72 PROCEEDINGS : BOSTON SOCIETY NATURAL HISTORY. 
system is therefore quite different from that of the Proteidae and is 
exactly like that of the larvae of the lungless Salamandrids, as the 
animal breathes by its external gills and probably to some extent by 
its skin. There are three external gill bushes on each side of the 
head, attached to the branchial arches, and three gill slits on each 
side, allowing the free passage of water through the pharynx. This 
arrangement of gill bushes and gill slits is closely similar to that 
found in the larval Spelerpes, but in the Proteidae the two pos¬ 
terior only of these gill slits appear. It is possible that the walls of 
the pharynx and esophagus may be also concerned in respiration, as 
is the case in the adult forms of Desmognathus and of Spelerpes as 
described by Miss Barrows (: 00 ) and Wilder (’ 96 ). 
Summary and Conclusion. 
Aside from its extraordinary proportions, which may be the 
result of its strange habitat and mode of life, as Stejneger suggests, 
it is evident from the preceding description of the various systems 
and their comparison with those of allied forms, that Typhlomolge 
in most of its structural peculiarities shows a marked resemblance to 
the Spelerpes larva. 1 It is equally true that these similarities are 
the very points in which it differs widely from the Proteidae. In 
color, in the possession of non-functional, subcutaneous eyes, and in 
slenderness of limb, Typhlomolge does resemble Proteus, but these 
common characteristics may be due to their similar surroundings 
and manner of life, and seem hardly fundamental enough to warrant 
the placing of both forms in the same family, especially when the 
very large number of important structural similarities to Spelerpes 
is considered. The following tabular summary may help to make 
these points clear. 
Spelerpes Larva and Typhlomolge. Necturus and Proteus. 
Number of gill slits, 3. Number of gill slits, 2. 
Number of fingers, 4. Number of fingers, 4 (N) ; 3 (P). 
Number of toes, 5. Number of toes, 4 (N) ; 2 (P). 
1 1 have compared it carefully with the Spelerpes ruber larva, and found the 
resemblance especially strong in that form. Specimens of the other species 
were not at hand, but from plates of Parker (' 82 ), Wiedersheim (’ 75 ), and 
others, the similarities are plainly apparent, though somewhat less marked. 
