220 FROCEEDIXGS OF THE NATIOXAL MUSEU}r. vol.53. 



rating the parts of each lateral half of the crests would have been 

 hidden by a deposit of cement; but a specimen of a tooth of G. 

 foridanum in a similar stage of development has already a rather 

 thick layer of this. 



The outer cones of each crest are furnished with buttresses which 

 effectually close up the transverse valleys. These buttresses are 

 mostly composed of subsidiary conules, which cling closely either to 

 the parts of the principal cone or to one another. The row of conules 

 forming the anterior buttress of the first crest runs into the anterior 

 cingulum. The posterior buttress is composed of two conules, the 

 hinder of Avhich is applied to a conule which forms the anterior 

 buttress of the second crest. The hinder buttress of this second 

 crest, apparently of three or four flattened conules, joins the anterior 

 buttress of the third crest; but it appears to be carried also inward 

 and downward by a row of flattened conules between the inner halves 

 of the second and third crests. The posterior buttress of the third 

 crest is similarly disposed, as is also that of the fourth crest. The 

 hinder buttress of the fifth crest is less completely developed. 



On wear the outer halves of the three anterior crests would pro- 

 duce each a trefoil: the succeeding two would form less complete 

 ones. 



There is a tendency for the formation of buttresses on the cones of 

 the inner half of the tooth. In the first transverse valley the inner 

 buttresses are nearly as complete as the outer ones. In the second 

 transverse valley there is an accessory conule attached to the front 

 of the third crest, but none to the rear of the second crest ; nor are 

 there buttresses on the succeeding crests. At the inner end of the 

 first transverse valley there is a very large tubercle. There was a 

 similar one at the inner end of the fourth valley. 



The injuries done to the tooth make it difficult to speak with con- 

 fidence regarding the cingulum, but this appears to have been present 

 along the outer face of the tooth. At the outer end of each valley 

 there is a cluster of small tubercles, and these appear to have met 

 across the base of the outer cones. 



The tooth here described resembles somewhat the corresponding 

 one of Gomyliotlienum tropicum (Cope), of the Blanco Pliocene: 

 but in the McPherson tooth there are a greater subdivision of the 

 principal cones and a larger number of accessory conules in the 

 valleys. It is also not greatly unlike the corresponding tooth of 

 Gomphotheriiim foridanum^ likewise a Pliocene species. The latter, 

 however, very larely develops buttresses on both halves of the crests: 

 and it appears to have had properly only four cross crests and a 

 talon. Likewise, the principal cones of G. f^oridanum appear to be 

 more obtuse than those of G. elegans. 



