338 



Belli 



The distinction of normal and semi-sterile plants is well marked through- 

 out these tables, especially in the numbers of full pods. The last 

 table proves that plants which have half aborted pollen always have 

 half their embryo-sacs aborted. The first of the two tables goes to 

 show that plants with normal pollen always have normal embryo-sacs. 

 The ovules of the parents of four of these families are compared 

 with those of their semi-sterile progeny iu the following table. 



The agreement is complete, when we consider that they were grown in 

 different years. This corroborates the evidence from the progenies of 

 the fertile parents, that the differences in the percentages of aborted 

 ovules in the ripe pods are due to the segregation of some genetic factors. 

 So far as the evidence goes, and it goes far, all the plants with semi- 

 sterile pollen abort half their embryo-sacs; and the percentages of seeds 

 and undeveloped ovules in the mature pods are determined by the dropping 

 of certain flowers and the lack of nutrition (in the widest sense of the 

 word) in certain young pods. 



To sum up: — In the third generation of the Lyon -Velvet cross, 

 the plants with normal parents have normal progeny, and the plants 

 with semi-sterile parents have progeny half normal and half seuii-sterile 

 iu both pollen-grains and ovules. Neglecting the diploid generation, we 

 may say: fertile haploid generations, after self-conjugation, 

 remain all fertile; and half-abortive haploid generations, after 

 conjugation, give equal numbers of fertile haploid families, 

 and of half-abort ive haploid families. 



Fertelity iu tlie Fourth aud Fifth (ifeueratioiüs. 



Six faiiiiiies of ferlil(! plants have been grown on a large scale iu 

 these generations. Hundreds of flowers and thousands of pods have 

 been examined. All wen' fertile. Three of these are agricultural plants 

 of value (Fig. 17). Their jjedigree numbers are, in order of earliness, 



