312 Belling. 



after infiltration with paraffin, into sections, 14 microns thick. Every 

 one of the ovules (42 in all) had a complete embryo-sac nearly filling 

 the nucellus (Fig. 5). Over 40 pollen-grains were found embedded in 

 the glutinous stigma of a Velvet-bean flower, and it may be supposed 

 that pollination is always in excess. But, doubtless from accidental 

 physiological causes, not every ovule of the Velvet bean produces a 

 mature seed. There are some ovules in some pods of every Velvet bean 

 plant I have examined which do not form seeds, but dry up, and are 

 seen in the ripe pods as black specks less than one iuillimeter across. 

 A Velvet bean plant, taken at random from a pedigreed row grown on 

 poles iu 1912, gave the following data from the dry pods. 



Seeds Aborted Percentage of 



" ■ (viable or dead) ovules ovules aborted 



V-174-41 427 2126 272 12 



The plant had 208 pods with no aborted ovules. It had 13 per cent 

 of aborted ovules in the distal halves of its pods, that is, in the first 

 three seed -places. It is probably more accurate to reckon with the 

 distal halves than with the whole pods, because this method disregards 

 the sometimes extensive abortion of basal ovules in certain, presumably 

 starved, pods. Considering the large number of pods borne by one 

 plant, and that the plants, in Florida, rarely escape an attack of cater- 

 pillars in September, I regard the small percentage of undeveloped 

 ovules as physiological, not genetic. 



The Lyon bean resembles the Velvet bean in physiological abortion 

 of ovules. A Lyon bean plant, taken at random out of a row grown 

 on poles in 1912, gave: 



Pedigree No. Pods Seeds , , . . j 



° ovules ovules aborted 



L-154-31 446 1912 483 20 



The plant had 165 pods with no aborted ovules. (I have found Velvet 

 bean plants with probably as high a percentage of aborted ovules; but 

 in these records the dead seeds were purposely reckoned with the aborted 

 ovules, and I cannot give the percentage of the latter.) This Lyon 

 bean plant had only 13 per cent of aborted ovules in the distal halves 

 of its pods. It belonged to the parent line of the two crosses. 



The Yokohama bean has about the same percentage of good seeds 

 as the Velvet and Lyon. I made several counts of the percentages of 

 good seeds, but only reckoned the aborted ovules in one plant out of 

 a row grown on poles in 1913. 



