750 



SCIENCE 



[N. S. Vol. XXVII. No. 697 



I have added two corollaries from my 

 studies on the titanotheres, in -which it ap- 

 pears to be highly probable that in different 

 parts of the body of highly complex verte- 

 brated animals, different evolution factors may 

 be operating coincidently to produce a coor- 

 dinated adaptive result, namely: 



First, that while inseparable from the 

 others, each process may in certain conditions 

 hecome an initiative or leading factor; second, 

 that in complex organisms one factor may te 

 initiative in one group of characters while 

 another factor may at the same time he initia- 

 tive in another group of characters, the insep- 

 arable action bringing about a continuously 

 ■harmonious result. 



(Fig. 1, A, PalcBOsyops) and dolichocephalic 

 forms (Fig. 1, C, Dolichorhinv^) ; the former 

 become increasingly brachycephalic, the latter 

 become increasingly dolichocephalic. This 

 change of proportion is brought about as fol- 

 lows : (1) there is a redistribution of materials, 

 (2) this effects a change in the entire propor- 

 tions of the skull, (3) the different component 

 bones are affected differently, because there 

 are distinct percentages of increment, in 

 breadth or in length, in the bones of each 

 region. 



First Experiment, Redistribution (Fig. 1). 

 This proves that a general redistribution of 

 materials will convert a brachycephalic into a 

 dolichocephalic type. I outline the broad 



A 



Fio. 1. Artificial Dolichocephaly. 



A, PalcEosyops, extreme brachycephalic type. 



B, outline produced artificially by stretching A to length of C. 

 G, extreme dolichocephalic type, DoUchorhinus. 



In a recent paper before the Zoological So- 

 ciety at New Haven these corollaries were 

 illustrated in the evolution of the titanotheres, 

 as shown in the accompanying Figs. 1, 2 and 3. 



1. Selection of Fluctuations 

 It is observed that in four or five contem- 

 poraneous phyla of Middle Eocene titano- 

 theres, unquestionably derived from a com- 

 mon lower Eocene ancestor, there is a tend- 

 ency to diverge into brachycephalic forms 



skull of Palaeosyops on a sheet of India rubber 

 and stretching the sheet lengthwise, produce 

 the Fig 1, B, thus artificially creating a skuU 

 of dolichocephalic type which approximately 

 resembles Fig. 1, C. This experiment iUua- 

 trates what may be done by a mere redistribu- 

 tion of materials. 



Second Experiment, Redistribution (Fig. 2). 

 This illustrates progressive dolichocephaly. 

 Here are represented the outlines derived by 

 stretching and slightly expanding the skull of 



