W. JB. Scott — Variations and Mutations. 363 



work must be admitted, provided only that the fundamental 

 assumption upon which the entire argument depends be 

 granted, namely, that individual variations form the starting 

 points of new species. 



Let us now proceed to examine these results in the light of 

 palaeontology, confining our attention principally to those mam- 

 malian phyla whose history is known with some degree of 

 completeness. Little confidence may, perhaps, be felt in 

 these phyla by those who have not especially studied them and 

 it may be imagined that more complete knowledge will 

 require them to be altogether remodelled. Granting all that 

 may be said with regard to the incompleteness of the record 

 and the danger of building ambitious hypotheses upon such 

 narrow foundations, nevertheless we cannot avoid seeing that 

 several facts stand out clearly and are not in the least likely to 

 be overthrown by new evidence. A certain limit of error in 

 the construction of phylogenies is established by the order of 

 geological succession. So far as the lacustrine Tertiary forma- 

 tions of North America are concerned, the evidence derived 

 from the actual superposition of strata fails at only three points. 

 As yet, no contact between the Uinta and White River, 

 White River and John Day, or between the Deep River and 

 the Loup Fork, has been observed, and yet in all these cases 

 indirect evidence of almost equal cogency is available and 

 there can be very little doubt that this order of geological 

 succession is known with sufficient accuracy. This order being 

 admitted, there is no room for any very radical reconstruction 

 of the phylogenies. By no possibility can Hyracotherium be 

 derived from Equus or Poebr other ium from Auchenia, and 

 whatever changes in the details may be necessitated by further 

 knowledge, such changes are not likely to revolutionize the 

 inferences to be drawn from a study of the phyla as now con- 

 structed. Few observers, if any, would now uphold the 

 arrangement of the equine phylum proposed by Kowalevsky, 

 namely, Palceotherium — Anchitherium — Hipparion — Equus, 

 and yet it is surprising to see how the general character of this 

 series and the inferences as to the modes of evolution drawn 

 from it agree with those deduced from a study of the equine 

 phylum as we now have it. Kowalevsky's mistake merely 

 consisted in putting certain representatives of side branches 

 into the main line of descent and that a similar error has been 

 made in phylogenies now accepted is not at all improbable. 

 But the correction of such errors will change the general 

 result but little and we may appeal with considerable confi- 

 dence to the inference from these phylogenies. 



