85 rnocEKDixGs of the 



behaviour cannot profitably be discussed, rurthermore, I do not 

 think that the analysis has been carried far enough in any case to 

 say whether segregation takes place or not. I do know that in cases 

 where de Vries saw only two forms in F, and F^, I saw several at 

 least in F^, without, however, being able to fix their limits. This 

 difficulty is due to the fact that the differences between these 

 yellow-flowered CEuotheras are, in their visible effects, more 

 of degree than of kind ; so that a Mendelian distribution of the 

 characters over the offspring would be very hard to recognise. 

 The large number of factors involved, the difficulty of sharply 

 limiting the visible characters caused by them, the possibilities of 

 coupling and repulsion, all tend to form an F^ which cannot 

 be entangled, and which one can, by proceeding arbitrarily, bring 

 to one, two, or several forms. Under these circumstances, I can- 

 not attribute much weight to the apparent exception de Vries's 

 crosses furnish *. 



It may yet be objected that other first-rate observers have 

 described stable species-hybrids. A very interesting case of this 

 has recently come to my knowledge. Darwin got from a friend 

 hybrid-geese from the cross of t]\e common goose with a Chinese 

 goose. From them he obtained five birds of the second generation, 

 which he says, " in every detail resembled their hybrid parents." 



I will try to repeat the experiment, but think it very likely that 

 the apparent uniformity is due to the small number of segregates 

 obtained. In this respect it is A^ery interesting to know that 

 Mr. Heribert Nilsson, who has allowed me to communicate this 

 here, has succeeded in showing, for some, at least, of Wichura's 

 stable »Sa??a--hybrids, that these are not stable, but segregate 

 very considerably. I have further details of these experiments, 

 which I will be glad to show anyone present interested in 

 them. 



Another objection which is offered again and again is that a 

 difference exists between species and varieties, inasmuch as the 

 first are mostly sterile after a cross, while the latter are not. 



]S"ow, if my contention is right, that only micro-species exist, 

 this statement can be transcribed to this sentence, that micro- 

 species different in many factors (as such referred to different 

 Linnean species are likely to be) will less easily cross than such 

 which differ in but few factors. Now I would see nothing 

 unexpected in this, and nothing which would tell against my 

 theory — no more, in fact, than the objection that certain chemical 

 substances show affinity and others do not, tells against the 

 theor}' that chemical compounds are formed by the union of other 

 chemical substances already present. But is the contention quite 

 true? Are species essentially so much more sterile inter se than 

 varieties ? By essentially I mean if we take into consideration 

 secondary difficulties against crossing, as differences in size between 

 the animals which have to copulate, poisonous substances on the 

 stigmata of flowers, etc. We must not forget that we have 

 cases where reproduction-cells so closely related as the poUen- 



* Renner has since shown that de Vriee's results can be explained in a 

 Mendelian way : see Flora, cyii. (1914) 115-150. 



