with the front talon, are broken off, the remaining part of the crown being composed of seven ridges and a slight posterior talon. These ridges are quite intact, and much more apart, than in E. primigenius, agreeing in this respect with E. Indicus. The digitations are well marked at the apex, forming distinct points, and in the last ridge their separation can be traced to a depth of nearly an inch, a condition which ordinarily involves a high degree of crimping of the enamel-plates. The crown is narrow in front, and widens so abruptly behind, as to have suggested to De Blainville the term 'subdidyme to characterize it; he describes it as resembling most the analogous tooth of the existing African Elephant. This peculiarity is best expressed by the dimensions, viz., length of crown 2.7 inches, width in front at the second ridge 1.1 inches, width behind 1.7; the length being to the extreme width in the ratio of about 3:2. The empty alveolus of the last milk-molar (m. m. 4) is distinctly visible in M. Le Clerc's specimen. The penultimate milk-molar thus yields, for its term in the ridge-formula, 8 colliculi besides talons. The specimen, so far as mineral condition is concerned, is well fossilized, like those from the Sewalik Hills, and the "Forest bed" of the Norfolk coast, being hard, heavy, and weathered, and not adherent to the tongue.

The specimen next to be noticed, is a detached and very finely preserved antepenultimate true molar (m. 1) of the lower jaw left side, No. 741a of the additions to the Cat. of Foss. Mamm. of the Royal College of Surgeons. It is a comparatively late acquisition (since 1855), and was brought from Mexico by Mr. Taylor. The crown and body of the tooth are quite perfect from end to end; the fangs are mostly broken off, but a portion of them still remains. The crown is composed of twelve colliculi, with front and hind talons. Of these the eight anterior divisions are worn, the rest being intact. The discs of wear are wide and open, wider than in the ordinary varieties of the existing Indian Elephant, and nearly approaching the width commonly presented by E. antiquus. But they differ from those of the latter species, in showing no angular expansion in the middle of the discs, and no outlying loop at the angles. In this respect they correspond more with the discs of the existing Indian Elephant. The edges (macharides) of the enamelplates are highly crimped with numerous close-set flexures; in this respect also maintaining a resemblance to the Indian Elephant, and differing from E. antiquus. Notwithstanding the distinctions here indicated, the aspect of the crown in the Mexican molar, bears a striking general resemblance to that of typical specimens of the same age of E. antiquus, the most obvious difference being, that the crown in the former is much wider in proportion to the length, than in the latter, in which the molars have narrow crowns, like those of the African Elephant. The specimen is represented by fig. 1 of Pl. II.

A notable peculiarity in the Mexican tooth is, that the body of the molar is very much bowed sidewise, i. e. concave on the outer side and convex on the inner. The amount of arcuation is much greater than I remember to .have seen in any other species of

Elephant, fossil or recent, in molars of corresponding age, viz., adolescent. Something of the same kind is seen in Mr. Bollaert's specimen, as figured in the Geologist': but in this case, in a minor degree, in consequence of the anterior third of the crown having been worn away. I believe that this peculiarity in the lower molars of E. Columbi is a constant character of the species, and that it bears a relation to the converging form of the rami of the jaw, to be noticed in the sequel.

The dimensions of the Mexican molar are:

It is hardly necessary to remark that the characters of the molar above described, differ entirely from those of the common form of the Mammoth of North America.

Von Meyer, in the Neues Jahrbuch' for 1840, briefly notices some fossil remains of Mastodon and Elephant, contained in the Mexican collection of Herr Uhde. Among these are an upper and lower molar, of a fossil Elephant, in which the enamel-plates were wider apart than in E. primigenius, in this respect having a closer resemblance to those of E. proboletes of Fischer de Waldheim, which Lartet conjecturally refers to E. meridionalis. The description would agree with that of E. Columbi, from the same region.

Sir Charles Lyell's Georgian specimen, from the Brunswick Canal, upon which my first knowledge of E. Columbi was founded, consists of the middle portion of the penultimate or last true molar, probably the latter (m. 3) lower jaw right side, broken off, both at the anterior and posterior ends. The fragment comprises ten complete ridges, with part of two others, of which the anterior seven are more or less worn. All the fangs are broken off, together with the basal mass of ivory. The summit of the crown is concave from back to front, and the tooth is also concave with a little torsion on the outside, and convex inwards, showing that it was considerably arcuated laterally, like the specimen last described. The discs of wear are of moderate width, as in the Indian Elephant, with a tendency in some of them to expansion in the middle. This is most pronounced in the second, where the expansion nearly attains half an inch. The plates of enamel are thicker than in the Mammoth, and about equal to those of the Indian Elephant; they present a considerable amount of parallel shallow plaiting, which is prominently shown where they rise above the level of the cement. The wear of the crown takes place in a succession of steps, from the

front backwards, which it is of importance to notice with reference to the inferred food of the species. These steps rise like a flight of stairs, each being composed of the whole mass of cement of one of the valleys, and the combined enamel plates and ivory of the ridge immediately behind it. There are five of these steps in the Georgian specimen, the posterior ridges being intact.

The dimensions are as follow:

Length of crown measured at the base 9.5 inch.
Ditto ditto at summit of crown

Width of crown in front

'Ditto at 4th remaining ridge

Ditto behind at widest part

6.9

Height of ditto at 8th plate where unworn 6.2
Ditto of anterior worn plate

2.5

Pl. I. represents a longitudinal section of the specimen, by which the specific distinction from the Mammoth is best shown. The plates converge from the convex base to the summit irregularly, but somewhat like the voussoirs of an arch; so that the same number of plates, diminishes from 9.5 inches at the base to about 7 at the crown. The ridges are not so high as in the Mammoth, and their constituent elements, i. e. the enamel, ivory, and cement are thicker. In the interval between the 10th and 11th ridges, the cement attains, near the base, the excessive thickness of six-tenths of an inch, being about twice as much as what is ordinarily seen in the section of the Mammoth. For the contrasted difference, I refer to the sections, pl. I. fig. 1, of the 'Fauna Antiqua Sivalensis.'

The specimen next to be noticed is No. 33,218 of the MSS. register, British Museum, Palæont. Gallery. It was acquired of Mr. Bollaert, and it bears a record of having been procured from San Felipe de Austin, on the Brazos river, in Texas. It is figured in the Geologist.' This superb morceau consists of the posterior threefourths of the last true molar, lower jaw, left side, well advanced in wear. The crown presents the remains of the posterior fourteen ridges and hind talon; the anterior portion had been ground down by use, and has disappeared. The two anterior ridges are worn to the base, and confluent in a common depression of ivory, upon which a slight islet of enamel remains. Of the succeeding ridges, the next seven are worn down into transverse discs, which are open, and bounded by highly crimped and thick plates of enamel, bearing a close resemblance in this respect to the corresponding teeth of the existing Indian Elephant. Some of the plates show a considerable amount of undulation in the general sweep of the macharides, but there is no tendency to the mesial expansion, or outlying loop, seen in E. antiquus. The five posterior ridges are all more or less affected by wear; the most of them present distinct annular discs on the tips of the digitations, which are seen to be of large size, as in E. planifrons

Vol. 5. p. 57, Pl. IV.

and E. meridionalis. The eighth ridge shows these annular discs semi-confluent into a transverse depression. The ninth presents five worn digitations; the tenth and eleventh six, and the twelfth five. There is no mark of pressure behind, proving the tooth to be the last of the molar series. The space occupied jointly by the first six distinct transverse discs, amounts to 5.4 inches, giving an average of nine-tenths of an inch to each ridge. This is considerably greater than that shown by the crown of the Georgian molar, but it is to be borne in mind that the difference is accounted for by the teeth being in different stages of wear.

The principal dimensions are:

Extreme length of crown

Width of ditto at first transverse disc

12.5 inch. 3.0 99

Making allowance for the part of the tooth borne upon the anterior fangs, which has been worn away, the entire molar must have been of very large size; and it indicates a species that attained colossal dimensions.

*

Other illustrations of E. Columbi, are furnished by a collection of fossil bones, part of which was purchased for the British Museum in 1847. They are stated to have been found by Mr. W. Huff on the banks of the Brazos River, near San Felipe de Austin in Texas. One of the specimens, a fine Bovine skull, (Bison latifrons, Leidy), is identifiable with a figure given by Dr. W. M. Carpenter, of New Orleans, who published the first account of these remains. Among them were numerous fragments of bones, said to have been of Elephant and Mastodon; the teeth of Elephants, especially, predominating. A proboscidian tusk measured eleven feet in length, and twenty-six inches in girth at the base; but no details are given regarding the molars of Elephas. In the series, belonging to the National Collection, reputed to be of this origin, (Nos. 20,701-5 MSS. Register) is a superb specimen (No. 20,702) of a last molar of the lower jaw right side, comprising the posterior three-fourths of the crown in fine preservation, the anterior part having been lost by a vertical fracture. Fifteen ridges are presented together with a talon-ridge. Of these, thirteen are more or less worn, the seven anterior into continuous transverse discs, which, measured along the surface of the crown, occupy a length of 4.6 inch. The two next (8th and 9th), are divided towards the outer side by a wide fissure into two unequal flattened elliptical discs; the 10th yields three discs, the 11th four, and the 12th five thick annular depressions. The rest are nearly intact, and present from four to five very thick digi

Silliman's Journal, 1846, 2nd Series, Vol. I. p. 245.

tations. The enamel-plates of the seven anterior bands, present irregular secondary wavy curves, but they are free, or nearly so, from crimping. In this respect, and in being perceptibly thinner, they differ considerably from the other Texan lower molar, No. 33,218 above described. Regarded sidewise, the ridges look very thick and massive, and they are retrofracted about half way up, by an abrupt flexure, somewhat like the Pourentrui molar figured by Blainville, (Elephants, p. 199, Pl. X fig. Va).

The dimensions above, yield an average of about 7 inch to the seven anterior discs, and .65 to the series of thirteen, being considerably less than in No. 33.218. The latter, also, in the thickness and undulation of the enamel-plates, resembles more the existing Indian Elephant. Although crimping is absent from the fossil under description, the great thickness of the ridges, and the limited number and massiveness of the digitations, remove it from E. primigenius, in which the digitations are slender, and double the number. The width of the crown is enormous, being nearly 5 inches, agreeing in this respect with the Alabama molar to be noticed in the sequel. Although with some doubt, I refer the specimen to E. Columbi. It is well fossilized and adheres strongly to the tongue.

Another specimen, of the same series, (No. 20,702), is a fragment comprising the posterior two-thirds of a left lower penultimate, (m. 2.), and including eleven ridges, the talon being wanting. Of these the anterior six are partly worn, but none of them into transverse discs; the first three, are in three divisions, each forming a flattened ellipse; the enamel is thick, but does not show any considerable amount of crimping. The ridge-plates are nearly vertical, and the intact digitations of the hinder ones are thick. The crown is well coated with cement.

The above dimensions yield an average of 7 inch to each ridge. The specimen agrees very closely, in every respect, with the corresponding molar of the Indian Elephant, and with the characters of the lower molar No. 33,218.

There are other Elephant molars in the Texan series, which belong to a different species, to be noticed in the sequel.

The late Dr. Warren, in his excellent monograph on the 'Masto