482 Causes and Course of Organic Evolution 



mentally are only distinguished from the horny teeth of Petro- 

 myzon in that they fail to form odontoblasts and so form no 

 calcified tissue. Thus vomerine, palatine, mandibular, max- 

 illary and premaxillary teeth may be formed, or may again 

 disappear. Amongst Batrachia the writer inclines to consider 

 that a suctorial mouth with horny teeth originated first in 

 the group, and that these were afterwards shed without replace- 

 ment by calcified teeth as is still seen in Siren and in many 

 Anura, where in change from a vegetable to an animal diet 

 the tongue and the lips are used to seize and hold the prey. 

 But in the Apoda and simpler Urodela one to several sets of 

 teeth are formed, of which one row along the premaxilla and 

 maxilla and another along the dentary have survived through 

 the higher Apoda and Urodela to the Mammalia. 



So amongst urodeles the originally numerous and diffuse 

 teeth on the vomero-palatine and other bones tend to become 

 more and more restricted till they form a row along upper 

 and lower jaw. But in Spelerpes, Plethodon, Desmognathus, 

 and others there is a distinct tendency toward heterodont 

 tooth localization, in that there is a subdivision of the crown 

 of some teeth into two equal or subequal cusps. Such for- 

 mations, taken in conjunction with Cope's observations {73: 

 318) on the mechanical origin of incisor, canine, premolar, 

 and molar teeth, suggest that with increasing tendency to 

 life on land, with greater use then on hard objects and at differ- 

 ent angles, also with specialization in the nerves and muscles 

 of the upper and lower jaw, the heterodont mammalian type 

 might well have evolved from the simpler and mainly homodont 

 batrachian. 



But study of such mammalian groups as the marsupials, 

 the toothed whales, the rodents, and others show that even 

 amongst these the teeth may be homodont or some that are 

 usually present may never be formed, as in the canines that 

 are entirely absent in rodents. 



The cervical vertebrae in batrachians and mammals suggest 

 a rather profound difference in number and relations, that 

 at the present day would diagnose the two groups more easily 

 than any other. In all existing batrachians the number is 

 two, of which the most anterior unites with the occipital region 

 of the skull, while the second is recognizably absent in Apoda 

 but forms the odontoid process in urodeles. The succeeding 

 vertebrae bear ribs, but, if one define as cer\4cals all in front 



