Vol. 65.] ANNIVERSARY ADDRESS OP THE PRESIDENT. CX1X 



stage of evolution to another -with such rapidity that they may 

 be equally distinguished as labile. But the labile stems do not 

 continuously develop at a steady rate : there are intervals in their 

 history, often prolonged, when they seem to pass into a resting 

 stage and their variability becomes latent for a while. Such appa- 

 rently was the case with the Mammalia during a great part of the 

 Mesozoic Era. 



The cases of rapid change are the most interesting, and as an 

 illustration of the value of even a rough chronologic scale we may 

 conveniently consider one or two instances in this connexion, 

 commencing with the ancestry of the horse, which still remains 

 the classic example of its kind, thanks to the investigations of 

 American palaeontologists, particularly Osborne and Gidley. 1 



Gidley, basing his observations on over a thousand specimens 

 preserved in the American Museum of Natural History, enumerates 

 some 60 species and a large number of genera ; many of these do 

 not lie in the direct line of descent, and of those that do the series 

 is interrupted by two well-marked gaps, one between the Hyraco- 

 theriinae and the Anchitheriinae, the other between the Anchi- 

 theriinae and the Protohippinae. The series is not, however, so 

 incomplete as to justify Deperet's assertion that ' the supposed 

 pedigree of the Equidae is a deceptive illusion '. 2 The species which 

 lie more or less in the direct line, extend from the opening days of 

 the Eocene to the middle of the Pliocene, or through a period, 

 according to our scale, of about five or six millions of years. The 

 horizons on which remains of the Equidae have been found number, 

 I suppose, some nine or twelve ; between them lie intervals, 

 seme half a million of years in average duration, which represent 

 absolute blanks in our knowledge. The Assyrian and Egyptian 

 civilizations begin some ten thousand years remote from us, a period 

 difficult enough to conceive : the successive stages by which we re- 

 construct the genealogy of the horse are separated from each other 

 by intervals fifty times as great. We may look at it in another 

 way : probably five years would be a fair estimate of the duration 

 of one generation of horses, if so, then a million generations 

 intervene between the latest species of horse and the Eohippus of 

 the early Eocene times. Our knowledge of these generations is 

 restricted to groups of ten or a dozen, perhaps a hundred, taken at 

 random, as it were, from each successive hundred thousand. ' The 

 argument of the insufficiency of palaeontological documents ' may be 



1 J. W. Gidley, ' Eevision of the Miocene & Pliocene Equidte of North 

 America ' Bull. Am. Mus. Nat. Hist. vol. xxiii (1907) p. 865. 



2 C. Deperet, ' Les Transformations du Monde Animal ' Paris, 1907, p. 107. 



