358 MR. H. M. BERNARD ON THE SYSTEMATIC [Aug. 1895, 



deduce Apus from a chaetopod I assumed the former presence of 

 such setse, and, finding the shell-gland opening on the dorsal branch 

 of the 5th head-limb, I suggested that this gland might be a de- 

 rivative of a setiparous gland. Here is not the place to enter upon 

 this rather difficult discussion, especially as I have gone into the 

 subject at length elsewhere. 1 I will only add that the discovery of 

 tufts of setae on the limbs of Triarthrus, which are in many respects 

 very primitive, in practically the very spot where, in Apus, I as- 

 sumed their former presence, may, I think, fairly be claimed as one 

 more slight confirmation of my general argument. 



Head-Segments. — Dr. Beecher corrects me in my suggestion that 

 Triarthrus may have had only four segments in its head-region. There 

 are five pairs of limbs apparently attached under the head-shield. 

 This would, however, make no difference to the general drift of mj r 

 argument that the gradual building-up of the crustacean head can 

 be traced in the trilobites. The gradual incorporation cf body- 

 segments, probably associated with the specialization and co-ordi- 

 nated movements of mouth-limbs, undoubtedly took place, and it 

 did not stop with the trilobites, but continued till in IAmulus and the 

 Eurypterids six segments fused to form a head-region. In the case 

 of IAmulus the whole of the rest of the body corresponds to a 

 pygidium with persistent phyllopodan appendages. In the Eury- 

 pterids, on the other hand, the posterior segments remained free, as 

 they were also in the earliest known trilobites (Olenellus). 



Phylogenetic Conclusions. — Summing up our comparison of the new 

 and important facts described by Dr. Beecher with the known con- 

 ditions of Apus, we find that the Crustacea can now be linked, 

 step by step, with the chaetopod annelids. The line of 

 development is practically that which I sketched in my former 

 paper, but we are now in a position to supplement that scheme by 

 additional details which bring it still closer to the actual order. 



The common ancestor of Apus and Triarthrus had a many-seg- 

 mented cylindrical body, the former still showing traces of more 

 than sixty segments. A large prostomium bent round towards 

 the ventral surface, so that the mouth opened downwards and back- 

 wards ; each segment except the first was provided with a pair of 

 swimming leaf-like appendages, richly provided with setee ; each such 

 appendage carried on its dorsal edge a gill and an exopodite. 

 The' appendages of the first segment were simply cirrus-like, and 

 they pointed downwards on each side of the ventrally-placed 

 prostomium. Such an animal is a typical chaetopod annelid, with 

 the first segment bent round ventrally. We gather from later 

 specializations that this bending was for the purpose of using the 

 parapodia nearest the mouth to push in the food. 



These free-swimming browsing annelids early developed pleurae 

 on a few of the most anterior segments. Such pleurae were formed 



1 ' Comparative Morphology of the G-aleodidfe,' § Excretion. Eead before 

 the Linnean Society, February 1895. (In the press.) 



