May 8, 1908] 



SCIENCE 



751 



a dolichocephalic animal into that of its still 

 more long-skulled descendant. Fig. 2, A, rep- 

 resents the palatal view of a skull which is 

 ancestral to the second skull represented in 

 Fig. 2, 0. An outline of the skull represented 



aly or from dolichocephaly to an intensified 

 dolichocephaly may be interpreted partly as a 

 mere redistribution of materials, all parts be- 

 ing stretched in the same proportion. But 

 this does not describe all that actvAilly occurs 



A B C 



Fig. 2. Artificial Dolichocephaly. 



A, a skull ancestral to G. 



B, outline produced artificially by stretching A to length and width of 0. 



C, extreme long-skulled type, Dolichorhinus. 



Fig. 3. Reetigradations. A, OroMppus, Middle Eocene horse; B, Palceosyops, Middle Eocene 

 titanothere. The circles indicate new euspules rising independently in these two phyla. 



in Fig. 1, A, is traced on India rubber and 

 stretched into the outline represented in Fig. 

 2, B, which it is seen gives us an approximate 

 approach to Fig. 3, G. 



The above two experiments prove that 

 transition from brachycephaly to dolichoceph- 



in nature, because in sTcull lengthening or 

 shortening each hone is affected somewhat 

 differently. 



I am inclined to regard dolichocephaly and 

 brachycephaly in the vertebrates generally as 

 caused by the natural selection of fluctuations 



