FLOWERS OF IMPA TIENS FULVA 1 53 



fulva some flowers were castrated just before the opening and were 

 sealed in paper en\'elopes. In some of these the pollen sacs had already- 

 ruptured. In pallida the results of five experiments were as follows: 

 one capsule with five seeds; one with four; one with three; and two> 

 failed to form fruit. Somewhat similar results were obtained with 

 fulva; unfortunately the records of these have been lost. The results 

 with these castrated flowers would indicate that spontaneous seK- 

 DoUination sometimes occurs before the opening of the flowers. 



The number and size of the seeds in cleistogamous and chasmogamous 

 capsules, Miss Riatt (39 )states to be the same in I. pallida. So far as 

 the number of seeds in the capsule is concerned, this does not agree with 

 my observations on either pallida or fulva In pallida, I have found 

 in the vicinity of Philadelphia the number of seeds in the chasmogamous 

 capsules is generally four or five; in the cleistogamous capsules two or 

 three, or not uncommonly on the lower branches only one. This num- 

 ber represents also the condition in fulva. My observation of the 

 sizes of the two types of seeds in fulva agrees with Miss Riatt's report 

 of pallida. The average size in the chasmogamous seed was 5.1 X 4 m. m. 

 and in the cleistogamous seed 5X4 m.m. The weights varied with the 

 condition of drying. In seeds gathered from the bursting pod the 

 average weight in 48 chasmogamous seeds was 0.0208 grams and in 

 59 cleistogamous seeds was 0.0195 grams. The size of capsules is 

 quite different in the two types of flowers, but the smaller capsule of 

 the cleistogamous flowers is accounted for by the smaller number of 

 seeds. 



There seems to be no general rule in the different species as to the 

 size and number of cleistogamous as compared with chasmogamous 

 seed. Shaw (45) says for Polygala polygavia that there is no difference in 

 the seed resulting from the different types of flowers. Helene Ritzerow 

 (41, p. 71) quotes Burck (11) to the effect that in Heteranthera spicata the 

 cleistogamous fruit is a half larger than the chasmogamous and in 

 Heteranthera Potamogeton, Solms, and Heteranthera Kotschyana, Fenzl, 

 it is more than twice as large with twice the number of seeds; in Com- 

 melina bengalensis (41 p. 172) the seeds of the cleistogamous flowers 

 are double the weight of those of the chasmogamous flower. In Van- 

 dellia nummular if olia (41, p. 201) the number of seeds in the cleisto- 

 gamous flowers is on the whole lower than in the chasmogamous. In 

 Houstonia caerulea (41, p. 204) the size of the seed is larger in the chas- 

 mogamous capsule, but the number of seeds per capsule is larger in 

 the cleistogamous capsules. In Specularia perfoliata (41, p. 208) the 



