MASTODON TEETH FROM PERIM ISLAND. 
3—151 
partially blocked by accessory columns (a, h)} When worn it is quite evident that 
the first inner column of the present tooth would present a trefoil-shaped dentine 
islet as in the corresponding ridges of the partially worn teeth figured in the first 
volume. Owing to the absence in the second and third valleys of a distinct posterior 
accessory column {h), the second and third ridges would not present such perfect 
trefoils of dentine. The tooth is readily distinguished from the upper molars of M. 
latidens (vol. I., pi. XXXVIII.) by its higher columns and less open transverse 
valleys ; and is also distinguished from those of M. sivalensis by the want of a 
completely alternate arrangement of the inner and outer columns of the ridges, as well 
as bv the distinct trefoil which would be formed on the first inner column when worn 
■/ 
down. The tooth is totally unlike the second upper true molar of the trilophodont 
M. falconer i (vol. I., pi. XXXII., fig. 1), and is readily distinguished from that of M. 
pandionis {ibid, pi. XXXVA.) by its comparatively smooth enamel, and the more 
distinct trefoil which would be formed on the first inner column when worn. 
Seeing, therefore, that the specimen under consideration agrees in all general 
characters with the molars of M. perimensis and differs from those of all the other 
Siwalik species of mastodons, it may be pretty safely referred to the former species : 
its size shows that it is the penultimate tooth. 
Assuming this reference to be correct, it would appear that the tooth figured in 
plate XLI., fig. 4 of the first volume as m. 1 of M. perimensis cannot really belong to 
that species^ ; its columns being arranged very alternately, the valleys completely 
blocked, and the whole tooth of a narrower type than the present specimen. In all 
these respects the former tooth agrees with M. sivalensis (compare mm. 4 of that 
species represented in fig. 2 of the same plate of the first volume), and may probably 
be referred to that species, being an example of a tooth in which the fifth ridge is 
not developed. 
Upper milk-molars? — With the exception of a tooth figured in plate XLI., fig. 3 
of the first volume of this work, and provisionally considered as mm. 2 / the milk- 
molars of the present species have not hitherto been known. The tooth figured in 
plate XVII., figs. 3, 3a, of the present volume is in the collection of the British 
Museum, to which it was presented by the late Mr. C. Falconer, and its mineral 
condition leaves little doubt that it came from Perim Island : it is labelled by Dr. 
Falconer as the penultimate left upper milk-molar of Mastodqn latidens. Of the 
correctness of the serial position there is no doubt, and the tooth, which is almost 
untouched by wear, evidently belongs to a tetralophodont species, as it contains three 
ridges and fore-and-aft talons, and is too small to be mm. 4 of a trilophodont species. 
1 There are more minor accessory columns in the last molar figured in vol. I. than in the first ; but this can only be 
regarded as an individual variation. 
2 It was at first supposed that the present tooth belonged to the same species as the one represented in vol. I., pi. XLI., 
fig. 4, and the last ridge and talon {ta) were accordingly restored from that figure. This restoration is, therefore, erroneous. 
3 The teeth of the milk -molar and premolar series are enumerated according to the typical series. Thus the ante- 
penultimate=mm. 2 ; the penultimate=mm. 3 ; the ultimate=mm. 4. 
4 Termed nma. 1 in vol. I. 
