THE BATRACHIA OF NORTH AMERICA. 123 



of 1.8'"'", which are without gills, it is a very miaute tubercle on the 

 outer metatarsus. In a little earlier stage it cau not but be wanting, 

 though this I have not seen, and I have little doubt that it is then a 

 Hemidactylium, unless, indeed, the parietal bones be not ossified. 



Another foetal condition rendered permanent is seen in the generic 

 character of the genus CEdipus, which differs from Spelerpes solely in 

 the fcetal non-separation of the digits which continues even after the 

 bones of the digits have been developed. In the larva of Sp. ruber the 

 digits are early entirely distinct, so that so far as this species is con- 

 cerned CEdipus presents an inexact parallelism, but they are also 

 more distinct than in the mature Spelerpes hellii, where, as might be sup- 

 loosed, the foetal union is delayed to maturity in other respects, as iu a 

 specimen from Orizaba, Mexico, of 16.5 lines in length. There the union 

 is about as e3:tensive as in CEdipus morio. In the young of Thorius 

 pemiatulus, the digits are not distinguished in specimens of .60 of the 

 full size, and otherwise entirely mature. In the adult they are distinct 

 for half their length. The digits in the young larva of GyrinopMlus 

 porphyriticus are as distinct as in those of Spelerpes rubra. In one exam- 

 ple I find the simple foot of earlier stages retained, resembling exactly 

 that of (Edipus, excepting that there are emarginations for but three 

 toes instead of five. Genera which have no premaxillary fontanelle 

 at maturity have it in the larval stage. Finally, closely allied genera, 

 which only differ in the degrees of ossi^cation of the parietal and pala- 

 tine bones, represent simply the relation between undeveloped and 

 developed conditions of the same form. 



The relations of the genera may be expressed as follows: Those of 

 the first or Plethodontine section are related to those of the second or 

 Spelerpine by an inexact parallelism^ excepting Autodax, whose pecul- 

 iarities exclude it from the comparison. Those in each section differing 

 in the union or separation of the premaxillary bones are related in 

 the same way to each other. The nearly allied genera in the Plethodon- 

 tine group are Hemidactylium and Plethodon, and Batrachoseps and 

 Stereochilus. In the first case we have only inexact parallelism, because 

 while Plethodon has the four digits of Hemidactylium, its parietal 

 bones are unossified, though an acceleration of development in these 

 respects would render the relations one of exact parallelism. This is 

 . apparently the relation between Batrachoseps and Stereochilus, for 

 with the foetal digits of the latter the former preserves also its foetal 

 cranium. It only remains to ascertain whether Stereochilus loses its 

 branchiae before or after acquiring the normal number of digits. From 

 the very small size of one at least of these the former case seems prob- 

 able, but I have not yet been able to prove it by direct observation. 

 Should it be so, we would have a case of exact parallelism. 



In the Spelerpine group the relation between Geotriton and Gyri- 

 nophilus is again one of inexact parallelism, since when the digits of 

 the latter are only separated at the extremities, as in the former, the 



