GENERA OF THE ARCIFEROUS ANURA. 97 



3. Hy lodes, probably. 



Cranium broad, rounded ; digital dilatations supported by a strong 

 cross-limb ; no parotoids ; xiphisternum with a distal, entire carti- 

 lage ; vomerine teeth ; muzzle and canthus, not marked, . . Batrachyla. 



4. Hylodes or Cy.stignathi? 



Head normal ; vomerine teeth ; parotoid gland ; digits not webbed ; 



thumb of male spurred ; minute posterior digital dilatations, . Plectromantis. 



5. To ? Cystignathi or Ceratophrydes. 



A cavum tympani. Xiphisternum? Sacrum? Cuneiform bone 

 little developed ; form elongate ; toes free ; parotoid glands ; no 

 vomerine teeth ; tongue extensively free ; cranium probably com- 

 plete above pupil ?......... Nattereria. 



Supplement on the osseous structures of the Types of the Urodela. 



It is as yet not always possible to distinguish differing plans of structure from 

 differing degrees of development of a single plan. The assumption of the faculty of 

 reproduction and cessation of developement at any of the various stages through 

 which most vertebrates pass, would give rise to a variety of genera and species. 

 That the comparative characters of existing genera, etc., do in very many cases illus- 

 trate in part such a hypothesis, is well known ; and the fact, which has been laid down* 

 as a rule of zoology, that " every character distinguishing suborders, families, and 

 genera, will he found among the individuals of some species, living or extinct, to mark 

 mere varieties or stages of growth" \s also in its favor; but it is opposed by the fre- 

 quent existence of superadded structures, which are of the nature of adaptations to 

 external circumstances. Both the more comprehensive '" plans " or " groups " and 

 the subordinate "genera," difier from each other in ways which have the above 

 two kinds of significance separate or together. 



With reference to these propositions, and the ultimate problem of the origin of 

 species, no group presents greater facilities for investigation than the Urodele Batra- 

 chia. A few points are collated in the following pages, with this in view. 



The Batrachia Urodela have been regarded as naturally distinguished into those 

 with persistent, and those with temporary branchias, — perennibranchiata and caduci- 

 brancbiata, — by almost all naturalists. The feature has been regarded either as of 

 primary importance, as by Wagler and Buonaparte, or as secondary, as by Cuvier, 

 Duraeril, and Muller. The types known as Perennibranchiates are Siren, Proteus, 

 Necturus and Siredon. The skeletal features presented by these are so diverse 



* Proc. Acad. Philada. 1862, 75 : ou p. 66 (middle) "generic " should be read genetic. 



25 



