Mr. Bennett found on the banks of the Wey that the plants which bore cleistogamic flowers alone were to those bearing perfect flowers as 20 to 1; but we should remember that this is a naturalised species. The perfect flowers are usually barren in England; but Professor Asa Gray writes to me that after midsummer in the United States some or many of them produce capsules.
Impatiens noli-me-tangere.
I can add nothing of importance to Von Mohl's description, excepting that one of the rudimentary petals shows a vestige of a nectary, as Mr. Bennett likewise found to be the case with I. fulva. As in this latter species all five stamens produce some pollen, though small in amount; a single anther contains, according to Von Mohl, not more than 50 grains, and these emit their tubes while still enclosed within it. The pollen-grains of the perfect flowers are tied together by threads, but not, so far as I could see, those of the cleistogamic flowers; and a provision of this kind would here have been useless, as the grains can never be transported by insects. The flowers of I. balsamina are visited by humble-bees (8/15. H. Muller 'Die Befruchtung' etc. page 170.), and I am almost sure that this is the case with the perfect flowers of I. noli-me-tangere. From the perfect flowers of this latter species covered with a net eleven spontaneously self-fertilised capsules were produced, and these yielded on an average 3.45 seeds. Some perfect flowers with their anthers still containing an abundance of pollen were fertilised with pollen from a distinct plant; and the three capsules thus produced contained, to my surprise, only 2, 2, and 1 seed. As I. balsamina is proterandrous, so probably is the present species; and if so, cross-fertilisation was effected by me at too early a period, and this may account for the capsules yielding so few seeds.
Drosera rotundifolia.
The first flower-stems which were thrown up by some plants in my greenhouse bore only cleistogamic flowers. The petals of small size remained permanently closed over the reproductive organs, but their white tips could just be seen between the almost completely closed sepals. The pollen, which was scanty in amount, but not so scanty as in Viola or Oxalis, remained enclosed within the anthers, whence the tubes proceeded and penetrated the stigma. As the ovarium swelled the little withered corolla was carried upwards in the form of a cap. These cleistogamic flowers produced an abundance of seed. Later in the season perfect flowers appeared. With plants in a state of nature the flowers open only in the early morning, as I have been informed by Mr. Wallis, who particularly attended to the time of their flowering. In the case of D. Anglica, the still folded petals on some plants in my greenhouse opened just sufficiently to leave a minute aperture; the anthers dehisced properly, but the pollen-grains adhered in a mass to them, and thence emitted their tubes, which penetrated the stigmas. These flowers, therefore, were in an intermediate condition, and could not be called either perfect or cleistogamic.
A few miscellaneous observations may be added with respect to some other species, as throwing light on our subject. Mr. Scott states that Eranthemum ambiguum bears three kinds of flowers,--large, conspicuous, open ones, which are quite sterile,--others of intermediate size, which are open and moderately fertile--and lastly small closed or cleistogamic ones, which are perfectly fertile. (8/16. 'Journal of Botany' London new series volume 1 1872 pages 161- 4.) Ruellia tuberosa, likewise one of the Acanthaceae, produces both open and cleistogamic flowers; the latter yield from 18 to 24, whilst the former only from 8 to 10 seeds; these two kinds of flowers are produced simultaneously, whereas in several other members of the family the cleistogamic ones appear only during the hot season. According to Torrey and Gray, the North American species of Helianthemum, when growing in poor soil, produce only cleistogamic flowers. The cleistogamic flowers of Specularia perfoliata are highly remarkable, as they are closed by a tympanum formed by the rudimentary corolla, and without any trace of an opening.