Chapter XVI: Part III (3)

In the discussion above I have placed great emphasis on the large size of the primitive pygidium, because, although there is nothing new in the idea, its significance seems to have been overlooked.

If the large pygidium is primitive, then multisegmentation in trilobites can not be primitive but is the result of adaptation to a crawling life. It is annelid-like, but is not in itself to be relied upon as showing relationship to the Chætopoda. Simple trilobites with few segments, like the Agnostidæ, Eodiscidæ etc., were, therefore, properly placed by Beecher at the base of his classification, and there is now less chance than ever that they can be called degenerate animals.

From the phylogeny of certain groups, such as the Asaphidæ, it is learned that the geologically older members of the family have more strongly segmented anterior and posterior shields than the later ones. That there has been a "smoothing out" is demonstrated by a study of the ontogeny of the later forms. From such examples it has come to be thought that all smooth trilobites are specialized and occupy a terminal position in their genealogical line. This has caused some wonder that smooth agnostids like _Phalacroma bibullatum_ and _P. nudum_ should be found in strata so old as the Middle Cambrian, and was a source of great perplexity to me in the case of _Weymouthia_ (Ottawa Nat., vol. 27, 1913) (fig. 35). This is a smooth member of the Eodiscidæ, and, in fact, one of the simplest trilobites known, for while it has three thoracic segments, it shows almost no trace of dorsal furrows or segmentation on cephalon or pygidium, and, of course, no eyes. Following the general rule, I took this to be a smooth-out eodiscid, and was surprised that it should come from the Lower Cambrian, where it is associated with _Elliptocephala_ at Troy, New York, and with _Callavia_ at North Weymouth, Massachusetts, and where it has lately been found by Kiær associated with _Holmia_ and _Kjerulfia_ at Tømten, Norway. It now appears it is really in its proper zone, and instead of being the most specialized, is the simplest of the Eodiscidæ.

What appears to be a still simpler trilobite is the form described by Walcott as Naraoia.

=Naraoia compacta= Walcott.

(Text fig. 36.)

Illustrated: Walcott, Smithson. Misc. Coll., vol. 57, 1912, p. 175,
pl. 28, figs. 3, 4.--Cleland, Geology, Physical and Historical, New
York, 1916, p. 412, fig. 382 F (somewhat restored).

This very imperfectly known form is referred by Walcott to the Notostraca on what appear to be wholly inadequate grounds, and while I do not insist on my interpretation, I can not refrain from calling attention to the fact that it _can_ be explained as the most primitive of all trilobites. It consists of two subequal shields, the anterior of which shows slight, and the posterior considerable evidence of segmentation. It has no eyes, no glabella, and no thorax, and is directly comparable to a very young _Phalacroma bibullatum_ (see Barrande 1852, pl. 49, figs. a, b). Walcott states that there is nothing to show how many segments there are in the cephalic shield, but that on one specimen fourteen were faintly indicated on the abdominal covering. The appendages are imperfectly unknown, as no specimen showing the ventral side has yet been described. The possible presence of antennas and three other appendages belonging to the cephalic shield is mentioned, and there are tips of fourteen legs projecting from beneath the side of one specimen. As figured, some of the appendages have the form of exopodites, others of endopodites, indicating that they were biramous.

_Naraoia_ is, so far as now known, possessed of no characteristics which would prevent its reference to the Trilobita, while the presence of a large abdominal as well as a cephalic shield would make it difficult to place in even so highly variable a group as the Branchiopoda. On the other hand, its only exceptional feature as a trilobite is the lack of thorax, and all study of the ontogeny of the group has led us to expect just that sort of a trilobite to be found some day in the most ancient fossiliferous rocks. _Naraoia_ can, I think, be best explained as a trilobite which grew to the adult state without losing its protaspian form. It was found in the Middle Cambrian of British Columbia.

Even if _Naraoia_ should eventually prove to possess characteristics which preclude the possibility of its being a primitive trilobite, it at least represents what I should expect a pre-Cambrian trilobite to look like. What the ancestry of the nektonic primitive trilobite may have been is not yet clear, but all the evidence from the morphology of cephalon, pygidium, and appendages indicates that it was a descendant of a swimming and not a crawling organism.

Since the above was written, the Museum of Comparative Zoology has purchased a specimen of this species obtained from the original locality. The shields are subequal, the posterior one slightly the larger, and the axial lobes are definitely outlined on both. The glabella is about one third the total width, nearly parallel-sided, somewhat pointed at the front. There are no traces of glabellar furrows. The axial lobe of the pygidium is also about one third the total width, extends nearly to the posterior margin, and has a rounded posterior end. The measurements are as follows: Length, 33 mm.; length of cephalon, 16 mm., width, 15 mm.; length of glabella, 11.5 mm., width, 5.5 mm.; length of pygidium, 17 mm., width, 15 mm.; length of axial lobe, 14 mm., width, 5.5 mm.

The species is decidedly _Agnostus_-like in both cephalon and pygidium, and were it not so large, might be taken for the young of such a trilobite. The pointed glabella is comparable to the axial lobes of the so-called pygidia of the young of _Condylopyge rex_ and _Peronopsis integer_ (Barrande, Syst. Sil., vol. 1, pl. 49).

The Ancestor of the Trilobites, and the Descent of the Arthropoda.

The "annelid" theory of the origin of the Crustacea and therefore of the trilobites, originating with Hatschek (1877) and so ably championed by Bernard (1892), has now been a fundamental working hypothesis for some years, and has had a profound influence in shaping thought about trilobites. This hypothesis has, however, its weak points, the principal one being its total inhibition of the workings of that great talisman of the palæontologist, the law of recapitulation. Its acceptance has forced the zoologist to look upon the nauplius as a specially adapted larva, and has caused more than one forced explanation of the protaspis of the trilobite. When so keen a student as Calman says that the nauplius must point in some way to the ancestor of the Crustacea (1909, p. 26), it is time to reëxamine some of the fundamentals. This has been done in the preceding pages and evidence adduced to show that the primitive features of a trilobite indicate a swimming animal, and that the adaptations are those which enabled it to assume a crawling mode of existence. It has also been pointed out that in Naraoia there is preserved down to Middle Cambrian times an animal like that to which ontogeny points as a possible ancestor of the trilobites. _Naraoia_ is not the simplest conceivable animal of its own type, however, for it has built up a pygidium of fourteen or fifteen somites. One would expect to find in Proterozoic sediments remains of similar animals with pygidia composed of only one or two somites, with five pairs of appendages on the cephalon, one or two pairs on the pygidium, a ventral mouth, and a short hypostoma. Anything simpler than this could not, in my opinion, be classed as a trilobite.

What the ancestor of this animal was is mere surmise. It probably had no test, and it may be noted in this connection that _Naraoia_ had a very thin shell, as shown by its state of preservation, and was in that respect intermediate between the trilobite and the theoretical ancestor. Every analysis of the cephalon of the trilobite shows that it is made up of several segments, certainly five, probably six, possibly seven. Every study of the trilobite, whether of adult, young, or protaspis, indicates the primitiveness of the lateral extensions or pleural lobes. The same studies indicate as clearly the location of the vital organs along the median lobe. These suggestions all point to a soft-bodied, depressed animal composed of few segments, probably with simple marginal eyes, a mouth beneath the anterior margin, tactile organs at one or both ends, with an oval shape, and a straight narrow gut running from anterior mouth to terminal anus. The broad flat shape gives great buoyancy and is frequently developed in the plankton. Inherited by the trilobites, it proved of great use to the swimmers among them.

The known animal which most nearly approaches the form which I should expect the remote ancestor of the trilobites to have had is _Amiskwia sagittiformis_ Walcott (Smithson. Misc. Coll., vol. 57, 1911, p. 112, pl. 22, figs. 3, 4). This "worm" from the Middle Cambrian is similar in outline to the recent _Spadella_, and is referred by Walcott to the Chætognatha. It has a pair of lateral expansions and a flattened caudal fin, a narrow median alimentary canal, and a pair of rather long simple tentacles. With the exception of a thin septum back of the head, no traces of segmentation are shown.

Some time in the late pre-Cambrian, the pre-trilobite, which probably swam by rhythmic undulations of the body, began to come into occasional contact with a substratum, and two things happened: symmetrically placed, i. e., paired, appendages began to develop on the contact surface, and a test on the dorsal side. The first use of the appendages may have been in pushing food forward to the mouth, and for the greater convenience in catching such material, a fold in front of the mouth may have elongated to form the prototype of the hypostoma. At this time the substratum may not have been the ocean bottom at all, but the animals, still free swimmers, may have alighted at feeding time on floating algæ from the surface of which they collected their food. While the dorsal test was originally jointed at every segment, the undulatory mode of swimming seems to have given way to the method of sculling by means of the posterior end only, or by the use of the appendages, and the anterior segments early became fused together.

The result of the hardening of the dorsal test was of course to reduce to that extent the area available for respiration, and this function was now transferred in part to the limbs, which bifurcated, one branch continuing the food-gathering process and the other becoming a gill. The next step may have been the "discovery" of the ocean bottom and the tapping of an hitherto unexploited supply of food. Upon this, there set in those adaptations to a crawling mode of existence which are so well shown in the trilobite. The crawling legs became lengthened and took on a hardened test, the hypostoma was greatly elongated, pushing the mouth backward, and new segments were added to produce a long worm-like form which could adapt itself to the inequalities of the bottom. That the test of the appendages became hardened later than that of the body is shown by the specimens of Neolenus, in which the dorsal shell as preserved in the shale is thick and solid, while the test of the appendages is a mere film.

The late Proterozoic or very earliest Cambrian was probably the time of the great splitting up into groups. The first development seems to have been among the trilobites themselves, the Hypoparia giving rise to two groups with compound eyes, first the Opisthoparia and later the Proparia. About this same time the Copepoda may have split off from the Hypoparia, continuing in the pelagic habitat. At first, most of the trilobites seem to have led a crawling existence, but about Middle Cambrian time they began to go back partially to the ancestral swimming habits, and retained some of the trunk segments to form a larger pygidium. The functional importance of the pygidium explains why it can not be used successfully in making major divisions in classification. Nearly related trilobites may be adapted to diverse methods of life.

EVOLUTION WITHIN THE CRUSTACEA.

The question naturally arises as to whether the higher Crustacea were derived from some one trilobite, or whether the different groups have been developed independently from different stocks. The opinion that all other crustaceans could have been derived from an _Apus_-like form has been rather generally held in recent years, but Carpenter (1903, p. 334) has shown that the leptostracan, _Nebalia_, is really a more primitive animal than _Apus_. He has pointed out that in Leptostraca the thorax bears eight pairs of simple limbs with lamelliform exopodites and segmented endopodites, while the abdomen of eight segments has six pairs of pleopods and a pair of furcal processes, so that only one segment is limbless. Contrasted with this are the crowded and complicated limbs of the anterior part of the trunk of _Apus_, and the appendage-less condition of the hinder portion. Further, a comparison between the appendages of the head of _Nebalia_ and those of _Apus_ shows that the former are the more primitive. The antennules of Nebalia are elongate, those of _Apus_ greatly reduced; the mandible of _Nebalia_ has a long endopodite, and Carpenter points out that from it either the malacostracan mandible with a reduced endopodite or the branchiopodan mandible with none could be derived, but that the former could not have arisen from the latter. The maxillæ of _Apus_ are also much the more specialized and reduced.

_Nebalia_ being in all else more primitive than _Apus_, it follows that the numerous abdominal segments of the latter may well have arisen by the multiplication of an originally moderate number, and the last trace of primitiveness disappears.

It is now possible to add to the results obtained from comparative morphology the testimony of palæontology, already outlined above, and since the two are in agreement, it must be admitted that the modern Branchiopoda are really highly specialized.

As has already been pointed out, _Hymenocaris_, the leptostracan of the Middle Cambrian, has very much the same sort of appendages as the Branchiopoda of the same age, both being of the trilobite type. Which is the more primitive, and was one derived from the other?

The Branchiopoda were much more abundant and much more highly diversified in Cambrian times than were the Leptostraca, and, therefore, are probably older. Some of the Cambrian branchiopods were without a carapace, and some were sessile-eyed. These were more trilobite-like than Hymenocaris. Many of the Cambrian branchiopods had developed a bivalved carapace, though not so large a one as that of the primitive Leptostraca. The present indications are, therefore, that the Branchiopoda are really older than the Leptostraca, and also that the latter were derived from them. It seems very generally agreed that the Malacostraca are descended from the Leptostraca, and the fossils of the Pennsylvanian supply a number of links in the chain of descent. Thus, _Pygocephalus cooperi_, with its brood pouches, is believed by Calman (1909, p. 181) to stand at the base of the Peracaridan series of orders, and _Uronectes_, _Palæocaris_, and the like are Palæozoic representatives of the Syncarida. Others of the Pennsylvanian species appear to tend in the direction of the Stomatopoda, whose true representatives have been found in the Jurassic. The Isopoda seem to be the only group of Malacostraca not readily connected up with the Leptostraca. Their depressed form, their sessile-eyes, and their antiquity all combine to indicate a separate origin for the group, and it has already been pointed out how readily they can be derived directly from the trilobite.

While the Copepoda seem to have been derived directly from the Hypoparia, the remainder of the Crustacea apparently branched off after the compound eyes became fully developed, unless, as seems entirely possible, compound eyes have been developed independently in various groups. Most Crustacea were derived from crawling trilobites (Lower Cambrian or pre-Cambrian Opisthoparia), for they lost the large pygidium, and also the major part of the pleural lobes. In all Crustacea, too, other than the Copepoda and Ostracoda, there is a tendency to lose the exopodites of the antennæ.

These modifications, which produced a considerable difference in the general appearance of the animal, are easily understood. As has been shown in previous pages, the trilobites themselves exhibit the degenerative effect on the anterior appendages of the backward movement of the mouth, and the transformation of a biramous appendage with an endobase into a uniramous antenna is a simple result of such a process. The feeding habits of the trilobites were peculiar and specialized, and it is natural that some members of the group should have broken away from them. In any progressive mode of browsing the hypostoma was a hindrance, so was soon gotten rid of, and the endobases not grouped around the mouth likewise became functionless. The chief factor in the development of the higher Crustacea seems to have been the pinching claw, by means of which food could be conveyed to the mouth. It had the same place in crustacean development that the opposable thumb is believed to have had in that of man.

An intermediate stage between the Trilobita and the higher Crustacea is at last exhibited to us by the wonderful, but unfortunately rather specialized _Marrella_, already described. It retains the hypostoma and the undifferentiated biramous appendages of the trilobite, but has uniramous antennæ, there are no endobases on the coxopodites of the thoracic appendages, the pygidium is reduced to a single segment, and the lateral lobes of the thorax are also much reduced. _Marrella_ is far from being the simplest of its group, but is the only example which survived even down to Middle Cambrian times of what was probably once an important series of species transitional between the trilobites and the higher Crustacea.

In this theory of the origin of the Crustacea from the Trilobita, the nauplius becomes explicable and points very definitely to the ancestor. According to Calman (1909, p. 23):

The typical nauplius has an oval unsegmented body and three pairs
of limbs, corresponding to the antennules, antennas, and mandibles
of the adult. The antennules are uniramous, the others biramous,
and all three pairs are used in swimming. The antennæ may have a
spiniform or hooked masticatory process at the base, and share with
the mandibles which have a similar process, the function of seizing
and masticating the food. The mouth is overhung by a large labrum
or upper lip and the integument of the dorsal surface of the body
forms a more or less definite dorsal shield. The paired eyes are as
yet wanting, but the median eye is large and conspicuous.

The large labrum or hypostoma, the biramous character of the appendages, especially of the antennæ, the functional gnathobases on the second and third appendages, and the oval unsegmented shield are all characteristics of the trilobites, and it is interesting to note that all nauplii have the free-swimming habit.

The effect of inheritance and modification through millions of generations is also shown in the nauplius, but rather less than would be expected. The most important modification is the temporary suppression of the posterior pairs of appendages of the head, so that they are generally developed later than the thoracic limbs. The median or nauplius eye has not yet been found in trilobites, and if it is, as it appears to be, a specialized eye, it has probably arisen since the later Crustacea passed the trilobite stage in their phylogeny.

The oldest Crustacea, other than trilobites, so far known are the Branchiopoda and Phyllocarida described by Walcott and discussed above. It is important to note that while the former have already achieved such modified characteristics that they have been referred to modern orders, they retain the trilobite-like limbs and some of them still have well developed pleural lobes.

Calman (1909, p. 101) says of the Copepoda:

On the hypothesis that the nauplius represents the ancestral type
of the Crustacea, the Eucopepoda would be regarded as the most
primitive existing members of the class, retaining as they do,
naupliar characters in the form of the first three pairs of
appendages and in the absence of paired eyes and of a shell-fold.
As already indicated, however, it is much more probable that they
are to be regarded as a specialized and in some respects degenerate
group which, while retaining, in some cases, a very primitive
structure of the cephalic appendages, has diverged from the
ancestral stock in the reduction of the number of somites, the loss
of the paired eyes and the shell-fold, and the simplified form of
the trunk-limbs.

If the Eucopepoda be viewed in the light of the theory of descent here suggested, it is at once seen that while they are modified and specialized, they more nearly approximate the hypothetical ancestor than any other living Crustacea. Compound eyes are absent, and it can not be proved that they were ever present, although Grobben is said to have observed rudiments of them in the development of _Calanus_. The "simplified limbs" are the simple limbs of the trilobite, somewhat modified. The absence of the shell-fold and carapace is certainly a primitive characteristic. Add to this the direct development of the small number of segments, and the infolded pleural lobes, and it must be admitted that the group presents more trilobite-like characteristics than any other. It seems very likely that the primitive features were retained because of the pelagic habitat of a large part of the group.

Ruedemann (Proc. Nat. Acad. Sci., vol. 4, 1918, p. 382, pl.) has recently outlined a possible method of derivation of the acorn barnacles from the phyllocarids. Starting from a recent _Balanus_ with rostrum and carina separated by two pairs of lateralia, he traces back through _Calophragmus_ with three pairs of lateralia to _Protobalanus_ of the Devonian with five pairs. Still older is the newly discovered _Eobalanus_ of the upper Ordovician, which also has five pairs of lateralia but the middle pair is reversed, so that when the lateralia of each side are fitted together, they form a pair of shields like those of _Rhinocaris_, separated by the rostrum and carina, which are supposed to be homologous with the rostrum and dorsal plate of the Phyllocarida. Ruedemann suggests that the ancestral phyllocarid attached itself by the head, dorsal side downward, and the lateralia were developed from the two valves of the carapace during its upward migration, to protect the ventral side exposed in the new position.

This theory is very ingenious, but has not been fully published at the time of writing, and it seems very doubtful if it can be sustained.

_Summary._

The salient points in the preceding discussion should be disentangled from their setting and put forward in a brief summary.

It is argued that the ancestral arthropod was a short and wide pelagic animal of few segments, which so far changed its habits as to settle upon a substratum. As a result of change in feeding habits, appendages were developed, and, due perhaps to physiological change induced by changed food, a shell was secreted on the dorsal surface, covering the whole body. Such a shell need not have been segmented, and, in fact, the stiffer the shell, the more reason for development of the appendages. Activity as a swimming and crawling animal tended to break up the dorsal test into segments corresponding to those of the soft parts, and, by adaptation, a floating animal became a crawling one, with consequent change from a form like that of _Naraoia_ to one like _Pædeumias_. (See figs. 36-40.) A continuation of this line of development by breaking up and loss of the dorsal test led through forms similar to _Marrella_ to the Branchiopoda of the Cambrian, in which not only is there great reduction in the test, but also loss of appendages. The origin of the carapace is still obscure, but Bernard (1892, p. 214, fig. 48) has already pointed out that some trilobites, Acidaspidæ particularly, have backward projecting spines on the posterior margin of the cephalon, which suggest the possibility of the production of such a shield, and in _Marrella_ such spines are so extravagantly developed as almost to confirm the probability of such origin. In this line of development two pairs of tactile antennæ were produced, while the anomomeristic character of the trilobite was retained. From similar opisthoparian ancestors there were, however, derived primitive Malacostraca retaining biramous antennæ, but with a carapace and reduced pleural lobes and pygidium. From this offshoot were probably derived the Ostracoda, the Cirripedia, and the various orders of the Malacostraca, with the possible exception of the Isopoda. I have suggested independent origins of the Copepoda and Isopoda, but realize the weighty arguments which can be adduced against such an interpretation.

It is customary to speak of the Crustacea and Trilobita as having had a common ancestry, rather than the former being in direct line of descent from the latter, but when it can be shown that the higher Crustacea are all derivable from the Trilobita, and that they possess no characteristics which need have been inherited from any other source than that group, it seems needless to postulate the evolution of the same organs along two lines of development.

I can not go into the question of which are more primitive, sessile or stalked eyes, but considering the various types found among the trilobites, one can but feel that the stalked eyes are not the most simple. While no trilobite had movable stalked eyes, it is possible to homologize free cheeks with such structures. They always bear the visual surface, and, in certain trilobites (_Cyclopyge_), the entire cheek is broken up into lenses. Since a free cheek is a separate entity, it is conceivable that it might lie modified into a movable organ.

EVOLUTION OF THE MEROSTOMATA.

It has been pointed out above that the Limulava (_Sidneyia_, _Amiella_, _Emeraldella_) have certain characteristics in common with the trilobites on the one hand and the Eurypterida on the other. These relationships have been emphasized by Walcott, who derives the Eurypterida through the Limulava and the Aglaspina from the Trilobita. The Limulava may be derived from the Trilobita, but indicate a line somewhat different from that of the remainder of the Crustacea. In this line the second cephalic appendages do not become antennæ and the axial lobe seems to broaden out, so that the pleural lobes become an integral part of the body. As in the modern Crustacea, the pygidium is reduced to the anal plate, and this grows out into a spine-like telson.

From the Limulava to the Eurypterida is a long leap, and before it can be made without danger, many intermediate steps must be placed in position. The direct ancestor of the Eurypterida is certainly not to be seen in the highly specialized _Sidneyia_, and probably not in _Emeraldella_, but it might be sought in a related form with a few more segments. The few species now known do suggest the beginning of a grouping of appendages about the mouth, a suppression of appendages on the abdomen, and a development of gills on the thorax only. Further than that the route is uncertain.

Clarke and Ruedemann, whose recent extensive studies give their opinion much weight, seem fully convinced that the Merostomata could not have been derived from the Trilobita, but are rather inclined to agree with Bernard that the arachnids and the crustaceans were derived independently from similar chætopod annelids (1912, p. 148).

The greater part of their work was, however, finished before 1910, and although they refer to Walcott's description of the Limulava (1911), they did not have the advantage of studying the wonderful series of Crustacea described by him in 1912. While the evidence is far from clear, it would appear that the discovery of animals with the form of Limiting and the eurypterids and the appendages of trilobites means something more than descent from similar ancestors. Biramous limbs of the type found in the trilobites would probably not be evolved independently on two lines, even if the ancestral stocks were of the same blood.

The Aglaspidæ, as represented by _Molaria_ and _Habelia_ in the Middle Cambrian, are quite obvious closely related to the trilobites easily derived from them, and retain numerous of their characteristics. That they are not trilobites is, however, shown by the presence of two pairs of antennæ, the absence of facial sutures, and the possession of a spine-like telson.

The Aglaspidæ have always been placed in the Merostomata, and nearer the Limulidæ than the Eurypterida. The discovery of appendages does not at all tend to strengthen that view, but indicates rather that they are true Crustacea which have not given rise to any group now known. The exterior form is, however, _Limulus_-like, and since it is known from ontogeny that the ancestor of that genus was an animal with free body segments, there is still a temptation to try to see in the Aglaspidæ the progenitors of the limulids.

The oldest known _Limulus_-like animal other than the Aglaspidæ is _Neolimulus falcatus_ Woodward (Geol. Mag., dec. 1, vol. 5, 1868, p. 1, pl. 1, fig. 1). The structure of the head of this animal is typically limuloid, with simple and compound eyes and even the ophthalmic ridges. Yet, curiously enough, it shows what in a trilobite would be considered the posterior half of the facial suture, running from the eye to the genal angle. The body is composed of eight free segments with the posterior end missing. _Belinurus_, from the Mississippian and Pennsylvanian, has a sort of pygidium, the posterior three segments being fused together, and _Prestwichia_ of the Pennsylvanian has all the segments of the abdomen fused together. So far as form goes, a very good series of stages can be selected, from the Aglaspidæ of the Cambrian through _Neolimulus_ to the Belinuridæ of the late Palæozoic and the Limulidæ of the Mesozoic to recent. Without much more knowledge of the appendages than is now available, it would be quite impossible to defend such a line. It is, however, suggestive.

EVOLUTION OF THE "TRACHEATA."

The trilobites were such abundant and highly variable animals, adapting themselves to various methods of life in the sea, that it appears highly probably that some of them may have become adapted to life on the land. The ancestors of the Chilopoda, Diplopoda, and Insecta appear to have been air-breathing animals as early as the Cambrian, or at latest, the Ordovician. Since absolutely nothing is yet known of the land or even of the fresh-water life of those periods, nothing can now be proved.

In discussing the relationship of the trilobites to the various tracheate animals, I have pointed out such palæontologic evidence as I have been able to gather. Studies in the field of comparative morphology do not fall within my province. I only hope to have made the structure of the trilobite a little more accessible to the student of phylogenies.

SUMMARY ON LINES OF DESCENT.

In order to put into graphic and concise form the suggestions made above, it is necessary to define and give names to some of the groups outlined. The hypothetical ancestor need not be included in the classification and for reasons of convenience may be referred to merely as the Protostracean.

The group of free-swimming trilobites without thoracic segments was probably a large one, and within it there were doubtless considerable variations and numerous adaptations. While the only known animal which could possibly be referred to this group, _Naraoia_, is blind, it is entirely possible that other species had eyes, and that the cephala and pygidia were variously modified. For this reason and because of the lack of all thoracic segments, it seems better to erect a new order rather than merely a family for the group, and _Nektaspia_ (swimming shields) may be suggested. The only known family is Naraoidæ Walcott, which must be redefined.

_Marrella_ and _Habelia_ are types of Crustacea which can neither be placed in the Trilobita nor in any of the established subclasses of the Eucrustacea. They represent a transitional group, the members of which are, so far as known, adapted to the crawling mode of life, though it may prove that there are also swimmers which can be classified with them. To this subclass the name _Haplopoda_ may be applied, the feet being simple.

The two known families, Marrellidæ Walcott and Aglaspidæ Clarke, belong to different orders, the second having already the name Aglaspina Walcott. The name _Marrellina_ may therefore be used for the other.

For _Sidneyia_, Walcott proposed the new subordinal name Limulava, placing it under the Eurypterida. While _Sidneyia_, _Emeraldella_, and _Amiella_ may belong to the group that gave rise to the Eurypterida, they are themselves Crustacea, and a place must be found for them in that group. The possession of only one pair of antennæ prevents their reception by the Haplopoda, and allies them to the Trilobita, but the modifications of the trunk and its appendages keep them out of that subclass, and a new one has to be erected for them. This may be known as the _Xenopoda_, in allusion to the strange appendages of _Sidneyia_.

_Synopsis._

Class Crustacea.

Subclass Trilobita Walch.

Crustacea with one pair of uniramous antennæ, and possessing facial sutures.

Order Nektaspia nov.

Trilobita without thoracic segments. Cephala and pygidia simple.

Family Naraoidæ Walcott.

Cephalon and pygidium large, both shields nearly smooth. Eyes absent. A single species: _Naraoia compacta_ Walcott, Middle Cambrian, British Columbia.

Subclass Haplopoda nov.

Crustacea with trilobate form, two pairs of uniramous antennæ, no facial sutures, sessile compound eyes present or absent, pygidium and pleural lobes generally reduced, large labrum present, appendages of the trunk biramous.

Order Marrellina nov.

Form trilobite-like, pleural lobes reduced, endobases absent from coxopodites of body, pygidium a small plate.

Family Marrellidæ Walcott.

Cephalon with long genal and nuchal spines. Eyes marginal. A single species: _Marrella splendens_ Walcott, Middle Cambrian, British Columbia.

Order Aglaspina Walcott.

Body trilobite-like, with few thoracic segments, and a spine-like telson. Appendages biramous.

Family Aglaspidæ Clarke.

Cephalon trilobate, with or without compound eyes, seven or eight segments in the thorax.

Genus _Aglaspis_ Hall.

Compound eyes present, seven segments in thorax. Upper Cambrian, Wisconsin.

Genus _Molaria_ Walcott.

Compound eyes absent, eight segments in thorax. Middle Cambrian, British Columbia.

Genus _Habelia_ Walcott.

Compound eyes absent. Not yet fully described. Middle Cambrian, British Columbia.

Subclass Xenopoda nov.

Crustacea with more or less eurypterid-like form, one pair of uniramous antennæ, biramous appendages on anterior part of trunk, modified endopodites on cephalon.

Order Limulava Walcott.

Cephalon with lateral or marginal eyes and large epistoma. Body with eleven free segments and a telson. Cephalic appendages grouped about the mouth.

Family Sidneyidæ Walcott.

Trunk probably with exopodites only, and without appendages on the last two segments. Telson with a pair of lateral swimmerets.

Genus _Sidneyia_ Walcott.

Third cephalic appendage a large compound claw. Gnathobases forming strong jaws. Middle Cambrian, British Columbia.

Genus _Amiella_ Walcott.

Middle Cambrian, British Columbia.

Family Emeraldellidæ nov.

Trunk with biramous appendages in anterior part, and appendages on all segments except possibly the spine-like telson.

Genus _Emeraldella_ Walcott.

Cephalic appendages simple spiniferous endopodites. Eyes unknown. Middle Cambrian, British Columbia.

Final Summary.

It is generally believed that the Arthropoda constitute a natural, monophyletic group. The data assembled in the preceding pages indicate that the other Arthropoda were derived directly or indirectly from the Trilobita because:

(1) the trilobites are the oldest known arthropods;

(2) the trilobites of all formations show great variation in the number of trunk segments, but with a tendency for the number to become fixed in each genus;

(3) the trilobites have a constant number of segments in the head;

(4) the position of the mouth is variable, so that either the Crustacea or the Arachnida could be derived from the trilobites;

(5) the trilobite type of appendage is found, in vestigial form at least, throughout the Arthropoda;

(6) the appendages of all other Arthropoda are of forms which could have been derived from those of trilobites;

(7) the appendages of trilobites are the simplest known among the Arthropoda;

(8) the trilobites show practically all known kinds of sessile arthropodan eyes, simple, compound, and aggregate;

(9) the apparent specializations of trilobites, large pleural lobes and pygidia, are primitive, and both suffer reduction within the group.

The ancestor of the trilobite is believed to have been a soft-bodied, free-swimming, flat, blind or nearly blind animal of few segments, because:

(a) the form of both adult and embryo is of a type more adapted for floating than crawling;

(b) the large pygidium is shown by ontogeny to be primitive, and the elongate worm-like form secondary;

(c) the history of the trilobites shows a considerable increase in the average number of segments in successive periods from the Cambrian to the Permian;

(d) the simplest trilobites are nearly or quite blind.