CAUSES OF THE EVOLUTION AND EXTINCTION OF THE TITANOTHERES 



843 



or diminish, the expression of a character — that is, we 

 can modify the germinal character itself. These 

 germinal quantitative variations (in color patterns 

 and pigmented areas) behave as simple units — that 

 is, biocharacters. Blending of size (Castle, 1916.1) is 

 demonstrated in rabbits, in which the offspring of a 

 large and a small rabbit are of intermediate size — that 

 is, neither small size nor large size dominating in the 

 cross. "Size," observes Castle, "is an imstable 

 character, ever varying." Slow changes in size can 

 be effected by artificial selection without any crossing 

 whatever. Progressive diminution in size may be 

 effected by crossing small breeds. 



More recently Jennings (1917.1, p. 305) observes: 



It appears to me that the experimental work in Mendel- 

 ism * * * is supplying a complete foundation for evolution 

 through the accumulation by [artificial] selection of minute 

 gradation. * * * A visible character may be modified in 

 the finest gradations by alterations in diverse parts of the 

 germinal material. The objections raised by the mutationists 

 to gradual change through [artificial] selection are breaking 

 down as a result of the thoroughness of the mutationists' own 

 studies. 



Jennings is of the opinion (op. cit., p. 306) that 

 Mendelian heredity acts as an accelerator to the 

 effectiveness of [artificial] selection. The sum of 

 Castle's and Jennings's observations and opinions is 

 that certain germinal characters may be altered by 

 artificially selecting in each generation the merely 

 quantitative variation that goes farthest in the desired 

 direction. 



Against this general experimental conclusion are 

 earlier (Punnett, 1911.1, p. 138; Davenport, 1910) 

 and more recent (Pearl, 1917.1, pp. 72, 73) observations 

 that in certain biocharacters fluctuations are not 

 heritable, and that it is impossible through cumulative 

 artificial selection of fluctuations to establish a new 

 quantitative mean. 



To sum up: The years of experiment since the 

 rediscovery of Mendelism in 1903 present consid- 

 erable but not yet conclusive evidence that certain 

 proportional characters as well as certain numerically 

 new characters can be accumulated through artificial 

 selection, and that certain other characters are not 

 cumulative through artificial selection of quantitative 

 variations but are dependent upon the saltation, or 

 sudden appearance, of new character genes or sudden 

 gradations of character in the germ. 



The observations (Castle, Jennings) in experimental 

 heredity that certain germinal quantitative varia- 

 tions of form and color are heritable and can be 

 guided in certain directions by artificial selection do 

 not necessarily prove that allometrons have arisen 

 by similar processes under natural selection but 

 render it probable that heritable fluctuations of 

 proportion of survival value are selected and accumu- 

 lated under natural selection, as in Darwin's theory of 

 the causes of the long neck of the giraffe. 



THEORETIC AND EXPERIMENTAL CAUSES OF THE 

 EVOLUTION OF ALLOMETRONS 



The above studies in experimental heredity and 

 selection have a very important bearing upon our 

 interpretations of heredity in the titanotheres and 

 other quadrupeds known from their fossilized remains. 

 It should not be assumed that aU allometrons are 

 due to the same causes. 



Among titanotheres adaptation through changes of 

 proportion has been shown to constitute the larger 

 part of the bodily evolution. In all mammals the 

 response of the musculature and the skeleton to 

 changes of environment and habitat is more conspicu- 

 ous in proportions than in any other character groups. 

 Allometric adaptation includes a continuous and 

 perfect adjustment of all the demands of weight and 

 speed in locomotion and of offense and defense in the 

 capture or warding off of enemies. It follows, as 

 already remarked, that evidence for the Lamarckian, 

 the Darwinian, and the tetrakinetic theories of causa- 

 tion must be most closely analyzed. The following 12 

 modes of allometric change, germinal, ontogenetic, 

 environmental, selectional, are observable: 



1. Germinal aUometrons, arising by continuous or gradual 



change. 



2. Germinal aUometrons, arising by sudden changes (saltations) . 



3. Certain germinal allometrons, uninfluenced by the direct 



action of environment. 



4. Germinal aUometrons, apparently influenced by direct 



action of environment. 



5. Germinal aUometrons, fluctuating and nonfluctuating. 



6. Certain germinal aUometrons, of high survival selection 



value. 



7. Other germinal aUometrons, of apparently no survival 



selection value. 



8. Ontogenetic aUometrons, experimentaUy influenced by 



changes of habit. 



9. Ontogenetic allometrons, experimentally influenced by 



changes of environment. 



10. Germinal aUometrons — prenatal, adolescent, adult, male 



sexual, and female sexual development. 



11. Ontogenic allometrons, influenced by glandular internal 



secretions, enzymes, and other organic catalyzers. 



12. Germinal (?) allometrons, influenced by organic catalyzers. 



GERMZNAI AUOMETRONS AEISIWG BY CONTINUOUS OK GKADUAI 

 CHANGE 



The continuity of proportional change observed in 

 all the phyla of titanotheres described in this mono- 

 graph, as well as in several other phyla of imgulates, 

 such as the rhinoceroses and the horses, constitutes 

 in itself very strong evidence against the adequacy of 

 the Darwinian theory of the natural selection of indi- 

 vidual fluctuations. 



Proportional continuity is in the nature of the con- 

 tinuous "mutations of Waagen" rather than of the 

 discontinuous "mutations of De Vries." The general 

 contrast is very ably presented by Scott (1894.7, p. 

 355) in his article entitled "Mutations and variations." 

 Progressive brachycephaly and progressive dolicho- 

 cephaly in the titanotheres point to the presence of 



