60 PROCEEDINGS OF THE ACADEMY OF [1889. 



While preparing the chapter on Polygonum arifolium for Native 

 Flowers and Ferns of the United States (1879, Vol. II, p. 154, 

 series I), a double system of flowers was noted — one smallei\than 

 the other — but as both of them seemed to have the sexual organs 

 perfect with anthers abundantly polliniferous, no reference was 

 made to the fact, but the whole matter was laid over for future in- 

 vestigation. 



Attention was drawn to the subject a few years subsequently, by 

 noting numbers of nectar-loving Vespce, as well as the common 

 honey-bee vigorously at work on the smaller flowers of Polygonum 

 Virginianum. These smaller flowers are pure white, and, after 

 mid-day, expand sufficiently to expose the whole sexual organs to 

 easy ocular examination. There are two series of stamens. The 

 exterior lean outwardly, the inner bend inwards and the abundantly 

 polliniferous anthers are in actual contact with the apparently 

 perfect stigmas. The whole gynoecial system seems perfect. Those 

 who speculate in these matters would surely say the flowers were 

 specialized for self-fertilization. At the base of the inner series of 

 stamens are glands which copiously exude nectar, attracting the 

 insects above noted. It was evident, however, that none of these 

 small, though apparently perfect, flowers, ever produced seeds. On 

 the same plant are other flowers, larger, and of a greenish purple 

 tint. From these flowers the styles with their hooked apices pro- 

 truding, extend considerably beyond the sepals. These flowers always 

 produce perfect seeds. Though the smaller flowers seem specialized 

 for self-fertilization, in view of the fact that they produce no seeds, 

 that they are abundantly nectariferous and polliniferous, and that 

 the larger flowers project the pistils beyond the closed sepals, appar- 

 ently out of reach of own-pollen, and yet are abundantly fertile, the 

 general conclusion that the whole arrangement is for cross-fertiliza- 

 tion would surely be pardonable. 



During the several years past in which continuous observations 

 were made, no insect has been noted visiting the fertile flowers. It 

 is safe to say they are not fertilized by insect agency. The nectar 

 in the infertile flowers doubtless serves some good purpose, but it 

 .certainly has no reference to cross-fertilization by insect agency. 



Setting aside as wholly out of the question the production of nec- 

 tar, and the visits of insects in connection with the fertilization of the 

 productive flowers, it still remained to be considered whether the 



