Herrick, Nerves of Silurotd Fishes. 223 
pore of the main line, when this union does not occur). These 
organs are innervated from the glossopharyngeus and corre- 
spond exactly with the middle pit line of Amia (Fig. 1, m. /). 
Regarding the structure of the canal organs I have noth- 
ing to add to the many previous descriptions of these organs. 
Those of Ameiurus conform to the usual teleostean type, such 
as CoLE figures for Gadus ('98, Fig. 4) and I have figured for 
-Menidia (00, Fig. 416). A very unsatisfactory figure of a 
canal organ of Ameiurus catus is given by WricuT (’84, Fig 7, 
Plate I) and the late Dr. BunKER (’97) thoroughly investigated 
the nerve terminations in the lateral canal organ of Ameiurus 
nebulosus (A. catus) finding the nerve fibers to terminate in 
free basket-like arborizations around the pear cells, just as in the 
sense organs of the internal ear. 
2. Comparison of lateral line canals of other siluroids. - 
Review of the Literature.—A comparison of the arrange- 
ments just described with PoLtarp’s account (’92) of African 
and South American siluroids, notably Clarias and Auchenaspis, 
shows a general agreement which is rather close in matters of 
morphological importance, though there is considerable variation 
in details. All of these forms agree also quite closely (except 
in the operculo-mandibular canal) with Chaetostomus, a heavily 
armored and probably very ancient form. This gives POLLARD 
ground for the assumption that the siluroid type of lateral line 
is exceedingly primitive, a conclusion which he supports by a 
comparison with PanpEr’s figure of the Devonian form Coccos- 
teus. In all of Portarp’s cases (Clarias, Auchenaspis, 
Chaetostomus, Callichthys, Trichomycterus) the first organ of 
the main line is supplied by a r. oticus, arising separately from 
the ganglion, as in Ameiurus and the second organ by the IX 
nerve. 
POLLARD attempted to determine the homologies of each 
of the organs and pores among the five species of siluroids ex- 
amined by him and with Amia, as described by ALuis. It is 
an open question how far into detail such homologies can be 
pushed with profit. In particular, PorLarp’s suggested 
