Cervus and functionally usurping the place of the main ones, and (2) the posterior being correspondingly reduced and thrown backward.

In an article published several years ago in a popular periodical, * in response to a solicitation for information, I expressed my views as to the morphology of the antlers of Elaphurus, based on the inspection of the illustrations of the antlers of different ages. It does not appear to me, in view of the known facts of development, that there can be a reasonable doubt as to the correctness of the interpretation offered, and the remarks then made are thus reproduced:

"They may be compared to those of the typical stags, so far as the main antlers (homologically speaking) are concerned, but with the brow antlers extraordinarily developed, branched, and usurping the place of the main antlers; in other words, the several elements of the horns, compared with those of the ordinary deer, are reversed, the direction of the growth being upward or forward instead of backward, so that what in most deer are the "main" antlers, are thrown out of axis and deflected backward, while those which correspond to the brow antlers of other deer take the place of the "main" antlers and develope two to four tines, † according to age. This mode of growth is unexampled in any other deer and apparently justifies the generic distinction of the species from the others, as M. Alphonse Milne-Edwards has proposed."

Of course it is to be understood that the growth is "unexampled" only in degree and not in kind.

II. In order to avoid the ambiguity and misconception that might be entailed by the use of the terms employed by sportsmen to designate the several prongs of the antlers of the stag, a uniform system of nomenclature seems to be desirable. Such is further needed to insure conciseness and precision of diagnosis. (1).The simple spikes of the first year and their after growths may be designated protoceres, (2) the

"On the Elaphure of China."<Forest and Stream, vol, 1, pp. 242-243, Nov. 27, 1873. + Misprinted "tines."

66

"The growing antlers of the young (three year old) exhibit approximately proportions like those of the three year old red deer-"Spayad [misprinted Spoxod"]-but they are much more robust, and the brow-antlers relatively larger; the difference then is the result of the disproportionate increase or hypertrophy of the "brow"-antlers, and corresponding atrophy of the "main" antlers.-Original note to communication to Forest and Stream.

anterior offshoots of the second year deuteroceres, and the succeeding (3) third, (4) fourth and (5) fifth anterior offshoots, respectively (3) tritoceres, (4) tetartoceres, and (5) pemptoceres. The antlers of Cervus, Elaphurus, and Cariacus may then be briefly defined as follows. In Cervus, the antlers have the main axis continuous in the protoceres and incurved backwards, the deuteroceres are procurrent and developed as "brow-antlers," and the tritoceres, the tetartoceres, and the pemptoceres diverge forwards from the protoceres.

In Elaphurus, the antlers have the main axis continued into the deuteroceres which are supra-current and bifurcate into an anterior larger (and often sub-divided) and a posterior smaller prong, the protoceres are deflected backwards, and the tritoceres are rudimentary or absent.

In Cariacus, the antlers have the main axis subspirally excurrent into the tritoceres which generally bifurcate anteriorly, the protoceres are abruptly supra current from the tritoceres, the deuteroceres arise from the inner surface of the protoceres near their bases, and tetartoceres are in some (Eucervus) developed, and in some (Cariacus s. s.) suppressed.

The duplication of the prongs need not be taken cognizance of in this connection.

It is but right to caution the general reader that the homologies above noted are not of the same definite nature as those between corresponding bones and are somewhat artificial. The chief benefit resulting from their recognition is in the conciseness and exactness of description attainable therefrom.

III. In 1873, the comparison of the skeletons of deer in the Smithsonian collection forced on my attention most of the differences signalized by Prof. Garrod, and also the fact that there were two entirely distinct types manifested in the structure of the feet. In a communication to the American Association for the advancement of science, at Hartford, in 1874, I communicated some of these results and the following report (less typographical errors) was published in the

* I consider these words as English and therefore the terminal "ceres" as monosyllabic.

ceres.

Not infrequently the main axis is procurrrent into the tetartoceres or pempto

Daily Tribune of New York and subsequently in an octavo extra containing a report of the proceedings. * (p 89.)

"The genera of deer have hitherto been distinguished almost solely on account of differences in the antlers, and have been consequently regarded with suspicion by some naturalists. A recent examination, however, of the osteology of the genera revealed many previously unrecorded differences. The most peculiar of these are inverse modifications and developments of the limbs of the several genera.

The most familiar genera of our fauna afford exemplifications of these modifications: e. g., on the one hand, the common deer [ Cariacus virginianus] has the lateral metacarpal † bones atrophied at their proximal extremities, and well developed at their distal, where they articulate with corresponding phalanges; on the other hand, the so-called elk or wapiti [Cervus canadensis] has the metacarpal bones developed at the proximal extremities and atrophied at the distal, the corresponding phalanges being connected with the median bones. The reindeer [Rangifer tarandus] and moose [Alces machlis] agree with the common deer, while the elk in this country is sui generis, although represented by numerous analogous forms elsewhere, and especially in Asia. All the genera referred to are further corroborated by numerous other characters co-ordinated with those described, and which are found at the base of the skull, especially the auditory bullae, paroccipital-bones, the relation of the maxillary and nasal bones, the lachrymal bones, the palatine bones, &c."

Since this communication was made, a memoir by Sir Victor Brooke on the genus Cervulus § has become known to me, in which the same facts respecting the metacarpal bones in the genera above mentioned and several additional forms are given, but the publication of Sir Victor Brooke was considerably anterior to my own and more detailed and replete with statements of facts. It is not apparent, however, what taxonomic value Sir Victor Brooke would attach to the character noticed. It appears to me that, at least after the exclusion of the genera Alces, Rangifer, and Hydropotes, the remainder of the Cervinae may

* American Deer. <Proc. Am. Assoc. Adv. Sc., 1874, reported in New York Tribune, Extra, No. 21, p. 89.

† Misprinted "metatarsal."

Misprinted "ordinary."

On the characteristics of the primary groups of mammals. for Adv. of Sc., 1871, vol. 20, pp. 288-289.

Proc. Am. Assoc

be naturally primarily differentiated into (1) those with the external metacarpal bones developed at their proximal extremities, and (2) those with the external metacarpal bones developed at their distal extremities. The geographical distribution is noteworthy.

So far as known all the Cervidae with distally developed lateral metacarpals are Arctogean (European † and North American) and South American, while the Cervidae with proximally developed lateral metacarpals are especially Asiatic.

It still remains to be proved, however, whether this division is the most natural.

IV. Like Prof. Garrod, I have long been perplexed by the inadequacy of the binomial nomenclature to express the various degrees of relationship and have chiefly availed myself of the method of dichotomous analysis. In an article published in 1871 * I used essentially the same mode of expression as has Prof. Garrod in his recent article, in the following remarks:

"A difficult problem is the arrangement in a linear series of forms so as to best express their relationship. This is perhaps most aptly effected by taking, in the first place, the most generalized type known. (a), and following that by the one (of two or more) most closely allied to it (a 1), then by the one nearest to that (a 2), and thus to the end of the series, wherever it may lead; then we may recommence with the one next most nearly related to (a) the first type, and project another series (b, b 1 &c). It will be evident that the last term of the first series (ax) will often be much less nearly related to the first term of its own series (a) than is the first term of the second series (b); and, of course, that it,—(ax) the last term of the first series,— so far from being intermediate between the two (a and b), must be the most remote from the first term, if we are right in the appreciation of the relative affinities of the succeeding series, since both are the descendants of the same original progenitor."

V. I was incautious enough to adopt Mr. Sclater's arrangement of the

* On Sclater's Muntjac and other species of the genus Cervulus. Proc. Zool. Soc. London, 1874, pp. 33-42.

The European realm of the older authors (:

of course meant.

sis.

Palearctic of some recent ones) is

The only American form is the near ally of the European Stag Cervus canaden

Cervidae in 1872, in my "Arrangement of the families of Mammals," but, being sceptical of its truth to nature, I especially indicated (by the words "Genera fide Sclater") that I did so entirely on his authority. I must regret this as subsequent investigation has convinced me that it is the most inapt and unscientific arrangement that has ever been proposed for the Cervinæ. I shall only add in explanation that I knew that Mr. Sclater was acquainted with the principal systematic attempts of previous authors, and supposed that he had duly weighed the values of the several differential characters, and had based his conclusions on valid considerations. Although I then recognized the comparative unimportance of the characters nominally employed by him to differentiate" Cervus" and "Dama," even because I did so, and their relative unimportance seemed so evident at first sight, I was induced to believe that Mr. Sclater's conclusions had been based on the examination of skeletons or other characters respecting which he preferred to be reticent until he could elaborate them more fully. The distinctions used were presumably external co-ordinates of previously unknown characters and simply employed as technical diagnostic marks. I was led to this belief by his statement that he mentioned for the genera "only their most obvious external characters." I am now forced to believe, however, after a careful examination of the characters of Cervus and Dama, that there are no other characters which can justify Mr. Sclater's classification and that it is utterly des titute of a scientific basis. In this opinion I am practically sustained by the conclusions of such capable observers and reasoners as Sir Vincet Brooke and Prof. Garrod. The latter has very justly remarked that "Dama vulgaris as well as Dama mesopotamica, from the shape of their antlers-neglecting the palmation, an evidently insignificant character [but which Mr. Sclater considers all important]—are intimately allied to the pseudaxine group." The antlers in two examples of Cervus canadensis in the Smithsonian Museum are actually palmated at their extremities, (but in both cases on one side only,*) and the two sides would certainly not be placed (at least when to. gether) even by Mr. Sclater, in different genera.

THEODORE GILL.

"Cornua palmata" is the only character used by Mr. Sclater to isolate Cervus Dama in a " genus" from the negative medley (Cornua non palmata) designated by him as the " genus Cervus"