The absence of antennules at any stage in development allies _Limulus_ so closely with the Arachnida and separates it so far from the Trilobita that in recent years there has been a tendency to give up the attempt to prove a relationship between the merostomes and trilobites, especially since Clarke and Ruedemann, in their extensive study of the Eurypterida, found nothing to indicate the crustacean nature of that group. A new point of view is, however, presented by the curious _Sidneyia inexpectans_ and _Emeraldella brocki_ described by Walcott from the Middle Cambrian.

=Sidneyia inexpectans= Walcott.

Ill.u.s.trated: Walcott, Smithson. Misc. Coll., vol. 57, 1911, p. 21, pl. 2, fig. 1 (not figs. 2, 3); pls. 3-5; pl. 6, fig. 3; pl. 7, fig. 1.

The body of this animal is elongate, somewhat eurypterid-like, but with a broad telson supplied with lateral swimmerets. The cephalon is short, with lateral compound eyes. The trunk consists of eleven segments, the anterior nine of which are conspicuously wider than the two behind them, and the telson consists of a single elongate plate.

On the ventral side of the head there is a large hypostoma and five, pairs of appendages. The first pair are multisegmented antennules. The second pair have not been adequately described. The third are large, complex claws, and the fourth and fifth suggest broad, stocky endopodites. Broad gnathobases are attached to the c.o.xopodites of the third to fifth pairs of appendages and form very strong jaws.

The first nine segments of the thorax have one pair each of broad filiform branchial appendages, suggestive of the exopodites of trilobites, but no endopodites have been seen. The tenth and eleventh segments seem to lack appendages entirely.

=Emeraldella brocki= Walcott.

Ill.u.s.trated: _Sidneyia inexpectans_ Walcott _partim_, Smithson.

Misc. Coll., vol. 57, 1911, pl. 2, figs. 2, 3 (not fig. 1);--Ibid., 1912, p. 206, text fig. 10.

_Emeraldella brocki_ Walcott, Ibid., 1912, p. 203, pl. 30, fig. 2; text fig. 8;--Ibid., vol. 67, 1918, p. 118 (correction).

_Emeraldella_ has much the same shape as _Sidneyia_ and the same number of segments, but instead of a broad flat telson, it has a long _Limulus_-like spine. The cephalon is about as wide as long, and eyes have not yet been seen. The body consists of eleven segments and a telson (Walcott says twelve and a telson but shows only eleven in the figures). Nine of the segments, as in _Sidneyia_, are broad, the next two narrow.

The ventral side of the cephalon has a long hypostoma, and five pairs of appendages. The first pair are very long multi segmented antennules and the next four pairs seem to be rather slender, spiniferous, jointed endopodites. Whether or not gnathobases were present is not shown by the figures, but owing to the long hypostoma the appendages are grouped about the mouth. All the segments of the body, unless it were the telson, seem to have borne appendages. On the anterior end, they were clearly biramous (1912, p. 206, text fig. 10), and that they were present along the body is shown by figure 2, plate 30, 1912.

The present state of knowledge of both these peculiar animals leaves much to be desired. The indications are that the cephalic appendages are not biramous, and that only one pair of antennae, the first, are developed as tactile organs. The thoracic appendages of _Emeraldella_ are biramous, and also possibly those of _Sidneyia_. In the latter, the last two abdominal segments seem to have been without appendages, while in _Emeraldella_ at least one branch of each appendage, and possibly both, is retained.

These animals, which may be looked upon as the last survivors of an order of pre-Cambrian arthropods, have the appearance of an eurypterid, but their dominant characteristics are crustacean. The features which suggest the Eurypterida are: elongate, obovate, non-trilobate, tapering body; telson-like posterior segment; marginal, compound, sessile eyes; claw-like third cephalic appendages; and, more particularly, the general resemblance of the test to that of an eurypterid like _Strabops_. In form, _Sidneyia_ agrees with the theoretical prototype of the Eurypterida reconstructed by Clarke and Ruedemann (Mem. 14, N. Y. State Mus., vol. 1, 1912, p. 124) in its short wide head with marginal eyes, and its undifferentiated body.

There is, moreover, no differentiation of the postcephalic appendages.

The crustacean characteristics are seen in the presence of five, instead of six, pairs of appendages on the head, the first of which are multisegmented antennules, and in the biramous appendages on the body of _Emeraldella_. It should be noted that these latter are typically trilobitic, each consisting of an endopodite with six segments and a setiferous exopodite.

Clarke and Ruedemann (1912, p. 406) have discussed _Sidneyia_ briefly, and conclude:

It seems to us probable that the Limulava [_Sidneyia_ and _Amiella_] as described are not eurypterids but const.i.tute a primitive order, though exhibiting some remarkable adaptive features. This order possibly belongs to the Merostomata, but is distinctly allied to the crustaceans in such important characters as the structure of the legs and telson, and is therefore much generalized.

The specialization of _Sidneyia_ consists in the remarkable development of a highly complex claw on each of the third cephalic appendages, and in the compound tail-fin, built up of the last segment and one or more pairs of swimmerets. These two characteristics seem to preclude the possibility of deriving the eurypterids from _Sidneyia_ itself, but it seems entirely within reason that they may have been derived from another slightly less specialized member of the same order.

That _Sidneyia_ is descended from any known trilobite seems highly improbable, but that it was descended from the same ancestral stock as the trilobites is, I believe, indicated by the presence of five pairs of appendages on the cephalon and trilobitic legs on the abdomen.

=Molaria= and =Habelia.=

Other so-called Merostomata found by Walcott in the Middle Cambrian are the genera _Molaria_ and _Habelia_, both referred to the Cambrian family Aglaspidae. These genera seem to conform with _Aglaspis_ of the Upper Cambrian in having a trilobite-like cephalon without facial sutures, a trilobite-like thorax of a small but variable (7-12) number of segments, and a _Limulus_-like telson. Neither of them has yet been fully described or figured, but (Walcott 1912 A, p. 202) _Habelia_ appears to have five pairs of cephalic appendages, the first two pairs of which are multisegmented antennae. The thoracic appendages are likewise none too well known, but they appear to have been biramous.

The endopodites are better preserved than the exopodites, but in at least one specimen of _Molaria_ the exopodites are conspicuous.

If these genera are properly described and figured, their appendages are typically crustacean, and fundamentally in agreement with those of _Marrella_. The relation to the Trilobita is evidently close, the princ.i.p.al differences being the absence of facial sutures and the presence of true antennae. I am therefore transferring the Aglaspidae from the Merostomata to a new subcla.s.s under the Crustacea.

ARANEae.

The spiders have the head and thorax fused, the abdomen unsegmented except in the most primitive suborder, and so appear even less trilobite-like than the insects. The appendages likewise are highly specialized. The cephalothorax bears six pairs of appendages, the first of which are the pre-oral chelicerae, while behind the mouth are the pedipalpi and four pairs of ambulatory legs. The posterior pairs of walking legs belong to the thorax, but the anterior ones are to be h.o.m.ologized with the maxillae of Crustacea, so that the spiders are like the trilobites in having functional walking legs on the head.

The chief likenesses are, however, seen in the very young. On the germ band there appear a pair of buds in front of the rudiments of the chelicerae which later unite to form the rostrum of the adult. At the time these buds appear, the chelicerae are post-oral, but afterward move forward so that both rostrum and chelicerae are in front of the mouth. The rostrum is therefore the product of the union of the antennules, and the chelicerae are to be h.o.m.ologized with the antennae.

There seems to be some doubt about the h.o.m.ology of the pedipalps with the mandibles, as at least one investigator claims to have found rudiments of a segment between the one bearing the chelicerae and that with the pedipalps.

Jaworowski (Zool. Anzeiger, 1891, p. 173, fig. 4) has figured the pedipalp from the germ band of _Trochosa singoriensis_, and called attention to the fact that it consists of a c.o.xopodite and two segmented branches which may be interpreted as exopodite and endopodite. He designated as exopodite the longer branch which persists in the adult, but since the ambulatory legs of Crustacea are endopodites, that would seem a more likely interpretation. As the figure is drawn, the so-called endopodite would appear to spring from the proximal segment of the "exopodite." If the two terms were interchanged, the h.o.m.ology with the limb of the trilobite or other crustacean would be quite perfect.

In the young, the abdomen is segmented and the anterior segments develop limb-buds, the first pair of which become the lung books and the last two pairs the spinnerets of the adult. There seems to be some question about the number of segments. Montgomery (Jour. Morphology, vol. 20, 1909, p. 337). reviewing the literature, finds that from eight to twelve have been seen in front of the a.n.a.l segment. The number seem to vary with the species studied. This of course suggests connection with the anomomeristic trilobites.

The oldest true spiders are found in the Pennsylvanian, and several genera are now known. The head and thorax are fused completely, but the abdomen is distinctly segmented. Some of the Anthracomarti resemble the trilobites more closely than do the Araneae, as they lack the constriction between the cephalothorax and abdomen. The spiders of the Pennsylvanian have this constriction less perfectly developed than do modern Araneae, and occupy an intermediate position in this respect.

In the Anthracomarti, the pedipalpi are simple, pediform, and all the appendages have very much the appearance of the c.o.xopodites and endopodites of trilobites. Chelicerae are not known, and pleural lobes are well developed in this group. Anthracomarti have not yet been found in strata older than the Pennsylvanian, but they seem to be to a certain extent intermediate between true spiders and the marine arachnid.

Insecta.

Handlirsch (in several papers, most of which are collected in "Die Fossilen Insekten," 1908) has attempted to show that all the Arthropoda can be derived from the Trilobita, and has advocated the view that the Insecta sprang directly from that group, without the intervention of other tracheate stock. At first sight, this transformation seems almost an impossibility, and the view does not seem to have gained any great headway among entomologists in the fourteen years since it was first promulgated. If an adult trilobite be compared with an adult modern insect, few likenesses will be seen, but when the trilobite is stripped of its specializations and compared with the germ-band of a primitive insect, the theory begins to seem more possible.

Handlirsch really presented very little specific evidence in favor of his theory. In fact, one gets the impression that he has insisted on only two points. Firstly, that the most ancient known insects, the Palaeodictyoptera, were amphibious, and their larvae, which lived in water, were very like the adult. Secondly, that the wings of the Palaeodictyoptera probably worked vertically only, and the two main wings were h.o.m.ologous with rudimentary wing-like outgrowths on each segment of the body. These outgrowths have the appearance of, and might have been derived from, the pleural lobes of trilobites.

He figured (1908, p. 1305, fig. 7) a reconstructed larva of a palaeodictyopterid as having biramous limbs on each segment, but so far as I can find, this figure is purely schematic, for there seems to be no ill.u.s.tration or description of any such larva in the body of his work.

That the insects arose directly from aquatic animals is of course possible, and Handlirsch's first argument has considerable force. It may, however, be purely a chance that the oldest insects now known to us happen to be an amphibious tribe. The Palaeodictyoptera are not yet known to antedate the Pennsylvanian, but there can be no doubt that, insects existed long before that time, and the fact that their remains have not been found is good evidence that the pre-Pennsylvanian insects were not aquatic. Comstock, who has recently investigated the matter, does not believe that the Palaeodictyoptera were amphibious (The Wings of Insects, Ithaca, N. Y., 1918, p. 91).

The second argument, that wings arose from the pleural lobes of trilobites, is exceedingly weak. Where most fully set forth (1907, p.

157), he suggests that trilobites may occasionally have left the water, climbed a steep bank or a plant, and then glided back into their native element, taking advantage of the broad flat shape to make a comfortable and gentle descent! This sport apparently became so engaging that the animal tried experiments with flexible wing tips, eventually got the whole of the pleural lobes in a flexible condition, and selected those of the second and third thoracic segments for preservation, while discarding the remainder. The pleural lobes of trilobites are not only too firmly joined to the axial portion of the test to be easily transformed into movable organs, but they are structurally too unlike the veined wings of insects to make the suggestion of this derivation even worthy of consideration.

Tothill (1916) has recently reinvestigated the possible connection between insects, chilopods, and trilobites, and, from the early appearance of the spiracles in the young, came to the conclusion that the insects were derived from terrestrial animals. He suggested that they may have come through the chilopods from the trilobites. The hypothetical ancestor of the insects, as restored by Tothill from the evidence of embryology and comparative anatomy, is an animal more easily derived from the Chilopoda than from the Trilobita. Five pairs of appendages are present on the head, and the trunk is made up of fourteen similar segments, each with a pair of walking limbs and a pair of spiracles.

Only the maxillae and maxillulae are represented as biramous. If the ancestor of the Insecta was, as seems possible, tracheate, this fact alone would rule out the trilobites. Among tracheates, the Chilopoda are certainly more closely allied to the Insecta than are any other wingless forms. If the ancestors of the insects were not actually chilopods, they may have been chilopod-like, and there can be little doubt that both groups trace to the same stock.

As to the ancestry of the Chilopoda, it is probable that they had the same origin as the other Arthropoda. Tothill has pointed out that in the embryo of some chilopods there are rudiments of two pairs of antennae and that the two pairs of maxillae and the maxillipeds are biramous. This would point rather to the Haplopoda than directly to the trilobites as possible ancestors, and may explain why the former vanish so suddenly from the geological record after their brief appearance in the Middle Cambrian. They may have gone on to the land.

There seem to be no insuperable obstacles to prevent the derivation, indirectly, of the insects from some trilobite with numerous free segments, and small pygidium. The antennules and pleural lobes must be lost, the antennas and trunk limbs modified by loss of exopodites.

Wings and tracheae must be acquired.

Handlirsch places the date of origin of the Insecta rather late, just at the end of the Devonian and during the "Carboniferous." By that time most families of trilobites had died out, so that the possibilities of origin of new stocks were much diminished. If the haplopod-chilopod-insect line is a better approximation to the truth, then the divergence began in the Cambrian.

Chilopoda.

The adult chilopod lacks the antennules, and all of the other appendages, with the exception of the maxillulae, are uniramous. The walking legs are similar to the endopodites of trilobites, and usually have six or seven segments. The appendages are therefore such as could be derived by modification of those of trilobites by the almost complete loss of the exopodites and shortening of the endopodites of the head. The position of the post-oral appendages, the posterior ones outside those closest the mouth, is perhaps foreshadowed in the arrangement of those of Triarthrus.

The Chilopoda differ from the Hexapoda in developing the antennae instead of the antennules as tactile organs, but this can not be used with any great effect as an argument that the latter did not arise from the ancestors of the former, since it is entirely possible that in early Palaeozoic times the pre-Chilopoda possessed two pairs of antennae. The first pair are still recognizable in the embryo of certain species.