202 MENDELISM 



it is scarcely possible to cite a case in which it is 

 definitely and certainly known that Mendel's law, 

 subject to the modifications already described, does 

 not hold good. Cases of various kinds are, indeed, 

 recorded, but these records are derived from experi- 

 ments either carried out before the bearings of the 

 Mendelian phenomena were at all fully appreciated, or 

 and this is the most frequent case without any 

 knowledge at all of Mender's discovery. 



Thus a considerable number of cases were formerly 

 described in which the first cross or heterozygote of F 1 

 bred true instead of segregating in F 2 . There is some 

 doubt whether any case of this kind will really stand 

 criticism ; Millardet's case, for example, which was 

 described at the end of the last chapter but one, has 

 never been confirmed. It is quite certain that among 

 all the numerous crosses studied during the last six 

 years no example of the kind has been substantiated. 

 The most recent cases to be described of a first cross 

 breeding true are those of de Vries, and at these we 

 are bound to pause, because de Vries is surpassed by 

 no recent observer in weight of authority. Neverthe- 

 less, de Vries' cases are of so complex a kind that we 

 have some hesitation in accepting them without further 

 study. For the rest, this is one of the problems which 

 remain for the future to deal with. 



We may now turn for a brief space to some of the 

 cases in which we have as yet no certain knowledge of 

 the manner in which inheritance proceeds. 



The most obvious extension of Mendel's law to 

 processes where it cannot be directly shown to hold 



