64 BERTRAM G. SMITH 



fully discussed in a later section; for the present it will be suffi- 

 cient to call attention to one of its leading aspects. From exist- 

 ing urodeles we may select a series of forms illustrating all stages 

 in a transition from an aquatic to a terrestrial mode of life, or 

 vice versa. In which direction should the series be read? Or 

 have we stated the question incorrectly, and have the urodeles 

 reached their present condition, some from an aquatic, some from 

 a terrestrial ancestry? 



In studying this aspect of the phylogenetic problem our atten- 

 tion cannot fail to be attracted by Cryptobranchus. For here 

 we have a urodele whose entire life is spent in the water, charac- 

 terized by persistent gill slits, the most primitive brain (Osborn, 

 '88), and external fertilization (Smith, '07). On the other hand 

 Cryptobranchus is known to possess, deciduous external gills, 

 functional lungs, and a method of locomotion by crawling on the 

 bottom which suggests a former terrestrial habit. Is Crypto- 

 branchus primitively aquatic, or does it come down to us bearing 

 evidence of a former land-living existence? An answer to this 

 question would go far in advancing our knowledge of the phylog- 

 eny of the entire group. 



In the solution of our phylogenetic problem comparative anat- 

 omy, paleontology and embryology must work together. It is 

 the embryological evidence that has hitherto been most conspicu- 

 ously lacking. Notwithstanding the important position of the 

 aquatic urodeles, it is here that we find one of the widest gaps in 

 our knowledge of comparative embryology. Not only has the 

 development of Cryptobranchus allegheniensis remained unde- 

 scribed, but little or nothing is known concerning the embryology 

 of most of its near relatives. Very recently, it is true, consider- 

 able progress has been made in the study of the embryology of 

 Cryptobranchus japonicus, but part of this work was done on very 

 scanty material, and the field is by no means exhausted. Of the 

 development of Amphiuma and Siren practically nothing is known. 

 Some results have been obtained with special problems in the 

 development of Necturus, but the life history has not been covered 

 in a comprehensive manner. For a study of the ])hylogenetic 

 relations of these forms a knowledge of the development in its 



