102 THE VOYAGE OF H.M.S. CHALLENGER. 



with a iDrotoplasmic substance, often containing nuclei. No doubt it is the protoplasmic 

 epithelium which fills these latter canals. A nerve passes through them, and terminates 

 at the surface of the cuticle in what Dohrn {loc. cit., p. 38) calls a " Borstenapparat." 

 Dohrn never observed a single seta' at the end of these canals ; but always two or more 

 (sometimes even a rosette of eight or nine) together. This observation of Dohrn's 

 seems to be most accurate ; as a rule I found the integument of the species of Nymiohon 

 furnished with forked setae (PL XVI. figs 1, 2, 4, 7), one of the setae being often split 

 again (fig. 3) ; that of Pallene australiensis shows also forked setas ; the integument of 

 Ascorhynchus glaber is also furnished with double setae, which are here extremely small 

 and rudimentary. One of the species of Phoxichilidium (P. imtagonicum) shows a 

 combination of four or five (PI. XVI. fig. 17), whUe in another species {Phoxichilidium 

 pilosum) two combined setae are always observed. 



In Colossendeis setae are totally wanting ; and thus in this respect the genus 

 Ascorhynchus, with its rudimentary setae, stands between Nymphon and Colossendeis. 



According to Dohrn,^ this pore-canal, which terminates in these forked sette, " often " 

 takes its origin from one of the integumentary cavities, which he saw filled up with glands. 

 Dohrn therefore considers these setae as tactile organs, having probably the special 

 function of causing on irritation, by reflex action, the secretion of a poisonous fluid by 

 the glands, which are situated in the conical cavities of the integument. I think this 

 a very ingenious supposition, but I wish to state in opposition to it, that according to 

 what I have seen of the matter — (l) as a rule the cylindrical and narrow pore-canals do 

 not originate in the conical cavities ; and (2) that the occurrence of the glands in the 

 conical cavities is the exception, and that, as a rule, these cavities are empty or partly 

 filled up with a protoplasmic substance, nuclei, blood-corpuscles, &c. In the difi"erent 

 species of Nymphon it hardly ever happens that the pore-canal with the set« takes its 

 orio-in in a conical cavity ; in the species of Phoxichilidium it does not seem to be so rare 

 (PI. XVI. fig. 17), even in these species, however, it is by no means the rule. 



With regard to the form of the conical cavities, in the first place it must be men- 

 tioned, that they have a most regular conical shape in the difi"erent species of the 

 genus Nymphon, and also in some species of Pallene. In Phoxichilidium they are of a 

 more elongated form, and often a small lateral branch passes from the main canal near 

 the extremity (PI. XVI. fig. 1 7, h). The genus Colossendeis shows these cavities of a 

 much more irregular shape. As a rule every cavity is bifid, and terminates in two 

 narrow pore-canals (PL XVI. fig. 12, c). The cavities are usually almost quite filled 

 up with protoplasmic substance. In this genus I once observed distinct cells, with large 



1 Huxley calls " seta; " all the hair-like processes from the flue microscopic clo-svn to stout spines, which are found 

 on the outer surface of the cuticle (Crayfish, Loiulon, 1880, p. 197). I use the word here, and on the following pages, 

 in a much more restricted sense ; having already used the words hairs and spines for the integumeutaiy appen- 

 dages, I call " setas " those which I consider as being more pwticularly of a sensory nature. 



2 Loc. cit., p. 38. 



