228 JOURNAL OF COMPARATIVE NEUROLOGY. 
are behind this pore three large pit organs. corresponding 
exactly to the line a. /. of the figure just referred to. 
The osseous support of the opercular canal is more com- 
plete than in A. melas. Between the mandible and the pre- 
operculum there is a rather long interval. The ventral end of 
the preoperculum slips under the quadrate, carrying the canal 
and causing it to appear as if the canal passed through the 
quadrate. Such is, however, not the case. From the dorsal 
end of the preoperculum the canal passes through a chain of 
three tubular ossicles, or supra-opercular bones, to the squamosal, 
within which it unites with the main canal. None of these 
tubular ossicles in my specimens was found to be joiued to 
the hyomadibular bone. The main canal between the squamosal 
and the post-temporal is wholly membranous, the accessory 
ossicle ESC. not being present. Behind the post-temporal 
there are several drainpipe-like ossicles, then the canal becomes 
incomplete, as CoLLiINGE describes. In my specimens I have 
been able to find no trace of an occipital commissure connect- 
ing the main canals of the two sides, such as COLLINGE figures, 
nor any other connection across the median line. On the sur- 
face, however, the middle pit line is obvious in the same rela- 
tions as in A. melas. 
A larger specimen 37 cm. long shows important variations. 
There are but two supra-opercular ossicles. These are both 
very long and the double pore (/ of main canal + /X of the 
operculo-mandibular) lies between them. The ossicle ESC. is 
present in exactly the same relations as shown by my sections 
of A. melas. It is a firmly ossified cancellous bone about 7 
mm. long, club-shaped, with the larger end fitting closely into 
the acute angle formed by the articulation of the dorsal limb of 
the post-temporal with the cranium. There is no appreciable 
interval between it and the cranium, but behind it the canal is 
unprotected for some 6 mm. before entering the post-temporal 
bone. Behind the latter there is but one tubular accessory 
ossicle and this is very thoroughly ossified. In this specimen, 
as in the smaller ones, there is no communication between the 
supra-orbital and infra-orbital canals near the nasal apertures, 
