160 BLASTOGENIC VARIATIONS. 



and M . glabra. Thus some of the characters, such as 

 the colour of the flower petals, remained fixed in the 

 hybrids. Also, on the whole, the hybrids had a greater 

 resemblance to the female than to the male parent. 

 In the second generation flowers of new colours, viz., 

 white and red, appeared, in addition to the yellow-white 

 and violet flowers exhibited by the parents. Correns 

 also made a number of crosses between undoubted 

 species (e. g., Cirsium palustre .+. spinosissimum, 

 Achillea macrophylla + moschata, Car ex echinata +, 

 fcetida, etc.) and he is doubtful whether one of these 

 hybrids showed a single really dominant character. It 

 was quite obvious that almost all the characters which 

 served to differentiate the parents were present, in 

 greater or less degree, in the hybrids. Correns con- 

 cludes, therefore, that almost without exception the 

 domination of a character shows itself only in crosses 

 between varieties, whilst the hybrids of true species 

 show the characters of both species, though in dimin- 

 ished degree. 



It has been pointed out by Weldon * that Mendel's 

 results are partly vitiated by the fact that he quite neg- 

 lected the ancestry of the plants with which he started 

 his cross-fertilisations. Weldon also adduces a con- 

 siderable body of evidence to show that the separation 

 of the seed characters into definite dominant and re- 

 cessive types by no means invariably holds good. The 

 offspring of cross-bred peas may continue to contain a 

 large percentage of intermediate forms, even as long as 

 25 generations after the crossing. f 



*Biometrica, i. p. 228, 1902. 



* For further evidence concerning the Law see Report to Evolution 

 Committee of Royal Society by Miss Saunders and W. Bateson, 1902. 



