448 bulletin: museum of comparative zoology. 



This animal is in appearance far more like Salamandra than like 

 Triturus, as it has a long rounded tail, and is marked with large 

 yellow spots. The dentition, however, is that of Triturus. Perhaps 

 this animal had better remain as Cope described it, the type of Neurer- 

 gus. Here also may belong Molge macrosema Boulenger, whose 

 habitat is unknown. 



A very natural group is formed by vulgaris, pahnatus, italicus, 

 moniandonii, boscai, vittatus of Gray, and the North American forms 

 allied to mridescens. These are all similar in size, coloration, shape of 

 head, and even agree in the possession of three longitudinal grooves 

 on the head. I do not see how these animals can be divided into 

 two genera. It may here be noted that in this group are two forms 

 known as vitiaius and two known as meridionalis. 



This necessitates a renaming of two forms, a course as regrettable 

 as it is necessary. 



The other species, including the European cristatus, viarmoratus, 

 and alpestris, the Asiatic woltcrstoffi, sinensis, j^yrrhogastcr, and eyisi- 

 caudus, and the American iorosus are more or less problematical in 

 their relationships ; the last, torosus, seems closer to the Asiatics than 

 to the Europeans, just as viridesccns is far closer to the Europeans of 

 the vulgaris group than to torosus. 



If the possession of a dorsal crest by the breeding males and the 

 lack of a bony temporal arch are considered primitive characters, it 

 is quite apparent that these two characters are practically restricted 

 to Europe, where the Salamandridae reach their greatest development 

 in number of forms, and which is the geometrical centre of the family's 

 range (especially if we consider viridesccns as derived from Europe and 

 torosus from Asia). 



In this case certainly one does not find the primitive forms on the 

 periphery of the range. On the other hand, if the bony arch is regarded 

 as primitive and the dorsal crest of the breeding male as specialized, 

 we find the primitive forms on the periphery of the range. But if we 

 regard the bony arch as primitive, then all the other salamanders 

 have lost it. 



There are certainly strong reasons why the presence of the post- 

 fronto-squamosal arch should be considered primitive. In the first 

 place the idea is quite erroneous that the more larval a salamander, 

 the more primitive it is. Early amphibians as well as the Crossop- 

 terygian ganoids from which they were almost certainly derived had a 

 practically complete bony temporal roof. This roof was present in 

 the Cotylosaur reptiles and persists in its most complete modern state 



