SPECIFICITY IN FERTILIZATION 213 



of all possible combinations of the offspring to one 

 another. The third set of experiments could not of 

 course be fully carried out. 



The parents and offspring were shown to be recip- 

 rocally compatible with the two unrelated plants. In 

 the second set of experiments he found that the off- 

 spring are divisible into two approximately equal classes 

 for each parent, viz., fertile with the parent or sterile, 

 including in the second class some that set seed very 

 scantily. The relation of a given offspring to one 

 parent is entirely independent of its relation to the 

 other. The offspring may therefore be divided into 

 four classes, viz.: 



1. Fertile with both parents B and G = bg. 



2. Fertile with parent B, sterile with G = bG. 



3. Fertile with parent G, sterile with B =Bg. 



4. Sterile with both parents=BG 



The stated result obviously indicates heritable char- 

 acters concerned in self-incompatibility. Correns as- 

 sumes that these are to be interpreted as stuffs that 

 inhibit the normal development of self-pollen, whether 

 in a positive sense or in the sense of the absence of a 

 stuff necessary to growth of the pollen tube. If this 

 is so, the behavior of the offspring among themselves 

 with reference to self-sterility should be predicable, and 

 Correns states from the results of crosses of twelve of 

 the offspring with all of the others that the expectation 

 given by the formulae is approximately realized. He 

 argues that the demonstration of the heritability of 

 incompatibility disproves Jost's theory that the self- 

 incompatibility is due to individual stuffs; that is, to 

 the chemical make-up of each id ; vidual being specific; 



