I04 



NATURE 



[March 23, i<;i 1 



the fifth edition of the "Origin," and in this then; 

 appears a passage of which it is «iifn<;.ni lo aiiol.- 

 the first sentence : — 



"It may be doubttd whi-lhcr sudden ;ind k'*|'' 

 deviations of struiture. such as we occasionally sre in 

 our domestic productions, mure tspecially with 

 plants, are ever nermanrntly jjropafjated in a state 

 of nature " (p. 4«)). 



But, as explained on p. 104. it was Fleeming 

 Jenkin, and not Wallace, who led him to minimise 

 the importance of "single variations, whether slight 

 or strongly marked " ; and as these are mutations, 

 Fleeming Jenkin 's argument cuts to the root of de 

 Vries's theory. 



But the real divergence between Darwin and de 

 Vries is not so much as regards variability, but selec- 

 tion, the importance of which the latter consistently 

 minimises. To Darwin, variability of any kind 

 merely supplied the field in which selection works, 

 and "sinks to quite a subordinate position in com- 

 parison." With de Vries it is exactly the opposite. 

 "Specific characters do not arise by selection." The 

 combinations of characters which arise from sexual 

 reproduction are, however, subject to it. With respect 

 to adaptations, which is the crucial point, de Vries 

 speaks with a more uncertain voice than in his 

 " Plant Breeding." " Everything points to the con- 

 clusion " that mutation " will explain adaptations just 

 as completely, or, rather, just as incompletely," as 

 natural selection acting on fluctuating variability. 

 This is a cryptic utterance at best. But if natural 

 selection cannot produce specific characters, it cannot 

 endow them, except at haphazard, with survival- 

 value; in that case, its inability to account for 

 adaptations seems a necessary consequence. He 

 finally sums up his position : — 



" I willingly admit that almost anything can be 

 squared with the theory in a very plausible way ; . . . 

 but this is not science." 



Some of us think that it is, and there we must 

 leave it. 



De Vries has devoted immense labour to the in- 

 vestigation of the cases, not infrequent, where seed- 

 lings occur with more than two cotyledons. Two is 

 a reduced whorl, and a w'horl may be explained as 

 due to the suppression of internodes. The number of 

 leaves in a whorl is often variable, though seldom 

 so in the case of opposite leaves; this is in accord- 

 ance with the principles of the " repetition of similar 

 parts." That two cotyledons should be the rule may 

 well be adaptive, seeing that they have to be packed 

 in the seed. That there should occasionally be more is 

 a not improbable mutation. It is interesting to note 

 that de Vries entirely failed to fix tricotyly by selec- 

 tion. Nor is tricotyly, although it appears to be not 

 uncommon, often follow-ed by trimerous leaves. De 

 Vries has been more fortunate than I have been with 

 the sycamore, as he raised two high trees with 

 "branches in trimerous whorls." He says nothing 

 about the leaves. I signally failed in raising a number 

 of tricotyledonous seedlings ; in the third year they 

 all reverted to the opposite-leaved arrangement. And 

 I have only come across a single case of the wild 

 maple with trimerous leaves. This serves as an 

 NO. 2160, VOL. 86] 



illustration, if one were wanted, of how little surviv;n 

 value mutations possess in nature. 



There can be no doubt that what we want is 

 purely empirical study of the variation under artifici 

 but precise conditions of some clean-cut species fr- 

 from any suspicion of hybrid origin. It would ha . 

 to be done on a considerable scale, and as it woul 

 have to extend over a long period of years it woul 

 be a tedious and laborious business. I have son 

 hopes that it may be undertaken in .America. Tl 

 horticultural papers are now full of what is takin 

 place with Primula obconica, which from a conditii 

 _of stability has passed into one of high variabilit} 

 But, as de Vries very justly remarks, the availal-i 

 records in such cases "lack precision." But h»- arri\. 

 at the interesting conclusion that 



"in horticulture . . . mutations are largely ol tl 

 retrogressive or degressive kind. Discontinuoi 

 formation of species on the progressive line is mu> 

 rarer " (p. 602). 



Here species is used in the de Vriesian, not tl 

 Linnean, sense. The meaning, I take it, is th.. 

 latent characters become active or the reverse. On 

 an earlier page (p. 4) he seems to imply that all tlv 

 cultivator can do is to evoke latent characters. "T! 

 first condition necessary for raising a novelty is lo 

 possess it." 



Some fifteen years ago 1 made a careful study > 

 what could be ascertained as to the cultural evoluti. 

 of Cyclamen latifolium. I may state the conclusion, 

 which 1 confess I was not prepared for ;— 



"The general tendency of a plant varying fre; 

 under artificial conditions seems to be atavistic, i.>-. 

 to shed adaptive modifications which have ceased to 

 be useful, and either to revert to a more generaUsed 

 type or to reproduce ' characters which are already 

 general in other members of the same group.'" 



As might have been expected, this had not escaped 

 Darwin, from whom ("Origin," sixth edition., 

 p. 127) 1 had quoted the concluding words of the 

 sentence. 



De Vries still maintains the singular distinction 

 which he draws between cultural variation in horti- 

 culture and in agriculture; perhaps on the Continent 

 the two arts are more distinct than with us, where it 

 can hardly be maintained. " Horticultural varieties 

 are generally constant" (p. 76). Agricultural races 

 "remain dependent on continued selection, and do not 

 really become constant " (p. 422). 



The Darwinian theory rests on a number of con- 

 verging lines of argument, and derives its probability 

 from their cumulative force; just as in a law court 

 each branch of evidence may be slender in itself, yet 

 the conclusion to which they all point has a higher 

 degree of probability than any one taken separate! 

 De Vries's attack on natural selection, I must confe- 

 would not shake my faith even if I found it mi 

 convincing than I do. But he is entitled to the meri 

 which he justly claims for himself, of having probed 

 variability by a rigorous experimental method. It is 

 much to be wished, but scarcely to be hoped, that 

 others will follow him in his lifelong devotion to t^*^ 

 laborious a research. 



W. T. Tiuselton-Dyek. 



