130 SAMUEL WENDELL WILLISTON— LULL. [U ™ 01K {\™™vn, 



shown in the figure, there is not a single thing to differentiate the form from an am- 

 phibian, unless it be the apparent absence of the eleithrum. The palate is different from that of 

 other known reptiles, though distinctly reptilian in structure. Williston continues (p. 236) : 



The American cotylosaurs, more especially the Diadectidw, Limnoscelidse, and Seymouriidse, show marked resem- 

 blances in many ways to the contemporary amphibians, in their short legs, broad feet, enormous humeral entocondyle, 

 digital fossa of the femur, pronounced adductor crest, as well as girdles: but I do not believe that these resemblances 

 were so much the result of phylogeny as of convergent evolution, the adaptation to similar environmental conditions 

 and similar habits. 



In "A new family of reptiles from the Permian," 1911, 105 Williston describes briefly the 

 contents of a collection of Permian fossils from New Mexico in the Marsh material at Yale, 

 made more than 30 years before. This was found in beds equivalent stratigraphically to the 

 lower or Wichita division of Texas. Certain forms are either closely allied to or identical with 

 those of the Texas beds, others are quite different; this Williston attributes to differing environ- 

 mental conditions, since the identical or allied forms of New Mexico are from the red clays and 

 red sandstones which are quite like those of the Texas deposits, while the unlike forms are from 

 sandstones and clays unlike anything in the latter State. There is also a complete absence of 

 concretions and of fish remains in New Mexico. While a full discussion of the Yale material 

 was published in "American Permian Vertebrates," a remarkable new form, Limnoscelis, was 

 described somewhat in extenso, as it is not only the finest thing of the collection, but is one of 

 the most notable specimens of a reptile ever obtained from the Permian deposits of America. 

 There is more than one individual represented, so that parts lacking in one are present in another; 

 but one specimen is practically perfect and is in essentially complete articulation, the skeleton 

 lying as the animal died. Limnoscelis paludis is a cotylosaur, representing the new f amil y 

 Limnoscelidse, a large form, 7 feet over all, powerful, of carnivorous habits, but a subaquatic 

 or marsh-dwelling reptile, with limbs strongly suggestive of the turtles. The relationship lies 

 most closely with Diadectes and Pareiasaurus, which, together with Propappus also, may per- 

 haps be placed in the same order of reptiles. A restoration of Limnoscelis by Williston was 

 described in 1912. 107 



A joint paper by Williston and Case, in the same year, 108 treats of the allied skulls of Dia- 

 dectes and Animasaurus. Perhaps their most remarkable feature is the differentiated denti- 

 tion, with strong chisel-shaped incisors and curious transversely elongated cheek teeth. They 

 have usually been considered as herbivorous, but the character of the incisors, the absence of 

 any power of trituration in the unworn maxillary teeth, and the possibility of the use of the 

 palatine processes of the maxillaries as accessory organs of mastication lead to the suspicion 

 that the animals were not exclusively, if at all, herbivorous, and that they may have included 

 the less well-protected invertebrates in their diet. 



A very important paper entitled "Primitive Reptiles, a Review," also appeared in 1912. 108 

 This summarizes our knowledge of the Permocarboniferous tetrapods of the world, correlating 

 those characteristics upon which phylogenies and classification must depend. Of the foreign 

 forms, Williston had no autoptic knowledge, but he had studied personally nearly all of the 

 known American material and was thus in position to speak authoritatively. The characters 

 of these creatures are grouped under those which are constant and those which are inconstant 

 and variable. It is conceivable that with a more complete knowledge of certain genera some 

 of the constant characters will be transferred to the variable list, but unless error has been made 

 the reverse will not be possible. These characters which are enumerated are the chief ones 

 upon which we must depend, for the present at least, for the classification of the known Permo- 

 carboniferous reptiles. 



Williston expresses emphatically his lack of faith in the DeVreesian "mutation theory" of 

 the origin of species, nor does he believe that any paleontologist can defend such a theory. And 

 he does not consider that any theory of the origin of species, or even of evolution, can get very 

 far when time is left out of account. If, in any series of phylogenetic forms, we find a gradual 

 transformation of structure, or the gradual acquisition of new characters, we do right in uniting 



