BRANCHIOPODA.

The early idea that the trilobites were closely related to the Branchiopoda was rejuvenated by the work of Bernard on the ApodidÆ (1892) and has since received the support of most writers on the subject. Fundamentally, a great deal of the argument seems to be that Apus lies the nearest of any modern representative of the class to the theoretical crustacean ancestor, and as the trilobites are the oldest Crustacea, they must be closely related. Most writers state that the trilobites could not be derived from the Branchiopoda (see, however, Walcott 1912 A), nor the latter from any known trilobite, but both subclasses are believed to be close to the parent stem.

Viewed from the dorsal side, there is very little similarity between any of the branchiopods and the trilobites, and it is only in the Notostraca, with their sessile eyes and depressed form, that any comparison can be made. The chief way in which modern Branchiopoda and Trilobita agree is that both have a variable number of segments in the body, that number becoming very large in Apus on the one hand and Mesonacis and PÆdeumias on the other. In neither are the appendages, except those about the mouth, grouped in tagmata. Other likenesses are: the Branchiopoda are the only Crustacea, other than Trilobita, in which gnathobases are found on limbs far removed from the mouth; the trunk limbs are essentially leaf-like in both, though the limb of the branchiopod is not so primitive as that of the trilobite; caudal cerci occur in both groups.

If the appendages be compared in a little more detail, the differences prove more striking than the likenesses.

In the Branchiopoda, the antennules are either not segmented or only obscurely so. In trilobites they are richly segmented.

In Branchiopoda, the antennÆ are variable. In the Notostraca they are vestigial, while in the males of the Anostraca they are powerful and often complexly developed claspers. Either condition might develop from the generalized biramous antennas of Trilobita, but the present evidence indicates a tendency toward obsolescence. Claus' observations indicate that the antennÆ of the Anostraca are developments of the exopodites, rather than of the endopodites.

The mandibles and maxillÆ of the Branchiopoda are greatly reduced, and grouped closely about the mouth. Only the coxopodites of the Trilobita are modified as oral appendages.

The trunk limbs of Apus are supposed to be the most primitive among the Branchiopoda, and comparison will be made with them. Each appendage consists of a flattened axial portion, from the inner margin of which spring six endites, and from the outer, two large flat exites (see fig. 34). This limb is not articulated with the ventral membrane, but attached to it, and, if Lankester's interpretation of the origin of schizopodal limbs be correct, then the limb of Apus bears very little relation to that of the Trilobita. In Apus there is no distinct coxopodite and the endobases which so greatly resemble the similar organs in the Trilobita are not really homologous with them, but are developments of the first endite. Beecher's comparison of the posterior thoracic and pygidial limbs of Triarthrus with those of Apus can not be sustained. Neither Triarthrus nor any other trilobite shows any trace of phyllopodan limbs. Beecher figured (1894 B, pl. 7, figs. 3, 4) a series of endopodites from the pygidium of a young Triarthrus beside a series of limbs from a larval Apus. Superficially, they are strikingly alike, but while the endopodites of Triarthrus are segmented, the limbs of Apus are not, and the parts which appear to be similar are really not homologous. The similarity of the thoracic limbs in the two groups is therefore a case of parallelism and does not denote relationship.

Geologically, the Branchiopoda are as old as the Trilobita, and while they did not have the development in the past that the trilobite had, they were apparently differentiated fully as early. Anostraca, Notostraca and Conchostraca, three of the four orders, are represented in the Cambrian by forms which are, except in their appendages, as highly organized as the existing species. Brief notes on the principal Middle Cambrian Branchiopoda follow:

Burgessia bella Walcott.

This is the most strikingly like the modern Branchiopoda of any species described by Walcott from the Middle Cambrian, and invites comparison with Apus. The carapace is long, loosely attached to the body, and extends over the greater part of the thorax. The eyes are small, sessile, and close to the anterior margin.

The appendages of the head consist of two pairs of antennÆ, and three pairs of slender, jointed legs. Both pairs of antennÆ are slender and many-jointed, the antennules somewhat smaller than the antennÆ. The exact structure of the limbs about the mouth has not yet been made out, but they are slender, tapering, endopodite-like legs, with at least three or four segments in each, and probably more.

There are eight pairs of thoracic appendages, each limb having the form of the endopodite of a trilobite and consisting of seven segments and a terminal spine. The proximal three segments of each appendage are larger than the outer ones, and have a flattened triangular expansion on the inner side. Walcott also states that "One specimen shows on seven pairs of legs, small, elongate, oval bodies attached near the first joint to the outer side of the leg. These bodies left but slight impression on the rock and are rarely seen. They appear to represent the gills." They are not figured, but taken in connection with the endopodite-like appearance of the segmented limbs, one would expect them to be vestigial exopodites.

A small hypostoma is present on the ventral side, and several of the specimens show wonderfully well the form of the alimentary canal and the hepatic cÆca. The main branches of the latter enter the mesenteron just behind the fifth pair of cephalic appendages.

Behind the thorax the abdomen is long, limbless, and tapers to a point. It is said to consist of at least thirty segments.

Compared with Apus, Burgessia appears both more primitive and more specialized. The carapace and limbless abdomen are Apus-like, but there are very few appendagiferous segments, and the appendages are not at all phyllopodan, but directly comparable with those of trilobites, except, of course, for the uniramous character of the cephalic limbs. A closer comparison may be made with Marrella.

Waptia fieldensis Walcott.

The carapace is short, covering the head and the anterior part of the thorax. The latter consists of eight short segments with appendages, while the six abdominal segments, which are similar to those of the thorax, are without limbs except for the last, which bears a pair of broad swimmerets. The eyes are marginal and pedunculate. The antennules are imperfectly known, but apparently short, while the antennas are long and slender, with relatively few, long, segments. The mandibles appear to be like endopodites of trilobites and show at least six segments. As so often happens in these specimens from British Columbia, the preservation of the other appendages is unsatisfactory. As illustrated (Walcott, 1912 A, pl. 27, fig. 5), both endopodites and exopodites appear to be present, and the shaft of the exopodite seems to be segmented as in Triarthrus.

Walcott considers Waptia as a transitional form between the Branchiopoda and the Malacostraca.

Yohoia tenuis Walcott.

This species, though incompletely known, has several interesting characteristics. The head shows, quite plainly in some specimens, the five segments of which it is composed. The eyes are small, situated in a niche between the first and second segments, and are described as being pedunculate. The eight segments of the thorax all show short triangular pleural extensions, somewhat like those of Remopleurides or Robergia. The abdomen consists of four cylindrical segments, the last with a pair of expanded caudal rami.

The antennules appear to be short, while the antennas are large, with several segments, ending in three spines, and apparently adapted for serving as claspers in the male. The third, fourth, and fifth pairs of cephalic appendages are short, tapering, endopodite-like legs similar to those of Burgessia.

The appendages of the thorax are not well preserved, and there seem to be none on the abdomen.

This species is referred by Walcott to the Anostraca.

Opabina regalis Walcott.

This most remarkably specialized anostracan is not well enough known to allow comparison to be made with other contemporaneous Crustacea, but it is worthy of mention.

There is no carapace, the eyes are pedunculated, thorax and abdomen are not differentiated, and the telson is a broad, elongate, spatulate plate. There seem to be sexual differences in the form of the anterior cephalic and caudal appendages, but this is not fully established. The most remarkable feature is the long, large, median cephalic appendage which is so suggestive of the proboscis of the recent Thamnocephalus platyurus Packard. The appendages are not well enough preserved to permit a determination as to whether they are schizopodal or phyllopodan.

Summary.

Walcott referred Burgessia and Waptia to new families under the Notostraca, while Yohoia and Opabina were placed with the Anostraca. Except for the development of the carapace, there is a striking similarity between Waptia and Yohoia, serving to connect the two groups.

The Branchiopoda were very highly specialized as early as Middle Cambrian time, the carapace of the Notostraca being fully developed and the abdomen limbless. Some (Burgessia) had numerous segments, but most had relatively few. The most striking point about them, however, is that so far as is known none of them had phyllopodan limbs. While the preservation is in most cases unsatisfactory, such limbs as are preserved are trilobite-like, and in the case of Burgessia there can be no possible doubt of the structure. Another interesting feature is the retention by Yohoia of vestiges of pleural lobes. The Middle Cambrian Branchiopoda are more closely allied to the Trilobita than are the modern ones, but still the subclass is not so closely related to that group as has been thought. Modern Apus is certainly much less like a trilobite than has been supposed, and very far from being primitive. The Branchiopoda of the Middle Cambrian could have been derived from the trilobites by the loss of the pleural lobes, the development of the posterior margin of the cephalon to form a carapace, and the loss of the appendages from the abdominal segments. Modern branchiopods can be derived from those of the Middle Cambrian by the modification of the appendages through the reduction of the endopodite and exopodite and the growth of the endites and exites from the proximal segments.

Carpenter (1903, p. 334), from his study of recent crustaceans, has already come to the conclusion that the Branchiopoda are not the most primitive subclass, and this opinion is strengthened by evidence derived from the Trilobita and from the Branchiopoda of the Middle Cambrian.

COPEPODA.

The non-parasitic Eucopepoda are in many ways much nearer to the trilobites than any other Crustacea. These little animals lack the carapace, and the body is short, with typically ten free segments and a telson bearing caudal furcÆ. The head is composed of five segments (if the first thoracic segment is really the fused first and second), is often flattened, and lacks compound eyes. Pleural lobes are well developed, but instead of being flattened as in the trilobite, they are turned down at the sides or even incurved. A labrum is present.

The antennules, antennÆ, and mandibles are quite like those of trilobites. The antennules are very long and made up of numerous segments. The antennÆ are biramous, the junction between the coxopodite and basipodite is well marked, and the endopodite consists of only two segments.

The mandibles are said to "retain more completely than in any other Crustacea the form of biramous swimming limbs which they possess in the nauplius." The coxopodites form jaws, while both the reduced endopodite and exopodite are furnished with long setÆ. The maxillulÆ are also biramous, but very different in form from those of the trilobite, and the maxillÆ are phyllopodan.

The first thoracic limb is uniramous and similar to the maxillÆ, but the five following pairs are biramous swimming legs with coxopodite, basipodite, exopodite, and endopodite. Both the exopodite and endopodite are shorter than in the trilobites, but bear setÆ and spines.

The last pair of thoracic limbs are usually modified in the male into copulatory organs. In some females they are enlarged to form plates for the protection of the eggs, in others they are unmodified. In still others they are much reduced or disappear. The abdomen is without appendages.

The development in Copepoda is direct, by the addition posteriorly to the larval form (nauplius) of segments, and the appendages remain nearly unmodified in the adult.

Altogether, the primitive Copepoda seem much more closely allied to the Trilobita than any other modern Crustacea, but unfortunately no fossil representative of the subclass has been found. This is not so surprising when one considers the habits and the habitat of most of the existing species. Many are parasitic, many pelagic in both fresh and marine waters, and many of those living on the bottom belong to the deep sea or fresh water. Most free-living forms are minute, and all have thin tests.

The eyes of copepods are of interest, in that they suggest the paired ocelli of the HarpedidÆ and TrinucleidÆ. In the Copepoda there are, in the simplest and typical form of these organs, three ocelli, each supplied with its own nerve from the brain. Two of these are dorsal and look upward, while the third is ventral. In some forms the dorsal ocelli are doubled, so that five in all are present (cf. some species of Harpes with three ocelli on each mound). In some, the cuticle over the dorsal eyes is thickened so as to form a lens, as appears to be the case in the trilobites. These peculiar eyes may be a direct inheritance from the Hypoparia.

ARCHICOPEPODA.

Professor Schuchert has called my attention to the exceedingly curious little crustacean which Handlirsch (1914) has described from the Triassic of the Vosges. Handlirsch erected a new species, genus, family, and order for this animal, which he considered most closely allied to the copepods, hence the ordinal name. Euthycarcinus kessleri, the species in question, was found in a clayey lens in the Voltzia sandstone (Upper Bunter). Associated with the new crustacean were specimens of Estheria only, but in the Voltzia sandstone itself land plants, fresh and brackish water animals, and occasionally, marine animals are found. The clayey lens seems to have been of fresh or brackish water origin.

All of the specimens (three were found) are small, about 35 mm. long without including the caudal rami, crushed flat, and not very well preserved. The head is short, not so wide as the succeeding segments, and apparently has large compound eyes at the posterior lateral angles. The thorax consists of six segments which are broader than the head or abdomen. The abdomen, which is not quite complete in any one specimen, is interpreted by Handlirsch as having four segments in the female and five in the male. Least satisfactory of all are traces of what are interpreted by the describer as a pair of long stiff unsegmented cerci or stylets on the last segment.

The ventral side of one head shield shows faint traces of several appendages which must have presented great difficulty in their interpretation. A pair of antennules appear to spring from near the front of the lower surface, and the remainder of the organs are grouped about the mouth, which is on the median line back of the center. Handlirsch sees in these somewhat obscure appendages four pairs of biramous limbs, antennÆ, mandibles, maxillulÆ, and maxillÆ, both branches of each consisting of short similar segments, endopodites and exopodites being alike pediform.

Each segment of the thorax has a pair of appendages, and those on the first two are clearly biramous. The endopodites are walking legs made up of numerous short segments (twelve or thirteen according to Handlirsch's drawing), while the exopodite is a long breathing and rowing limb, evidently of great flexibility and curiously like the antennules of the same animal. The individual segments are narrow at the proximal end, expand greatly at the sides, and have a concave distal profile. A limb reminds one of a stipe of Diplograptus. Both branches are spiniferous.

No appendages are actually present on the abdomen, but each segment has a pair of scars showing the points of attachment. From the small size of these, it is inferred that the limbs were poorly developed.

This species is described in so much detail because, if it is a primitive copepod, it has a very important bearing on the ancestry of that group and is the only related form that has been found fossil.

The non-parasitic copepods have typically ten (eleven) free segments, including the telson, and the four abdominal segments are much more slender than the six in front of them. In this respect the agreement is striking, and the presence of five pairs of appendages in the head and six free segments in the thorax is a more primitive condition than in modern forms where the first two thoracic segments are apparently fused (Calman, 1909, p. 73).

The large compound eyes of this animal are of course not present in the copepods, but as vestiges of eyes have been found in the young of Calanus, it is possible that the ancestral forms had eyes.

The greatest difficulty is in finding a satisfactory explanation of the appendages. The general condition is somewhat more primitive than in the copepods, for all the appendages are biramous, while in the modern forms the maxillipeds are uniramous and the sixth pair of thoracic appendages are usually modified in the male as copulatory organs. In the copepods the modification is in the direction of reduction, both endopodites and exopodites usually possessing fewer segments than the corresponding branches in the trilobites. The endopodite of Euthycarcinus, on the contrary, possesses, if Handlirsch's interpretation is correct, twice as many segments as the endopodite of a trilobite. If the Copepoda are descended from the trilobites, as everything tends to indicate, then Euthycarcinus is certainly not a connecting link. The only truly copepodan characteristic of this genus is the agreement in number and disposition of free segments. The division into three regions instead of two, the compound eyes, and the structure of the appendages are all foreign to that group.

With the Limulava fresh in mind, one is tempted to compare Euthycarcinus with that ancient type. The short head and large marginal eyes recall Sidneyia, and the grouping of the appendages about the mouth also suggests that genus and Emeraldella. In the Limulava likewise there is a contraction of the posterior segments, although it is behind the ninth instead of the sixth. There is no likeness in detail between the appendages of the Limulava and those of Euthycarcinus, but the composite claws of Sidneyia show that in this group there was a tendency toward the formation of extra segments.

If this fossil had been found in the Cambrian instead of the Triassic, it would probably have been referred to the Limulava, and is not at all impossible that it is a descendant from that group. As a connecting link between the Trilobita and Copepoda it is, however, quite unsatisfactory.

OSTRACODA.

The bivalved shell of the Ostracoda gives to this group of animals an external appearance very different from that of the trilobites, but the few appendages, though highly modified, are directly comparable. The development, although modified by the early appearance of the bivalved shell within which the nauplius lies, is direct. Imperfect compound eyes are present in one family.

The antennules are short and much modified by functioning as swimming, creeping, or digging organs. They consist of eight or less segments. The antennas are also locomotor organs, and in most orders are biramous. The mandibles are biramous and usually with, but sometimes without, a gnathobase. The maxillulÆ are likewise biramous but much modified.

The homology of the third post-oral limb is in question, some considering it a maxilla and others a maxilliped. It has various forms in different genera. It is always much modified, but exopodite and endopodite are generally represented at least by rudiments. The fourth post-oral limb is a lobed plate, usually not distinctly segmented, and the fifth a uniramous pediform leg. The sixth, if present at all, is vestigial.

Very little comparison can be made between the Ostracoda and Trilobita, other than in the ground-plan of the limbs, but the presence of biramous antennÆ is a primitive characteristic.

CIRRIPEDIA.

Like the ostracod, the adult cirriped bears little external resemblance to the trilobite. The form of the nauplius is somewhat peculiar, but it has the typical three pairs of appendages, to which are added in the later metanauplius stages the maxillÆ and six pairs of thoracic appendages. In the adult, the antennules, which serve for attachment of the larva, usually persist in a functionless condition, while the antennas disappear. The mandibles, maxillulÆ, and maxillÆ are simple and much modified to form mouth parts, and the six pairs of thoracic appendages are developed into long, multisegmented, biramous appendages bearing numerous setÆ which serve for catching prey. Paired eyes are present in later metanauplius stages, but lost early in the development. The relationship to the trilobite evidently is not close.

MALACOSTRACA.

1. Phyllocarida.

The oldest malacostracans whose appendages are known are species of Hymenocaris. One, described as long ago as 1866 by Salter, has what seem to be a pair of antennÆ and a pair of jaw-like mouth-parts. Another more completely known species has recently been reported by Walcott (1912 A, p. 183, pl. 31, figs. 1-6). This latter form is described as having five pairs of cephalic appendages: a pair of minute antennules beside the small pedunculated eyes, a pair of large uniramous antennÆ, slender mandibles and maxillulÆ, and large maxillÆ composed of short stout segments. There are eight pairs of biramous thoracic limbs, the exopodites setiferous, the endopodites composed of short wide segments and ending in terminal claw-like spines. These appendages are like those of trilobites.

Hymcnocaris belongs to the great group of extinct ceratocarid Crustacea which are admitted to the lowest of the malacostracan orders, Phyllocarida, because of their resemblance to Nebalia, Paranebalia, Nebaliopsis, and Nebaliella, the four genera which are at present living. The general form of the recent and fossil representatives of the order is strikingly similar. The chief outward difference is that in many of the fossils the telson is accompanied by two furcal rami, while in the modern genera it is simple. It now becomes possible to make some comparison between the appendages of Hymcnocaris of the Middle Cambrian and the NebaliidÆ of modern seas.

In both there are five pairs of cephalic and eight of thoracic appendages, while those of the abdomen of Hymenocaris are not known.

In both, the antennules are less developed than the antennÆ. In the NebaliidÆ the antennules show evidence of having been originally double (they are obviously so in the embryo), while they are single in Hymcnocaris. In both, the antennÆ are simple. The remaining cephalic organs are too little shown by the specimen from the Middle Cambrian to allow detailed comparison. The mandibles, maxillulÆ, and maxillÆ of Nebalia are, however, of types which could be derived from the trilobite.

In three of the genera of the NebaliidÆ, the eight pairs of thoracic limbs are all similar to one another, though those of the genera differ. All are biramous. The limbs of Hymcnocaris can apparently be most closely correlated with those of Nebalia antarctica, in which the endopodite consists of short flattened segments, and the exopodite is a long setiferous plate. Epipodites are present in both Nebalia and Hymcnocaris.

So far as the appendages of Hymenocaris are known, they agree very well with those of the NebaliidÆ, and since they are of the trilobite type, it may safely be stated that the Trilobita and Malacostraca are closely related.

2. Syncarida.

Walcott (1918, p. 170) has compared the limbs of Neolenus with those of the syncarid genera Anaspides and Koonunga. These are primitive Malacostraca without a carapace, but as they have a compressed test and Anaspides has stalked eyes, their gross anatomy does not suggest the trilobite. The thoracic appendages are very trilobite-like, since the endopodite has six segments (in Anaspides) and a multisegmented setiferous exopodite. The coxopodites, except of the first thoracic segment, do not, however, show endobases, and those which are present are peculiar articulated ones. The cephalic appendages are specialized, and the antennules double as in most of the Malacostraca. External epipodites are very numerous on the anterior limbs.

This group extends back as far as the Pennsylvanian and had then probably already become adapted to fresh-water life. It may be significant that the PalÆozoic syncarids appear to have lacked epipodites. While differing very considerably from the Trilobita, the Syncarida could have been derived from them.

3. Isopoda.

Since the earliest times there has been a constant temptation to compare the depressed shields of the trilobites with the similar ones of isopods. Indeed, when Scrolls with its Lichadian body was first discovered about a hundred years ago, it was thought that living trilobites had been found at last. The trilobate body, cephalic shield, sessile eyes, abdominal shield, and pleural extensions make a wonderful parallel. This similarity is, however, somewhat superficial. The appendages are very definitely segregated in groups on the various regions of the body, and while the pleopods are biramous, the thoracic legs are without exopodites (except in very early stages of development of one genus). The Isopoda arose just at the time of the disappearance of the Trilobita, and there seems a possibility of a direct derivation of the one group from the other. It should be pointed out that while the differences of Isopoda from Trilobita are important, they are all of a kind which could have been produced by the development from a trilobite-like stock. For example:

Isopoda have a definite number of segments. There is less variation in the number of segments among the later than the earlier trilobites.

Isopoda have no facial suture. In at least three genera of trilobites the cheeks become fused to the cranidium and the sutures obliterated.

Isopoda have one or two segments of the thorax annexed to the head. While this is not known to occur in trilobites, it is possible that it did.

Most Isopoda have a fairly stiff ventral test. The ventral membrane of trilobites would probably have become stiffened by impregnation of lime if the habit of enrollment had been given up.

In Isopoda the antennÆ are practically uniramous sensory organs. The second cephalic appendages of trilobites are capable of such development through reduction of the exopodite.

In the Isopoda the coxopodites are usually fused with the body, remaining as free, movably articulated segments only in a part of the thoracic legs of one suborder, the Asellota. Endobases are entirely absent. This is of course entirely unlike the condition in Trilobita, but a probable modification.

In Isopoda there is a distinct grouping of the appendages, with specialization of function. The trilobites show a beginning of tagmata, and such development would be expected if evolution were progressive.

In both groups, development from the embryo is direct. Rudiments of exopodites of thoracic legs have been seen in the young of one genus.

The oldest known isopod is Oxyuropoda ligioides Carpenter and Swain (Proc. Royal Irish Acad., vol. 27, sect. B, 1908, p. 63, fig. 1), found in the Upper Devonian of County Kilkenny, Ireland. The appendages are not known, but the test is in some ways like that of a trilobite. The thorax, abdomen, and pygidium are especially like those of certain trilobites, and there is no greater differentiation between thorax and abdomen than there is between the regions before and behind the fifteenth segment of a PÆdeumias or Mesonacis. The anal segment is directly comparable to the pygidium of a Ceraurus, the stiff unsegmented uropods being like the great lateral spines of that genus.

The interpretation of the head offered by Carpenter and Swain is very difficult to understand, as their description and figure do not seem to agree. What they consider the first thoracic segment (fused with the head) seems to me to be the posterior part of the cephalon. and it shows at the back a narrow transverse area which is at least analogous to the nuchal segment of the trilobite. If this interpretation can be sustained, Oxyuropoda would be a very primitive isopod in which the first thoracic segment (second of Carpenter and Swain) is still free. According to the interpretation of the original authors, the species is more specialized than recent Isopoda, as they claim that two thoracic segments are fused in the head. The second interpretation was perhaps made on the basis of the number of segments (nineteen) in a recent isopod.

Marrella splendens Walcott.

Among the most wonderful of the specimens described by Doctor Walcott is the "lace crab." While the systematic position was not satisfactorily determined by the describer, it has been aptly compared to a trilobite. The great nuchal and genal spines and the large marginal sessile eyes, coupled with the almost total lack of thoracic and abdominal test, give it a bizarre appearance which may obscure its real relationships.

The cephalon appears to bear five pairs of appendages, antennules, and antennÆ, both tactile organs with numerous short segments, mandibles, and first and second maxillÆ. The last three pairs are elongate, very spinose limbs, of peculiar appearance. They seem to have seven segments, but are not well preserved. These organs are attached near the posterior end of the labrum.

There are twenty-four pairs of biramous thoracic appendages, which lack endobases. The endopodites are long and slender, with numerous spines; the exopodites have narrow, thin shafts, with long, forward pointed setÆ. The anal segment consists of a single plate.

Further information about this fossil will be eagerly awaited. None of the illustrations so far published shows biramous appendages on the cephalon. This, coupled with the presence of tactile antennÆ, makes its reference to the Trilobita impossible, but the present interpretation indicates that it was closely allied to them.

Restoration of Marrella.

(Text fig. 32.)

The accompanying restoration of the ventral surface of Marrella is a tentative one, based on Doctor Walcott's description and figures. The outline is taken from his plate 26, figure 1; the appendages of the head from plate 26, figures 1-3, 5, and plate 25, figures 2, 3; the endopodites, shown on the left side only, from figures 3 and 6, plate 25. I have not studied actual specimens, and the original description is very incomplete. The restoration is therefore subject to revision as the species becomes better known.