42 STUDIES IN INDIAN TOBACCOS. 



easily blown or knocked off. For this reason half of them were 

 enclosed in large well-perforated bags as a protection. Some of 

 the capsules obtained a fair size, but were not so large as the 

 normal ones. On examination they were, however, found to be 

 quite empty, the ovules not having developed. On only three 

 plants did seed set in all the thousands of flowers which had 

 been castrated, and the total number of capsules was five. 

 In 1910, on a plant of Tj^pe 9, about one hundred flowers Avere 

 castrated and one fully formed capsule was found, the seed of 

 which germinated and produced plants similar to Type 9. In 

 1911, again, on a plant of Type 9, one capsule containing seed 

 was found in about one hundred castrated flowers. In this type 

 the stamens are so much shorter than the style, that if enclosed 

 in a bag the flowers normally set no seed and the majority of the 

 capsules drop. It has always been found necessary to self this 

 type in order to obtain sufficient pure seed to maintain the 

 culture. If the seeds in these two capsules were due to apogamy, 

 this method of seed production must be theexception and not the 

 rule. It is of course possible that the two capsules were due to errors 

 in .castration. The other three capsules were found on one plant of 

 another type, but by an accident these were not examined. 

 They were large, well-formed capsules, apparently containing 

 seed, but unfortunately were destroyed before this fact had been 

 definitely ascertained. 



Considering the great number of flowers examined and 

 the fact that every stimulus to apogamous seed formation 

 had been given to the plants by pruning and capsule removal, 

 the results obtained are exceedingly small, even if we assume 

 that all five capsules contained parthenogenetically formed 

 seed. 



In addition, in all the first generations (nine in number) 

 which have been raised at Pusa during the last five years, each 

 culture containing about one hundred plants, no individuals 

 resembling the mother plant have been detected. The cultures 

 have been absolutely uniform and the reciprocals identical. I 



