MARYLAND GEOLOGICAL SURVEY 159 



Fork beds he did not look upon this as the direct ancestor of Mammul 

 americcmus, but considered the line of descent to have been through some 

 European or Asiatic species, like Mammut borsoni which is undeniably 

 a near relative of our common mastodon. 



It will be readily seen that we have much to learn in regard to the 

 species of American elephants and their derivation, but it should be 

 remembered that until the discoveries of Andrews and Beadnell in 

 Africa, we had no hint of the origin of the Proboscidians themselves, and 

 the energy with which paleontological research is being carried on in the 

 United States may at any time bring to light the evidence needed for the 

 solution of the foregoing problems. 



Teeth. — It is customary in noting differences between teeth of ele- 

 phants to give the number of ridges on a tooth, but this plan is unsatisfac- 

 tory, not only because perfect teeth are rare, but because it is frequently 

 impossible to tell the number of ridges concealed by a thick coating of ce- 

 ment. The number of ridges on the grinding surface changes continually 

 with the wear and displacement of the teeth, so that this is also of little 

 use for comparative purposes. For these reasons the number of ridges 

 to an inch has been taken for purposes of comparison, and as many teeth 

 as possible have been measured in order to get the best possible average. 

 In giving the characters of the species the average number of ridges in 

 10 inches has been taken because this is practically equal to 25 centi- 

 meters and makes a convenient standard of comparison. From the ma- 

 terial available, which includes a very considerable number of teeth in 

 the U. S. National Museum, the number of cross folds in this space, in 

 the upper molars, is as follows : Elephas imperalor, 12 ; Elephas columbi_ 

 18; Elephas primigenius, 24. The number of cross ridges in the lower 

 molars is slightly less than this. 



It is necessary to explain that the term ridge, as here used, means loop 

 or ellipse — the worn face of each enamel pocket forming a greatly elon- 

 gated ellipse. There are occasional abnormalities in the enamel plates 

 and in rare instances a single plate is intercalated between two complete 

 ellipses and may unite with the edge of an adjoining loop. The enamel 

 plates are also by no means equi-distant from one another throughout 



