1899. J NATURAL SCIENCES OF PHILADELPHIA. 93 



diiriug the Avhole mouth. I cannot say that these millions of violet 

 ilowers had no insect visitors during that period, but it is safe to 

 «ay that if any were visited by insects the number would be so 

 iinconsiderable as to have no influence on the result. 



Numberless flowers were opened from day to day. All the 

 anther-sacs were abundantly supplied with perfect pollen. From 

 the earliest period the upper jjortion of the style, with its stigma, 

 would be in advance of the stamens. AYhen the pollen would be 

 projected from the anther-cells, the membranous apj)endages of the 

 anthers would be so closely appressed to the style that it would be 

 very difficult for pollen to get through to reach the stigma. 

 Though pollen grains could get through this close-fitting sheath, the 

 :stigma is so thoroughly encased by the mass of hair in the throat, 

 that it is inconceivable that pollen should ever reach it. As a 

 matter of fact, I could never find a grain of pollen on the jiortion 

 of the style above the membranous sheath, and of course none in 

 the stigma, though the ovarium clasped by the anthers themselves 

 would be completely pollen-dusted. 



I was satisfied from these observalions that the flo Wei's could 

 not prove fertile, and was prepared to believe that so far as Viola 

 -eucullata is concerned, the popular impression regarding infertilty 

 in this species was correct. I started to examine several other 

 species in the same relations. Viola tricolor is conceded to have its 

 early perfect flowers fertile. I examined some flowers from the 

 wild species grovving in my garden. I found the stigma barely 

 •extending beyond the membrane sheath, which, by the way, was not 

 closely clasping, nor Avas there a mass of hair on the petals to 

 guard the receptive portion from an attack by the pollen grains. 

 It did seem that the fecundity of the early flowers could be ac- 

 counted for. But again I could not find a single pollen grain on 

 the apex of the style, or stigma, as we would terminologically say! 



By the time the first week in May was reached, I was surprised 

 to find an abundant crop of seed vessels on the earlier flowers in 

 the orchard, the peduncles curving toward the ground just as they 

 •do in the cleistogamic condition. A number, probably one-third, 

 had failed to perfect seed, but they had advanced considerably 

 before withering, showing that their early demise was not from 

 non-pollination, but from a failure of nutrition, just as in many 

 fruit trees we find the ground beneath their branches strewed with 



