84 BULLETIN OF THE TORREY CLUB [vOL. 52 



varied among flowers open on the same day, and it varied from 

 day to day. 



Naturally the question arises as to whether plants of the new 

 form are highly self-compatible, or are more decidedly or more 

 uniformly so than are plants of the old form. In them, self- 

 pollinations from pollen of the longer set of stamens are "legit- 

 imate." Legitimate pollination is not possible in self-pollina- 

 tions of any individuals of the three old forms. For them, legit- 

 imate pollinations are all inter-form cross-pollinations which as 

 a rule appear to be more productive of fruit and seeds than are 

 illegitimate self- or cross-pollinations in accordance with the re- 

 sults obtained by Darwin (1865 and 1877). 



The first plant of the semi-homomorphic form was highly 

 self-compatible not only in the large number of capsules pro- 

 duced but in the high average of seed per capsule and in the 

 viability of the seeds. Twelve plants of its offspring having the 

 new form of flower were tested by controlled self-pollinations; 

 three were highly self-fruitful, five were somewhat self-fertile, 

 and four appeared to be self-incompatible. Another of these 

 plants was grown in isolation and it yielded fine capsules for 

 nearly every flower that bloomed. But also four of the six 

 plants with intermediate flowers which were tested were self- 

 fruitful and one of these was very highly so. 



The semi-homomorphic form appears to be highly self- 

 fertile, but the degree varies for individual plants and some are 

 self-sterile. This condition has also been found (Stout, 1923) 

 for other forms and especially for the mid-styled form. The 

 new form appears from the data at hand to be at least as highly 

 self-fruitful as is the mid-styled form. 



The appearance of only one plant of the new form among a 

 considerable number of plants together with the fact that this 

 form has not been reported previously suggests a mutation. Its 

 form of flower re-appears in a considerable nimiber of its off- 

 spring, so there is an hereditary value to the new character. It 

 does not immediatel}' breed true, so the new form is not at 

 first homozygous. 



The new form presents a correlated modification of both the 

 pistil and at least one set of stamens. In the size of the stig- 

 matic lobes there is decided resemblance to the long-pistils; the 

 lengths of both jiistil and stamens are modified; in the color of 



