chap, xxv.] CONCLUDING REFLEXIONS. 573 



an expectation is created in the mind that the distinctness of these 

 forms of varieties, all living [and probably breeding] together, may 

 be of the same nature as the distinctness of Species ; and since it 

 is clear that the distinctness of the varieties is not the work of 

 separate Selection we cannot avoid the suspicion that the same 

 may be true of the specific differences too. 



An error more far-reaching and mischievous is the doctrine 

 that a new variation must immediately be swamped, if I may use 

 the term that authors have thought fit to employ. This doctrine 

 would come with more force were it the fact that as a matter of 

 experience the offspring of two varieties, or of variety and normal, 

 does usually present a mean between the characters of its parents. 

 Such a simple result is, I believe, rarely found among the facts of 

 inheritance. It is true that with regard to this part of the problem 

 there is as yet little solid evidence to which we may appeal, but in 

 so far as common knowledge is a guide, the balance of experience 

 is, I believe, the other way. Though it is obvious that there are 

 certain classes of characters that are often evenly blended in the 

 offspring, it is equally certain that there are others that are not. 



In all this we are still able only to quote case against case. 

 No one has found general expressions differentiating the two 

 classes of characters, nor is it easy to point to any one character 

 that uniformly follows either rule. Perhaps we are justified in the 

 impression that among characters which blend or may blend evenly, 

 are especially certain quantitative characters, such as stature; while 

 characters depending upon differences of number, or upon quali- 

 tative differences, as for example colour, are more often alternative 

 in their inheritance. But even this is very imperfectly true, and 

 as appeared in the case of Earwigs (p. 40) there may be a definite 

 dimorphism in respect of a character which to our eye is simply 

 quantitative. Nevertheless it may be remembered that it is 

 especially by differences of number and by qualitative differences 

 that species are commonly distinguished. Specific differences are 

 less often quantitative only. 



But however this may be, whatever may be the meaning of 

 alternative inheritance and the physical facts from which it results, 

 and though it may not be possible to find general expressions to dis- 

 tinguish characters so inherited from characters that may blend, it 

 is quite certain that the distinctness and Discontinuity of many 

 characters is in some unknown way a part of their nature, and is 

 not directly dependent upon Natural Selection at all. 



The belief that all distinctness is due to Natural Selection, and 

 the expectation that apart from Natural Selection there would be 

 a general level of confusion, agrees ill with the facts of Variation. 

 We may doubt indeed whether the ideas associated with that 

 flower of speech, " Panmixia," are not as false to the laws of life as 

 the word to the laws of language. 



