320 



Belling. 



If this abortion of ovules is due to the random abortion of 50 per 

 cent of the enibrj'o-sacs in the flowers, then those young ovaiies with 

 all their embrj'o-sacs aborted (which comprise theoretically between V32 

 and Vei of the total ovaries), and probably also some of those with 

 many embryo -sacs aborted, will fall in the flower stage, or as small 

 pods. Tliis would of course lower the percentage of aborted ovules in 

 the ripe pods below the theoretical 50. Now there is abundant oppor- 

 tunity for the dropping of such ovaries or pods. The average number 

 of flowers on the largest racemes of 38 plants of Velvet beau grown 



Fig. 12. Selected pod uf the first generation of the Velvet by China cross. 

 Natural size. 



on poles in 1912, was 3()-0. The average number of ripe pods on the 

 largest racemes of the same plants was only 14' 1. The average number 

 of flowers on the largest racemes of the 11 Fi plants of the Velvet- 

 Yokohama cross (grown alongside the Velvet plants) was 33-0. The 

 average number of ripe pods on these racemes was 14'3. Tlius only 

 47 per cent of the Velvet-bean flowers produced ripe pods, and only 43 

 per cent of the hybrid flowers. A similar loss of a large proportion of 

 the flowers or young pods occurs in the Lyon, Yokohama, and China 

 beans, and in most of the segregates from the crosses, under the con- 

 ditions in which thev are ijfrown here. 



