34 



SCIENCE. 



and only two, appear to be imaginable. The one assumes 

 that these successive forms of equine animals have come 

 into existence independently of one another. The other 

 assumes that they are the result of the gradual modifica- 

 tion undergone by the successive members of a contin- 

 uous line of ancestry. As I am not aware that any 

 zoologist maintains the first hypothesis, I do not feel 

 called upon to discuss it. The adoption of the second, 

 however, is equivalent to the acceptance of the doctrine 

 of evolution so far as horses are concerned, and in the 

 absence of evidence to the contrary, I shall suppose that 

 it is accepted. Since the commencement of the eocene 

 epoch, the animals which constitute ihe family of the 

 equicUe have undergone processes of modification of 

 three kinds: — r, there has been an excess of develop- 

 ment of one part of the oldest form over another; 2, cer- 

 tain parts have undergone complete or partial suppression; 

 3, parts originally distinct have coalesced. Employing the 

 term " law " simply in the sense of a general statement 

 of facts ascertained by observation, I shall speak of these 

 three processes by which the eohippus form has passed 

 into equus as the expression of a three-fold law of evo- 

 lution. It is of profound interest to remark that this law 

 or generalized statement of the nature of the ancestral 

 evolution of the horse, is precisely the same as that 

 which formulates the process of individual development 

 in animals generally, from the period at which the broad 

 characters of the group to which an animal belongs are 

 discernible onwards. After a mammalian embryo, for ex- 

 ample, has taken on its general mammalian characters, 

 its further progress towards its special form is affected 

 by the excessive growth of one part or relation to an- 

 other, by the arrest or suppression of parts already 

 formed, and by the coalescence of parts primarily dis- 

 tinct. This coincidence of the laws of ancestral and in- 

 dividual development creates a strong confidence in the 

 general validity of the former, and a belief that we may 

 safely employ it in reasoning deductively from the known 

 to the unknown. The astronomer who has determined 

 three places of a new planet calculates its place at any 

 epoch, however remote; and, if the law of evolution is to 

 be depended upon, the zoologist who knows a certain 

 length of the course of that evolution in any given case 

 may with equal justice reason backwards to the earlier 

 but unknown stages. Applying this method to the 

 case of the horse, 1 do not see that there is any 

 reason to doubt that the eocene equidae were preceded 

 by mesozoic forms, which differed from eohippus in the 

 same way as eohippus differs from equus. And thus we 

 are ultimately led to conceive of a first form of the equine 

 series, which, if the law is of general validity, must need 

 have been provided with five sub-equal digits on each 

 plantigrade foot, with complete sub-equal antebrachial 

 and crural bones, with clavicles, and with, as at present, 

 44 teeth, the cheek-teeth having short crowns and simple 

 ridged or tuberculated patterns. Moreover, since Marsh's 

 investigations have shown that the older forms of any 

 given mammalian group have less developed cerebral 

 hemispheres than the later, there is a primd facie 

 probability that this primordial hippoid had a low form 

 of brain. Further, since the existing horse has a diffuse 

 allantoic placentation, the primary form could not have 

 presented a higher, and may have possessed a lower con- 

 dition of the various modes by which the foetus derives 

 nourishment from the parent. Such an animal as this, 

 however, v/ould find no place in any of our systems of 

 classification of the mammalia. It would come nearest to 

 the lemuroidea and the insectivora, though the non-pre- 

 hensile pes would separate it from the former, and the 

 I lacentation from the latter group. A natural classifica- 

 tion is one which associates together all those forms 

 which are closely allied and separates them from the rest. 

 Hut, whether in the ordinary sense of the word "alli- 

 ance," or in its purely morphological sense, it is impossi- 

 ble to imagine a group of animals more closely allied 



than our primordial hippoids are with their descendants' 

 Yet, according to existing arrangements, the ancestors 

 would have to be placed in one order of the class of 

 mammalia, and their descendants in another. It may be 

 suggested that it might be as well to wait until the pri- 

 mordial hippoid is discovered before discussing the diffi- 

 culties which will be created by its appearance. But the 

 truth is that that problem is already pressing in another 

 shape. Numerous "lemurs," with marked ungulate 

 characters are being discovered in the older tertiaries of 

 the United States and elsewhere; and no one can study 

 the more ancient mammals with which we are already 

 acquainted without being constantly struck with the in- 

 sectivorous characters which they present. In fact, there 

 is nothing in the dentition of either primates, carnivores, 

 or ungulates, which is any means of deciding whether 

 a given fossil skeleton, with skull, teeth, and limbs almost 

 complete, ought to be ranged with the lemurs, the 

 insectivores, the carnivores, or the ungulates. In 

 whatever order of mammals a sufficiently long 

 series of forms has come to light, they illustrate the 

 three-fold law of evolution as clearly, though, perhaps, 

 not so strikingly, as the equine series does. Carnivores, 

 artiodactyles, and perisscdactyles all tend, as we trace 

 them back through the tertiary epoch, towards less modi- 

 fied forms which will fit into none of the recognized or- 

 ders, but come closer to the insectivora than to any 

 other. It would, however, be most inconvenient and 

 misleading to term these primordial forms insectivora, 

 the mammals so-called being themselves more or less 

 specialized modifications of the same common type, and 

 only, in a partial and limited sense, representatives of 

 that type. The root of the matter appears to me to be 

 that the palaeontological facts which have come to light 

 in the course of the last ten or fifteen years have com- 

 pletely broken down existing taxonomical conceptions, 

 and that the attempts to construct fresh classification 

 upon the old model are necessarily futile. The Cuvieran 

 method, which all modern classifiers have follow ed, has 

 been of immense value in leading to the close investiga- 

 tion and the clear statement of the anatomical characters 

 of animals. But its principle, the association into sharp 

 logical categories defined by such characters, was sapped 

 when Von Baer showed that, in estimating the likenesses 

 and unlikeness of the animals, development must be fully 

 taken into account; and if the importance of individual 

 development is admitted, that of ancestral development 

 necessarily follows. If the end of all zoological classifi- 

 cation is a clear and concise expression of the morpho- 

 logical resemblances and differences of animals, then all 

 such resemblances must have ataxonomic value. But they 

 fall under three heads : — (1) those of adult individuals : 

 (2) those of successive stages of embryological develop- 

 ment or individual evolution; (3) those of successive stages 

 of the evolution of the species, or ancestral evolution. 

 An arrangement is " natural," that is, logically justifiable, 

 exactly in so far as it expresses the relations of likenesses 

 and unlikenesses enumerated under these heads. Hence, 

 in attempting to classify the mammalia, we must take into 

 account not only their adult and embryogenetic charac- 

 ters, but their morphological relations, in so far as the 

 several forms represent different stages of evolution. And 

 thus, just as the persistent antagonism of Cuvier and his 

 school to the essence of Lamarck's teachings (imperfect 

 and objectionable as these often were in their accidents) 

 turns out to have been a reactionary mistake, so Cuvier's 

 no less definite repudiation of Bonnet's "Schelle " at the 

 present day, the existence of a " scala animantium," is a 

 necessary consequence of the doctrine of evolution, and 

 its establishment constitutes, I believe, the foundation of 

 scientific taxonomy. Many years ago, in my lectures at 

 the Royal College of Surgeon, Is particularly insisted on 

 the central position of the insectivora among the higher 

 mammalia ; and further study of this order and of the 

 rodentia has only strengened my conviction that any one 



