually at its upper part, immediately beneath the stomach of the crab (Fig. 69), and from this point it proceeds to throw out roots in all directions, and as it grows to extend its main bulk, called the central tumour (c.t), towards the lower part of the intestine. As the posterior border of the central tumour grows down towards the hind-gut, the future organs of the adult Sacculina become differentiated in its substance; the mantle-cavity being excavated and surrounding the rudiment of the visceral mass, while as the central tumour grows downwards it leaves behind it an ever extending system of roots. When the central tumour in process of differentiation has reached the unpaired diverticulum 98of the crab’s intestine, at the junction between thorax and abdomen, all the adult organs are laid down in miniature, and the whole structure is surrounded by an additional sac formed by invagination known as the perivisceral space (Fig. 70). The young “Sacculina interna” remains in this position for some time, and being applied to the ventral abdominal tissues of the crab just at the point where thorax and abdomen join, or a little below it, it causes the crab’s epithelium to degenerate, so that when the crab moults, a little hole is left in this region of the same size as the body of the Sacculina, owing to the failure of the epithelium to form chitin here; and thus the little parasite is pushed through this hole and comes to the exterior as the adolescent “Sacculina externa.” From this point onwards the crab, being inhibited in its growth through the action of the parasite, never moults again; so that the Sacculina occupies a safe position protruding from the crab’s abdomen, which laps over it and protects it. The remarkable features of this development are, firstly, the difficulty of understanding how the developing embryo is directed in its complicated wanderings so as always to reach the same spot where it is destined to come to the exterior; and, secondly, the loss after the Cypris stage of all the organs and the resumption of an embryonic undifferentiated state from which the adult is newly evolved. A certain parallel to this history is found in that of the Monstrillidae, described on pp. 64–66.

Fig. 69.—The mid-gut of Inachus mauritanicus with a young Sacculina overlying it, × 2. c.t, “Central tumour” of the parasite; d.i, d.s, inferior and superior diverticula of alimentary canal of host; n, “nucleus,” or body-rudiment of Sacculina; r, its roots; x, definitive position of the parasite.

Fig. 70.—Later stage in the development of the “Sacculina interna,” × 2. b, Body of Sacculina; c.t, “central tumour”; d.i, d.s, inferior and superior diverticula of alimentary canal of host; o, opening of perivisceral cavity of Sacculina; r, its roots.

99

Fig. 71.—Fourteen Cypris larvae fixed round the mantle-opening (o) of a young Sacculina externa, × 20.

The Rhizocephala are hermaphrodite with the possible exception of Sylon, which appears to be female and perhaps parthenogenetic, no male having been seen; but unlike most other hermaphrodite Cirripedes, they reproduce by a continual round of self-fertilisation. This is the more remarkable in that the vestiges of what appears to be a male sex are still found in Sacculina and Peltogaster; certain of the Cypris larvae in these genera, instead of fixing on and inoculating other crabs, become attached round the mantle-openings of young parasites of the same species as themselves, which have recently attained to the exterior of their hosts (Fig. 71). These larvae, which remind us of the complemental males in Scalpellum, etc., never produce spermatozoa, but rapidly degenerate where they are fixed, and appear never to play any rôle in the reproduction of their species. The nature of this remarkable phenomenon, together with the sexual condition of the Cirripedes in general, will be discussed in the next section.

Much remains to be elucidated in the life-histories of these curious animals, and it seems probable that intermediate stages may exist, showing us how the extreme discontinuity of development has been reached. Suggestive in this respect is the newly discovered parasite of the Isopod, Calathura, which the author has named Duplorbis calathurae.^[73] This animal does not appear 100to possess a root-system, but is attached to its host by a tube which passes right through the mesentery and opens into the mantle-cavity of the parasite. It may be suggested that this tube corresponds to the stalk of the normal Cirripede, but its exact mode of formation would certainly throw much light on the question of Rhizocephalan development.

Phenomena of Growth and Sex in the Crustacea.

In the foregoing account of the Cirripedia we have met with certain peculiar sexual relations in which closely allied species exhibit marked differences in regard to the distribution of the qualities of sex among their individuals; we have seen that the majority of species are hermaphrodite, unlike most Crustacea which, with the other exception of the parasitic Isopoda, are normally dioecious; and that in some species complemental males exist side by side with the hermaphrodites, while, in yet others, the individuals are either females or dwarf males.

Before examining the causes of these conditions, it will be opportune to consider a number of facts which throw light on the question of sex and hermaphroditism in general. We may then return to the discussion of the hermaphroditism found in particular in the Cirripedia and Isopoda.

Parasitic Castration.—Giard^[74] was the first to observe that a number of parasites exert a remarkable influence on the sexual characters of their host, such that the generative glands become reduced, or may completely degenerate, while the secondary sexual characters become materially altered. This was proved to occur in the most widely different hosts, affected by the most widely different parasites (e.g. Crustacea, Insecta, Worms). Moreover, it was apparent that the affection does not consist in the parasite merely destroying the generative organs, with which it often does not come into contact, but rather in the general disturbance of the metabolism set up by its presence.

The most completely studied cases of parasitic castration are those of the Rhizocephalous Sacculina neglecta, parasitic on the spider-crab, Inachus mauritanicus,^[75] and of Peltogaster curvatus 101on the Hermit-crab, Eupagurus excavatus, var. meticulosa.^[76] The ordinary males of I. mauritanicus have the appearance shown in Fig. 72, A. The abdomen is small and bears a pair of copulatory styles, while the chelipedes are long and swollen. In the female (B) the abdomen is much larger and trough-shaped, and carries four pairs of ovigerous appendages; the chelae are small and narrow.

Fig. 72.—Illustrating the effect of parasitic Sacculina neglecta on Inachus mauritanicus, nat. size. A, Normal male; Inachus; B, normal female; C, male infested by Sacculina (final stage); D, abdomen of infested female; E, infested male in an early stage of its modification.

Now it is found that in about 70 per cent of males infected with Sacculina the body takes on to varying degrees the female characters, the abdomen becoming broad as in the female, with a tendency to develop the ovigerous appendages, while the chelae become reduced (Fig. 72, C). This assumption of the female characteristics by the male under the influence of the parasite may be so perfect that the abdomen and chelae become typically female in dimensions, while the abdomen develops not only the 102copulatory styles typical of the male, but also the four pairs of ovigerous appendages typical of the female. The parasitised females, on the other hand, though they may show a degenerate condition of the ovigerous appendages (Fig. 72, D), never develop a single positively male characteristic. On dissecting crabs of these various categories it is found that the generative organs are in varying conditions of degeneration and disintegration.

The most remarkable fact in this history is the subsequent behaviour of males which have assumed perfect female external characters, if the Sacculina drops off and the crabs recover from the disease. It is found that under these circumstances these males may regenerate from the remains of their gonads a perfect hermaphrodite gland, capable of producing mature ova and spermatozoa. The females appear quite incapable, on the other hand, of producing the male primary elements of sex on recovery, any more than they can produce the secondary. Exactly analogous facts have been observed in the case of the hermit-crabs parasitised by Peltogaster, but here the affected males produce small ova in their testes before the parasite is got rid of. Here, too, the females seem incapable of assuming male characters under the influence of the parasite.

To summarise shortly the conclusions to be deduced from these facts—certain animals react to the presence of parasites by altering their sexual condition. This alteration consists in the female sex in an arrest of reproductive activity, in the male sex in the arrest of reproductive activity coupled with the assumption of all the external characters proper to the female. But in these males it is not merely the external characters that have been altered; their capacity for subsequently developing hermaphrodite glands shows that their whole organisation has been converted towards the female state. That this alteration consists in a reorganisation of the metabolic activities of the body is clearly suggested, and in the succeeding paragraph we furnish some further evidence in support of this view.

103

Fig. 73.—Inachus mauritanicus, × 1. A, Low male; B, middle male; C, high male; the great chela of the right side is the only appendage represented.

Partial and Temporary Hermaphroditism. High and Low Dimorphism.

The reproductive phases of animals are frequently rhythmic, periods of growth alternating with periods of reproduction. This is well exemplified in the case of the ordinary males of Inachus mauritanicus, of some other Oxyrhynchous crabs, and of the Crayfish Cambarus.^[77] During the breeding season the males of I. mauritanicus fall into three chief categories: Small males with swollen chelae (Fig. 73, A), middle-sized males with flattened chelae (B), and large males with enormously swollen chelae (C). On dissecting specimens of the first and third categories it is found that the testes occupy a large part of the thoracic cavity and are full of spermatozoa, while in the middle-sized males with female-like chelae the testes appear shrivelled and contain few spermatozoa. These non-breeding crabs are, in fact, undergoing a period of active growth and sexual suppression before attaining the final state of development exhibited by the large breeding males. This phenomenon is obviously parallel to the “high and low dimorphism”^[78] so common in Lamellicorn beetles, where the males of many species are divided into two chief categories, viz. “low males” of small size in which the secondary sexual characters are poorly developed, and “high males” of large size in which these characters are proportionately 104much more highly developed than in the low males. The only difference between the two cases is that whereas in the beetles growth ceases on the attainment of maturity in the low degree, in the Crustacea the low male passes through a period of growth and sexual suppression to reach the high degree of development.

The condition of the middle-sized males may be looked upon as one of partial hermaphroditism, indications of the female state being found in the flattened chelae and in the reduced state of the testes. This interpretation is greatly strengthened by the state of affairs observed in the life-history of the male Sand-hoppers, Amphipods of the genus Orchestia.^[79] In the young males of several species of this genus, at the time of year when they are not actively breeding, small ova are developed in the upper part of the testes of more than half of the male individuals, these ova being broken down and reabsorbed as the breeding season reaches its height. Nor is this phenomenon confined to this genus; in the males of a number of widely different Crustacea these small ova are found in the testes at certain periods of the life-history (e.g. Astacus^[80]), when the animal is not breeding.

The foregoing facts indicate unmistakably that the males of a number of Crustacea under certain metabolic conditions, i.e. when a stage of active growth as opposed to a stage of reproductive activity is initiated, alter their sexual constitution in such a way that the latent female characteristics are developed, and the organism appears as a partial hermaphrodite. In the preceding paragraph we saw that the males of a number of animals, especially Crustacea, react to the metabolic disturbance set up by the presence of a parasite in exactly the same way, i.e. by developing into partial or total hermaphrodites. From these two converging bodies of facts we may conclude, firstly, that sex and metabolism are two closely connected phenomena; and, secondly, that the male sex is especially liable to assume hermaphrodite characters whenever its metabolic requirements are conservative, assimilatory, or in a preponderating degree anabolic, as when a phase of active growth is initiated, or the drain on the system, due to the presence of a parasite, is to be made good.

The above-mentioned groups contain the only normally hermaphrodite Crustacea, and since they are in most respects highly specialised, we may be certain that they have been secondarily derived from dioecious ancestors. They both lead a sessile or parasitic life, and it is noteworthy that this habit is often associated with hermaphroditism, e.g. in Tunicates. A sessile or parasitic mode of life is one in which the metabolic functions are vegetative and assimilatory rather than actively kinetic or metabolic. It is in this state that we have seen the males of a number of Crustacea taking on a temporary or partial hermaphroditism. We may, therefore, inquire, whether in these cases of normal hermaphroditism there is any evidence to show that here too the hermaphroditism has been acquired by the male sex as a response to the change in the metabolic conditions. In the parasitic Isopoda Epicarida (see pp. 129–136) the hermaphroditism is of a very simple kind; all the individuals are at first males, whose function it is to fix on and fertilise the adult parasites. These subsequently develop into females which are in their turn cross-fertilised by the young larvae derived from a previous generation. All the individuals being alike, it seems probable that they have been derived from one sex, and the general nature of hermaphroditism deduced above may lead us to suppose that that sex was originally male, the female having been suppressed. In certain Cirripedia, e.g. most species of Scalpellum, there exist, besides the hermaphrodite individuals, complemental males, so that here a superficial conclusion might be drawn that the hermaphrodites represent the female sex. But if we can suggest that the complemental males are in reality similar in derivation to the hermaphrodite individuals, we shall be in a position to claim that the hermaphrodite Cirripedes are similar to the Isopoda Epicarida, and have probably also been derived from the male sex. There is decided evidence pointing to this conclusion. In the first place, the complemental males of at least one species of Scalpellum, S. peronii, do show an incipient hermaphroditism^[81] in the presence 106of small ova in their generative glands, which, however, never come to maturity.

The condition of the degenerate males in the Rhizocephala may also be interpreted in the same manner. These never pass beyond the Cypris stage of development, in which they resemble in detail the Cypris larvae of the ordinary hermaphrodite individuals, and they are quite useless in the propagation of their species.

It is more reasonable to suppose that these Cypris larvae, which fix on the mantle-openings of adult parasites, are in reality identical with the ordinary Cypris which infest crabs and develop into the hermaphrodites, than that they represent a whole male sex doomed beforehand to uselessness and degeneration. If we suppose that the Cirripedes have passed through a state of protandric hermaphroditism similar to that of the Isopoda Epicarida, it is plain that all the larvae must have originally possessed the instinct of first fixing on the adult parasites, and we may suppose that this instinct has been retained in the Rhizocephala, but is now only actually fulfilled by a certain proportion of the larvae, which, under existing circumstances, are useless and fail to develop further; while the rest of the larvae, not finding an adult parasite to fix upon, go straight on to infect their hosts and develop into the adult hermaphrodites.

The same explanation would apply to the complemental males in Scalpellum, etc., these individuals being also potential hermaphrodites, which are arrested in development, though not so completely as in the Rhizocephala, owing to the position they have taken up.

This theory throws light on another dark feature of Cirripede life-history, namely, the gregarious instinct. The associations of Cirripedes are not formed by a number of Cypris larvae fixing together on the same spot, but rather by the Cypris larvae seeking out adolescent individuals of their own species and fixing on or near them. Now, if we suppose that the Cirripedes have passed through a condition of protandric hermaphroditism similar to that of the Isopoda Epicarida, it is clear that a slight modification of the sexual instinct of the larvae would lead to the gregarious habit, while its retention in some individuals in its original form accounts for their finding their way to the 107mantles of adult individuals and developing into the so-called complemental males.

Certain Cirripedes, viz. certain species of Scalpellum and Ibla and all the Acrothoracica, are dioecious. It is impossible to decide at present whether these species retain the primitive dioecious condition of the ancestral Cirripedes, or whether they too have been derived from an hermaphrodite state, but in the present state of knowledge they hardly affect the validity of the theory that has been proposed to account for the nature of the complemental males and the hermaphrodite individuals.

Order IV. Ostracoda.

The Ostracoda are small Crustacea, the body consisting of very few—about eight—segments, and being completely enclosed in a carapace, which has the form of a bivalve shell. Development is direct, without a Nauplius stage.

The Ostracoda^[82] are marine and fresh-water animals that can be divided into several families, differing slightly in habits and in structures correlated with those habits.

Fig. 74.—Candona reptans. A, Natural size; B, X 15. a, 1st antennae; b, 2nd antennae; c, walking legs. (After Baird.)

The Cypridae and Cytheridae include all the fresh-water and a vast majority of marine genera, adapted for a sluggish life among water-plants, though some can swim with considerable activity. The common Cypris and Candona of our ponds and streams are familiar instances. The movements of these animals are effected by means of the two pairs of uniramous pediform antennae which move together and in a vertical straight line. In the Cypridae (Fig. 74) there are, besides the mandibles, two pairs of maxillae, a pair of walking legs, and, lastly, a pair of appendages, which are doubled up into the carapace, and are used for cleaning purposes. In the marine Cytheridae there is only one maxilla, the last three appendages 108being pediform and used in walking. The telson in the Cytheridae is rudimentary, but is well developed in the Cypridae. The heart is altogether absent.

In many of the fresh-water forms, e.g. common species of Candona and Cypris, males are never found, and parthenogenetic reproduction by the females appears to proceed uninterruptedly. Weismann^[83] kept females of Cypris reptans breeding parthenogenetically for eight years. He also remarks on the fact that these, and indeed all parthenogenetic female Ostracoda, retain the receptaculum seminis, used normally for storing the spermatozoa derived from the male, unimpaired.

Some of the Cytheridae occur in deep water. Thus Cythere dictyon was frequently taken by the Challenger in depths of over 1000 fathoms, but the majority prefer shallow water.

Fig. 75.—Asterope oblonga, ♀, removed from its carapace, × 25. A, Alimentary canal; A₁, A₂, 1st and 2nd antennae; E, eye; G, gills; G.O, generative opening; H, heart; M, mandible; T, 6th appendage; T′, last appendage (cleaning foot). (After Claus.)

The Halocypridae and Cypridinidae comprise marine genera of a pelagic habit. The first antennae are chiefly sensory, but the second antennae are biramous, and they do not merely move up and down, as in the preceding families, but sideways like 109oars, the valves of the shells being excavated to admit of free movements. There are two pairs of maxillae; the succeeding limbs differ in the two families. In the Cypridinidae, e.g. Asterope (Fig. 75), the first leg (T) is lamelliform and is used as an accessory maxilla, while the second leg (T’) is turned upwards into the shell as a cleaning organ. In the Halocypridae the first leg is pediform, and differs in the two sexes, while the second leg is rudimentary and points backwards. In Asterope peculiar branchial organs (G) are present on the back. Both families possess a heart; the Halocypridae are blind, while the Cypridinidae possess eyes.

The Polycopidae and Cytherellidae are curious marine families of a pelagic habit, with biramous second antennae well adapted for swimming, and very broad. The first maxilla in the Polycopidae is also employed in swimming, while the second is modified into a branchial organ; the maxillae of the Cytherellidae are more normal in structure, but both carry branchial lamellae. The posterior limbs are altogether absent in Polycopidae, and in the Cytherellidae are only represented by the copulatory organs of the male.

SUB-CLASS II.—MALACOSTRACA.

The Malacostraca are generally large Crustacea, and they are characterised by the presence of a definite and constant number of segments composing the body. In addition to the paired eyes we can distinguish two pairs of antennae, a mandibular segment, and two maxillary segments composing the head region proper; there then follow eight thoracic segments, the limbs belonging to the anterior thoracic segments being often turned forwards towards the mouth, and modified in structure to act as maxillipedes, while at any rate the last four are used in locomotion and are termed “pereiopods.”^[84] The abdomen is composed of six segments, which typically carry as many pairs of biramous “pleopods,” and the body terminates in a telson. Not counting the paired eyes or the telson, there are present nineteen segments. The excretory organs in the adult open at the bases of the second antennae, and are known as “green glands,” but in the larva maxillary glands may be present homologous to those which persist in the adult Entomostraca. This is the typical arrangement, but sometimes the maxillary glands persist in adult Malacostraca, e.g. Nebalia, Anaspides, and some Isopods.

The hepato-pancreatic diverticula are directed posteriorly, and not anteriorly as in most Entomostraca, and the stomach is often furnished with chitinous teeth and ridges forming an elaborate gastric mill, especially in the larger Decapods.

Division. Phyllocarida.

Fig. 76.—Nebalia geoffroyi, ♀, × 20. A.1, A.2, 1st and 2nd antennae; Ab.1, Ab.6, 1st and 6th abdominal appendages; A.G. antennary gland; C, half of caudal fork; E, eye; G, ventral ganglionic chain; H, heart; I, intestine; L, upper liver-diverticulum; M, adductor muscle of halves of carapace; MX, palp of 1st maxilla; O, ovary; R, rostrum. (After Claus.)

The small shrimp-like Crustacean Nebalia, which is found burrowing in the superficial layers of sand in the littoral and sometimes the deeper regions of most seas, has been regarded, ever since its anatomy was made out by Claus,^[85] as a connecting link between Entomostraca and Malacostraca, and has been placed in a separate group Leptostraca.

The segmentation of the body is Malacostracan, save that two extra segments are present in the abdomen, and the paired compound eyes are borne upon stalks. The eight thoracic limbs are all very similar; they are built on the typical biramous plan, and each carries a bract; they have been compared, owing to their flattened, expanded shape, to the 112foliaceous limbs of the Phyllopods. The abdominal appendages are also biramous. The heart is greatly elongated, stretching through thorax and abdomen; there are present both the antennary excretory glands characteristic of adult Malacostraca and the maxillary glands characteristic of adult Entomostraca, and both the posterior and anterior livers characteristic of the two Orders respectively are present. This combination of characters justifies the belief that Nebalia represents a primitive form, standing to some extent in an intermediate position between Entomostraca and Malacostraca, but it may be doubted if the special relationship to the Phyllopoda, claimed on the strength of the foliaceous appearance of the thoracic limbs, can be legitimately pressed.

Nebalia shows the clearest signs of relationship to the other primitive Malacostraca, and especially to the Mysidae, which it resembles not only in general form and in the essentially biramous character of its appendages, but also in many embryological points and in the similarity in development of the brood-pouch.^[86]

A large number of very ancient palaeozoic fossils are known which are placed provisionally with Nebalia in the Division Phyllocarida, and some of these are no doubt closely related to the existing isolated genus. Hymenocaris from the Cambrian.

SERIES 2. EUMALACOSTRACA.

Before entering on a description of the members of this Series it is necessary to introduce and justify a new scheme of classification which has been proposed by Dr. W. T. Calman. This scheme necessitates the abandonment of the old Order Schizopoda, and also ignores the distinction which used to be considered fundamental between the sessile-eyed Crustacea (Edriophthalmata) and the stalk-eyed forms (Podophthalmata).

The old group of Schizopoda, to which Nebalia and the isolated form Anaspides, to be considered later, are undoubtedly related, represent very clearly the stem-forms from which the various branches of the Malacostracan stock diverge. No doubt they are themselves specialised in many directions, since they are a dominant group in present day seas, but their organisation is 113fundamentally of a primitive type. We see this especially in the comparative absence of fusion or reduction of the segments of the body externally and of the nervous system internally, and in the simple undifferentiated character of the trunk-limbs, all of which conform to the primitive biramous type. The most anterior thoracic limbs of the Schizopods are of particular interest. In the higher Malacostraca three of these limbs are usually turned forwards towards the mouth to act as maxillipedes, and the most anterior of all, the first maxillipede, is apt, especially in the Decapoda, to take on a flattened foliaceous form owing to the expansion of the basal segments to act as gnathobases (see Fig. 1, A, p. 10). Now this appendage in the Schizopods preserves its typical biramous character, and resembles the succeeding thoracic limbs, but in many of the species the basal joints show a tendency to be produced into biting blades (Fig. 1, E, p. 10), thus indicating the first step in the evolution of the foliaceous first maxillipede of the Decapoda. The primitive character of the Schizopods is also indicated by the fact that most of the Decapoda with uniramous limbs on the five hinder thoracic segments pass through what is known as the “Mysis stage” in development, when these limbs are biramous, the exopodites being subsequently lost in most cases.

The “Schizopoda” include a very large number of pelagic Crustacea of moderate size, which superficially appear to resemble one another very closely. The slender, elongated body, the presence of biramous limbs on all the thoracic and abdominal segments, and the possession of a single row of gills at the bases of the thoracic limbs, are, generally speaking, typical of the families Mysidae, Lophogastridae, Eucopiidae, and Euphausiidae, which go to make up the old Order Schizopoda.

It has, however, been pointed out first by Boas,^[87] and subsequently by Hansen and Calman,^[88] that the Euphausiidae are in many respects distinct from the other three families, and agree with the Decapoda, while the Eucopiidae, Lophogastridae, and Mysidae agree with the Cumacea, Isopoda, and Amphipoda.

It has, therefore, been suggested by these authors that the classification of the Malacostraca should be revised, and Calman (loc. cit.) has brought forward the following scheme:—

114The division Peracarida, including the Eucopiidae, Lophogastridae, and Mysidae (= Mysidacea), the Cumacea, Isopoda, and Amphipoda, is characterised by the fact that when a carapace is present it leaves at least four of the thoracic segments free and uncoalesced: by the presence of a brood-pouch formed from the oostegites on the thoracic limbs of the female: by the elongated heart: by the few and simple hepatic caeca: by the filiform spermatozoa: and by the direct method of development without a complicated larval metamorphosis. The biting face of the mandible has a movable joint, the “lacinia mobilis.”^[89]

The division Eucarida, on the other hand, including the Euphausiidae and the Decapoda, shows the converse of these characters. The carapace coalesces with all the thoracic segments, there is never a brood-pouch formed from oostegites, the hepatic caeca are much ramified, the heart is short, the spermatozoa are spherical with radiating pseudopodia, the development is indirect with a complicated metamorphosis, and the mandible is without a lacinia mobilis.

Corresponding divisions are made by Calman to receive the other Malacostraca, namely, the Phyllocarida for Nebalia, the Syncarida for Anaspides, and the Hoplocarida for the Stomatopoda or Squillidae.

The important array of characters which separates the Euphausiidae from the other Schizopods and unites them with the Decapoda can no longer be neglected, and the consideration of Anaspides and its allies will further emphasise the extreme difficulty of retaining the Schizopoda as a natural group. In the sequel Calman’s proposed scheme will be adopted.

DIVISION 1. SYNCARIDA.

There is no carapace, and all the eight thoracic segments may be free and distinct. Eyes may be pedunculate or sessile. The mandible is without a lacinia mobilis. There is no brood-pouch, the eggs being deposited and hidden after fertilisation. The spermatozoa are filiform, the hepatic caeca very numerous, and the heart tubular and elongated, with ostia only in one place in 115the anterior thoracic region. The auditory organ is at the base of the first antennae.

Order. Anaspidacea.

Fam. 1. Anaspididae.—The mountain-shrimp of Tasmania, Anaspides tasmaniae, was first described by Thomson^[90] in 1893 from specimens taken in a little pool near the summit of Mount Wellington; it was redescribed by Calman,^[91] who drew attention to its remarkable resemblance to certain Carboniferous fossils of Europe and N. America (Gampsonyx, Palaeocaris, etc.).

The creature appears to be confined to the deep pools of the rivers and tarns on the mountains of the southern and western portions of Tasmania.^[92] The waters in which it occurs are always cold and absolutely clear, and there is no record of its living at altitudes much below 2000 feet, while it frequently occurs at 4000 feet. The body may attain upwards of two inches in length; it is deeply pigmented with black chromatophores, and it is held perfectly horizontal without any flexure. The animal rarely swims unless disturbed, usually walking about on stones and water-plants at the bottom of deep pools. In walking the endopodites of the thoracic limbs are chiefly instrumental, but they are assisted by the exopodites of the abdominal limbs.

When frightened the shrimp can dart rapidly forwards or sideways by the strokes of its powerful tail-fan, but it never jumps backwards as do the other Malacostraca. It appears to browse upon the algal slime covering the rocks and on the submerged liver-worts and mosses, but it does not refuse animal food, even feeding on the dead bodies of members of its own species. The thoracic limbs, which are all biramous except the last pair, carry a double series of remarkable plate-like gills on their coxopodites. The slender and setose exopodites of the thoracic limbs are respiratory in function, being kept in continual motion even when the animal is at rest, and serving to keep up a current of fresh water round the gills.

Anaspides shows a remarkable combination of structural characters, some of which are peculiar, while others are possessed in common with the Peracarida or Eucarida. The chief peculiar 116characters are the entire absence of a carapace, and the freedom of the eight thoracic segments, with eight free thoracic ganglia in the nerve-cord; the peculiar double series of plate-like gills; the structure of the alimentary canal; and the fact that the eggs, instead of being carried in a brood-pouch, or affixed to the abdominal limbs, are deposited under stones and among water-plants.^[93]

Fig. 77.—Anaspides tasmaniae in natural position for walking, × 1. The last two pereiopods point backwards and are overlapped by the first two pleopods.

The Peracaridan features, uniting it especially with the Mysidacea, are the structure of the elongated heart, the filiform spermatozoa, and the fact that no complicated metamorphosis is passed through, the young hatching out in a condition similar to, though possibly not identical with, the adult form.

The Eucaridan, especially Decapodan, features are the presence of an auditory sac on the basal joint of the antennules, 117and the modification of the endopodites of the first two abdominal appendages in the male to form a copulatory organ.

A type of a new genus of this family was found by me in the littoral zone of the Great Lake of Tasmania at an elevation of 3700 feet, and named Paranaspides lacustris.

This little shrimp (Fig. 78), which does not appear to grow to more than an inch in length, is totally different in appearance from Anaspides, being pale green and transparent, with a very marked dorsal hump as in Mysis, to which it bears a very striking superficial resemblance. It leads a more active swimming life than Anaspides, and with this habit is correlated the flexure of the body and the greater size of the tail-fan and the scale of the second antenna. The mandible is peculiar in being furnished with a four-jointed biramous palp, while that of Anaspides is three-jointed and uniramous, and the first thoracic appendage is provided with a setose biting lobe on the antepenultimate joint, thus more resembling a maxillipede. In other respects it agrees essentially in structure with Anaspides.

Fig. 78.—Paranaspides lacustris, × 4. a¹, a², First and second antennae; Ab.1, first abdominal segment; ep, epipodites or gills on the thoracic legs; md, mandible; Pl.1, first pleopod; T, telson; Th.8, eighth free thoracic segment; U, uropod, or sixth pleopod.

Fam. 2. Koonungidae.—The sole representative of this family, Koonunga cursor, has been recently described by Mr. O. A. Sayce,^[94] of Melbourne University, from a small stream some 118miles to the west of Melbourne. Although plainly belonging to the Anaspidacea, this interesting little animal, which only measures a few millimetres in length, and follows a similar habit to Anaspides, running about with its body unflexed, differs from all the other members of the Division in possessing sessile instead of stalked eyes, in the first thoracic segment being fixed to the head, and in a number of minor anatomical points.

It is impossible at present to assign the Carboniferous forms (Gampsonyx, Palaeocaris, etc.) to their exact position in the Division, but it seems that they agreed more closely with Anaspides than with the other two genera. From the position in which the fossils are preserved, it would appear that they followed a similar walking habit to Anaspides, and that the body was unflexed.

DIVISION 2. PERACARIDA.

The carapace, when present, leaves at least four of the thoracic somites distinct; the first thoracic segment is always fused with the head. The eyes are pedunculate or sessile.

The mandible possesses a lacinia mobilis. A brood-pouch is formed in the female from oostegites attached to the thoracic limbs. The hepatic caeca are few and simple; the heart is elongated and tubular; the spermatozoa are filiform, and development takes place without a complicated metamorphosis.

Order I. Mysidacea.

The Mysidacea, although pelagic, are not very often met with in the true plankton on the surface; they generally swim some way below the surface, going down in many cases into the abysses. For this reason they thrive excellently in aquaria, and the common Mysis vulgaris is often present in such numbers in the tanks at the Zoological station at Naples as to damage the other inmates by the mere press of numbers. The Mysidacea, like the majority of the Peracarida, undergo a direct development, and hatch out with the structure of the adult fully formed.

Many of the Mysidacea bear auditory sacs upon the sixth pair of pleopods, a characteristic not found in the Euphausiacea.

Fam. 1. Eucopiidae.—The curious form Eucopia australis 119(Fig. 79) described by Sars,^[95] may be chosen as an example of the Mysidacea.

The peculiarity of this form consists chiefly in the immense elongation of the endopodites of the fifth, sixth, and seventh thoracic appendages. Characteristic of the Mysidacea is the freedom of the hinder thoracic segments from fusion with the carapace, otherwise this animal is seen closely to resemble the Euphausia figured (Fig. 102). Eucopia australis, like so many of the Mysidacea, is a deep-sea animal, being brought up with the dredge from over 1000 fathoms; it is very widely distributed over the Atlantic Ocean.

Fig. 79.—Eucopia australis, young female, × 3. A, 1st antenna; Ab.1, 1st abdominal segment; Ab.6, 6th abdominal appendage; E, eye; T, telson; Th, 5th thoracic appendage. (After Sars.)

Fam. 2. Lophogastridae.—The members of this family (Lophogaster, Gnathophausia) agree with the Eucopiidae in the possession of branched gills on some of the thoracic limbs, in the absence of auditory sacs on the sixth pair of pleopods, in the presence of normally developed pleopods in both the male and female, and in the brood-lamellae being developed on all seven of the thoracic limbs. The endopodites of the posterior thoracic limbs are, however, of a normal size.

Fig. 80.—Dorsal view of male Diastylis stygia, × 12. A, 2nd antenna; Ab.6, 6th abdominal appendage. (After Sars.)

Fam. 3. Mysidae.—These differ from both the foregoing families in the absence of gills, in the presence of an auditory sac on the sixth pleopods, in the reduction of the other pleopods in the female, and in the brood-lamellae being developed only on the more posterior pairs of thoracic limbs. A number of closely 120related genera compose this family, of which Mysis, Boreomysis, and Siriella may be mentioned. Mysis oculata, var. relicta, is a fresh-water form from the lakes of northern and central Europe.

Order II. Cumacea.^[96]

The Cumacea are a group of small marine animals rarely attaining an inch in length, which agree with the Mysidacea in the characters noted above as diagnostic of the Division Peracarida; they possess, however, in addition a number of peculiar properties, and Sars believes them to be of a primitive nature showing relationship to Nebalia, and possibly to an ancestral Zoaea-like form. They follow a habit similar to that of the Mysidacea, being caught either in the surface-plankton or in great depths, many of the deep-sea forms being blind. They are, however, not true plankton forms, and they appear to attain a greater development both in point of variety and size in the seas of the northern hemisphere. The thoracic limbs may be biramous, but there is a tendency among many of the genera to lose the exopodites of some of the thoracic legs, an exopodite never being present on the last few thoracic limbs of the female and on the last in the male. In the Cumidae the four posterior pairs in both sexes have no exopodites. The first three thoracic appendages following the maxillae are distinguished as maxillipedes; they are uniramous, and the first pair carries an 121epipodite and a large gill upon the basal joints. Pleopods are only developed in the male sex.

The flagellum of the second antennae in the male may be enormously elongated, as in the Atlantic deep-sea species shown in Fig. 80, so as to exceed in length the rest of the body.

Fam. 1. Cumidae.—No sharp demarcation between thorax and abdomen. Four posterior pairs of legs in both sexes without exopodites. Male with five well-developed pleopods in addition to the uropods. Telson wanting. Cuma, Cyclaspis, etc.

Fam. 2. Lampropidae.—Body-form resembles that of Cumidae. All the thoracic limbs except the last have exopodites. The male has three pairs of pleopods. Telson present. Lamprops, Platyaspis, etc.

Fam. 3. Leuconidae.—Body-form similar to above. Male has only two pairs of pleopods. Mouth-parts peculiar, much less setose than in other families. Telson absent. Leucon, Eudorella.

Fam. 4. Diastylidae.—Anterior part of thorax sharply marked off from posterior part. Male has two pairs of pleopods. Telson present. Diastylis (Fig. 80). D. goodsiri from the Arctic ocean measures over an inch in length.

Fam. 5. Pseudocumidae.—Rather similar to Diastylidae, but differ in reduced size of telson and presence of exopodites on third and fourth thoracic legs of female. This family is represented by three very similar marine forms of the genus Pseudocuma; but, as Sars has shown,^[97] the Caspian Sea contains thirteen peculiar species, only one of which can be referred to the genus Pseudocuma, while the rest may be partitioned among four genera, Pterocuma, Stenocuma, Caspiocuma, Schizorhynchus.

Order III. Isopoda.

The Isopoda and the Amphipoda are frequently classed together as Arthrostraca or Edriophthalmata, owing to a number of features which they share in common, as, for instance, the sessile eyes which distinguish them from the podophthalmatous Schizopoda and Decapoda, the absence of a carapace, and the thoracic limbs which are uniramous throughout their whole existence. For the rest, in the presence of brood-plates and the other diagnostic 122characters, they are plainly allied to the other Peracarida, and an easy transition is effected from the Mysidacea to the Isopoda through the Chelifera or Anisopoda. Only one thoracic segment is usually fused with the head, the appendage of this segment being the maxillipede; in the Chelifera among Isopoda, and the Caprellidae among Amphipoda, two thoracic segments are fused with the head.

The Isopoda are distinguished from the Amphipoda by the dorso-ventral flattening of the body, as opposed to the lateral flattening in the Amphipoda, by the posterior position of the heart, and by the branchial organs being situated on the abdominal instead of on the thoracic limbs.

The Isopoda, following Sars’^[98] classification, fall into six sub-orders—the Chelifera, Flabellifera, Valvifera, Asellota, Oniscoida, and Epicarida,—to which must be added the Phreatoicidea.

Sub-Order 1. Chelifera.

The Chelifera, including the families (1) Apseudidae and (2) Tanaidae, are interesting in that they afford a transition between the ordinary Isopods and the Mysidacea. The important features in which they resemble the Mysidacea are, first, the fusion of the first two thoracic segments with the head, with the coincident formation of a kind of carapace in which the respiratory functions are discharged by a pair of branchial lamellae attached to the maxillipedes; and, second, the presence of very small exopodites on the first two thoracic appendages of the Apseudidae.

The second pair of thoracic limbs, i.e. the pair behind the maxillipedes, are developed both in the Apseudidae and Tanaidae into a pair of powerful chelae, and these frequently show marked sexual differences, being much more highly developed in the males than in the females. The biramous and flattened pleopods are purely natatory in function, and the uropods or pleopods of the sixth pair are terminal in position and slender.

123

Fig. 81.—Apseudes spinosus, ♂, × 15. A, 1st antenna; Ab, 6th abdominal appendage; T, 2nd thoracic appendage. (After Sars.)

Both families, of which the Apseudidae contain the larger forms, sometimes attaining to an inch in length, are littoral in habit, or occur in sand and ooze at considerable depths, many of the genera being blind. Many Tanaids (e.g. Leptochelia, Tanais, Heterotanais, etc.) live in the algal growths of the littoral zone, and being highly heliotropic they are easy to collect if a basinful of algae is placed in a strong light. The females carry the eggs about with them in a brood-pouch formed, as is usual in the Peracarida, by lamellae produced from the bases of the thoracic limbs. The males on coming to maturity do not appear to grow any more, or to take food, their mouth-parts frequently degenerating and the alimentary canal being devoid of food. They are thus in the position of insects which do not moult after coming to maturity; and, as in Insects, the males are apt to show a kind of high and low dimorphism—certain of the males being small with secondary sexual characters little different from those of the females, while others are large with these characters highly developed. Fritz Müller, in his Facts for Darwin, observes that in a Brazilian species of Leptochelia, apparently identical with the European L. dubia, the males occur under two totally distinct forms—one in which the chelae are greatly developed, and another in which the chelae resemble those of the female, but the antennae in this form are provided with far longer and more numerous sensory hairs than in the first form. Müller suggested that these two varieties were produced by natural selection, the characters of the one form compensating for the absence of the characters of the other. A general consideration of the sexual dimorphism in the Tanaidae^[99] 124lends some support to this view, since the smaller species with feeble chelae do appear to be compensated by a greater development of sensory hairs on the antennae, but the specific differences are so difficult to appreciate in the Tanaidae that it is possible that the two forms of the male in Müller’s supposed single species really belonged to two separate species.

Sub-Order 2. Flabellifera.

The Flabellifera include a number of rather heterogeneous families which resemble one another, however, in the uropods being lateral and not terminal, and being expanded together with the telson to form a caudal fan for swimming. The pleopods are sometimes natatory and sometimes branchial in function. Some of the families are parasitic or semi-parasitic in habit.

Fam. 1. Anthuridae.—These are elongated cylindrical creatures found in mud and among weeds upon the sea-bottom; their mouth-parts are evidently intended for piercing and sucking, but whether they are parasitic at certain periods on other animals is not exactly known. Anthura, Paranthura, Cruregens.

Fig. 82.—Gnathia maxillaris. A, Segmented larva, × 10; B, Praniza larva, × 5; C, gravid female, × 5; D, male, × 5.

Fam. 2. Gnathiidae.^[100]—These forms appear to be related to the Anthuridae; they are ectoparasitic on various kinds of fish during larval life, but on assuming the adult state they do not feed any more, subsisting merely on the nourishment amassed during the larval periods. The larvae themselves are continually leaving their hosts, and can be taken in great numbers living freely among weeds on the sea-bottom. The larvae, together with the adults of Gnathia maxillaris, are extremely abundant among the roots of the sea-weed Poseidonia cavolinii in the Bay of Naples. The young larvae hatch out from the body of the female in the state shown in Fig. 82, A. This minute larva fixes upon a fish, and after a time it is transformed into the so-called Praniza larva (B), in which the gut is so distended with the fluid sucked from the host that the segmentation in the hind part of the thorax is entirely lost. When this larva moults it may, however, reacquire temporarily its segmentation. After a certain period of this parasitic mode of life the Praniza finally abandons its host, and becomes transformed into the adult male or female. This may take place at very different stages in the 125growth of the larva, the range of variation in size of the adults being 1–8 mm., and it must be remembered that when once the adult condition is assumed growth entirely ceases. What it is that determines the stage of growth in each individual when it shall be transformed into the adult is not known. The males and females differ from one another so extraordinarily that it was for long denied that they were both derived from the Praniza larvae. This is nevertheless the case. The change from the Praniza to the female (Fig. 82, C) is not very great. The ovary absorbs all the nourishment in the gut and comes to occupy the whole of the body, all the other organs degenerating, including the alimentary canal and mouth-parts. Indeed, only the limbs with their muscles and the nervous system remain. The change to the male (D) is more radical. The food is here stored in the liver, which increases in the male just as the ovary does in the female. The segmentation is reacquired, and the massive square head is formed from the hinder part of the head in the Praniza, the anterior portion with its stylet-like appendages being thrown away. The powerful nippers of the male are not formed inside the cases of the old styliform mandibles, but are independent and possibly not homologous organs. The meaning of the marked sexual dimorphism and the use of the males’ nippers are not in the least known, though the animals are easy to keep under observation. In captivity the males never take the slightest notice of either larval or adult females.

126Fam. 3. Cymothoidae.^[101]—This is a group of parasites more completely parasitic than the foregoing, but their outer organisation does not differ greatly from an ordinary Isopodan form. A great many very similar species are known which infest the gill-chambers, mouths, and skin of various fishes. The chief interest that attaches to them is found in the fact that a number of them, and perhaps all, are hermaphrodite, each individual acting as a male when free-swimming and young, and then subsequently settling down and becoming female. This condition is exactly the same as that occurring universally in the great group of parasitic Isopoda, the Epicarida, to be considered later. There is no evidence that the Cymothoidae are phyletically related to the Epicarida, so that the similar sexual organisation appears to be due to convergence resulting from similar conditions of life. The general question of hermaphroditism in the Crustacea has been shortly discussed on pp. 105–106. Cymothoa.

Fam. 4. Cirolanidae.—In this family is placed the largest Isopod known—the deep-sea Bathynomus giganteus, found in the Gulf of Mexico and the Indian Ocean, sometimes measuring a foot long by four inches broad. A common small littoral form is Cirolana.

Fam. 5. Serolidae.^[102]—The genus Serolis comprises flattened forms bearing a curious resemblance to Trilobites, which Milne Edwards considered more than superficial. The genus is confined to the littoral and deep waters of the southern hemisphere.

Fam. 6. Sphaeromidae.^[103]—These are flattened, broad-bodied forms, most commonly met with in the Mediterranean and warmer seas. Without being actually parasitic, they are frequently found as scavengers in decaying material, and they show some relationship to the parasitic Cymothoidae. In some of the genera, e.g. Cymodoce, the ovigerous female shows a degenerate condition 127of the mouth-parts, while the maxillipedes undergo an enlargement, and are used for causing a current through the brood-chamber.

Sub-Order 3. Valvifera.

Fig. 83.—Munnopsis typica (Munnopsidae), ♂, × 2. A, 2nd antenna; Ab, abdomen; T, 5th thoracic appendage or 4th leg. (After Sars.)

The Valvifera, illustrated by the Idotheidae and Arcturidae, are characterised by the uropods being turned back and expanded to form folding doors covering up the delicate pleopods, which are mostly respiratory in function, though the anterior pairs may serve as swimming organs. Arcturus is a typically deep sea genus, many species, remarkably furnished with spiny processes, having been taken by the Challenger in the southern hemisphere. The Idotheidae are more littoral forms, several species of Idothea being commonly met with off the British coasts, occasionally penetrating into brackish or even fresh water.

Sub-Order 4. Asellota.

In this group the abdominal segments are fused dorsally to form a shield-like caudal region; the pleopods are respiratory in function and reduced in numbers, the first pair being often expanded and produced backwards to form an operculum covering the rest. Several of the Asellota are fresh-water, Asellus aquaticus 128(Asellidae) being extremely abundant all over Europe in weed-grown ditches, the mud of slowly-moving streams, and even on the shores of large lakes. They are mostly sluggish in habit, but the marine Munnopsidae (Fig. 83, Munnopsis) are expert swimmers, the swimming organs being fashioned by the expansion and elongation of the thoracic legs.

Sub-Order 5. Oniscoida.

The Oniscoida^[104] are terrestrial forms in which the abdomen is fully segmented, the pleopods are respiratory, their endopodites being delicate branchiae, while their exopodites are plate-like and form protective opercula for the gills, and the uropods are biramous and not expanded. The epimera of the segments are greatly produced. The terrestrial Isopods, although air-breathers,^[105] are dependent on moisture, and are only found in damp situations. It seems probable that they have been derived from marine Isopods, since the more generalised of them, e.g., Ligia (Fig. 84), common on the English coasts, are only found in damp caves and crannies in the rocks.

Fig. 84.—Ligia oceanica, ventral and dorsal views, × 1. (From original drawings prepared for Professor Weldon.)

129The related Ligidium is found far inland, but always in the neighbourhood of water. These two genera may be distinguished by the numerous joints in the flagellum of the second antennae, the flagellum being in all cases the portion of the antenna succeeding the long fifth joint. Philoscia muscorum occurs usually near the coast, but it is also found inland in England under trees in damp moss. This genus and the common Oniscus, found in woods, are distinguished by the presence of three joints in the flagellum of the second antenna. Philoscia can be distinguished from Oniscus by its narrower body and the pretty marbled appearance of its back. The genus Trichoniscus has four joints in the flagellum; various species are found in woods. In Porcellio and Armadillidium there are only two joints in the flagellum, while Armadillidium, the common garden wood-louse, can be distinguished from all others by the flattened shape of the uropods, and the habit of rolling up into a ball like an Armadillo.

There is also a very peculiar species, Platyarthrus hoffmannseggii, which occurs in England and Northern Europe, and always lives in ants’ nests. It is supposed that they serve as scavengers for the ants, which tend them carefully, and evidently treat them as domestic animals of some kind. The small creature is quite white and blind, and has exceedingly short antennae.

Sub-Order 6. Epicarida.

The Epicarida include an immense number of Isopods, parasitic upon other Crustacea. In the adult state they become greatly deformed, and offer very few characters of classificatory value, but they all pass through certain highly characteristic larval stages which are essentially similar in the different families. All the species are protandric hermaphrodites, each individual being male while in a larval state, and then losing its male organisation and becoming female as the parasitic habit is assumed.

Two series of families are recognised according to the larval stages passed through, the Cryptoniscina, in which the adult male organisation is assumed in the Cryptoniscus stage, and the female condition is imposed directly upon this form, and the Bopyrina, in which the Cryptoniscus passes into a further larval stage, the Bopyrus, which performs the function of the 130male, and upon which the female organisation is imposed as the parasitic habit is assumed.

The following is a list of the Epicarida with the Crustacea which serve as their hosts^[106]:—

Fig. 85.—Epicaridian larva, probably belonging to one of the Cryptoniscina. A, 2nd antenna; Ab, abdominal appendages; T, thoracic appendages. (From Bonnier, after Hansen.)

In all cases the first larval form which hatches out from the maternal brood-pouch is called the Epicaridian larva (Fig. 85).

This little larva has two pairs of antennae, a pair of curious frontal processes, and a pair of mandibles. The other mouth-parts are missing; there are only six thoracic limbs, but the full complement of six biramous pleopods are present, and at the end of the body there may be a long tube of unknown function.

As a type of the Cryptoniscina we may take the Liriopsidae,^[107] parasitic on the Rhizocephala, which are, of course, themselves parasitic on the Decapoda, the whole association forming a very remarkable study in Carcinology.

Almost every species of the Rhizocephala is subject to the attacks of Liriopsids, the latter fixing either on the Rhizocephala themselves, or else on the Decapod host at a point near the fixation of the Rhizocephalous parasite. An exceedingly common Liriopsid is Danalia curvata, parasitic on Sacculina neglecta, 131which is itself parasitic on the spider-crab, Inachus mauritanicus, at Naples. The adult Danalia is a mere curved bag full of eggs or developing embryos, and without any other recognisable organs except two pairs of spermathecae upon the ventral surface where the spermatozoa derived from the larval males are stored.

Fig. 86.—Inachus mauritanicus, ♀, × 1, carrying two Sacculina neglecta (a, b), and a Danalia curvata (c), the latter bearing two dwarf males.

In Fig. 86 is represented a female of Inachus mauritanicus which carried upon it two Sacculinae and a Danalia curvata, and upon the latter are seen two minute larval males in the act of fertilising the adult Danalia. The eggs develop into the Epicaridian stage, after which the larva passes into the Cryptoniscus stage (Fig. 87). In this larval form the segments are clearly delimited; the only mouth-parts present are the mandibles, but there are seven pairs of thoracic limbs and the full number of pleopods. This Cryptoniscus stage is found in all the Epicarida, and only differs in detail in the various families.

Fig. 87.—Ventral view of Cryptoniscus larva of Danalia curvata, ♂, × 25.

In the Cryptoniscina the Cryptoniscus larva is the male, and at this stage possesses a pair of large testes in the thorax. The ovaries are also present at this stage as very small bodies applied to the anterior ends of the testes. The larval males in this state seek out adult fixed Danaliae and fertilise them; and, when this is accomplished, they themselves become fixed to the host and begin to develop into the adult female condition. The limbs are all lost, and out of the mouth grows a long proboscis (Fig. 88, P), which penetrates the tissues of the host. The ovaries begin to grow, and a remarkable process of absorption in the testes takes place. These organs, when fixation occurs, are never empty of spermatozoa, and are frequently crammed with them. After fixation some 132large cells at the interior borders of the testes begin to feed upon the remains of these organs and to grow enormously in size and to multiply by amitosis. These phagocytes, as they really are, attain an enormous size, but they are doomed to degeneration, the chromatin becoming dispersed through the cytoplasm, and the nuclei dividing first by amitosis and then breaking up and disappearing. As the parasite grows, the heart at the posterior end of the body ceases to beat; the ovaries increase enormously at the expense of the alimentary canal, and on the ventral surface two pairs of spermathecae are invaginated ready to receive the spermatozoa of a larval male. In the adult condition, after fertilisation has taken place and the ovaries occupy almost the whole of the body, the remains of the phagocytic cells can be seen on the dorsal surface in a degenerate state. They evidently are not used as food, and their sole function is to make away with the male organisation when it has become useless.^[108]

Fig. 88.—Side view of Danalia curvata, × 15, shortly after fixation and loss of larval appendages. A, Alimentary canal; E, eye; H, heart; N, phagocytic cells; O, ovary; P, proboscis.

Fig. 89.—Optical section (dorsal view) of Danalia curvata, in the same stage as Fig. 88. A, Alimentary canal; Ec, ectoderm; H, heart; N, phagocytic cells; O, ovaries; P, proboscis.

In the series Bopyrina, after the free-living Epicaridian 133and Cryptoniscus stages, a further larval state is assumed, called the Bopyrus, which is the functional male, and, after performing this function, passes on to the adult female condition.

The family Bopyridae is parasitic in the branchial chamber of Decapoda, especially Macrura and Anomura. When one of these Decapods is infested with an adult Bopyrid the gill-chamber in which it is situated is greatly swollen, as shown in Fig. 90. A very common Bopyrid is Bopyrus fougerouxi, parasitic in the gill-chambers of Palaemon serratus. The Bopyrus larva or functional male has the appearance shown in Fig. 91. It differs from the Cryptoniscus stage in possessing a rudimentary pair of anterior thoracic limbs and seven pairs normally developed, while the abdominal limbs are plate-like and branchial in function. The male can often be found attached to the female beneath the last pair of incubatory lamellae.

Fig. 90.—Galathea intermedia, with a Pleurocrypta microbranchiata under its left branchiostegite (B), × 1. (After Sars.)

Fig. 91.—Ventral view of male Bopyrus fougerouxi, × 30. A, 1st and 2nd antennae; T, 8th (last) thoracic appendage. (After Bonnier.)

The adult female condition, which is assumed after the Bopyrid stage is passed through, is illustrated in Fig. 92. The body acquires a remarkable asymmetry, due to the unequal pressure exerted by the walls of the gill-chamber. The antennae and mandibles (Fig. 92, B) are entirely covered up by the largely expanded maxillipedes; maxillae are, as usual, entirely 134absent. Very large lamellae grow out from the bases of the thoracic limbs to form a brood-pouch, and in this manner the adult condition is attained.

The final complication in the life-histories of these Isopoda is reached by the family Entoniscidae, which are parasitic when adult inside the thoracic cavity of Brachyura and Paguridae. The cephalothorax of a Carcinus maenas, which contains an adult Portunion maenadis (P), is shown in Fig. 93. The parasite is of a reddish colour when alive.

Fig. 92.—Bopyrus fougerouxi. A, Ventral view of female carrying a male (M) between her abdominal appendages, × 8; B, ventral view of part of head of female, the maxillipedes and the left mandible having been removed. A.1, A.2, 1st and 2nd antennae; M, male; Mn, right mandible; Mx, left maxillipede; O, oostegite; T, left 4th thoracic appendage or 3rd leg. (After Bonnier.)

Fig. 93.—Cephalothorax of Carcinus maenas, seen from the ventral side, containing a parasitic Portunion maenadis (P), × ½. (After Bonnier.)

The Entoniscidae pass through a free living Epicaridian and Cryptoniscus stage, and become adult males in the Bopyrus stage. It is stated, however, by Giard and Bonnier^[109] that these individuals, which actually function as males, never grow up 135into adult females, though all the adult females have passed through a male stage in which the male genital ducts are not formed. The hermaphroditism, therefore, in these animals at any rate is absolutely useless from a reproductive point of view, and this justifies our looking for some other explanation of it, such as was suggested on p. 105.

Fig. 94.—Portunion maenadis, ♀:—A, Young, × 10; B, older, × 5; C, adult, before the eggs are laid, × 3. A, 2nd antenna; Ab, abdomen; B, anterior lobe of brood-pouch; B′, its lateral lobe; H, head; 1, 2, 1st and 2nd incubatory lamellae (oostegites). (After Giard and Bonnier.)

The Bopyrus fixes in the gill-chamber of the host and becomes converted into the adult female by a series of transformations. As these changes take place it invaginates the wall of the gill-chamber and pushes its way into the thoracic cavity of the crab, though it lies all the time enveloped in the invaginated wall of the gill-chamber, and not free in the body-cavity of the crab. The transformations which it undergoes are shown in Fig. 94. The body first assumes a grub-like appearance (A), and two pairs of incubatory lamellae (1, 2) grow out from the first and second thoracic segments. In the next stage (B) these lamellae assume gigantic proportions, and four pairs of branchiae grow out from 136the abdominal segments (Ab). In the final stage (C) the incubatory lamellae have further increased in size, and constitute the main bulk of the body; the enormous mass of eggs is passed into the incubatory pouch, and all that remains of the rest of the body is the small head (H) and the abdomen (Ab), furnished with its branchiae. Communication with the external world is kept up through an aperture which leads from the brood-pouch into the gill-chamber of the host, and through this aperture the young are hatched out when they are developed sufficiently.

The presence of these parasites, although they are never in actual contact with the internal organs of the crab, calls forth the same phenomenon of parasitic castration as was observed in the Rhizocephala. A remarkable association is also found to exist between the Entoniscidae and Rhizocephala, of such a kind that, on the whole, a crab infested with a Rhizocephalan is more likely to harbour an Entoniscid than one without. The explanation of this association is probably that a crab with a Sacculina inside it is prevented from moulting as often as an uninfected crab, and, in consequence, the larval stages of the Entoniscid in the crab’s gill-chamber are more safely passed through.

Sub-Order 7. Phreatoicidea.^[110]

The members of this sub-order, although agreeing with the Isopoda in the essentials of their anatomy, resemble the Amphipoda in being rather laterally compressed, and in having the hand of the first free thoracic limb enlarged and subchelate. The abdomen is greatly produced laterally by expansions of the segments. In fact, the shape of the body and of the limbs is very Amphipodan.—Phreatoicus from New Zealand, Southern Australia, and Tasmania. Phreatoicopsis,^[111] a very large form from Gippsland, Victoria. Only one family exists, Phreatoicidae.

Order IV. Amphipoda.

In this order the body is flattened laterally, the heart is anterior in position, and the branchial organs are attached to the thoracic limbs.

There are three well defined sub-orders, (i.) the Crevettina, including 137a vast assemblage of very similar animals, of which the common Gammarus and Orchestia may serve as examples; (ii.) the Laemodipoda or Caprellids, and (iii.) the Hyperina.

We cannot do more than touch on the organisation of these sub-orders.

Sub-Order 1. Crevettina.

In this sub-order only one thoracic segment is fused with the head; the basal joints of the thoracic limbs are expanded to form broad lateral plates, and the abdomen is well developed, with six pairs of pleopods, the last three pairs being always turned backwards, and stiffened to act as uropods.

This group has numerous fresh-water representatives, e.g. Gammarus of several species, the blind well-shrimp Niphargus, and the S. American Hyalella; but the vast majority of the species are marine, and are found especially in the littoral zone wherever the rocks are covered with a rich growth of algae, Polyzoa, etc. The Talitridae or “Sand-hoppers” have deserted the waters and live entirely in the sand and under rocks on the shore, and one common European species, Orchestia gammarellus, penetrates far inland, and may be found in gardens where the soil is moist many miles from the sea.

The Rev. T. R. R. Stebbing, in his standard work^[112] on this group, recognises forty-one families, and more than 1000 species, so that we can only mention a few of the families, many of which, indeed, differ from one another in small characters.

Fam. Lysianassidae.—The first joint of the first antenna is short, with an accessory flagellum. Mandible with a palp, and with an almost smooth cutting edge. The third joint of the second gnathopod is elongated. This family is entirely marine, comprising forty-eight genera, with species distributed in all seas. One genus, Pseudalibrotus, inhabits the brackish water of the Caspian Sea. Lysianassa has several common British and Mediterranean species.

Fam. Haustoriidae.—The members of this family are specially adapted for burrowing, the joints of the hinder thoracic limbs being expanded, and furnished with spines for digging. Some of the species are common on the British coasts, e.g. Haustorius arenarius. Pontoporeia has an interesting distribution, 138one species, P. femorata, being entirely marine, in the Arctic and North Atlantic, P. affinis inhabiting the Atlantic, and also fresh-water lakes in Europe and North America, P. microphthalma being confined to the Caspian Sea, and P. loyi to Lakes Superior and Michigan.

Fam. Gammaridae.—Includes fifty-two genera. The first antennae are slender, with the accessory flagellum very variable. The mandibles have a dentate cutting edge, spine-row, and molar surface, and a three-jointed palp. The first two thoracic limbs are subchelate. This family includes a few marine, but mostly brackish and fresh-water species. Crangonyx is entirely subterranean in habitat, as is Niphargus, N. forelii occurring, however, in the deep waters of Lake Geneva. Both these genera are blind. Gammarus has thirty species, G. locusta being the common species on the North Atlantic coasts, and G. pulex the common fresh-water species of streams and lakes in Europe. A number of Gammaridae inhabit the Caspian Sea, e.g. Boeckia, Gmelina, Niphargoides, etc., while the enormous Gammarid fauna of Lake Baikal, constituting numerous genera, showing a great variety of structure, some of them being blind, belong to this family, e.g. Macrohectopus (Constantia), Acanthogammarus, Heterogammarus, etc.

Fig. 95.—Gammarus locusta, ♂ (above) and ♀ (below), × 4. Abd.1, First abdominal segment; T, telson; Th, seventh free thoracic segment (= 8th thoracic segment); U, third uropod. (After Della Valle.)

139Fam. Talitridae.—This family may be distinguished by the absence of a palp on the mandible, and by one ramus of the uropods being very small or wanting. The various kinds of “Sand-hoppers” belong here, familiar creatures on every sandy coast between tide-marks. The genera Talitrus and Talorchestia always frequent sand, while Orchestia is generally found under stones and among weed. Some species of Orchestia, e.g. O. gammarellus, live inland in moist places at some distance from the sea; one species of Talitrus (T. sylvaticus) occurs at great elevations in forests in Southern Australia.

Hyale is a coastal genus, and is also found on floating objects in the Sargasso Sea. Hyalella is confined to Lake Titicaca and the fresh waters of South America. Chiltonia from S. Australasia.

Fam. Corophiidae.—The members of this family have a rather flattened body and small abdomen, and the side-plates on the thorax are small. The uropods are also small and weak. Some species of the genus Corophium are characteristic of the Caspian Sea.

Sub-Order 2. Laemodipoda.

Fig. 96.—Caprella grandimana, × 4. a, Abdomen; g, gills; t, 3rd (first free) thoracic segment; t′, 8th thoracic segment. (After P. Mayer.)

Fam. 1. Caprellidae^[113] are also chiefly littoral forms, swarming among rocks covered by algae, though they are by no means so easy to detect as the Gammaridae and Tanaidae which haunt similar situations. In a basinful of algae or Polyzoa taken from the rocks fringing the Bay of Naples, the latter are easily collected, the Tanaidae always crawling out of the weeds in the direction of the light, while the Gammarids dart about in all directions; but the Caprellidae, with their branching stick-like forms, 140harmonise so well with their surroundings that it requires an experienced eye to detect them. The body is elongated and thin, resembling that of a stick-insect. The first two thoracic segments are more or less completely fused with the head; the second and third thoracic limbs end in claws; the two following thoracic limbs are normal in the genus Proto, rudimentary in Protella, and absent in the remaining genera, though their gills remain as conspicuous flabellate structures. The three hind legs are normal, and the abdomen is reduced to a tiny wart at the hind end of the greatly elongated thorax.

P. Mayer has described cases of external hermaphroditism as being fairly common in certain species, e.g. Caprella acutifrons, and this is interesting if we take into consideration the frequent partial hermaphroditism exhibited by the gonad of Orchestia at certain times of year (see p. 104).

Fam. 2. Cyamidae.—These are closely related to the Caprellidae in the form of the limbs and the reduced state of the abdomen. Cyamus ceti, which lives ectoparasitically on the skin of whales, has the body expanded laterally instead of being elongated, as in the Caprellids.

Sub-Order 3. Hyperina.

Fig. 97.—Phronima sedentaria, ♀, in a Pyrosoma colony, × 1. (After Claus, from Gerstaecker and Ortmann.)

These are an equally distinct and curious group of Amphipods, characterised by the large size of the head and the transparency of the body. Instead of haunting the littoral zone they are pelagic in habit, and many of them live inside transparent pelagic Molluscs, Tunicates, or Jellyfish. A well known form is Phronima sedentaria, which inhabits the glassy barrel-like cases of the Tunicate Pyrosoma in the Mediterranean. The female is often taken in the plankton together with her brood in one of these curious glass houses; the zooids of the Pyrosoma colony are completely eaten away and the external surface of the case, instead of being rough with the tentacles of the zooids, is worn to a smooth, glass-like surface. It has been observed that the female actively navigates her house upon the surface of the sea; she clings on with her 141thoracic legs inside, while the abdomen is pushed out through an opening of the Pyrosoma case behind, and by its alternate flexion and extension drives the boat forwards, the water being thus made to enter at the front aperture and supply the female and her brood with nourishment.

DIVISION 3. HOPLOCARIDA.

The carapace leaves at least four of the thoracic somites distinct. The eyes are pedunculate. The mandibles are without a lacinia mobilis; there are no oostegites, the eggs being carried in a chamber formed by the maxillipedes. The hepatic caeca are much ramified, the heart is greatly elongated, stretching through thorax and abdomen, with a pair of ostia in each segment. The spermatozoa are spherical, and there is a complicated and peculiar metamorphosis.

Order. Stomatopoda.

Fig. 98.—Lateral view of Squilla sp., × 1. A.1, A.2, 1st and 2nd antennae; Ab.1, 1st abdominal segment; Ab.6, 6th abdominal appendage; C, cephalothorax, consisting of the head fused with the first five thoracic segments; E, eye; M, 2nd maxillipede; T, telson. (After Gerstaecker and Ortmann.)

The Stomatopoda are rather large animals, occasionally reaching a foot in length, all of which exhibit a very similar structure; Squilla mantis and S. desmaresti are found on the south coast of England not very frequently; but they are very common in the Mediterranean, living in holes or in the sand within the littoral zone of shallow water. They differ from all the other Malacostraca 142by a combination of characters, and Calman proposes the term Hoplocarida for a division equivalent to the Peracarida, Eucarida, etc.

The abdomen is very broad and well developed, ending in a widely expanded telson. There is a carapace which covers the four anterior thoracic segments, leaving the four posterior segments free. The portion of the head carrying the stalked eyes constitutes an apparently separate segment articulated to the head. The antennae, mandibles, and maxillae are normal; there then follow five pairs of uniramous thoracic limbs turned forwards as maxillipedes and ending in claws; the second pair of these is modified into a huge raptorial arm, exactly resembling that of a Praying Mantis (cf. vol. v. p. 242), by means of which the Squilla seizes its prey. The last three thoracic limbs are small and biramous. The pleopods are powerful, flattened, biramous swimming organs with small hooks or “retinaculae” upon their endopodites, which link together each member of a pair in the middle, and with large branching gills upon the exopodites.

The internal anatomy exhibits several primitive features. The nervous system is not at all concentrated, there being a separate ganglion for each segment; and the heart stretches right through thorax and abdomen, with a pair of ostia in each segment. There are also ten hepatic diverticula given off segmentally from the intestine.

The female has the curious habit of carrying the developing eggs in a chamber improvised by the apposition of the maxillipedes, so that it looks rather as if she were in the act of devouring her own brood.

The metamorphosis of the larvae, despite the work of Claus^[114] and Brooks,^[115] is not very accurately known, especially uncertain being the identification of the different larvae with their adult forms. The chief interest consists in the fact that certain of the anterior thoracic limbs develop in their normal order and degenerate, to be reformed later, just as in the Phyllosoma larva of the Loricata (see pp. 165, 166).

In one series of larvae, probably not of Squilla itself, but of related genera, the young hatch out as “Erichthoidina” (Fig. 99), 143with the thoracic appendages developed as biramous organs as far as the fifth pair, and with a single abdominal pair of limbs.

Fig. 99.—Erichthoidina larva of a Stomatopod, with five pairs of maxillipedes, and the first pair of abdominal appendages, × 10. (From Balfour, after Claus.)

The abdominal series of limbs is next completed; the second thoracic limb assumes its adult raptorial structure, but the succeeding three limbs become greatly reduced and may entirely degenerate, leaving the posterior six thoracic segments without limbs.

Usually the anterior three pairs are only reduced, and then redevelop side by side with the small posterior limbs as they appear. This larva is then termed the “Erichthus” (Fig. 100); but when they completely disappear the larva is called a “Pseudozoaea,” owing to its resemblance to the Zoaea stage of the Decapoda, which is also characterised by the suppressed development of the thoracic segments.

Fig. 100.—Older Erichthus larva, with six pairs of abdominal appendages, × 15. (From Balfour, after Claus.)

The so-called “Alima” larva of Squilla is also a Pseudozoaea, but it is apparently arrived at directly without the previous formation and degeneration of the anterior thoracic limbs, the larva hatching out from the egg in the Pseudozoaeal stage.

Fam. Squillidae.—Of the six known genera none extend into the cold subarctic seas; the majority are characteristic of the warm or tropical seas (Gonodactylus), some of the species having very wide ranges, e.g. G. chiragra, which is completely circumtropical, and appears to have entered the Mediterranean at some period, though it is very rare there.

DIVISION 4. EUCARIDA.

The carapace fuses with all the thoracic segments. The eyes are pedunculate. The mandible is without a lacinia mobilis. There are no oostegites, the eggs being attached to the endopodites of the pleopods. The hepatic caeca are much ramified, the heart is abbreviated and saccular, the spermatozoa are spherical with radiating pseudopodia, and development is typically attended by a complicated larval metamorphosis.

Order I. Euphausiacea.

Fig. 101.—Calyptopis larva of Euphausia pellucida, × about 20. A.1, 1st antenna; Ab.6, 6th abdominal segment; E, eye; M, maxillipede. (After Sars.)

The Euphausiidae^[116] agree with the Decapoda in passing through a complicated larval metamorphosis. The young hatch out as Nauplii, with uniramous first antennae and biramous second antennae and mandibles. In the next stage, or “Calyptopis” (Fig. 101), which corresponds exactly to the Zoaea of the Decapoda, two pairs of maxillae and a pair of biramous maxillipedes are added; the hinder thoracic segments are undifferentiated, but the abdomen is fully segmented, 145and the rudiments of the sixth pair of pleopods are already visible.

In the next stage (“Furcilia”) the other abdominal pleopods are added, the whole series being completed before the thoracic appendages number more than two or three. This stage corresponds to the Metazoaea of the Decapoda, and the interference in the orderly differentiation of the segments with their appendages from before backwards is a phenomenon which we shall meet again when we treat of Decapod metamorphosis. It is evidently a secondary modification, furnishing the larva precociously with its most important swimming organs so as to enable it to lead a pelagic existence. The frequent violation of the law of metameric segmentation, that the most anterior segments being the first formed should be the first to be fully differentiated, leads us to suppose that the larval stages of the Eucarida at any rate do not represent phylogenetic adult stages through which the Malacostraca have passed. Nor do they, perhaps, even represent primitive larval stages, but have been secondarily acquired from an embryonic condition which used to be passed through within the egg-membranes, as in Nebalia and the Mysidacea, when the order of differentiation of the segments was normal. The case is a little different with the Nauplius larva. This larval form, in an identical condition, is found both in the Entomostraca as a general rule, and again in certain Malacostraca, viz. the Euphausiidae and the Peneidea. Whatever its phylogenetic meaning may be, we may be quite certain that the ancestor of the two great divisions of the Crustacea had a free-swimming Nauplius larva, and this conclusion is confirmed by the probable presence of a Nauplius larva in Trilobites.

The Euphausiidae, in contradistinction to the Mysidae, are frequently met with in the surface-plankton. Euphausia pellucida (Fig. 102) is of universal distribution, and is frequently taken at the surface as well as at considerable depths.

Many noteworthy features in Euphausiid organisation are brought out in Fig. 102. The shrimp-like appearance of the carapace and antennae indicate the special Decapodan affinities of the family; noteworthy, also, are the single series of gills and the biramous thoracic and abdominal limbs, similar to those of the Mysidacea. The Euphausiidae also possess phosphorescent organs of a highly developed kind, and these are usually situated, 146as in the type figured, upon the outer margins of the stalked eyes, on the bases of the second and seventh thoracic limbs, and on the ventral median line on the first four abdominal segments. These organs are lantern-like structures provided with a lens, a reflector, and a light-producing tissue, and they are under the control of the nervous system. Their exact use is not known, any more than is the use of phosphorescence in the majority of organisms which produce it; but in certain cases it appears that the Euphausiids make use of their phosphorescent organs as bull’s eye lanterns for illuminating the dark regions into which they penetrate or in which some of them permanently dwell. At any rate, associated with the presence of these organs in some deep-sea Euphausiids are remarkable modifications of the eyes; and we may perhaps here fittingly introduce a short discussion of these visual modifications in deep-sea Crustacea, and the conditions which call them forth.

Fig. 102.—Euphausia pellucida, female, × 5. G, Last gill; L, luminous organ of first leg; L′, luminous organ of 2nd abdominal segment; T, biramous thoracic appendages. (After Sars.)

Fig. 103.—A, Sections (diagrammatic) of Crustacean compound eye, A, with pigment in light-position for mosaic vision; B, with pigment in dark-position for refractive vision. c, Corneal lens; c.g, corneagen cells; cr, crystalline cone; f, basal membrane, or membrana fenestrata; ip, irido-pigment; n, nerve; r, retinula; rh, rhabdom; rp, retino-pigment; v, vitrella.

The compound eyes of Crustacea resemble those of Insects in that they are composed of a very large number of similar elements or “ommatidia,” more or less isolated from one another by pigment. Each ommatidium consists typically of a corneal lens (Fig. 103, c), secreted by flat corneagen cells (c.g) below; beneath the corneal lens is a transparent refractive body called the “crystalline cone” (cr), which is produced by a number of cells surrounding it called the “vitrellae” (v). Below the crystalline cone comes the “rhabdom” (rh), produced and nourished by “retinulacells” 147(r). The rhabdom is a transversely striated rod, constituting the true sensory part of each ommatidium, and is in connexion at its lower end with a nerve-fibre (n), passing to the optic ganglion. The rhabdoms rest upon a membrane (f) called the “membrana fenestrata.” Each ommatidium is isolated from its fellows which surround it by a complete cylinder of pigment, part of which is especially crowded round the crystalline cone, and is known as “irido-pigment” (ip), while the part which surrounds the rhabdom is called “retino-pigment” (rp).

When the pigment is arranged in this way, as in Fig. A, only those rays of light which strike an ommatidium approximately at right angles to the corneal surface can be perceived, since only these can reach the top of the rhabdom; the others pass through the crystalline cones obliquely, and are absorbed by the cylinder of pigment surrounding each ommatidium, so that they neither reach the rhabdom of the ommatidium which they originally entered, nor can they penetrate to the rhabdom of neighbouring ommatidia. This gives rise to what is known as “mosaic vision,” that is to say, each ommatidium only perceives the rays of light which are parallel to its long axis, and in this way an image is built up of which the various points are perceived side by side by means of separate eye-elements. The distinctness and efficiency of this mode of vision depends chiefly upon the number of ommatidia present, and the completeness with which they are isolated from one another by the pigment. Now this form of vision, depending as it does 148upon the absorption of a great number of the light-rays by pigment, and the transmission of only a limited number to the sensory surface, is only possible when there is a strong light, and there is no need for economising the light-rays. The most important discovery was made by Exner,^[117] that the majority of animals with compound eyes had the power of so arranging the pigment in their eyes as to enable them to see in two ways. In bright light the pigment is situated as in Fig. 103, A, so as completely to isolate the rhabdoms from one another (day-position); but in the dusk the pigment actively migrates, the irido-pigment passing to the surface (B) near the tops of the crystalline cones, and the retino-pigment passing interiorly to rest on the membrana fenestrata at the bases of the rhabdoms (night-position). When this happens the rays of light which strike the ommatidia at all sorts of angles, instead of being largely absorbed by the pigment, are refracted by the crystalline cones and distributed over the tops of the rhabdoms, passing freely from one ommatidium to another. In this way the eye acts on this occasion, not by mosaic vision, but on the principle of refraction, as in the Vertebrate eye. Of course the distinctness of vision is lost, but an immense economy in the use of light-rays is effected, and the creature can perceive objects and movements dimly in the dusk which by mosaic vision it could not see at all. The pigment is contained in living cells or chromatophores, and it is carried about by the active amoeboid movements of these cells with great rapidity.

Now, besides the active adaptability to different degrees of light brought about in the individual by these means, we find Crustacea living under special conditions in which the eyes are permanently modified for seeing in the dusk, and this naturally occurs in many deep-sea forms.

Doflein^[118] has examined the eyes of a great number of deep-sea Brachyura dredged by the Valdivia Expedition, and as the result of this investigation he states that the eyes of deep-sea Brachyura are never composed of so many ommatidia, nor are they so deeply pigmented as those of littoral or shallow water forms. At the same time an immense range of variation occurs among deep-sea forms which are apparently subjected to similar 149conditions of darkness, a variation stretching from almost normal eyes to their complete degeneration and the fusion of the eye-stalks with the carapace; and this variation is very difficult to account for. A very frequent condition for crabs living at about 100 fathoms, and even more, is for either the irido-pigment or the retino-pigment to be absent, for the number of ommatidia to be reduced, and for the corneal lenses to be greatly arched. There can be little doubt that these crabs use their eyes, not for mosaic vision, but to obtain the superposition-image characteristic of the Vertebrate eye. In deeper waters, where no daylight penetrates at all, this type of eye is also met with, and also further stages in degeneration where all pigment is absent, and the ommatidia show further signs of reduction and degeneration, e.g. Cyclodorippe dromioides. In a few forms, e.g. Cymonomus granulatus among Brachyura, and numerous Macrura, the ommatidia may entirely disappear, and the eye-stalks may become fused with the carapace or converted into tactile organs.

Progressive stages in degeneration, correlated with the depth in which the animals are found, are afforded by closely related species, or even by individuals of apparently the same species. Thus in the large Serolidae of Antarctic seas, Serolis schytei occurs in 7–128 metres, and has well-developed eyes; S. bronleyana, from 730 to 3600 metres, has small and semi-degenerate eyes; while S. antarctica in 730–2920 metres is completely blind. Lispognathus thompsoni is a deep-water spider-crab, and the individuals taken at various depths are said to exhibit progressive stages in degeneration according to the depth from which they come.

At the same time many anomalies occur which are difficult to explain. In the middle depths, i.e. at about 100 fathoms, side by side with species which have semi-degenerate or, at any rate, poorly pigmented eyes, occur species with intensely pigmented eyes composed of very numerous ommatidia, e.g. the Galatheid Munidopsis and several shrimps, while in the true abysses many of the species have quite normal pigmented eyes. This is especially the case with the deep-sea Pagurids, of which Alcock describes only one species, Parapylocheles scorpio, as having poorly pigmented eyes. An attempt to account for this was made by Milne Edwards and Bouvier,^[119] who pointed out that the truly deep-sea forms with well-developed eyes were always 150Crustacea of a roving habit, which were perhaps capable of penetrating into better lit regions, and to whom well-developed eyes might be useful, while the degenerate forms were sluggish. This explanation cannot be held to account for the phenomenon, as too many deep-sea forms with fairly normal eyes are known which are never taken outside deep waters. Doflein (loc. cit.) points out that in the Brachyura of the deep sea there is a remarkable correlation between the degree of degeneration of the eye and the size of the eggs—the large-egged forms having unpigmented and degenerate eyes, while the species with small eggs have pigmented eyes. He supposes that the species with large eggs undergo a direct development without pelagic free-swimming larvae, and that since they never reach the surface their eyes never meet with the necessary stimulus of light for the development of pigment; whereas the small-egged species undergo a pelagic larval existence when this stimulus is present and gives the necessary initiative for the development of the pigment.

Another factor enters into the question of eye-degeneration in the Crustacea. The great majority of deep-sea animals, including many deep-sea Crustacea, are phosphorescent, and it is certain that although daylight never penetrates into the abysses of the ocean, yet there is considerable illumination derived from the phosphorescence of the inhabitants of these regions.

Alcock^[120] points out in this connexion that the Pagurids, which are conspicuous in the great depths as animals with normally developed eyes, carry about anemones with them, and these organisms are very frequently phosphorescent to a high degree. It may well be, therefore, that the Pagurids are enabled to use their eyes in the normal manner owing to the phosphorescent light which they carry about with them, and this use of phosphorescent light may apply to a number of deep-sea Crustacea whose eyes are not at all or only partially degenerate.

An extremely interesting case of the use of phosphorescent light is given by Chun.^[121] In a number of Euphausiids occurring in deep waters each compound eye is divided into two parts—a frontal and ventro-lateral—which differ from one another very greatly in the nature and disposition of their ommatidia.

Fig. 104.—Section of eye of Stylocheiron mastigophorum. A, Frontal portion; B, ventro-lateral portion; C, phosphorescent organ; D, entrance of optic nerve; c, corneal lens; cr, crystalline cone; pg, pigment; ret, retinula; rh, rhabdom. (After Chun.)

In the frontal portion (Fig. 104, A) the ommatidia are few in 151number and long, the corneal lenses are highly arched, and the pigment is reduced to a few clumps in the iris. This part of the eye is evidently adapted for forming a vague superposition-image in the dusk. The ventro-lateral part (B), on the other hand, is composed of numerous small ommatidia, the crystalline cones of which can be completely isolated from one another by the irido-pigment. Immediately below this part of the eye is a phosphorescent organ (C) provided with a lens and tapetum. Chun suggests that the ventro-lateral part of the eye is used for obtaining a clear mosaic image of objects illuminated by the phosphorescent organ, while the frontal part of the eye is used for obtaining general visual impressions in dimly lit regions. This curious differentiation of the eye into two parts apparently only occurs in predaceous animals, which capture their prey alive upon the bottom, and to whom a clear vision of moving organisms is a necessity.

Another instance of Crustaceans making use of their own light is given by Alcock,^[122] who found two deep-sea prawns, Heterocarpus alphonsi and Aristaeus coruscans, at about 500 fathoms in the Indian Ocean. These animals produce a highly phosphorescent substance which they eject from the antennary glands, and they possess very large, deeply-pigmented eyes.

The whole subject of the modification of the pigment and structure of Crustacean eyes is an interesting one, because it presents us with one of those cases in which the direct response to a stimulus acting within the lifetime of the individual seems to run parallel to the fixed adaptations of a whole species, which have become hereditary and apparently independent of the external stimulus of light or of the absence of light. As far 152as is known, however, the direct response of the individual to the absence of light is limited to the reduction or disappearance of the pigment, and does not extend to those structural changes in the ommatidia which are characteristic of so many deep-sea forms.

Order II. Decapoda.^[123]

The Decapoda, together with the Euphausiidae, make up the Division Eucarida, the members of which differ from the Orders hitherto described in a number of characters, e.g. the presence of a carapace covering the whole of the thorax, the absence of a brood-pouch formed of oostegites, the presence of a short heart, of spermatozoa with radiating pseudopodia, and of a complicated larval metamorphosis, of which the Zoaea stages are most prominent.

The Decapoda differ from the Euphausiidae chiefly in the anterior three thoracic limbs being turned forwards towards the mouth to act as maxillipedes, and in the five succeeding thoracic limbs being nearly always uniramous and ambulatory or chelate; there are typically present three serial rows of gills attached to the thoracic segments, an upper series (“pleurobranchiae”) attached to the body-wall above the articulation of the limbs, a middle series (“arthrobranchiae”) attached at the articulation of the limbs, and a lower series (“podobranchiae”) attached to the basal joints of the limbs. These gills are enclosed in a special branchial chamber on each side of the thorax, formed by lateral wings of the carapace known as “branchiostegites.” The gills of each series are never all present in the same animal, the anterior and posterior members showing a special tendency to be reduced and to disappear. In this manner “branchial formulae” can be constructed for the various kinds of Decapods, which differ from the ideal formula in a manner distinctive of each kind. The second maxilla is always provided with an oar-like appendage on its outer margin (exopodite), known as the “scaphognathite,” which, by its rhythmical movement, keeps up a constant current of water through the gill-chamber.

A complicated auditory organ is present on the basal joint of the first antennae; this is a sac communicating with the exterior and lined internally with sensory hairs. The animal is 153said to place small pieces of sand, etc., in its ears to act as otoliths. Anaspides (see p. 116) is the only other Crustacean which has an auditory organ in this position.

The larval histories of the Decapods^[124] are of great interest, and will be given under the headings of the various groups. The first discoverer of the metamorphosis of the Decapoda was the Irish naturalist J. V. Thompson, certainly one of the ablest of British zoologists. In 1828, in his Zoological Researches, he describes certain Zoaeas of the Brachyura and proves that these animals are not an adult genus, as supposed, but larval forms. But Rathke, in 1829, described the direct development of the Crayfish; and Westwood, after describing the direct development of Gecarcinus, utterly denied Thompson’s assertions concerning metamorphosis. Thompson replied in the Royal Society Transactions for 1835, and described the Megalopa stage of Cancer pagurus. Rathke,^[125] although previously an opponent of Thompson, subsequently made confirmatory observations upon the larvae of the Anomura; and Spence Bate clinched the matter by describing Brachyuran metamorphosis with great accuracy in the Philosophical Transactions for 1859. Since then a mass of work has been done on the subject, though much detail still remains to be elucidated.

The Decapoda fall into three sub-orders, which graduate into one another—(i.) the Macrura, including the Lobsters, Crayfishes, Shrimps, and Prawns; (ii.) the Anomura, including the Hermit-lobsters and Hermit-crabs; and (iii.) the Brachyura or true Crabs.

Sub-Order 1. Macrura.

This sub-order^[126] is characterised by the large abdomen, furnished with five pairs of biramous pleopods, and ending in a powerful tail-fan composed of the telson and the greatly expanded sixth pair of pleopods, the whole apparatus being locomotory. The second antennae are furnished with very large external scales, representing the exopodites of those appendages. Some of the Shrimps and Prawns closely resemble the “Schizopods,” but the pereiopods are nearly always uniramous.^[127] Several subdivisions of the Macrura are recognised.

This tribe includes the Lobsters and Crayfishes, animals well known from their serviceableness to man. There are three families, which will be treated separately.

Fam. 1. Nephropsidae. The podobranchs are not united with the epipodites, and the last thoracic segment is fixed and fused to the carapace. The chelae are generally asymmetrical. The most important Lobsters are the European and the American species—Homarus vulgaris (= Astacus gammarus) and H. americanus respectively; these animals engage a large number of people in the fisheries. It is estimated that in America about £150,000 are spent every year on Lobsters.

The genus Nephrops contains the small Norwegian lobster and other forms.

Herrick^[128] gives some interesting particulars with regard to the life-history of the American species. The largest recorded specimen weighed about twenty-five pounds, and measured twenty inches from rostrum to tail; similar European specimens have been recorded, but, on the average, they are not so large as the American forms.

The Lobster, like all Crustacea, undergoes a series of moults as the result of increase in size, shedding the whole of the external integument in one piece. This is accomplished by a split taking place on the dorsal surface at the junction of thorax and abdomen; through the slit so formed the Lobster retracts first his thorax with all the limbs, and then his abdomen. When first issuing from the old shell the animal’s integument is soft and pulpy, but the increase in size of the body is already manifest; this increase per moult, which is approximately the same in young and adult animals, varies from 13 to 15 per cent of the animal’s length. According to this computation, a Lobster 2 inches long has moulted fourteen times, 5 inches twenty times, and 10 inches twenty-five times, and it may be roughly estimated that a 10–inch Lobster is four years old. Young Lobsters probably moult twice a year, and so do adult males, but females only moult once a year soon after the young are hatched out.

The process of moulting or ecdysis is an exceedingly 155dangerous one to the Lobster and to Crustacea in general, and is very frequently fatal. There is, first of all, the danger of the act not being accomplished skilfully, when death always ensues. The Lobster remains soft and unprotected for about six weeks after the ecdysis, and is very apt to fall a prey to the predaceous fish, such as Sharks, Skates, Cod, etc., which feed upon it. There are, however, some peculiar adaptations connected with the process which are of interest. In order to facilitate the ecdysis, areas of absorption are formed upon the dorsal and ventral surfaces of the carapace, on the narrower parts of the chelipedes, and at other places; in these areas the calcium carbonate is absorbed, and the old shell becomes elastic and thin, so as to allow a more easy escape for the moulting Lobster. It has been noticed that while this is taking place large concretions of calcium carbonate are formed at the sides of the stomach, known as “gastroliths,” which perhaps represent the waste lime that has been abstracted from the areas of absorption. After moulting the Lobster is in great need of lime for stiffening his shell, and it has been noticed that on these occasions he is very greedy of this substance, even devouring his own cast-off skin.

The male Lobster is especially prized on account of his larger chelae, but in both sexes the chelipedes are differentiated into a smaller cutting pincer and a larger crushing one. Lobsters may be right or left handed, with the large crushing claw on the right or left hand, and sometimes specimens occur with the smaller cutting pincers on both chelipedes, and very rarely, indeed, with crushing claws on both sides. Crustacea very commonly have the power of casting off a limb if they are seized by it or if it is injured, and of regenerating a new one. In the Lobster a so-called breaking joint is situated on each leg at the suture between the fused second and third segments; a membrane being pushed inwards from the skin, which not only serves to form a weak joint where rupture may easily take place, but also to stop excessive bleeding after rupture. In the newly-hatched larvae there is a normal joint between the second and third segments; and autotomy, or the voluntary throwing away of a limb, never occurs until the fourth larval stage, when the breaking joint is formed. Autotomy is a reflex act under the control of the segmental ganglion; if a Crab or Lobster be anæsthetised, and then a limb be injured or broken off below the 156breaking joint, the animal forgets to throw the injured leg or stump off at the breaking joint, a proceeding which always occurs under normal conditions. The regeneration of a limb starts from a papilla which grows out of the breaking joint, and after a number of moults acquires the specific form of the limb that has been lost. A number of interesting observations have been made upon the regeneration of the limbs in Crustacea. It was in the Hermit-crab that Morgan^[129] proved that regeneration and the liability to injury do not always run parallel, as Weismann held they should, since the rudimentary posterior thoracic limbs, which are never injured in nature, can regenerate when artificially removed as easily as any others. Przibram^[130] has shown that in the shrimp Alpheus, whose chelipedes are highly asymmetrical, if the large one be cut off, the small one immediately begins to grow and to take on the form of the large one, while the regenerated limb is formed as the small variety. This remarkable inversion in the symmetry of the animal clearly ensures that, if the large chela is injured and thrown away, the least amount of time is wasted in providing the shrimp with a new large claw.

To return to the Lobster; for the majority of the individuals there is a definite breeding season, viz. July and August, but a certain proportion breed earlier or later. A female begins to “berry” at about eight inches in length, and to produce more and more eggs up to about eighteen inches, when as many as 160,000 eggs are produced at a time; after this there is a decline in numbers. A female normally breeds only once in two years. Strict laws are enforced forbidding the sale of Lobsters and Crabs “in berry” in both England and America. The period of incubation, during which the developing eggs are attached to the swimmerets of the female, lasts about ten or eleven months, so that the larvae are hatched out approximately in the following June. On hatching, the larva, which measures about one-third of an inch, and is in the Mysis stage (i.e. it possesses all the thoracic limbs in a biramous condition, but is without the abdominal limbs), swims at first on the surface. After five or six months of this life, during which the abdominal pleopods are added from before backwards, it sinks to the bottom, loses the 157exopodites of the thoracic limbs, and is converted into the young Lobster, measuring about half an inch in length. The little Lobster starts in deepish water, and gradually crawls towards the shore; here it passes its adolescence, but on coming to maturity it migrates out again into the deep water.

Fam. 2. Astacidae.—In this family, which includes all the European and North American Crayfishes, Astacus (Potamobius) and Cambarus, the podobranchs are united with the epipodites, the last thoracic segment is free, there is only one pleurobranch or none at all, the gills have a central lamina, but the filaments are without terminal hooks, and the endopodites of the first two pairs of abdominal appendages in the male serve as copulatory organs. For the distribution, etc., of these forms see p. 213.

Fam. 3. Parastacidae.—This family includes the Crayfishes of the Southern Hemisphere, viz. Parastacus from South America, Astacopsis and Engaeus from Australia, Paranephrops from New Zealand, and Astacoides from Madagascar. These genera agree with the Potamobiidae in the union of the podobranchs with the epipodites, and in the free condition of the last thoracic segment, but there are generally four pleurobranchs, the gills are without a lamina, the filaments have terminal hooks, and there are no sexual appendages in the male. For distribution, etc., see also p. 213.

The larval development in the Crayfishes is still more abbreviated than in the Lobsters, the Mysis stage being passed through within the egg-membranes. The young, when they hatch out, are furnished with hooks upon the chelipedes, by which they anchor themselves to the pleopods of the mother.

Tribe 2. Eryonidea.

Fig. 105.—Willemoesia inornata, × ⅓. (From a figure prepared for Professor Weldon.)

These are remarkably archaic animals of great rarity, though they were common enough in Triassic seas, and have come down to us as fossils from those times, being thus among the oldest Decapoda known. They only survive now as deep sea species, and the genus discovered by the Challenger,^[131] Willemoesia (Fig. 105), confirmed the expectations of the Challenger naturalists that the abysses of the ocean would contain relics from older periods which 158had managed to survive where the competition was not so keen. The genus Willemoesia is very widely distributed, being dredged up from below a thousand fathoms in the Indian Ocean, the Mediterranean, North and South Atlantic, and the Pacific oceans. All the walking legs are chelate, and the animal is quite blind, as are all the Eryonidea, the eye-stalks being fused with the carapace.

Only a single family Eryonidae is recognised.

Tribe 3. Peneidea.—Tribe 4. Caridea.

We will now consider the Shrimps and Prawns, since in them occurs the most complete metamorphosis found in the Decapoda. The Peneidea are distinguished from the ordinary Prawns and Shrimps (Caridea) by having the first three instead of the first two pereiopods chelate. The genus Peneus affords several species which are of commercial value as objects of food; the edible Prawns of the Mediterranean belong to this genus, while in the North Sea two of the Caridea, viz. the Shrimp, Crangon vulgaris, and the Prawn, Palaemon serratus, are the forms very commonly eaten. Both subdivisions are well represented in the deep sea fauna from all parts of the world. Glyphocrangon spinulosa (Fig. 110, p. 164) is a deep sea Shrimp with eyes that have lost their pigment, and with the body covered with spines, while the last abdominal segment is fused with the telson to form a sharp bayonet-like process at the hind end of the body. Some of the deep-sea Prawns of the Indian Ocean 159are described by Alcock^[132] as possessing peculiar secondary sexual characters. Thus Parapeneus rectacutus ♂ has one lash of the first pair of antennae peculiarly bent to form a clasping organ, while Aristaeus crassipes has a hook on the end of the third maxillipede. In the latter the females have much longer rostra than the males, and are in general more powerfully built, so that they seem to have usurped the proper functions of the male, and probably engage in combats with one another over his person.

Fig. 106.—Nauplius larva of Peneus, sp. × 25. (From Balfour, after F. Müller).

As a general rule the Shrimps and Prawns occur in large shoals in the shallow waters of the littoral zone, and they have a remarkable power of adapting their colours to the surroundings in which they happen to be at any particular moment.^[133] This is brought about by the variously coloured chromatophores, which contract and expand in obedience to a stimulus transmitted through the eyes of the animal. A number of the Palaemonidae go up rivers into fresh water, while one family, the Atyidae, live in the completely fresh water of rivers and inland lakes. The Peneidea undergo a very complete metamorphosis which is primitive in respect to the order of formation of the segments from before backwards. The larva hatches out as a Nauplius (Fig. 106), which by the orderly addition of segments 160behind is converted into the Protozoaea (Fig. 107), possessing two pairs of biramous maxillipedes. It should be noted that the maxillae, which are foliaceous in the adult, are laid down in this condition in the larva, and this principle holds good throughout Crustacean metamorphosis, viz. that when a limb is foliaceous in the adult it is foliaceous in the larva, and when biramous in the adult it is biramous in the larva. Whilst the rest of the thoracic limbs are still rudimentary, the sixth pair of pleopods are being precociously developed (Fig. 108), being the only precociously formed limbs in the Peneidea, though the abdominal segments are fully marked off before the thoracic segments, and so must be considered as precocious in development. When the biramous thoracic limbs are completed the abdominal biramous pleopods are added, beginning from in front backwards. Thus the Mysis stage (Fig. 109) is reached, which resembles in all particulars the adult condition of the Schizopoda. The adult Prawn develops from this stage by the loss of some or all of the exopodites on the thoracic pereiopods.

Fig. 107.—Protozoea larva of Peneus, sp. × 25. (From Balfour, after F. Müller.)

Fig. 108.—Zoaea larva of Peneus, sp. × 25. A, A′, 1st and 2nd antennae; Ab.6, 6th abdominal appendage; Mxp, 2nd maxillipede; T, 4th–8th thoracic appendages (future walking legs). (After F. Müller.)

161Some of the Peneid larvae take on very peculiar forms, e.g. the Zoaeae of the Sergestidae,^[134] which often develop the most wonderful spines all over the body.

Fig. 109.—Mysis stage in the development of Peneus, sp. A.2, 2nd antenna; Ab.6, 6th abdominal appendage; T, telson; Th, the biramous thoracic appendages. (After Claus.)

The Caridea have a greatly abbreviated metamorphosis, the larva hatching out at a late Zoaea stage with all three pairs of maxillipedes fully formed and with a fully segmented abdomen. The succeeding thoracic limbs are added in order from before backwards, though the sixth pair of pleopods appear precociously as in the Peneidea. The other swimmerets do not begin to develop until the thoracic limbs are complete. Some Caridea show a yet more abbreviated metamorphosis, e.g. the fresh-water Palaemonetes varians of S. Europe, which hatches out at the Mysis stage.

We see, therefore, in the metamorphosis of the Macrura several apparently primitive features. In the first place, a free swimming Nauplius stage is preserved in certain forms, identical 162in all respects with the Nauplius of the Entomostraca. Secondly, the thoracic limbs when they are first developed are biramous, thus giving rise to the characteristic Mysis stage which links the Macrura on to the “Schizopoda.” Thirdly, the order of differentiation of the segments is typically from in front backwards, the only precociously developed appendage being the sixth abdominal. None of these characters are reproduced in the higher Decapoda in which there is never a free-living Nauplius, the first larval stage being the Zoaea; a number of the thoracic pereiopods, and usually all of them, are uniramous from the start; and the whole of the abdominal segments with their limbs tend to be precociously developed before the hinder thoracic segments make a distinct appearance.

Tribe 3. Peneidea.^[135]

The third legs are chelate except in genera in which the legs are much reduced. The third maxillipedes are seven-jointed, the second maxillipedes have normal end-joints, and the first maxillipedes are without a lobe on the base of the exopodite. The pleura of the first abdominal segment are not overlapped by those of the second. The abdomen is without a sharp bend. The branchiae are usually not phyllobranchs.

Fam 1. Peneidae.—The last two pairs of legs are well developed, and there is a nearly complete series of gills. Cerataspis,^[136] a pelagic form. Parapeneus, Peneus, Aristaeus, etc.

Fam. 2. Sergestidae.—The last or last two pairs of legs are reduced or lost. The gill-series is incomplete or wanting. Sergestes possesses gills, and the front end of the thorax is not greatly elongated. Lucifer has no gills, and the front of the thorax is greatly elongated, giving a very anomalous appearance to the animal. All the members of this family are pelagic in habit.

Fam. 3. Stenopodidae.—One or both legs of the third pair are longer and much stouter than those of the first two pairs. On a number of small anatomical points this family, including the littoral genus Stenopus from the Mediterranean and other warmer seas and Spongicola commensal with Hexactinellid sponges from Japan, is separated by some authors in a Tribe by itself.

The third legs are not chelate. The third maxillipedes are 4–6 jointed, the end-joint of the second maxillipede nearly always lies as a strip along the end of the joint before it, and the first maxillipedes have a lobe on the base of the exopodites. The pleura of the second abdominal segment overlap those of the first. The abdomen has a sharp bend; the branchiae are phyllobranchs.

Fam. 1. Pasiphaeidae.—In this family the end-joint of the second maxillipedes is normally formed, and exopodites are usually present on all the thoracic limbs. Rostrum small or wanting. Rather numerous genera are known, most of which inhabit the deep sea, though a few come into the littoral zone. Pasiphaea chiefly in the deep sea, Leptochela in the tropical littoral zone.

Fam. 2. Acanthephyridae.—The end-joint of the second maxillipede is modified as in other Caridea, and the rostrum is very strong and serrate, but in the presence of exopodites, and in the form of the mouth-parts, this family agrees with the preceding. It is also a characteristic deep-sea family. Acanthephyra, Hymenodora, Nematocarcinus, etc.

Fam. 3. Atyidae.—This is an entirely fresh-water family, especially characteristic of the rivers and lakes of the tropics, some of the forms being exceedingly large and taking the place of the Crayfishes in these waters. Characteristic of this family is the fact that the fingers of the chelae are spoon-shaped, and carry peculiar tufts of bristles. Exopodites are present on the thoracic limbs of some of the genera (Troglocaris, Xiphocaris from Australia and the Malay Islands, Atyephyra from S. and W. Europe), but are absent in others. Caridina, widely spread and common in Indo-Malay and Africa; Atya from West Indies, West Africa, and Pacific Islands.

Fam. 4. Alpheidae.^[137]—The exopodites are absent, and the rostrum is absent or very feeble. The chelae are powerful, and usually very asymmetrically developed. Alpheus has an enormous number of species which live chiefly in the tropical seas, where they haunt especially the coral-reefs, making their homes among the coral or in sponges, etc. Although occurring in the Mediterranean they penetrate very rarely into colder seas.

164Fam. 5 Psalidopodidae.—This family, characterised by the absence of chelae on the second thoracic limbs, which carry instead a terminal brush of hairs, and by the rudimentary condition of the eyes, is represented by the genus Psalidopus from the deep waters of the Indian Ocean.

Fam. 6. Pandalidae.—The first thoracic limb is without chelae, only six-jointed. The rostrum is large and toothed. The genus Pandalus has numerous representatives in the northern littoral, P. annulicornis being one of the prawns most commonly met with in the fish-markets.

Fam. 7. Hippolytidae.—The first and second thoracic limbs bear chelae, the carpus of the second being divided into two or more segments. The first pair of chelae are not distinctly stronger than the second. Virbius has many species in the littoral zone of all seas, and one species, V. acuminatus, is pelagic. Hippolyte also has numerous littoral forms distributed all over the world, but chiefly in the arctic or subarctic seas. H. varians, common on the English coasts, shows interesting colour-reactions to its surroundings.^[138]

Fig. 110.—Glyphocrangon spinulosa, from the right side, × 1. (From an original drawing prepared for Professor Weldon.)

Fam. 8. Palaemonidae.—The first two pairs of legs are chelate, the carpus of the second not being subdivided. Palaemon serratus, a very common prawn in the British littoral. Palaemonetes in the brackish and fresh waters of Europe and N. America.

Fam. 9. Glyphocrangonidae.—The first pair of legs are subchelate, the carpus of the second pair is subdivided, and the rostrum is long. Glyphocrangon (Fig. 110) with numerous species entirely confined to deep water.

Fam. 10. Crangonidae.—The first pair of legs are subchelate, the carpus of the second pair is not subdivided, and the rostrum is short. Crangon vulgaris is the common Shrimp of the North Sea.

Fig. 111.—Dorsal view of Scyllarus arctus, × ½. (From an original figure prepared for Professor Weldon.)

Fig. 112.—Embryonic area of developing Palinurus quadricornis. Ab.1, 1st abdominal segment; E, compound eye; E′, median simple eye; L, upper lip; L′, lower lip; M, mandible; Mx.1, Mx.2, 1st and 2nd maxillae; Mxp.1, 1st maxillipede; T, 6th (antepenultimate) thoracic appendage. (After Claus.)

The Loricata include the Langouste (Palinurus) of the Mediterranean coasts, which replaces there the Lobster of the North Sea as an article of food, and the peculiarly shaped Scyllarus arctus (Fig. 111), which is also prized in the Mediterranean as a delicacy. The bright red “Crayfishes,” Panulirus and Iasus, of the Australian coasts are also largely used as food. Besides its peculiarity in shape, S. arctus has remarkable scales on the second antennae in place of flagella. The larva hatches out as the so-called Phyllosoma, 166which must be regarded as a greatly flattened and modified^[139] Mysis stage.

Fig. 113.—Phyllosoma larva of Palinurus, sp. × 5. Ab, Abdomen; Mxp, 3rd maxillipede; T, antepenultimate (6th) thoracic appendage. (After Claus.)

In the embryo of Palinurus just before hatching (Fig. 112) we can recognise the limbs of the head and thorax normally developed in order. There are present three thoracic limbs, besides the maxillipedes. When the Phyllosoma hatches out the first maxillipedes have become quite rudimentary, and the second much reduced, while the second antennae and second maxillae are also reduced in size. The metamorphosis is completed by the re-development of the limbs and segments that have been secondarily suppressed during larval life, and by the appearance of the pleopods.

This process is again met with in the Squillidae (p. 143), but it resembles the suppression, in so many Decapodan metamorphoses, of anterior limbs and the precocious development of segments and limbs lying posteriorly. In the ordinary Decapoda, however, the suppressed limbs are merely not formed till later; while in the Loricata the limbs develop in the correct order, and subsequently degenerate. It is natural to wonder whether the condition of affairs in the Loricata represents the primitive process, and whether the precocious development of segments in the other Decapoda owes its origin to these animals having once had the direct mode of development when the segments were formed in the proper order, and to their having subsequently acquired the larval stages first of all by the degeneration, and then by the suppression of certain segments which were not of use during larval life. The complete metamorphosis, however, of the Peneidea, in which the segments and limbs appear in the 167right order, rather goes to show that this is the primitive mode of development in the Decapoda, and that the disarrangement in the order of appearance of the segments, both in the Squillidae and in the Loricata and other Decapods, has been independently acquired in the two cases to meet the needs of the larval existence.

Fam. 1. Palinuridae.—The cephalothorax is subcylindrical, the eyes are not enclosed in separate orbits formed by the edge of the carapace, and the second antennae possess flagella. Palinurus, with P. elephas, the European Rock Lobster or Langouste. Iasus with two species in the Antarctic littoral; Panulirus in the tropical littoral.

Fam. 2. Scyllaridae.—The cephalothorax is depressed, the eyes are enclosed in separate orbits formed by the edge of the carapace, and the second antennae have flat scales in the place of flagella. Scyllarus (Fig. 111), with the European S. arctus; Ibacus in rather deep water with several species, chiefly found in the southern hemisphere.

Tribe 6. Thalassinidea.

This tribe is included by some authors in the Anomura, and held to be closely related to the Galatheidea, but the unreduced abdomen is carried straight and unflexed, and gives a very Macrurous appearance to the animal. The Anomurous characters are the frequent reduction or absence of the antennal scale, the fact that only the first two pairs of pereiopods are ever chelate, and the reduced series of gills. The body is symmetrical, but the first pair of chelae is always highly asymmetrical. The posterior pairs of pereiopods, although small, are not characteristically reduced as in the Anomura. The animals belonging to this Tribe attain two or three inches in length, and generally burrow in sand or mud either in the littoral zone or in deeper waters; at the same time they can swim with considerable activity by means of the pleopods.

Fam. Callianassidae.—Callianassa subterranea is common at Naples, Gebia littoralis in the North Sea.

Sub-Order 2. Anomura

In this division are included the so-called Hermit-lobsters and Hermit-crabs, in which the condition of the abdomen is roughly intermediate between that of the Macrura and that of the Brachyura. 168It is not much reduced in size, and the pleopods of the sixth pair are fairly well developed, but it is usually carried flexed towards the thorax, and is never a powerful locomotory organ as in the Macrura. The antennal scale, if present at all, is a mere spine, not the large leaf-like structure of the Macrura; and there is never a partition between the two first antennae as in the Brachyura.

The last or last two pairs of pereiopods are reduced, and are turned on to the dorsal surface or carried inside the branchial chamber; but this curious character is met with again in certain Brachyura (Dromiacea and Oxystomata).

Tribe 1. Galatheidea.^[140]

Fig. 114.—Dorsal view of Munidopsis hamata, × ½. (From an original figure prepared for Professor Weldon.)

These are symmetrical crabs with a long carapace; the abdomen, which is as broad as the carapace, is always carried flexed under the thorax, and the sixth pair of pleopods are expanded to form with the telson a fan-like tail. The most anterior pereiopods are always much elongated and chelate; while the last pair are much reduced, and either turned up on to the dorsal surface, or else carried in the branchial chamber. The exact meaning of this last characteristic in these forms is doubtful; some of the species are said to carry shells temporarily upon their backs, a proceeding probably assisted by the last pair of thoracic limbs, while in others their limbs may be used for cleaning out the branchial chamber. Most of the Galatheidea, for instance, the common Porcellana and 169Galathea, are littoral animals, and may be found hiding under stones and in crevices on the shore; but a number occur in deep water, e.g. Munida and Munidopsis.

Fig. 115.—Zoaea of Porcellana, × 20. T, Telson. (After Claus.)

The shallow-water species have ordinarily developed eyes; the various species of Munida, which occur in fairly deep but by no means abyssal regions, have usually very large and highly pigmented eyes; while in Munidopsis, which is characteristic of very deep water, the eyes are degenerate and colourless, as shown in Fig. 114.

The Zoaeae, or young larval stages of the Galatheidea, are characterised by the immense length of the spines upon the carapace (Fig. 115). The young Zoaea which hatches out from the egg resembles in other respects that of the Brachyura. The Metazoaea, however, differs from that of the Brachyura in the fact that the third maxillipede is first present as a biramous swimming organ, and at its first appearance is not developed in its definitive form. The other thoracic limbs are not schizopodous when they appear, and indeed in nearly all respects the development proceeds as in the Brachyura.

Fam. 1. Aegleidae.—The gills are trichobranchiae, and there are eight arthrobranchs. There are no limbs on the second abdominal segment of the male. The abdomen is not carried folded on to the thorax. The first two characteristics separate this family from all the other Galatheidea. Aeglea laevis, a fresh-water species from the rivers of temperate S. America, is the sole representative.

Fam. 2. Galatheidae.—The abdomen is not folded against the thorax. The members of this family are often littoral in 170habit (Galathea, Fig. 116), but often go down into great depths (Munidopsis, Fig. 114).

Fig. 116.—Dorsal view of Galathea strigosa, × ½. (From an original figure prepared for Professor Weldon.)

Fam. 3. Porcellanidae.—The abdomen is folded against the thorax, and the body has a crab-like form. These are always littoral in habit, never descending into the depths. Pachycheles in the tropics, Porcellana with numerous species in all seas, P. platycheles being a common British species.

Tribe 2. Hippidea.

The Mole-crabs have the habit of burrowing in sand, and their limbs are peculiarly modified into digging organs for this purpose (see Fig. 117). In other respects they are seen to be closely related to the Galatheidea by the form of the carapace, the condition of the abdomen, and the reduced last thoracic limbs.

In Albunea, which is found in the Mediterranean, the first antennae^[141] are greatly lengthened and apposed to one another, and by means of a system of interlocking hairs they form a tube down which the water is sucked for respiration. The object of this arrangement is to ensure a supply of clear water, filtered from particles of sand, when the crab is buried beneath the surface, on these occasions the tip of the antennal tube being protruded above the surface of the sand. An exactly similar tube is used by the true Crab Corystes cassivelaunus, which has similar burrowing habits, but here the tube is formed from the second antennae and not from the first, so that the tubes in the two cases afford beautiful instances of analogous or homoplastic structures between which there is no homology (see p. 189).

171Fam. 1. Albuneidae.—The first legs are subchelate; the carapace is flattened, without expansions covering the legs. Albunea with several species in the Mediterranean, West Indies, and Indo-Pacific.

Fig. 117.—Remipes scutellatus, dorsal and ventral views, × 1. (From original drawings prepared for Professor Weldon.)

Fam. 2. Hippidae.—The first legs are simple, the carapace is subcylindrical with expansions covering the legs. Remipes (Fig. 117) and Hippa in tropical or sub-tropical seas.

Tribe 3. Paguridea.^[142]

The ordinary Hermit-crabs, common on the English as on every coast, are characterised by the fleshy asymmetrical abdomen from which all the hard matter has disappeared, and which is carried tucked away in an empty Gasteropod shell. The abdomen is spirally wound in accordance with the shape of the shell, and a firm attachment is effected by means of the sixth pair of pleopods, especially that of the left side, which is fashioned into the form of a hook and is curled round the columella of the shell; this attachment is so secure that in trying to pull a Hermit-crab out of its shell the body is torn apart before the hold gives way. The other pleopods are in a much reduced condition, being generally altogether absent from the right side of the abdomen, 172and often greatly reduced on the left side, especially in the male, though in the female they are still used for the attachment of the eggs.

The last two pereiopods are much reduced and are concealed inside the shell, which they help to carry. The great chelae are usually asymmetrically developed, that on the right side being much larger than that on the left, and often serving the purpose of shutting the entrance to the shell when the crab is withdrawn inside.

The constant association of a large group of animals like the Hermit-crabs with the appropriated empty houses of another group is sufficiently curious, but it does not stop there. In almost every case there are present one or more Sea-anemones growing on the outside of the shell, and each kind of Hermit-crab generally carries a special kind of Anemone. Thus at Plymouth, Eupagurus bernhardus is generally symbiotic with Sagartia parasitica, or else with a colony of Hydractinia echinata, while E. prideauxii is usually associated with Adamsia palliata. In the latter case the shell is frequently absorbed, so that the Anemone comes to envelop the crab like a blanket. Instead of Anemones carried turret-like and imposing aloft, or enveloping the inmate of the shell like a blanket, some of the Hermits have Sponges, an unexpected association; and it is a common sight at Naples to find the little red round Sponge, Suberites, running around animated by its Hermit within. It is held that Anemone and crab mutually assist one another, that the Anemone stings the crab’s enemies, and that the Hermit-crab carries the Anemone to new feeding-grounds. It is also said that when a crab grows too big for its shell, and is forced to seek another, it persuades the Anemone to loosen its attachment to the deserted shell and to be transplanted to the new one, and that there is something mesmeric in its power, because nobody else can pull an Anemone off a shell without either cutting it off at the base or tearing it to pieces. Other animals as well sometimes enter into this partnership. At Plymouth a Polychaet worm, Nereis fucata, frequently inhabits the Whelk’s shell, together with Eupagurus bernhardus, and puts out its head for a share of each meal; and at Naples the Amphipod Lysianax punctatus is almost always present in the shells of Eupagurus prideauxii.

173

Fig. 118.—Pylocheles miersii, × 1. A, End view of a piece of mangrove or bamboo, the opening of which is closed by the great chelae (c) of the Pagurid; B, the animal removed from its house. (After Alcock.)

Besides the ordinary twisted Pagurids which inhabit Gasteropod shells, there are a few which preserve the symmetry of the body. The interesting Pylocheles miersii^[143] (Fig. 118), taken by the Investigator in the Andaman Sea at 185 fathoms, inhabits pieces of bamboo; it is perfectly symmetrical, with well-developed pleopods and symmetrical chelae, which, when the animal is withdrawn, completely shut up the entrance to its house (Fig. 118, A).

It is doubtful whether this animal ever inhabited a spiral shell or not in its past history; but there is no doubt that a number of peculiar crabs, which caused the older systematists much trouble, are Pagurids, derived from asymmetrical shell-haunting ancestors that have secondarily taken to a different mode of life, and lost, or partially lost those characteristics of ordinary Hermit-crabs which are associated with life in a spiral shell. These are the Lithodidae and the “Robber-crab,” Birgus latro, of tropical coral islands.

Although the Robber-crab and the Lithodidae bear a certain superficial resemblance to one another in that they lead a free existence, and have reacquired to a great extent their symmetry, yet it is clear that they have been independently derived from different groups of asymmetrical Hermit-crabs, and that their resemblance to one another is due to convergence.

Birgus latro (Fig. 119), a gigantic crab, frequently over a foot in length, lives on land, and inhabits the coasts of coral islands in the Indian and Pacific Oceans where cocoa-nut trees 174grow. It feeds on the pulp of the cocoa-nut, which it extracts by hammering with its heavy chela on the “eye-hole” until room is made for the small chela to enter and extract the pulp. There is not the slightest doubt that the animal often ascends the cocoa-nut trees for the purpose of picking the nuts, a fact illustrated by a fine photograph by Dr. Andrews, exhibited in the Crustacean Gallery in the Natural History Departments of the British Museum. It uses the husk of the nut to line its burrow, and it is said to have the habit of putting its abdomen into the nut-shell for protection and carrying it about with it. Owing to its terrestrial mode of life, the branchial chamber is highly modified, being divided into two portions—a dorsal space, the lining of which is thrown into vascular ridges and folds for aerial respiration, and a lower portion where the rudimentary branchiae are situated. Although the Robber-crab lives ordinarily on land, it must be supposed that these branchiae are of some service; the young are hatched out as ordinary Zoaeas in the sea, and go through a pelagic existence before seeking the land. At the present time the Robber-crab is confined to the Pacific and the islands of the Indian Ocean, wherever the cocoa-nut grows. It seems, however, that its association with the cocoa-nut is a comparatively modern one. Mr. C. Hedley, of Sydney, who has had great experience of the Pacific Islands, informs me that the cocoa-nut is not, as is usually supposed, a native of these coral islands, but has been introduced, probably from Mexico, by the Polynesian mariners before the discovery of America by Columbus. Before the introduction of the cocoa-nut the Robber-crab must have fed on some other tree, possibly the Screw Pine, Pandanus.

The abdomen is full of oil, and is much prized as a delicacy by the natives, who tell many strange legends about the creature, but the philosopher may well find its structure more strange than fiction, and the consideration of its morphology an intellectual feast.

The appearance of the thorax and of the thoracic limbs is thoroughly Pagurid; the structure of the abdomen is highly peculiar.

From the ventral surface (Fig. 119) we can see at the tip of the tail three small calcified plates, which represent the fifth and sixth terga and the telson. Attached to the sixth segment are 175the much reduced and rudimentary pleopods of that segment, and on the left hand side of the body in the female are three well-developed pleopods of the first, second, and third segments, which are used for carrying the eggs. The extraordinary asymmetry of these limbs compared with the complete symmetry of the abdomen itself is only explicable on the hypothesis that these animals are descended from Hermit-crabs which had lost the pleopods on the right side.

Fig. 119.—Birgus latro, ♀, × ⅙, ventral view. Ab, First pleopod; T, last pereiopod.

These appendages are entirely absent in the male. The ventral surface of the abdomen is curiously warty and rugose, and is very soft and pulpy owing to the immense store of oil which it contains.

Fig. 120.—Dorsal view of abdomen, A, of Cenobita, sp.; B, of Birgus latro. T, Telson; 1–6, 1st–6th abdominal segments.

If we look at the dorsal surface of the abdomen we find that, unlike that of the Hermit-crabs, it is completely protected by a number of hard plates (Fig. 120, B). Beneath the carapace can be seen a number of small plates belonging to the last thoracic 176segment; following these there are four large plates (1–4) representing the terga of the first four abdominal segments; the fifth, sixth, and the telson are, as has been stated, carried on the under side of the abdomen, but they are represented diagrammatically (5, 6, T) in the dorsal view. Besides the large terga, there are a number of small plates laterally, usually two to each segment, but they show a tendency to subdivide and increase in the largest specimens. This condition of affairs is very different to that in the naked fleshy abdomen of an ordinary Pagurid, but it can easily be deduced from that of the genus Cenobita, ordinary Hermit-crabs found in the Indo-Pacific Oceans, from which the Robber-crab has evidently descended. In Cenobita (Fig. 120, A) we see the same system of plates upon the dorsal surface of the abdomen, but they are much smaller, and the lateral plates are not so numerous; indeed, the greater part of the abdomen remains fleshy and uncalcified. The under surface of the abdomen shows the same rugosity as is found in Birgus, and from a number of other anatomical characters it is evident that the Robber-crab is a highly modified Cenobita that has deserted its shell and developed a symmetrical abdomen protected by expanded and hardened plates which represent those found in a reduced condition in Cenobita. The species of Cenobita although they inhabit shells and have normal branchiae, live on the shore, and have not been seen to descend actually into the sea.

The Lithodidae, which are found in temperate seas, especially on the Northern Pacific coasts (though Lithodes maia occurs in the North Sea, and certain species inhabit deep water in the Indian Ocean), have a deceptively Brachyuran appearance, the thorax being much shortened and the abdomen being much 177reduced and carried tightly flexed on to the ventral surface of the thorax. They live a free, unprotected existence, and are highly calcified. They are, however, certainly Pagurids, as is evidenced by a number of anatomical characters, but most clearly by the asymmetry of the abdomen, especially in the female, which is not only markedly asymmetrical in the arrangement of its dorsal plates (Fig. 121), but also in the presence of three pleopods upon the left side only, as in Birgus. The male is without these appendages, and the sixth pair of pleopods is absent in both sexes. The remarkable calcified plates upon the abdomen bear a superficial resemblance to those in Birgus, but their evolution is traced, not from a Cenobite, but from an Eupagurine stock.^[144]

Fig. 121.—Lithodes maia, ♀, in ventral view, × ¼. The abdomen is flexed on the thorax, so that its dorsal surface is seen. l.3, Lateral plates of third abdominal segment; l.5, left lateral plate of fifth abdominal segment; m, marginal plate; T, brush-like last pereiopod; Te.6, telson and sixth abdominal segment.

In some of the Eupagurinae, e.g. Pylopagurus, feebly calcified plates are present upon the segments of the abdomen (Fig. 122, A).

In the most primitive of the Lithodidae we witness the 178reduction (Fig. 122, B) and disappearance (C) of these original plates, their place being taken first by a number of irregularly situated small spines and warts, which, however, subsequently fuse up to form definite segmental plates. In Lithodes maia, ♂ (D), there are a series of lateral and marginal plates, while in Acantholithus (E) a number of median plates appear, presumably by the fusion of the small spines present in the median line in Lithodes maia; finally, a fusion of the marginal and lateral plates may take place, so that each abdominal segment is covered by a median and two paired lateral plates.

Fig. 122.—Diagrams of abdomen: A, of Pylopagurus, sp.; B, of Hapalogaster cavicauda; C, of Dermaturus hispidus; D, of Lithodes maia, ♂; E, of Acantholithus hystrix. c, Central plates; l, lateral plates; m, marginal plates; T, telson; 1–6, 1st–6th abdominal segments. (After Bouvier.)

It is to be noted that the males and females of the various species do not follow a parallel course of development, the plates in the male being symmetrical, while those of the female are often highly asymmetrical (compare Figs. 122, D, and 121), thus giving the strongest evidence of a Pagurid ancestry.

Birgus and the Lithodidae, then, are Pagurids which have given up living in shells, and have become adapted to a free existence, protecting their soft parts by the development of 179hard plates, and re-acquiring, to a greater or less degree, a secondary symmetry of form. But the story of Pagurid evolution does not apparently stop here. The genus Paralomis, from the West Coast of America, superficially resembles Porcellana, and is held to be descended from such forms as Pylocheles, while isolated species are known (though not well known), such as Tylaspis, described in the Challenger Reports,^[145] which appear to be Pagurids that have deserted their shells.

Fig. 123.—Four stages in the development of Eupagurus longicarpus or E. annulipes, × 20. A, Ventral view of Zoaea; B, lateral view of Metazoaea; C, dorsal view of Glaucothoe; D, dorsal view of adolescent stage. Ab.6, 6th abdominal appendage; Mxp.1, Mxp.3, 1st and 3rd maxillipedes. (After M. T. Thompson.)

The metamorphosis of the Hermit-crabs has recently been studied by M. T. Thompson.^[146]

The Zoaea (Fig. 123, A) differs from that of the Galatheidea mainly in the absence of the long spines. It possesses the usual 180appendages characteristic of the Zoaea, namely, the first and second antennae, mandibles, first and second maxillae, and two pairs of biramous swimming maxillipedes and small third maxillipedes. In the Metazoaea (B), as in the Anomura generally, the third maxillipedes develop into biramous swimming organs, a thing they never do in the Brachyura, and the rudiments of the thoracic segments put in a first appearance. The abdominal segments are already fully formed in the Zoaea stage, so that here as in all other Zoaeas, the order of development from in front backwards is disturbed by the precocious differentiation of the abdominal segments. The next stage is the “Glaucothoe” (Fig. 123, C), which corresponds to the Megalopa of Brachyura (Fig. 125, p. 183). It differs from the adult Hermit-crab in the perfect symmetry of its body, the segmented abdomen, and the presence of five pairs of normal biramous pleopods. At this stage, which lasts four or five days, it resembles closely a little Galatheid. The asymmetry of the adult (Fig. 123, D) is now imposed upon this larva by the migration of the liver, gonads, and green glands into the abdomen, and by the shifting of the posterior lobes of the liver on to the left side of the intestine, which is displaced dorsally and to the right. The gonad lies entirely on the left side. The pleopods of the right side now degenerate, more completely in the male than in the female, and this degeneration is not completed until the little crab has found a shell and lived in it for some time. If a shell is withheld from it, the degeneration of the pleopods is much retarded, so that although the Hermit-crab assumes its asymmetry without the stimulus of the spiral shell, yet this stimulus is necessary for the normal completion of the later stages.

Fam. 1. Pylochelidae.—The abdomen is macrurous and symmetrical, with all the limbs present. Pylocheles (Fig. 118, p. 173).

Fam. 2. Paguridae.—The abdomen is asymmetrical, with some of the limbs lost. The antennal scale is well developed, and the flagella of the first antennae end in a filament.

Sub-Fam. 1. Eupagurinae.—The third maxillipedes are wide apart at the base, and the right chelipedes are much larger than the left. Parapagurus from deep-sea, Eupagurus from temperate, especially north temperate seas. Pylopagurus.

Sub-Fam. 2. Pagurinae.—The third maxillipedes are approximated 181at the base; the chelipedes are equal or subequal, or the left is much larger. Chiefly in the warm and tropical seas, but Clibanarius and Diogenes also in the Mediterranean.

Fam. 3. Cenobitidae.—The abdomen is as in Paguridae. The antennal scale is reduced, the flagella of the first antennae end bluntly. The members of this family are characteristic of tropical beaches, where they live on the land. Cenobita, with about six species, in the West Indies and Indo-Pacific, living in Mollusc shells; Birgus (Fig. 119) on Indo-Pacific coral islands.

Fam. 4. Lithodidae.—The abdomen is bent under the thorax, and the body is crab-like and calcified. The rostrum is spiniform, and the sixth abdominal appendages are lost.

Sub-Fam. 1. Hapalogasterinae.—Abdomen not fully calcified, and without complicated plates. Hapalogaster and Dermaturus in the North Pacific littoral.

Sub-Fam. 2. Lithodinae.—Abdomen fully calcified, with a complicated arrangement of plates. Lithodes (Fig. 121) practically universal distribution, littoral and deep sea. Acantholithus, deep littoral of Japan; Paralomis, west coast of America. This last genus should probably be placed in a separate family.

Sub-Order 3. Brachyura.^[147]

The abdomen is much reduced, especially in the male, and is carried completely flexed on to the ventral face of the thorax so as to be invisible from the dorsal surface. The pleopods in the male are only present on the two anterior segments, and are highly modified as copulatory organs; the pleopods in the female are four in number and are used simply for carrying the eggs; the pleopods of the sixth pair are always absent in both sexes. The first antennae and the stalked eyes can be retracted into special pits excavated in the carapace.

182

Fig. 124.—A, Zoaea, × 24, and B, Metazoaea, × 13, of Corystes cassivelaunus. Ab, 3rd abdominal segment; An, 1st antenna; E, eye; G, gills; M, 1st maxillipede; T.8, last thoracic appendage. (After Gurney.)

Fig. 125.—Later stage (Megalopa) in the development of Corystes cassivelaunus, × 10. A, Antenna; Ab, 3rd abdominal segment; C, great chela; T.8, last thoracic appendage. (After Gurney.)

The larva hatches out as a Zoaea^[148] (Fig. 124, A) very similar to that of the Anomura; it is furnished with an anterior and posterior spine on the carapace. It is characteristic of the Brachyuran Zoaea that the third maxillipede is fashioned from the beginning in its definitive expanded form, and is never a biramous swimming organ as in the Anomura. The only exception to this rule is found in the Dromiacea, the most primitive of the Brachyura, to be soon considered, in which not only the third maxillipede, but also the first pair of pereiopods may be developed as biramous oars, a condition taking one back to the Mysis stage of the Macrura. The Metazoaea (Fig. 124, B) has the rudiments of the thoracic limbs developed and crowded together at the back of the carapace; they are all laid down in their definitive forms, and the abdomen has the pleopods precociously developed. These Zoaeal stages are of course pelagic, but the Metazoaea next passes into the Megalopa stage (Fig. 125), in which the little crab forsakes its pelagic life and assumes the ground-habits of the adult; the Megalopa, which corresponds exactly to 183the Glaucothoe of the Pagurids, resembles a small Galathea or Porcellana, the abdomen being still large and unflexed and furnished with normal pleopods. From this stage the adult structure is soon achieved, though, owing to the continued growth of the Crustacea even after maturity is reached, there is often a slight progressive change in structure, especially in the male, at each successive moult of the individual. The Megalopa of Corystes cassivelaunus is peculiar in the immense production of the second antennae, which act as a respiratory tube (Fig. 125).

The Brachyura must be considered under the following subdivisions:—

Tribe 1. Dromiacea.

All authorities are agreed that these^[149] are the most primitive of the Brachyura. In them the abdomen is much less reduced in both sexes than in other Brachyura; there is a common orbitoantennary fossa, into which eyes and antennae are withdrawn, instead of a separate one on each side for each organ; the carapace is often much elongated as in the Macrura and Anomura, and a number of other anatomical characters might be mentioned which characterise the Dromiacea as intermediate between the true Brachyura and the lower forms. There are, however, two views as to the relationship of the Dromiacea; Claus held that they proceeded from a Galatheid 184stock, and hence that the development of the Brachyura ran through an Anomurous strain; but Huxley, and latterly Bouvier,^[150] adopt the view that the Dromiacea are descended, not from the Galatheidae, but direct from the Macrura, and especially from the Nephropsidea. Special resemblances are found between the Jurassic Nephropsidae and certain present day Dromiacea, e.g. Homolodromia paradoxa, the detailed form of the carapace in the two cases being very similar. It is, however, a little strange that in the Dromiacea we meet with the same reduction and dorsal position of the last, or last two pairs of thoracic limbs which we saw to be such a characteristic feature of the Anomura, especially of the Galatheidae. In the Dromiacea these limbs may be chelate, and they are used for attaching shells and other bodies temporarily to the back. Must we suppose that this resemblance to the Anomura is due to convergence, or that the Nephropsidae, which gave rise to perhaps both Galatheidae and Dromiacea, had this character, and that it has been subsequently lost in the Macruran stock? We have already mentioned that the Metazoaea of Dromia has not only a well-developed swimming third maxillipede, but also a biramous first pereiopod, a character which speaks strongly for Macruran affinities.

Fig. 126.—Dromia vulgaris, × 1. (After Milne Edwards and Bouvier.)

Fam. 1. Dromiidae.—The eyes and antennules are retractile into orbits. The last two pairs of thoracic limbs are small, and held dorsally. The sixth pair of pleopods are rudimentary or absent. Homolodromia from West Indies, deep-sea. Dromia, widely dispersed. D. vulgaris (Fig. 126) occurs on the English coasts.

Fam. 2. Dynomenidae.—Similar to the preceding family, but only the last pair of thoracic limbs is small, and held dorsally. The sixth pair of pleopods are reduced, but always present. Dynomene in the Indo-Pacific.

Fam. 3. Homolidae.—The eyes and antennules are not retractile into orbits. Only the last pair of thoracic limbs are reduced, the sixth pair of pleopods altogether absent. Homola 185and Latreillia, widely distributed, occur in the Mediterranean. Latreillopsis from the Pacific. L. petterdi,^[151] a magnificent species, with the carapace nearly a foot long, and with very long legs like a Spider-crab, has been dredged from 800 fathoms east of Sydney, New South Wales.

Tribe 2. Oxystomata.

This group comprises Crabs whose carapace is more or less circular, while the mouth, instead of being square as in the remaining Brachyura, is triangular with the apex pointing forward, and the third maxillipedes are not expanded into the flattened, lid-like structures found in other Crabs. There is the same tendency in some of the genera for the posterior thoracic limbs to be reduced and carried dorsally, as in the Galatheidae and Dromiacea. The well-known Dorippe from the Mediterranean has this feature, and frequently carries an empty shell upon its back, and Cymonomus^[152] presents the same peculiarity.

Fig. 127.—Cymonomus granulatus, × 1. A.1, A.2, 1st and 2nd antennae; E, eye-stalk; S, extra-orbital spine of carapace. (After Lankester.)

Cymonomus granulatus (Fig. 127) is an abyssal form that has been dredged from the Mediterranean and North Atlantic, in which the eye-stalks are curiously tuberculated, and the ommatidia of 186the eye are entirely unpigmented and degenerate, though a few corneal facets are still recognisable. This species is replaced by C. quadratus in the Caribbean Sea and by C. normani on the East African coast, in which the alteration of the eye-stalks into thorny, beak-like projections becomes progressively marked, and all traces even of the corneal facets disappear. This remarkable genus was mentioned in the excursus on Crustacean eyes on p. 149.

Fig. 128.—Calappa granulata, from in front, × ½. C, Hand of chelipede; T, walking legs. (After Garstang.)

The Oxystomata, like the Cyclometopa, to be considered later, live in sandy and gravelly regions, and burrow to a greater or less extent, and we find in both groups admirable adaptations for securing a pure stream of water, uncontaminated by particles of sand, for flushing the gills. Perhaps the most remarkable of these adaptations is afforded by Calappa.^[153] This animal has the chelipedes wonderfully modified in structure, and when it is reposing in the sand it holds them apposed to the front of the carapace, as shown in Fig. 128, so that the spines upon their edges, together with the hairy margin of the carapace, form a most efficient filter for straining off sand and grit from the stream of water which is sucked down between the closely-fitting chelipedes and carapace, to enter the branchial chambers at their sides. The exhaled current of water passes out anteriorly through a tube formed by a prolongation of the endopodites of the first maxillipedes. The exhalant aperture is shown in Fig. 128 by the two black cavities below the snout in the middle line.

A similar method is pursued by the related Matuta banksii^[153] (Fig. 129), a swimming and fossorial Crab found in the Indo-Pacific. In this Crab the chelipedes also fit against the carapace to form a strainer, and their function is assisted by the enlargement of the posterior spine, which acts as a kind of elbow-rest to keep 187the chelipedes properly in position. The inhalant openings are situated just in front of the chelipedes. It is a most remarkable fact that among the Cyclometopa, Lupa hastata (Fig. 131) has an exactly similar arrangement. Apparently we have here another instance of convergence, similar to that of Corystes and Albunea, but the case is complicated by the fact that some of the Oxystomata, and among them Matuta, show a certain amount of relationship to the Cyclometopous Portunids, so that it is just conceivable that the resemblances in the respiratory arrangement are due to a common descent and not to convergence.

Fig. 129.—Dorsal view of Matuta banksii, × 1. (From an original drawing prepared for Professor Weldon.)

In the Leucosiidae, of which the Mediterranean Ilia nucleus (Fig. 130) is an example, the inhalant aperture is situated between the orbits, and leads into gutters excavated in the “pterygostomial plates” flanking the mouth, which are furnished with filtering hairs and are converted into closed canals by expansions of the exopodites of the third maxillipedes. Thus these Crabs possess a filtering apparatus independent of the chelipedes and of the margin of the carapace.

Fam. 1. Calappidae.—Cephalothorax rounded and crab-like. The abdomen is hidden under the thorax, the antennae are small, and the legs normal in position. The afferent openings to the gill-chambers lie in front of the chelipedes. Male openings on coxae of last pair of legs. Calappa (Fig. 128) circumtropical, 188and extending into the warmer temperate seas. Matuta (Fig. 129) from the Indo-Pacific.

Fig. 130.—Dorsal view of Ilia nucleus, × 1. (From an original drawing prepared for Professor Weldon.)

Fam. 2. Leucosiidae.—Similar to the above, but the afferent openings to the gill-chambers lie at the bases of the third maxillipedes. Male openings on the sternum. This family contains a great number of forms, with headquarters in the tropical littoral, but extending into the temperate seas. Ilia in the European seas. I. nucleus (Fig. 130) common in the Mediterranean. Ebalia in the Atlantic, North Sea, and Indo-Pacific. Leucosia in Indo-Pacific.

Fam. 3. Dorippidae.—Cephalothorax short and square. The abdomen is not hidden under the thorax; the antennae are large, and the last two pairs of legs are held dorsally, and have terminal hooked claws. Dorippe, littoral in Mediterranean and Indo-Pacific. Cymonomus (Fig. 127) from deep-sea of Atlantic and Mediterranean.

Fam. 4. Raninidae.—Similar to Dorippidae, but the cephalothorax is elongated, and the legs usually have the last two joints very broad. Several genera, chiefly in the deeper littoral zone. Ranina dentata in the Indo-Pacific.

Tribe 3. Cyclometopa.

In these Crabs the carapace is circular rather than square; its frontal and lateral margins are produced into spines and there is no pointed rostrum. The mouth is square, and the third maxillipedes are greatly flattened and form a lid-like expansion over the other oral appendages. This group includes the common Shore-crab of our coasts (Carcinus maenas), the swimming Crabs with expanded pereiopods (Portunus, Lupa, etc.), the Edible Crab (Cancer pagurus), and many others.

Corystes cassivelaunus is a Crab of doubtful affinities. It is sometimes placed among the Oxyrhyncha, but, as Gurney^[154] has 189pointed out, the Megalopa shows Portunid characters, and the resemblance to the Oxystomata in the front of the carapace and in the mouth may be secondary. The respiratory arrangement of this Crab has already been mentioned in comparing its structure with that of the Mole-crab Albunea. The form of the antennal tube can be gathered from the figure of the Megalopa stage (Fig. 125, p. 183). It should be noted that when the Crab is buried in the sand with only the tip of the antennal tube projecting, the water is sucked down and enters the branchial cavities anteriorly, the antennal tube being continued by a tube formed from the third maxillipedes and the forehead; the water is exhaled at the sides of the branchial cavities beneath the branchiostegites. Thus in Corystes the normal direction of the current is reversed, but when the Crab is not buried, and is moving over the surface, it breathes in the usual manner, taking in the water at the sides of the branchiostegites and exhaling it anteriorly by the tube. The related Atelecyclus, found like Corystes very commonly at Plymouth, uses two methods of breathing: when it is in the surface-layers of sand it makes use of its antennal tube, which is, however, much shorter than in Corystes; but when it burrows deeper, where the antennal tube is no use, it folds its chelipedes and also its other legs, which are densely covered with bristles, so as to form a reservoir of pure water underneath it free from sand, which it passes through the gill-chambers in the usual manner (see Garstang, loc. cit. p. 186).

The respiratory adaptations in Lupa hastata and their convergence towards those of the Oxystomatous Matuta have been already touched upon (pp. 186, 187).

In this connexion must be mentioned the interesting experiments of W. F. R. Weldon^[155] upon the respiratory functions of Carcinus maenas at Plymouth, since these were the first noteworthy observations directed towards the exact measurement of the action of natural selection upon any animal, a field of observation in which Weldon will always be looked upon as a pioneer. An extended series of measurements by Weldon and Thompson on male specimens of Carcinus maenas of various sizes between the years 1893 and 1898 showed a steady decrease in the ratio of carapace breadth to length; the Crabs appeared to be becoming steadily narrower across the frontal margin, and 190the same thing, though not to the same extent, was happening in female Crabs. Weldon supposed that this change might be correlated with the silting up of Plymouth Sound and the consequent fouling of the water. To test this hypothesis he kept a very large number of male Crabs in water to which fine porcelain clay was added and kept in continual motion. In the course of the experiments the survivors and the dead were measured, and it was found that the mean carapace breadth of the survivors was less than that of those that succumbed. The experiment was repeated with the fine sand that is deposited and left at low water upon the stones on Plymouth beach, and the same result was observed. It was also noticed that the individuals which died had their gills clogged with the sand, while those that survived had not. As a further confirmation, a great many young male Crabs were isolated and kept in pure filtered water, and they were measured before and after moulting; these measurements, when compared with measurements of the frontal breadth in Crabs of the same size taken at random upon the beach, were found to show a greater breadth than the wild Crabs, thus indicating that a selection of narrow Crabs was taking place in Nature which did not take place when the Crabs were protected from the effects of fine sand in the water.

The whole chain of evidence goes to show that the carapace breadth in Carcinus maenas in Plymouth Sound is being influenced by the rapid change of conditions occurring in the locality. Various objections have been urged against this conclusion, but, though they merit further investigation, they do not appear very weighty.

The fresh-water Crab, Thelphusa fluviatilis, common in the South of Europe and on the North coast of Africa, belongs to the Cyclometopa, and is interesting from its direct mode of development without metamorphosis.

Fam. 1. Corystidae.—The orbits are formed, but, unlike all the other families of the Cyclometopa, are incomplete. The body is elongate and oval, and the rostrum and front edge of the mouth rather as in the Oxyrhyncha, in which Tribe they are sometimes included. Corystes, with a few species in European seas. C. cassivelaunus at Plymouth.

Fam. 2. Atelecyclidae.—Perhaps related to the foregoing. 191The carapace is sub-circular, and the rostrum short and toothed. Atelecyclus, European seas.

Fam. 3. Cancridae.—The carapace is broadly oval or hexagonal, and the flagella of the second antennae are short and not hairy as in the foregoing. The first antennae fold lengthwise. Carcinus maenas on English and North European coasts. This crab has become naturalised in some unexplained manner in Port Phillip, Melbourne. Cancer in North Atlantic, North Pacific, and along the west coast of America into the Antarctic regions. C. pagurus is the British Edible Crab.

Fig. 131.—Dorsal view of Lupa hastata, × 1. (From an original drawing prepared for Professor Weldon.)

Fam. 4. Portunidae.—The legs are flattened and adapted for swimming. The first antennae fold back transversely. Portunus, Atlantic and Mediterranean. Neptunus, Indo-Pacific. Callinectes, C. sapidus, the edible blue Crab of the Atlantic coasts of America. Lupa (Fig. 131).

Fam. 5. Xanthidae.—The first antennae fold transversely, but the legs are not adapted for swimming; the body is usually transversely oval. This family is especially characteristic of the tropical littoral, where it is very widely represented. Xantho, Actaea, Chlorodius, Pilumnus, Eriphia, with E. spinifrons, common in the Mediterranean.

Fam. 6. Thelphusidae (Potamonidae).—Fresh-water crabs, with the branchial region very much swollen. Thelphusa (or Potamon) has nearly a hundred species distributed from North Australia, through Asia, Japan, the Mediterranean region, and throughout Africa. Potamocarcinus in tropical America.

Tribe 4. Oxyrhyncha.

This section includes the Spider-crabs and related genera, in which the carapace is triangular, with the apex in front 192formed by a sharply-pointed rostrum. There are two chief series, the one comprising the Spider-crabs, with much elongated walking legs, e.g. the huge Maia squinado of European seas, the yet more enormous Macrocheira kämpferi from Japan, supposed to be the largest Crustacean in existence, and sometimes spanning from outstretched chela to chela as much as eleven feet, and the smaller forms, such as Inachus, Hyas, and Stenorhynchus, which are so common in moderate depths off the English coasts. The other series is represented by genera like Lambrus (Fig. 133), in which the legs are not much elongated, but the chelipedes are enormous.

The Spider-crabs do not burrow, and their respiratory mechanism is simple; but since they are forms that clamber about among weeds, etc., upon the sea-bottom, they often show remarkable protective resemblances to their surroundings, which are not found in the burrowing Cyclometopa. Alcock^[156] gives a good account and figure of Parthenope investigatoris, one of the short-legged Oxyrhyncha, the whole of whose dorsal surface is wonderfully sculptured to resemble a piece of the old corroded coral among which it lives.

But besides this, the long-legged forms, such as Inachus, Hyas, etc., have the habit of planting out Zoophytes, Sponges, and Algae upon their spiny carapaces, so that they literally become part and parcel of the organic surroundings among which they live. It may, perhaps, be wondered what are the enemies which these armoured Crustacea fear. Predaceous fish, such as the Cod, devour large quantities of Crabs, which are often found in their stomachs; and Octopuses of all sorts live specially upon Crabs, which they first of all paralyse by injecting them with the secretion of poison-glands situated in their mouth. The poison has been recently found by Dr. Martin Henze at Naples to be an alkaloid, minute quantities of which, when injected into a Crab, completely paralyse it. When the Crab is rendered helpless the Octopus cuts out a hole in the carapace with its beak, and sucks all the internal organs, and then leaves the empty shell.

Many of the Oxyrhyncha are found in the abysses; among them are Encephaloides armstrongi (Fig. 132), dredged by Alcock from below the 100–fathom line in the Indian Ocean, which has 193the gill-chambers (G) greatly swollen and enlarged to make up for the scarcity of oxygen in these deep regions.

Fig. 132.—Encephaloides armstrongi, × 1. The long walking legs are omitted. C, Great chela; G, one of the greatly swollen gill-chambers. (After Alcock.)

Fam. 1. Maiidae.—The chelipedes are not much larger than the other legs, but are very mobile. Orbits incomplete. A very large family, including all the true Spider-crabs, very common in the Atlantic and Mediterranean littoral. Inachus, Pisa, Hyas, Stenorhynchus, Maia, Encephaloides (Fig. 132).

Fam. 2. Parthenopidae.—The chelipedes are much larger than the other legs. Orbits complete. Lambrus (Fig. 133), Parthenope.

Fig. 133.—Lambrus miersi, × 1. (After Milne Edwards and Bouvier.)

Fam. 3. Hymenosomatidae. The carapace is thin and flat; the chelipedes are neither very long nor especially mobile. There are no orbits, and the male openings are on the sternum. Characteristic of the Antarctic seas. Hymenosoma, Trigonoplax.

Tribe 5. Catometopa.

These Crabs resemble the Cyclometopa in general appearance, but the carapace is very square in outline, and its margins are never so well provided with spines as in the Cyclometopa. The position of the male genital openings is peculiar, since they lie upon the sternum, and are connected with the copulatory appendages upon the abdomen by means of furrows excavated in the sternum. The Catometopa are either littoral or shallow water forms, or else they live entirely on land. The Grapsidae are marine Crabs, Pachygrapsus marmoratus (Fig. 134) at Naples being exceedingly common on rocks at high-water mark, over which it scuttles at a great rate; in the Mediterranean it takes 194the place of our common Garcinus maenas, which is not found there.

Fig. 134.—Dorsal view of Pachygrapsus marmoratus, × ⅓. (From an original drawing prepared for Professor Weldon.)

Among the land genera are Ocypoda, Gelasimus, and Gecarcinus of tropical lagoons and coastal swamps. Ocypoda often occurs in vast crowds in these regions, and digs burrows in the sand.

Fig. 135.—Gelasimus annulipes, × 1. A, Female; B, male. (After Alcock.)

Gelasimus (Fig. 135) is remarkable for the enormous size of one of the chelipedes, generally the right, in the male, which may actually exceed in size the rest of the body. It is not known what purpose this organ serves in the various species. In Gelasimus it is supposed that the male stops up the mouth of the burrow with it when he and the female are safely inside. It is also used as a weapon in sexual combats with other males; but Alcock, from observations made in the Indian Ocean, believes that the males use it for exciting the admiration of the females in courtship, as the huge chela is bright red in colour, and the males brandish it about before the females as if displaying its florid beauty.

The species of Ocypoda are exclusively terrestrial, and cannot live for a day in water. The gills have entirely disappeared, and the branchial chambers are converted into air-breathing lungs with highly vascular walls, the entrances into which are situated as round holes between the bases of the third and fourth pairs of walking legs. As their name implies, they can run with 195astonishing rapidity, and they seem to be always on the alert, directing their eyes, which are placed on exceedingly long stalks, in all directions.

Some of the Grapsidae, e.g. Aratus pisonii, are partially adapted for life on land. Fritz Müller, in his Facts for Darwin, alludes to this creature as “a charming lively crab which ascends mangrove bushes and gnaws their leaves.” The carapace can be elevated and depressed posteriorly, apparently by means of a membranous sac, which can be inflated by the body-fluids. This Crab retains its gills and can breathe under water in the ordinary way.

A great many other Catometopa are land-crabs; but we may specially mention the genus Gecarcinus, related to the marine Grapsidae, which has representatives in the West Indies and West Africa. The Crabs of this genus may live in sheltered situations several miles from the sea, but in spring the whole adult population rushes down in immense troops to the shore, where breeding and spawning take place; and when this is completed they migrate back again to the land. The young pass through the normal larval stages in the sea and then migrate inland.^[157]

Fam. 1. Carcinoplacidae.—The carapace is rounded and broader than long, usually with toothed front margin. The orbits and eyes are normal, and not much enlarged. Geryon, in the deep littoral of the northern hemisphere. Euryplax, Panoplax, etc., in the American coastal waters. Typhlocarcinus, etc., in the Indo-Pacific.

Fam. 2. Gonoplacidae.—The carapace is square, with the antero-lateral corners produced into spines. The orbits are transversely widened, and the eye-stalks long. Gonoplax, widely distributed in the littoral zone. G. rhomboides in British and European seas.

Fam. 3. Pinnotheridae.—Carapace round, with indistinct frontal margin. Orbits and eyes very small, often rudimentary. The members of this family live symbiotically or parasitically in the shells of living Bivalve Molluscs, corals, and wormtubes in all seas except the Arctic. Pinnotheres pisum is fairly 196commonly met with off the English coasts in the mantle-cavity of Cardium norwegicum.

Fam. 4. Grapsidae.^[158]—Carapace square, the lateral margins either strictly parallel or slightly arched. The orbits and eyes are moderately large, but the eye-stalks are not much lengthened. Littoral, fresh-water, and land. Pachygrapsus marmoratus (Fig. 134), the common shore-crab of the Mediterranean. Sesarma, with fresh-water and land representatives in the tropics of both hemispheres. Cyclograpsus, marine in the tropical littoral.

Fam. 5. Gecarcinidae.—Carapace square, but much swollen in the branchial region. Orbits and eyes moderately large. Typically land forms, which only occasionally visit the sea or fresh water. Cardisoma is a completely circumtropical genus, with species in tropical America, West and East Africa, and throughout the Indo-Pacific. Gecarcinus in West Indies and West Africa.

Fam. 6. Ocypodidae.—Carapace square or rounded, generally without teeth on the lateral margins. The orbits transversely lengthened, eye-stalks usually very long. The members of this family generally inhabit the mud-flats and sands of tropical coasts; in the southern hemisphere they extend far into the temperate regions. Macrophthalmus, with numerous species, in Indo-Pacific. Gelasimus (Fig. 135), in the tropics of both hemispheres. Ocypoda, with similar distribution.

A. Marine.

The great majority of the Crustacea are inhabitants of the sea. From a Zoogeographical point of view we divide the sea into three chief regions, each of which is characterised by a special kind of fauna—the littoral, the pelagic, and the abyssal regions.

The littoral region, which comprises all the shallow coastal waters down to about 100 fathoms, varies very greatly in its physical character according to the nature of the coast, its geological constitution, latitude, etc., but, on the whole, it is characterised by variability of temperature and salinity, by the presence of sunlight, and by the continuous motion of its waves. On the shores of the large oceans this region is also greatly affected by the tides. It is inhabited by a vast assemblage of Crustacea, all of which are dependent upon a solid substratum, either of rock or sand, or of vegetable or animal growth, upon which they may wander in search of food, or in which they may hide themselves. In consequence, the character of the Crustacea on any shore is largely determined by its geological nature.

Although a certain number of Entomostraca such as Copepoda (Harpacticidae and Cyclopidae), Ostracoda (Cypridae and Cytheridae), and a few Operculata are littoral in habit, it is the Malacostraca, from their larger size and variety of form, which give the character to coastal waters.

On rocky coasts, especially those affected by tides, a great many kinds of Shore-crab are found, which hide at low tide in 198the rock-pools and under stones. Carcinus maenas is characteristic of the rocky coasts of the North Sea, while it is replaced in warmer seas and all round the tropics by Crabs of the family Grapsidae, which are typical rock-livers, and exceedingly agile in clambering over tide-washed rocks. Porcellanidae are also very common under stones at low tide on rocky beaches. Such typical Shore-crabs as these are remarkably resistant to desiccation, and can live out of water for an astonishing time; nor do they require a change of water provided they have access to the air. The edible crab (Cancer pagurus) and the lobsters (Homarus and Palinurus) are dependent on rocks, but they rarely come close in-shore, preferring depths of a few fathoms.

Sandy coasts are preferred by Shrimps and Prawns, which haunt the shallow coastal waters in shoals; and in the sand are found all the Crabs whose respiratory mechanism is specially adapted for life in these regions, e.g. Hippidea or Mole-crabs, Corystes, Matuta, Calappa, etc.

Characteristic of sandy bottoms are also the Thalassinidea, such as Callianassa, which excavate galleries in the sand. On tropical sandy shores various species of Ocypoda and Gelasimus are conspicuous, which have deserted the sea, and live in burrows which they excavate on the shore. Gelasimus is especially abundant in the muddy sand of tropical mangrove swamps.

Besides the rocky and sandy coasts we must distinguish the muddy shores and bottoms which support a large amount of vegetable and animal growth. These, besides harbouring the greater number of Amphipods and Isopods, are also the natural home of the Dromiacea and Oxyrhyncha, or Spider-crabs, among which the habit is common of decking themselves out with pieces of weed or animal growth in order to harmonise better with their surroundings. Pagurids are also especially abundant in the deeper waters of these coasts.

Coral-reefs support a characteristic Crustacean fauna. In the growing coral at the reef-edge a number of small Cyclometopa are found, e.g. Chlorodius, Actaea, Xantho, which are finely sculptured and often coloured so as to harmonise with the coral. Alpheidae also, Shrimp-like Macrura with highly asymmetrical claws, which can emit a sharp cracking sound with the larger claw, are commonly found in pools on the reef. In the coralshingle 199formed by abrasion from the reef-edge at a few fathoms depth, Leucosiidae are found, in which, again, respiratory mechanisms for filtering sand from the gills are present.

Besides the geological nature of the coast, latitude has a very important bearing upon the distribution of littoral Crustacea. Indeed, the present distribution of littoral Crustacea appears to be far more determined by the temperature of the coastal waters than by the presence of any land-barriers, however formidable. We may distinguish an Arctic, Antarctic, and Circumtropical zone.

The Arctic zone includes the true Arctic seas, and stretches right down through boreal regions towards the sub-tropical seas. Almost all the truly Arctic forms penetrate fairly far south, the Arctic seas being characterised more by the absence of temperate forms than by the presence of forms peculiar to itself. At the same time it must be noted that the individuals from the coldest regions often grow to an enormous size, a characteristic which is physiologically unexplained.

A great many of the Crustacea characteristic of this region are circumpolar, i.e. they are not restricted in range to either the Atlantic or Pacific. This is especially true of the extremely northern types, e.g. Crangonidae and Hippolytidae, but it is also true of a number of Crustacea which do not now occur as far north as Greenland or Bering Strait, so that there is no longer any free communication for them between Pacific and Atlantic. This gives rise to a discontinuous distribution in the two oceans, exemplified in the common Shrimp, Crangon vulgaris, which is found on the temperate European coasts and on the Pacific coasts of Japan and Eastern America. The same is true of Eupagurus pubescens and E. bernhardus.

At the same time the boreal Atlantic and Pacific have their peculiar forms. Thus the European and American Lobsters are confined to the Atlantic, while the North Pacific possesses a very rich array of Lithodinae, which cannot be paralleled in the Atlantic.

We may explain the community of many littoral forms to both the North Atlantic and Pacific coasts by the continuous coast-line uniting them, which in former times possibly did not lie so far north, or else was not subjected to so rigorous a climate as now.

200In the Antarctic zone we are presented with very different relations, since the great continents are drawn out to points towards the south, and are isolated by vast tracts of intervening deep sea. Nevertheless, certain littoral forms are circumpolar, e.g. the Palinurid Iasus and the Crabs Cyclograpsus and Hymenosoma. The genus Dromidia is common to Australia and South Africa, though it is apparently absent from South America.

The Isopod genus Serolis is confined to Antarctic seas. The majority are littoral species, and they are distributed round the coasts of Patagonia, Australia, and Kerguelen in a manner that certainly suggests a closer connection between these shores in the past. These facts are, on the whole, evidence in favour of the former existence of an Antarctic continent stretching farther north and connecting Australia, Africa, and S. America—a supposition that has been put forward to account for the distribution of the Penguins, Struthious birds, Oligochaets, Crayfishes, etc., in these regions (see pp. 215–217).

In considering the Arctic and Antarctic faunas the supposed phenomenon of bipolarity must be mentioned, i.e. the occurrence of particular species in Arctic and Antarctic seas, but not in the intermediate regions. This discontinuous type of distribution was upheld for a variety of marine animals by Pfeffer, Murray, and others, but it has been very adversely criticised by Ortmann.^[159] As far as the Arctic and Antarctic Decapod fauna in general are concerned, the north polar forms are quite distinct from the south polar. Typical of the former are Hippolyte, Sclerocrangon, Hyas, Homarus, etc.; of the latter, Hymenosoma, Dromidia, Iasus. It appears, however, that in certain special cases, bipolarity of distribution may be produced owing to the operation of peculiar causes. Two such cases seem to be fairly well established. Crangon antarcticus occurs at the two poles, and apparently not in the intermediate regions; but, as Ortmann points out, it is represented right down the West American coast by a very closely related form, C. franciscorum. The waters on the tropical western coasts both of Africa and America are exceedingly cool, and it appears that in this way the Crangon may have migrated across the tropical belt, leaving a slightly modified race to represent it in this intermediate region. The other case of bipolarity is afforded by the “Schizopod,” Boreomysis 201scyphops, which occurs at both poles, but is not known from the tropics. This is a pelagic species, and we know that the Mysidae often descend to considerable depths. We also know that the Mysidae are dependent on cold water, only occurring in boreal or temperate waters. We may safely suppose, therefore, that the migration of this species has taken place by their forsaking the surface-waters as the tropics were approached, and passing down into the depths where the temperature is constantly low even in the tropics.

The dependence of Crustacea upon the temperature of the water is also illustrated by the distribution of the Lithodinae. The headquarters of this family are in the boreal Pacific, with a few scattered representatives in the boreal Atlantic. The cool currents on the western coasts of America, however, have permitted certain forms to migrate as far south as Patagonia, where they still have a littoral habit. In the tropical Indo-Pacific, where a few species occur, they are only found in deep waters. Thus at these various latitudes, by following cool currents or migrating into deep water, they are always subjected to similar conditions of temperature. The same kind of thing is observed in Arctic seas, where deep-sea forms are apt to take on secondarily a littoral habit owing to the temperature of the depths and of the shore being the same.

Despite the impassable barriers of land which now sever the tropical oceans, we can yet speak of a circumtropical zone possessing many species common to its most widely separated parts. Such circumtropical species, occurring on both the Atlantic and Pacific coasts of tropical America, on the West African coast, and in the Indo-Pacific, are various Grapsidae, Calappa granulata and its allies, and certain Albunea. The most striking instance of all is that of the Land-crabs. Of Ocypoda, the greater number of species occur in the Indo-Pacific, but representatives are also found on the tropical Eastern and Western American coasts and on the West African coast, and the same is true of Gelasimus. The genus Cardisoma, belonging to a different group of Land-crabs, is also typically circumtropical.

For this community of the circumtropical species we may certainly advance in explanation the comparatively recent formation of the Isthmus of Panama. Besides the resemblance of the Crustacea on the east and west coasts of the isthmus, we have an 202actual identity of species in several cases, e.g. Pachycheles panamensis and Hippa emerita, and the same thing has been observed for the marine fish.

Another connexion, at any rate during early tertiary times, which probably existed between now isolated tropical coasts, was across the Atlantic from the West Indies to the Mediterranean and West African coasts. Numerous facts speak for this connexion. Species of Palinurus and Dromia occur in the West Indies and the Mediterranean, which only differ from one another in detail, and a connexion between these two regions has been urged from the minute resemblances of the late Cretaceous Corals of the West Indies with those of the Gosau beds of S. Europe, and also of the Miocene land-molluscs of S. Europe with those at the present time found in the West Indies.

To account, then, for the present distribution of littoral Crustacea we must imagine that great changes have taken place during comparatively recent times in the coast-lines of the ocean, but the guiding principle in both the past and present has been temperature, and this factor enables us, despite the immense changes in the configuration of the globe that must have taken place, to divide the coasts latitudinally into Arctic, Antarctic, and Circumtropical zones.

Pelagic Crustacea belong chiefly to the Copepoda (Calanidae, Centropagidae, Candacidae, Pontellidae, Corycaeidae), a few Ostracoda (Halocypridae and Cypridinae), and among Malacostraca a few Amphipoda (Hyperina), some “Schizopoda,” and among Decapoda only the Sergestidae, if we except the few special forms which live on the floating weeds of the Sargasso Sea, e.g. the Prawns Virbius acuminatus and Latreutes ensiferus, and the Brachyura Neptunus sayi and Planes minutus. Besides these Crustacea which are pelagic as adults, there is an enormous host of larval forms, both among Entomostraca and Malacostraca, which are taken in the surface-plankton.

In dealing with the Copepoda we have already mentioned the vast pelagic shoals of these organisms which occur at particular times of the year, and have an important influence on fishing industries. Anomalocera pattersoni (Fig. 27, p. 60) is a good instance of this. It is a large Heterarthrandrian, about 3 mm. long, with the body of a fine bluish green colour; it has a remarkable power of springing out of the water, so that a shoal has the 203appearance of fine rain upon the surface of the sea. It occurs in the open Atlantic and Mediterranean, but comes into the coasts during violent storms; the Norwegian fishermen hail its presence in the fjords as the sign of the approach of the summer herring.

It was Haeckel^[160] who first clearly distinguished between “neritic” plankton, the species of which have their centres of distribution in shallow coastal waters and die out gradually as the open ocean is approached, and “oceanic” plankton which is habitually found in the open sea, and though it may invade the coasts is not dependent on the sea-bottom in any way. It appears that although these two kinds of plankton may get mixed up by currents and storms, they are always recruited by new generations from the neritic or oceanic stations proper to each kind.

Common oceanic species, found chiefly in the open Atlantic and in the North Sea, are Anomalocera pattersoni, Calanus finmarchicus, Centropages typicus, Metridia lucens, Oithona plumifera, etc. Common neritic species in the Channel and other coastal waters are Centropages hamatus, Euterpe acutifrons, Oithona nana, Temora longicornis, etc. It was found by Gough^[161] that although the true oceanic species invade the Channel from the open Atlantic to the west, they become rarer and rarer as they advance up the Channel. Thus the plankton midway between the Lizard and Ushant at all times of year is about 70 per cent. oceanic, while at the line drawn from Portland to the Cap de la Hague it is about 35 per cent. Seasonal changes in the salinity of the Channel water, chiefly due to the influx of oceanic water from the Atlantic, as observed by Matthews,^[162] do not clearly influence the distribution of oceanic and neritic forms. The influx of highly saline water from the Atlantic was most marked during the winter months up to February. From February to May the highly saline water receded, and during the summer months at the line drawn between Portland and the Cap de la Hague the salinity was rather low. This was increased in November by a patch of oceanic water being cut off from the main mass and passing up Channel, and it is noteworthy that during this month 204the highest percentage of oceanic forms was taken in the plankton of this region.

Calanus finmarchicus affords a clear instance of the way in which the plankton may be carried about for great distances by means of currents. This species has its home in the subarctic seas, but is carried down in the spring by the East Icelandic Polar stream to its spawning-place south of Iceland; the enormous shoals produced here are carried back, continually multiplying, along the coasts of Norway during the summer and autumn.

Besides these great migrations, the plankton organisms perform daily movements, the majority of the Crustacea avoiding the surface during the day, and often going down to as much as seventy fathoms or more, and only coming up to the surface at night. Others, however, e.g. Calanus finmarchicus, behave in the converse manner, preferring the sunlit surface to swim in.

Owing to their dispersal by means of oceanic currents the pelagic Crustacea do not offer any very striking features in regard to their distribution, and the possibility of always finding congenial temperatures by passing into the upper or under strata of water enables them to live in almost all seas. The tropical species of Sergestidae are mostly circumtropical, i.e. unhindered by the present barriers of land.

The Abyssal regions of the sea contain many of the most interesting Crustacea. Families entirely confined to the abyss are the Eryonidae, Pylochelidae, and certain Caridean Prawns (Psalidopodidae, etc.), but there are a great number of normally littoral genera which have representatives in deep water. If we draw the limit between the littoral and abyssal regions at about 200 metres, we can characterise the latter as absolutely dark except for the presence of phosphorescent organisms, with the temperature at a little above zero, and with a comparative lack of dissolved oxygen in the water. These conditions bring about remarkable modifications in the structure and life-histories of the inhabitants of the deep sea; we have already touched on the modifications of the visual organs and on the presence of phosphorescence in many of the animals; other points to be noticed are the usually uniform yellowish or bright red coloration, the frequent delicacy of the tissues without much calcification, variations in the structure of the breathing organs, e.g. in 205Bathynomus giganteus and Encephaloides armstrongi, and the loss of the larval development. Owing to the similarity of conditions in the deep sea all over the globe most of its inhabitants are universally distributed. It is also a striking fact that species are found in the deep sea of the tropics whose nearest allies occur, not in the littoral seas of the tropics, but in those of the temperate region. This fact has already been alluded to in dealing with the distribution of the Lithodinae. Alcock^[163] remarks that between 50–500 fathoms in the Indian Ocean are found Crabs such as Maia, Latreillia, and Homola, regarded as characteristic of the north temperate seas; the lobster Nephrops andamanica, taken at 150–400 fathoms, is closely allied to the Norwegian N. norwegica; and nine species of “Schizopoda,” which are certainly temperate forms, occur in the Indian Ocean at depths of 500–1750 fathoms.

B. Fresh-Water.^[164]

If we except the Crayfishes and River-crabs, the Crustacean fauna of running water is exceedingly poor, but in all standing fresh-water, from the smallest pond to the large lakes and inland seas, Crustacea, especially Entomostraca, are abundant and characteristic, and form an important item in the food of fresh-water fishes. In small ponds a vast assemblage of Cladocera is met with; these animals multiply with great rapidity by parthenogenesis, especially during spring and summer, but on the advent of untoward conditions sexual individuals are produced, which lay fertilised winter-eggs which lie dormant until favourable conditions again arise. As Weismann first pointed out, the frequency with which sexual individuals are produced in the various species is closely correlated with the liability of the water in which they live to dry up; so that the Cladocera which inhabit small ponds usually have at least two “epidemics” of sexual individuals, one during early summer and the other before the onset of winter.

Besides Cladocera, the Phyllopoda (e.g. Apus, Artemia, etc.) 206inhabit small pools; and also a great number of Cyclopidae. Of the other fresh-water families of Copepoda, viz. Centropagidae and Harpacticidae, inhabitants of small pieces of water are Diaptomus castor, as opposed to the other species of Diaptomus which are pelagic, and a number of Harpacticidae (Canthocamptus), the members of this family living in the weed or mud of either small ponds or else on the shores of the larger lakes. The greater number of Ostracoda are found in similar situations.

A district like the Broads of Norfolk, which consists partly of slowly-moving streams and partly of extensive stretches of shallow water, supports a Crustacean fauna intermediate in character between that found in small ponds and the truly pelagic fauna characteristic of deep lakes. A very complete list of the Crustacea of the Norfolk Broads, with an interesting commentary on their distribution, is given by Mr. Robert Gurney.^[165] We miss here the pelagic Cladocera, such as Leptodora, Bythotrephes, Holopedium, etc., which form so characteristic a feature of large lakes; at the same time, besides a rich development of the Cladocera, Cyclopidae, and Harpacticidae, which haunt the weeds and mud of shallow waters, we find such species as Polyphemus pediculus and Bosmina longirostris among Cladocera, which are otherwise confined to large bodies of water, and a few pelagic Diaptomus, e.g. D. gracilis. The fauna is also complicated in this district by the proximity to the sea and the frequently high salinity of the water, which allows a number of typically marine Copepods to pass up the estuaries and intermingle with typically fresh-water species; such are Eurytemora affinis among the Centropagidae, and several species of Harpacticidae (see p. 62).

The large lakes of the world, such as the continental lakes of Europe and America, or of our own Lake District, reproduce on a small scale the varied conditions which appertain to the ocean—as in the ocean, we can recognise in these lakes a littoral, a pelagic, and an abyssal region. Our knowledge of the physiography of lakes is largely due to the classical work of Forel,^[166] and the following account of the physical conditions in the various regions is condensed from his book.

The littoral region is sharply marked off from the others by the relative instability of its physical conditions, owing to the 207agitation of its waters, the affluence of streams and drainage, and the constant changes of temperature. The water in this region generally contains a good deal of solid matter in suspension, while the shelving banks of the lake support a wealth of vegetable growth, both of Algae and of Phanerogams, down to about 20–25 metres. At this depth the daylight does not penetrate sufficiently to admit of the growth of green plants, so that this region marks the limit, both physical and biological, between the littoral and the abyssal zones. In this littoral region there flourish a great quantity of Entomostraca, most of which are also found in small ponds where similar conditions of life prevail—the pelagic species only penetrating rarely, and by accident, into its waters. At the beginning of July Mr. H. O. S. Gibson and myself found that the weedy littoral region of Grasmere contained almost entirely large quantities of the Cladoceran Eurycercus lamellatus, and a number of Cyclops fuscus and C. strenuus. In the littoral zone of large lakes, Amphipods, Isopods, and fresh-water shrimps may also be met with, but this applies more to the lakes of the Tropics and of the Southern Hemisphere.

The pelagic^[167] region is distinguished from the littoral by the greater purity and transparency of its waters, and by the relative stability of the temperature, the annual range, even at the surface, in Geneva being from 4°–20° C., while at 100 metres the water has a uniform temperature of 4° or 5° C. The upper strata are, of course, brightly illuminated, but at 20 metres there is hardly sufficient light for green plants to grow, and at 100 metres it is completely dark. The inhabitants of this region, known collectively as plankton, spend their whole life swimming freely in the water, sometimes at the surface and sometimes in the deeper strata. They consist chiefly of Diatoms, Protozoa, Rotifera, and Crustacea. The pelagic Crustacea, especially the Cladocera, are often the most curiously and delicately built creatures. Leptodora hyalina, which is quite transparent, is the largest of them, attaining to three-quarters of an inch in length, though Bythotrephes longimanus is nearly as large if we include the immense spine which terminates the body. Holopedium gibberum, which is 208the commonest of all in Grasmere lake, but not so frequently met with in the other English lakes, is peculiar in that its body is enveloped in a spherical mass of transparent jelly, sometimes a quarter of an inch in diameter, so that the contents of a tow-net jar full of Holopedium have something of the consistency of boiled sago. The enormous quantities in which these animals often occur during summer is very astonishing; but to be truly appreciated tow-nettings should be taken at the surface of the lake either during night-time when there is not much moonlight, or else on a dark still day when there is a quiet drizzle falling on the surface of the water. In bright sunshine the plankton passes below the surface into the lower strata, and can be usually taken by sinking the tow-net some 10–20 feet, or to even greater depths in the water. The exact reason of these periodic migrations out of the light, and their dependence on other physical conditions, such as temperature and the agitation of the water, is not clearly understood. It appears, however, that when the water is rough, plankton always passes into the deeper regions. Besides the species mentioned, the minute Bosminidae, whose trunked heads are suggestive under the microscope of elephants, and Polyphemus pediculus are among the commonest pelagic Cladocera, though neither Polyphemus nor Bythotrephes ever form shoals. The above-mentioned genera are characteristic of the larger lakes in the Northern Hemisphere. Our knowledge of the Crustacean plankton of tropical lakes and of those of the Southern Hemisphere is limited (but see p. 216).

A very important constituent of lake-plankton is furnished by the Copepoda, especially of the genus Diaptomus. With the exception of Holopedium, by far the commonest Crustacean in Grasmere during July was found by Mr. Gibson and myself to be D. caeruleus.

At the same time a number of Cyclopidae, e.g. Cyclops strenuus, may occur in the pelagic region in considerable quantities, though they were never found by us in such numbers as Diaptomus.

The life-cycle of the pelagic Entomostraca has been studied in both the Cladocera and the Copepoda. In some of the Cladocera Weismann at first supposed that males had altogether disappeared, and that reproduction was entirely parthenogenetic. It appears, however, that all the pelagic species have at least one sexual period, namely, in the autumn, when resting eggs are 209produced which lie dormant during the winter. The pelagic Copepods may either produce resting eggs for the winter (Diaptomus), or else the winter is passed through in the Nauplius stage, the larvae hibernating in the mud until the spring (Cyclopidae).

We have so far only dealt with fresh-water Entomostraca. There are, in addition, a number of Malacostraca which inhabit fresh water, and some of these are found in the abyssal region of the great lakes, which must now be considered.

The physical conditions of the abyssal region are still more stable than those of the pelagic region, since the water is never disturbed, the bottom is always composed of a fine mud, the temperature is constant at 4°–5° C., and there is a total absence of light. It was hardly expected that animals would inhabit this region, until Forel discovered Asellus aquaticus in a depth of forty metres in the Lake of Geneva, and subsequently showed that quite a number of animals, including a Hydra, several worms, Molluscs, Crustacea, and larval Insects, may be found in these or even much greater depths.

The Crustacea of the abyssal region are two in number, and these have been found in a number of European lakes; Niphargus puteanus, a blind Amphipod closely allied to Gammarus; and Asellus forelii, allied to A. aquaticus and A. cavaticus, which may be either quite blind or else retain the rudiments of eyes.

These two Crustacea, under a practically identical form, are also found in the subterranean waters of Europe, and Forel considers that they have arrived in the abysses of the lakes from the subterranean channels, and are not derivatives of the littoral fauna.^[168]

Having completed our short review of lacustrine Crustacea, we may deal with the subterranean and cave Crustacea,^[169] which, as far as light and temperature are concerned, are subjected to very similar conditions to those dwelling at the bottom of deep lakes. The inhabitants of the subterranean waters have been chiefly brought to light in Artesian wells, etc., while the cavedwellers 210have been investigated especially at Carniola and in the American caves.

A number of species of Cyclopidae and Cypridae, many of which are blind and colourless, have been brought up in well-water. The Amphipod Niphargus puteanus has long been known from a similar source in England^[170] and all over Europe, and several other blind Gammarids inhabit the subterranean waters and caves in various parts of the world. Among Isopods, Asellus cavaticus is recorded from wells and caves in various parts of Europe, Caecidotea stygia and C. nickajackensis from the Mammoth and Nickajack Caves in America, and two very remarkable blind Isopods are described by Chilton from the subterranean waters of New Zealand, viz. Cruregens fontanus, whose nearest allies are the marine Anthuridae, and the Isopods Phreatoicus typicus and P. assimilis, which bear an extraordinary resemblance superficially to Amphipods. Besides these, a small number of subterranean Decapoda are known which retain the eye-stalks but are without functional ommatidia. These are Troglocaris schmidtii, in Hungary, related to the fresh-water Atyid Xiphocaris of East Indian and East Asiatic fresh waters rather than to the South European Atyephyra; Palaemonetes antrorum, from artesian wells in Texas; and several species of Cambarus from the Eastern United States. A blind species of Cambarus, C. stygius, has been described from the caves of Carniola, and if this determination is correct, is the sole Cambarus occurring outside America.

It will be seen from the above account that the subterranean Crustacea are an exceedingly interesting and, in many respects, archaic group, many of which have survived in these isolated and probably uncompetitive districts, while many secular changes were going on in the quick world overhead.

The remaining fresh water Malacostraca may be mentioned under the headings of the groups to which they belong.

Only one “Schizopod,” apart from Paranaspides, is known from fresh-water lakes, viz. Mysis relicta, which was discovered in 1861 by Lovén in the Scandinavian lakes, and has since been found in the Finnish lakes, the Caspian Sea, Lake Michigan, and other localities in N. America, and Lough Erne in Ireland. This species is closely related to Mysis oculata of Greenlandic seas.

211In the Southern Hemisphere we have a species of Anaspides, A. tasmaniae, occurring in mountain streams and tarns in Tasmania, a related form which haunts the littoral zone of the Great Lake in Tasmania, and a small species, Koonunga cursor, occurs in a little stream near Melbourne.

Of the Isopoda certain genera, viz. Asellus and Monolistra, are confined to fresh water, others, such as Sphaeroma, Idothea, Alitropus, and Cymothoa, have occasional fresh-water representatives. Packard^[171] describes a remarkable blind Isopod, Caecidotea, from the Mammoth Cave of Kentucky, which occupies a very isolated position, and in the same work gives a very complete exposition of the cave-fauna of North America and Europe.

The Phreatoicidae are a curious family of Isopods confined to the fresh waters of Australia and New Zealand, which bear a remarkable resemblance to Amphipods, being laterally compressed and possessing a subchelate hand on the anterior thoracic leg. These Isopods are exceedingly common in small mountain pools and in streams in Tasmania, and in the Great Lake in that country I have recently found a number of species which, together with some species of Amphipods, make up the dominant feature in the Crustacean fauna. One of these species may grow to fully an inch in length. The family is confined to the temperate regions, and is usually found on mountains. A number of species are known from the mainland of Australia, one coming from a high elevation on Mount Kosciusko, and another (Phreatoicopsis) from the forests of Gippsland attaining a great size, and living among damp leaves, etc.

The fresh-water Amphipoda all belong to the families Talitridae, Gammaridae, and Haustoriidae (see p. 137).

Among the Talitridae, or Sand-hoppers, Orchestia and Talitrus have marine as well as fresh-water and land representatives, while the American Hyalella is entirely from fresh water, most of the species being peculiar to Lake Titicaca. Many of these animals are partly emancipated from an aquatic life. Thus Orchestia gammarellus, which is common on the sea-shore of the Mediterranean, frequently penetrates far inland, and was found in large numbers by Kotschy near a spring 4000 feet up on Mount Olympus.

212Talitrus sylvaticus is very common among fallen leaves and decaying timber in Tasmania and Southern Australia, many miles from the sea, and often at an elevation of several thousand feet.

Among the Gammaridae, certain genera, e.g. Macrohectopus (Constantia), from Lake Baikal, are purely fresh-water. An enormous development of Gammaridae was discovered by Dybowsky in Lake Baikal, comprising 116 species, and lately a number more have been found by Korotneff.^[172] The majority of these were originally placed in the genus Gammarus, but Stebbing has rightly created a number of peculiar genera for them. Certain species are, however, placed in more widely distributed genera, e.g. Gammarus and Carinogammarus, which is also represented in the Caspian Sea. Korotneff found some remarkable transparent pelagic forms (Constantia) swimming in the abyssal regions at about 600 metres depth, the majority of them being blind, but some possessing rudimentary eyes, often on one side only.

Besides various species of Gammarus, a number of other Gammaridae are frequently found in brackish water. Among Haustoriidae Pontoporeia has representatives in both the oceans and inland lakes of the northern hemispheres (see p. 137).

Of the Decapoda, seven families are typically fresh-water in habitat—the Aegleidae, containing the single species Aeglea laevis, related to the Galatheidae, which inhabits streams in temperate S. America; the Atyidae, a family of Prawns from the tropical rivers and lakes of the New and Old World, and in the Mediterranean region. A number of Palaemonidae are found in fresh water, e.g. Palaemonetes varians in Europe and N. America, while several species of Palaemon occur in lakes, streams, and estuaries of the tropical Old and New World.

The expeditions of Moore and Cunnington to Lake Tanganyika brought back an exceedingly rich collection of Prawns, comprising twelve species, all of which are peculiar to the lake,^[173] and this is all the more surprising since Lakes Nyasa and Victoria Nyanza are only known to contain one species, Caridina nilotica, which ranges all over Africa and into Queensland and New Caledonia. The Tanganyika species, however, all belong to purely fresh-water genera, and do not afford any suggestion 213that they are part of a relict marine fauna. It would appear that they have been differentiated in the lake itself during a long period of isolation.

Two groups of Brachyura, viz. the Thelphusidae and the Sesarminae (a sub-family of the Grapsidae), are fresh-water. Thelphusa fluviatilis is an inhabitant of North Africa, and penetrates into the temperate regions of the Mediterranean, and is said to be exceedingly common in the Alban Lake near Rome. Both these families have representatives on land, e.g. Potamocarcinus in Central and South America, and certain species of Sesarma, and the closely related Gecarcinidae of the West Indies.

The remaining families to be dealt with are the two Crayfish families—the Astacidae and the Parastacidae—which live in rapidly moving rivers and streams, and occasionally in lakes. A few species of both families have taken to a subterranean mode of life, and excavate burrows in the earth, e.g. the Tasmanian Crayfish, Engaeus fossor. The distribution of the Crayfishes has long engaged the attention of naturalists. It was first seriously studied by Huxley,^[174] and has subsequently been followed up, especially in North America, by Faxon^[175] and Ortmann,^[176] but our knowledge of the South American and Australian forms is still very incomplete. The Astacidae inhabit the northern temperate hemisphere, the Parastacidae the southern temperate hemisphere, the tropical belt being practically destitute of Crayfishes. Of the Astacidae the genus Astacus (Potamobius), including the common Crayfishes of Europe, occurs in Europe and in North America west of the Rockies. The genus Cambaroides, which in certain respects approaches Cambarus, is found in Japan and Eastern Asia. The very large genus Cambarus, on the other hand, only occurs in North America east of the Rockies, so that Cambaroides occupies a very isolated position. The occurrence of a Cambarus, C. stygius, in the caves of Carniola, has been recorded by Joseph, so that it would appear that this genus had a much wider range formerly than now.

Of the southern temperate Parastacidae, Australia and Tasmania have the genera Astacopsis and Engaeus; New Zealand 214has Paranephrops, South America Parastacus, and Madagascar Astacoides. The last named genus is rather isolated in its characters, possessing a truncated rostrum and a highly modified branchial system, but it agrees with all the other Parastacine genera, and differs from the Astacidae in the absence of sexual appendages on the first abdominal segment, and in the absence of a distinct lamina on the podobranchiae. The largest crayfish in the world is Astacopsis franklinii, found in quite small streams on the north and west coast of Tasmania. Specimens have been caught weighing eight or nine pounds, and rivalling the European Lobster in size. Crayfishes appear to be entirely absent from Africa.

It seems reasonable to suppose that the two families of Crayfishes characteristic respectively of the northern and southern hemispheres have been independently derived from marine ancestors, which have subsequently become extinct. Their complete absence in the tropics is striking, and Huxley drew attention to the fact that it is exactly in those regions where the Crayfishes are absent that the other large fresh-water Malacostraca are particularly well developed, and vice versa. Thus the large fresh-water Prawns are typically circumtropical in distribution, while the South African rivers abound with River-crabs, which, in general, are found wherever Crayfishes do not occur.

A few of the more interesting features in connection with the distribution of fresh-water Crustacea have now been touched upon. With regard to the origin of this fauna, we can see that a number of the species are comparatively recent immigrants from the sea, working their way up the estuaries of rivers, a proceeding which can be observed to be taking place to-day in a district like the Broads of Norfolk. Others, again, but these are few, appear to be true relict marine animals left stranded in arms of the sea that have been cut off from the main ocean, and have been gradually converted into fresh-water lakes and seas. Such are, perhaps, Mysis relicta and the rich Gammarid fauna of Lake Baikal, a lake that, in the presence of Seals, Sponges, and other marine forms, has clearly retained some of the characters of the ocean from which it was derived. The majority of the fresh-water species, however, have probably been evolved in situ, and their origin from marine ancestors is lost in an obscure past. The Crustacean fauna of the Caspian 215Sea^[177] shows us in an interesting manner the effects of isolation and changes in salinity, etc., on the inhabitants of a basin which once formed part of the ocean. The waters of the Caspian Sea are not fresh, but they are on the average about one-third as salt as that of the open ocean. The Crustacea, described by Sars, belong to undoubtedly marine groups, e.g. the Mysidae, Cumacea, and Amphipoda Crevettina, but the remarkable feature of these Caspian Crustacea is the great variety of peculiar species representing marine genera which are very poorly represented in the sea, thus indicating that the variety of the fauna is not due to a great variety of species having been shut up in the Caspian Sea to begin with, but rather that, after the separation from the sea, the isolated species began to vary and branch out in the most luxuriant way—whether from lack of competition or owing to the changing conditions of salinity it is difficult to say. As an example, the Cumacea of the Caspian Sea are ten in number, all belonging to peculiar genera related to Pseudocuma, except one species which is included in that genus. These Caspian forms make up the Family Pseudocumidae, which contains in addition only two marine forms of the genus Pseudocuma (see p. 121). A very similar condition is found in the numerous Amphipods of the Caspian Sea. Considering the enormous changes that must have taken place in the distribution of land and water even during Tertiary times, it is astonishing that the fresh waters of the world do not contain more species in common with the ocean, but it must be considered that the limited area and comparatively uniform conditions of fresh-water lakes and streams would only permit a limited number of these forms to survive which could most easily adapt themselves to the changed conditions. And these would in all probability be the littoral species that were in the habit of passing up into the brackish waters of estuaries and lagoons, so that the uniform and limited nature of the fresh-water fauna can be accounted for to a certain extent by this hypothesis.

We have seen in dealing with the marine Crustacea of the littoral zone that the chief condition determining their distribution is temperature, and that the world may be divided into three chief 216areas of distribution for these animals, viz. the north temperate hemisphere, the tropics, and the south temperate hemisphere. It seems that the same division holds good for fresh-water Crustacea. We have already seen that the Crayfishes follow this rule, being practically absent from the tropics, and represented in the two temperate hemispheres by two distinct families, the Astacidae in the north and the Parastacidae in the south. Characteristic of the tropical belt are the absence of Crayfishes and the great development of Prawns and River-crabs. In the case of Entomostraca the great majority of the genera are cosmopolitan, especially those which live in small bodies of water liable to dry up, because these forms always have special means of dissemination in the shape of resting eggs which can be transported in a dry state by water-birds and other agencies to great distances; but those genera which inhabit large lakes are more confined in their distribution. The Copepod genus Diaptomus, characteristic of lake-plankton, ranges all over the northern hemisphere and into the tropics, but it is almost entirely replaced in the southern hemisphere by the related but distinct genus Boeckella,^[178] which occurs in temperate South America, New Zealand, and southern Australia, and was found by the author to be the chief inhabitant in the highland lakes and tarns of Tasmania, Diaptomus being entirely absent. Of the Cladocera there are a number of pelagic genera (e.g. Leptodora, Holopedium, Bythotrephes) entirely confined to the lakes of the northern hemisphere. The distribution of Bosmina is interesting. This genus is distributed all over the north temperate hemisphere in lakes and ponds of considerable size, not liable to desiccation; in the New World it passes right through the tropics into Patagonia,^[179] the chain of the Andes doubtless permitting its migration. In the tropics of the Old World it is unknown, but it turns up again, as the author recently found, as a common constituent in the plankton of the Tasmanian lakes. There is another instance of a group of Crustacea, characteristic of the north temperate hemisphere, being entirely absent from the tropics, at any rate of the Old World, but reappearing in the temperate regions of Australasia. The commonest fresh-water Amphipods in this region belong to the genus Neoniphargus, intermediate in its characters between the 217northern Niphargus and Gammarus, but grading almost completely into the latter. Both Niphargus and Gammarus are absolutely unknown from the tropics, but whether, like Bosmina, they occur in the Andes and temperate South America is not known; it seems, however, probable that they have reached Southern Australia by way of South America rather than through the tropics of Asia and Australia, where there is no range of mountains to permit the migration of a group of animals apparently dependent on a temperate climate. The other common fresh-water Amphipod in temperate Australia and New Zealand is Chiltonia, whose nearest ally is Hyalella from Lake Titicaca on the Andes, and temperate South America.

The Anaspidacea and Phreatoicidae, which are so characteristic of temperate Australia, and are generally of an Alpine habit, have never been found in South America, but the Anaspidacea are represented by numerous marine forms in the Permian and Carboniferous strata of the northern hemisphere, so that it is probable that this group reached the southern hemisphere from the north through America.

The distribution of the fresh-water Crustacea, therefore, in the temperate southern hemisphere affords strong evidence in favour of the view that the chief land-masses of this hemisphere, which are at present separated by such vast stretches of deep ocean, were at no very remote epoch connected in such a way as to permit of an intermixture of the temperate fauna of New Zealand, Australia, and South America. While this connexion existed, a certain number of forms characteristic of the northern hemisphere, which had worked through the tropics by means of the Andes, were enabled to reach temperate Australia and New Zealand. The existence of a coast-line connecting the various isolated parts of the southern hemisphere would, of course, also account for the community which exists between their littoral marine fauna. It is impossible to enter here into the nature of this land-connexion which is becoming more and more a necessary hypothesis for the student of geographical distribution, whatever group of animals he may choose, but it may be remarked that the connexion was probably by means of rays of land passing up from an Antarctic continent to join the southernmost projections of Tierra del Fuego, Tasmania, and New Zealand.

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 222period 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 223body 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.

224

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 225anterior 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 226been 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 227fixed 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.

228Three 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.

2292. 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 230alternating 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 231found 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. 232Many 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 233the 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, 234have 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 235succeeding 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 236pygidium 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, 237each 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 238the 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 239indications 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 240to 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 241on 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 242some 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 243it 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.

2442. 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 245the 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 246projection 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, 247or 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 248connected 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.)

249Fam. 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.

250

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.)