360 PROCEEDINGS OF THE OHIO ACADEMY OF SCIENCE. 



lines are of great length, and the evolution of a flower or of 

 an insectivorous plant is a way with many side paths. It is 

 the sieve that keeps evolution on the main line, killing all, or 

 nearly all that try to go in other directions. By this means nat- 

 ural selection is the one directing cause of the broad lines of 

 evolution". "Only a sieve", — a brilliant figure which would 

 have been welcomed by Darwin ; "the one directing cause", — 

 selection carried to an ultra-Darwinian extreme. 



No, the mutation theory is an ally, not the opponent, of 

 the main central thought of Darwinian natural selection. But 

 the mutation theory would restrict the material upon which nat- 

 ural selection can work to the occasional discontinuous muta- 

 tions instead of the ever-present fluctuations. Selection, as ap- 

 plied to the minute variations so universally present in animals 

 and plants, would be unavailing. The striking results which can 

 be reached by selecting from a field crop of beans or a wild 

 aggregate of Paramecium is due not to the accumulation of 

 fluctuating variations, but to the isolation of a number of dis- 

 tinct races or "pure lines", of which such a "mixed population" 

 is composed. If a "pure line" (produced by the closest self- 

 fertilization or by asexual reproduction), be taken as the start- 

 ing point, the results will be different. Johannsen with his beans 

 and Jennings with his parmecia have published similar con- 

 clusions, — ■ the progeny of the bigge&t bean in a given pure line 

 will average no bigger than the progeny of the smallest bean in 

 the same pure line ; the offspring of the smallest paramecium will 

 equal in average size the offspring of the largest within the same 

 pure line. Size variation is taken as typical of quantitative varia- 

 tion in general. 



Probably all zoologists would agree with de Vries in ascrib- 

 ing an enormous importance to mutations; the majority, perhaps, 

 would follow the theory to its extreme of universal application. 

 But there has always been a respectable minority calling for 

 caution. Castle, for example, has consistently supported the 

 view that many changes of evolutionary value result from the 

 summation of continuous variations, and cites de Vries' own 

 experiments on the buttercup, in which the average petal number 



