Kincssury, Columella Aurts and N. Factalis. 323 
Turning to the larval form for an interpretation of the con- 
dition in the adult, we find in a specimen 21 mm. in length, 
that the suspensorio-opercular connection is at this stage cellu- 
lar and extends from the cephalic border of the operculum to 
the squamosum as a dense cord of cells. It has the same rela- 
Fig. 6.4'Same,’three sections (75 yz) farther forward; g.—cartilaginous (colum- 
ellar) process of the quadrate. 
tion to the jugular nerve and vein as in Spelerpes, though it 
does not come into as close contact with either as in that form. 
Its cephalic end is rather difficult to determine (Fig. 7), since 
the anlage is continued forward to join the subsquamosal pro- 
cess (of the quadrate) which at this stage is a distinct bone, so 
that it may be said to be connected with both bones. There 
is, however, no direct connection with the (cartilaginous) 
quadrate, and from} the conditions in Necturus and Spelerpes, 
we are warranted, I think, in emphasizing the connection with 
the squamosum rather than that with the subsquamosal process 
of the quadrate which, in fact, is not as direct. Ina larva 33 
