26l 
to suggest the more ancient elephants of India and Europe, such as 
E. planifrons, E. meridionalis, E. antiquus, and E. prisciis (FohHg, 
non Goldfuss). Indeed, this tooth might ahnost be described as inter- 
mediate in character between Pohlig's two groups, Arckidiscodonta 
and the Loxodonta. The third lower molar is very large, almost 
equalling in size that of the most gigantic modern African elephants, 
and in shape is elongate and rather narrow, though broader relatively 
than in E. africainis, so much so as to verge upon the laticoronate 
type. It will be observed from the table of measurements that while 
the breadth of crown equals the maximum recorded, even of the upper 
molars, for any individual of the existing African or Indian species, 
the length is considerably less than in these exceptionally gigantic 
individuals. The degree of abrasion is such that the height of the 
crown can be only estimated, but, judging from the posterior, unworn 
portion, this dimension was less than in E. africanus, though probably 
not sufficiently so to deserve the name Tapinodiscal. This is a dis- 
tinguishing feature from E. . planifrons and E. meridionalis. The 
tooth has a strong antero-posterior curvature, with the concavity 
turned outward and the convexity inward. 
In the anterior part of the tooth abrasion has progressed so far as 
completely to remove all traces of the talon and several laminee, and 
thus the ridge formula cannot be determined with certainty. It is 
plain, however, that the number of ridges, not including the talons, 
cannot have been less than 12, and may have been 13. There is 
some doubt as to the maximum number of laminte which occurs in 
E. africanus. Adams (1877, p. 35) says : " I can find no record nor" 
" discover any ultimate molar of E. africanus with a larger ridge formula " 
" than X13X; indeed, in far the greatest number of specimens it seldom " 
"exceeds xiix." Pohlig, on the other hand, questions this (1888, 
p. 242), and says : "The maximum, which is rarely attained, is xi ix." 
It thus appears probable that in E. zulu the number of ridges was 
normally greater than in E. africanus, as might be inferred from the 
narrower shape of the laminae themselves. 
More significant than the number of the ridges is their shape, and 
one is immediately struck by the fact that in the fossil the pattern is 
decidedly less loxodont than in the recent Af rican species ; the median 
expansion of each lamina is less, and hence the successive ridges are 
more widely separated, while in the less abraded ridges the loxodont 
pattern is not displayed at all, the two enamel walls of each lamina 
being quite parallel. The enamel is very thick (pachyganal) and very 
strongly crimped. The lateral terminations of the laminae are either 
rounded or trifoliate, and each ridge has a feebly curved or crescentic 
shape, with the horns directed forward. The three posterior ridges 
have only the points of the digitations exposed, and of these there are 
four to each ridge, and those of successive plates are arranged in 
long-itudinal rows in a manner suaftrestive of E. meridionalis. 
In several respects these teeth of E. zulu resemble the curious 
molars of E. antiquus, which were originally described by Falconer as 
E.priscus: "The discs of wear present an unmistakable resemblance" 
