PENTALOPIIODON. 87 



phodori) Sivalensis, figured in the ' Fauna Antiqua', 1 presents 

 the first indication in that direction. In the ' intermediate 

 molars ' of this form, both upper and lower, besides the usual 

 anterior ' talon ' and four large ridges, there is a fifth ridge, 

 somewhat reduced in size, but exactly corresponding with 

 the other in form, composed of several large mammillary 

 tubercles, separated from the next ridge by a valley, and 

 throwing off an outlying tubercle, which reduces the valley, 

 as in M. (Tetralophodori) Arvemensis, to lateral gorges. This 

 fifth ridge is not a mere offset from, or subordinate appendage 

 to, the fourth ridge after the ordinary manner of a 'talon.' 

 It is supported directly by the last fang, and is separated, 

 both on the outer and inner sides, from the latter by the in- 

 tervening valley. In most of the species of Mastodon having 

 alternate rnarainillse, the hind ' talon ' in the upper molars 

 (and conversely in the lower) forms a crenulated ' bourrelet,' 

 which is given off from the inner posterior mammilla, de- 

 scending obliquely around the base of the outer division, and 

 generally more or less effaced by the pressure of the next 

 posterior molar during its progress forwards. In a fine 

 specimen of a penultimate upper molar of Mastodon Sivalensis, 

 which is now before me, the fifth ridge, although well de- 

 veloped and attaining the height of the fourth, bears no trace 

 of a ' talon ' appended to it ; while an antepenultimate lower> 

 which I have also before me, shows distinctly five ridges, the 

 last differing in no respect of complexity or development from 

 the others, except in being a little smaller, and it bears a 

 distinct crenulated adpressed ' talon ' appendage, having the 

 appearance of a terminal ' bourrelet.' 



In the preceding part, when discussing the conditions of 

 the ' ridge-formula ' in Trilophodon and Tetralophodon, it was 

 stated that while the penultimate milk molar always presents 

 one ridge less than the ' intermediate molars,' the last true 

 molar presents one ridge more. Conformably, the last true 

 molar in M. Sivalensis presents six ridges, besides the hind 

 ' talon,' thus maintaining throughout, so far as the dentition 

 is known, the numerical characters to be inferred from the 

 ridge-formula, as ascertained in Trilophodon and Tetralopho- 

 don." 1 I consider it sufficient, on the present occasion, to call 

 attention to this as a point of some interest and importance 

 in the systematic and palseontological relations of the Pro- 

 boscidean family, in reference to the indications they present 

 of an order of successive serial development, without entering 

 in detail upon the evidence in support of the view here taken. 

 That the species is a distinct form is abundantly borne out 



1 Op. cit. PI. xxxvi. figs. 1-6. (See vol. i. p. 467.— Ed.) 



2 Soe vol. i. PI. vii. fig. i. — Ed. 



