Vol. 51.] POSITION OF THE TRILOBITES. 357 



dorsal skin of the leg passed proximally into the under surface of 

 the pleura, and, on the other hand, that the gnathobase was joined 

 to the body somewhat farther in towards the median line. 



In this connexion it is impossible to avoid noticing the remark- 

 able length and filiform character of these gnathobases on the 

 trunk-limbs of Triarihrus. If the appearances are not deceptive, it 

 looks as if some differentiation had already taken place between the 

 jaw-pieces of the limbs nearer the mouth and those of the limbs 

 farther off ; this supposition being further justified by the fusion of 

 a certain number of segments to form a head-region. The filiform 

 character of the gnathobases of the thorax requires further elucida- 

 tion. 



There appear to be no traces of gills on the limbs of Triarihrus. 

 Their absence seems to allow of the exopodites being attached 

 higher up on the limb than is the case in Apus, in which a well- 

 developed gill persists. In Limulus both gills and exopodites have 

 vanished from the walking-limbs, except from the last pair, which 

 have rudimentary exopodites. But in this case the gill portions 

 of the phyllopodan abdominal (pygidial) appendages are greatly 

 specialized. It is possible that, when the pygidial or the transi- 

 tional limbs of Triarihrus are further unravelled, gills may be dis- 

 covered. This expectation is justified by the presence of larval 

 phyllopodan limbs at the extreme end of the body, indistinguishable, 

 so far as their ventral edges go, from the larval limbs of Apus, 

 which we know had gills inserted on their dorsal edges. It is- 

 true that gills are not shown on the limbs figured by Claus and 

 copied by Beecher, 1 but they are quite distinct on the smallest 

 rudimentary limb on a specimen of Apus productus (' Apodidse/ 

 fig. 10). The presence of typical phyllopodan appendages necessi- 

 tates the assumption that gills have somewhere to be taken into 

 account. It seems hardly likely that they should have completely 

 vanished along the whole length of the body. Somewhere in the 

 series between the developed crawling-legs and the larval phyllo- 

 podan limbs, gills of some kind, however rudimentary, must surely 

 have existed. In some trilobites they may have been better deve- 

 loped than in Triarihrus, and Walcott's interpretation of his sections- 

 is, in this respect, by no means necessarily incorrect. 



Former presence of Seta; on the Appendages. — The appendages of 

 Triarihrus show very distinct traces of the former presence of setae. 

 Those, however, which are personally of most interest to me are 

 figured by Dr. Beecher as tufts of stiff setae on the dorsal edges of 

 the exopodites, near their attachments to the coxopodite on the 

 4th and 5th head-limbs. In view of the very primitive annelidan 

 characters of the limbs above insisted upon, these setae, allowing for 

 some shiftiugs of position, may well be the derivatives of the para- 

 podial setae of the chaetopodan ancestral form. In my endeavours to 



1 ' The Appendages of the Pygidium of Triarthrus, 1 Am. Journ. Sci. ser. 3, 

 vol. xlvii. (1894) pi. vii. fig. 4. 



