420 JOURNAL OF THE ROYAL HORTICULTURAL SOCIETY. 



Delila (Red and White), and White. These four forms appeared approxi- 

 mately in the ratio of 9 : 3 : 3 : 1, which, from the Mendelian point of 

 view, suggested that two pairs of unit-characters were concerned in the 

 original cross of White x Red. De Vries suggests that these two pairs 

 were Flesh-coloured + White and Delila + White, and that the original 

 Dark Red was a compound character composed of Flesh and Delila, 

 which components were resolved after the crossing of the Red with 

 the White. 



" In other terms, if the Dark Red colour is composed of Flesh and 

 Delila, and the antagonistic character of both these elements is White, 

 the crossing of Dark Red with W T hite must give Dark Red hybrids which 

 in the following generation must split up into four types according to 

 the given distribution." 



On this interpretation of the facts, in the third generation the 

 White forms would breed true ; the Flesh and Delila forms would be 

 part constant, and part hybrid (giving Whites as well) ; the Dark Red 

 forms would split up into Red and Flesh forms, Red and Delila forms, 

 and so on, in accordance with the Mendelian principles. These results 

 were obtained in the experiments. Further, by crossing Flesh with 

 Delila all Dark Reds like the original grandparent were obtained. 



From these results de Vries draws the following conclusions : — 



1. It is possible to split up the colours of some flowers by crossing 

 the coloured type with the White variety. 



2. The constituents arrived at by this splitting often follow Mendel's 

 laws. 



3. By crossing the appropriate constituents the original compound 

 colours may be rebuilt. 



4. Instances of atavism may in this way be artificially produced. 



The above results of de Vries' experiments with Antirrhinum have 

 been reviewed and interpreted by Bateson * and Cuenot t on the basis 

 of the resolution of compound characters, much on the same lines as de 

 Vries' own interpretation. 



The reviewer, however, believes that the facts may be more simply 

 interpreted, without reference to the question of the resolution of com- 

 pound characters. 



The first noteworthy point in de Vries' results is that in the second 

 generation of the White x Red the proportion of Whites was only one 

 in sixteen. This suggests that the original " White " parent was not a 

 simple albino, but that the "White " was made up of two distinct unit- 

 characters. If the " White " had been a simple albino the proportion of 

 Whites in the second generation would have been one in four, instead of 

 one in sixteen, as certain experiments by the reviewer and others have 

 already proved. Reference to the original text of de Vries' " Mutations- 

 theorie," vol. ii., p. 197, confirms the idea that the " White " form was 

 not a simple albino, for it is described there as " Weiss mit oft deutlichem, 

 sehr blassrothem Anhauch." In other words, this flesh-tinted White form 

 would appear to be a Flesh and White bicolor rather than a simple albino. 



* Proc. Camb. Phil. Soc, 1902, xii. pp. 50-54. 

 f ISAnnte Biologique, 1902, p. lxx. 



