ON ARTIFICIAL ATAVISM. 21 



In Other terms, if the dark-red color is composed of flesh-colored and 

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

 crossing of dark-red and white must give dark-red hybrids, which in the fol- 

 lowing generation must split up in four types according to the given distribu- 

 tion. I found, as above said : 



1st Gen. 2nd Gen. 



Dark-red 51% 58% 56.25% 



Flesh-colored 16% 17% 18.75% 



Delila 31% 20% 18.75% 



White 2% 4% 6.25% 



The accordance is such as might be expected in view of the small number 

 of individuals counted. 



If we look closer at the given combination series, we at once see that the 

 white individuals have no other character besides this, and therefore must not 

 be counted in their offspring. But this is not the case with the three other 

 groups which contain partly constant types, and partly such as for one or 

 the other character or for both are still hybrids and as such may split up again 

 in the following generation. 



The series above given leads to the following conclusion : 



Number. Constant. Hybrid. 



Dark-red 918 



Flesh-colored 312 



Delila 312 



White I I o 



In other terms, if the dark-red, the flesh-colored, and the Delila children 

 of the first hybrids are self-fertilized, and their progeny is studied when flower- 

 ing, we may expect to find some of them constant and others dividing and 

 showing a certain amount of variation in the colors of their flowers. Only 

 this variation will be limited as the combination series indicates. The flesh- 

 colored hybrids do not contain the Delila characters, and therefore can only 

 split up flesh-colored and white, according to the mono-hybrid rules. Conse- 

 quently the Delila hybrids will only produce Delilas and Whites. But the 

 dark-red may be hybrids as to both characters, or only one of them, the other 

 being constant. In their progeny this constant character will always unite with 

 the split-up element, and it will in this way be possible to calculate the whole 

 composition of the progeny. 



I will now return to my experiments. We have already seen that among 

 the dark-red hybrids of the second generation some divide up in the same 

 manner as those of the first. As I said, I fertilized a series of hybrids of this 

 second generation with their own pollen and had the following results : 



A plant of the Delila type gave 361 flowering children, all without excep- 

 tion of the same type as the mother, which therefore was to be regarded as 

 constant. In the sanie way I had from the seed of a flesh-colored specimen 260 

 children, all with flesh-colored flowers only. Both of the possible constant new 

 combinations were thus gained by crossing, and this undoubtedly is one of the 

 chief points in Mendel's interpretations of the di-hybrids. 



The figures as given above show that hybrids of the types of Delila and 

 the flesh-colored forms may also split up in the next generation into their own 

 color and white. According to the law of mono-hybrids, there are to be ex- 

 pected in these cases ^ of colored and ^ of white offspring. I had only two 



