370 W. B. Scott — Variations and Mutations. 



of time. Even Hipparion and the modern rhinoceroses retain 

 a rudiment of the fifth metacarpal, though in neither line has 

 this digit been functional since the Oligocene, and in other 

 series similar facts have been observed. 



(7) No trustworthy distinction can be drawn between those 

 variations which resemble the normal condition of other species 

 and those which do not. Bateson has been criticised for not 

 making this distinction, but his critics have overlooked the 

 fact that he has discussed the question and shown that such a 

 discrimination cannot but be arbitrary. This emphasizes the 

 difference between the facts of variation and those of phylo- 

 genetic change. No normal form of extinct mammal ever 

 developed six digits, or three phalanges on pollex or hallux, or 

 four cuneiforms in the tarsus. The possible modes of connec- 

 tion between the carpus and tarsus, on the one hand, and the 

 metapodials, on the other, are strictly limited and when they 

 change, they do so very gradually, allowing every step of the 

 way to be identified. The change from what Ivowalevsky has 

 called the "inadaptive" to the " adaptive" mode of such con- 

 nection may be shown to have occurred independently in all of 

 the following artiodactyl groups ; the Pecora, Tylopoda (proba- 

 bly) Tragulina (probably), Suidai and Oreodontidce, and when 

 the phylum has been completely recovered, the change is inva- 

 riably found to be very gradual. To say that natural selection 

 has permitted only certain variations to persist, eliminating the 

 others, is simply to beg the whole question and to cheat our- 

 selves with an " explanation " which is only an assumption. If 

 variations are actually the material out of which new species 

 are made, then as Bateson well shows, the " perfection " of 

 such variations is not the work of selection. The study of a 

 complete phylum, each stage of which is represented by large 

 numbers of individuals, does not favor such a view of evolu- 

 tion. The direct, unswerving way in which development pro- 

 ceeds, however slowly, is not suggestive of many trials, and 

 failures in all directions save one. 



If we may express in one sentence the most strikingly appar- 

 ent difference between the facts of variation and those of phy- 

 logenetic change, it will be found in the comparatively lawless 

 and uncontrolled character of the former. Addition of parts, 

 coalescence or subdivision of parts, almost anything may happen, 

 and if the facts of variation are " the test of phylogenetic 

 possibility" then such possibilities are far wider and less 

 rigidly controlled than we have hitherto been accustomed to 

 believe. In phylogenetic change, on the contrary, we are 

 impressed by the orderly advance toward the final goal, 

 deviating very little from the direct line, though not always 

 progressing at the same rate. By this it is not meant that the 



