'Bulletin Soc. Bot. de France' tome 7 1860 page 468.) They there produce flowers not offering any peculiarity in structure, excepting that their corollas, though properly coloured, are deformed. These flowers may be ranked as cleistogamic, as they are developed, and not merely drawn, beneath the ground.
Ononis columnae.
Plants were raised from seeds sent me from Northern Italy. The sepals of the cleistogamic flowers are elongated and closely pressed together; the petals are much reduced in size, colourless, and folded over the interior organs. The filaments of the ten stamens are united into a tube, and this is not the case, according to Von Mohl, with the cleistogamic flowers of other Leguminosae. Five of the stamens are destitute of anthers, and alternate with the five thus provided. The two cells of the anthers are minute, rounded and separated from one another by connective tissue; they contain but few pollen-grains, and these have extremely delicate coats. The pistil is hook-shaped, with a plainly enlarged stigma, which is curled down, towards the anthers; it therefore differs much from that of the perfect flower. During the year 1867 no perfect flowers were produced, but in the following year there were both perfect and cleistogamic ones.
Ononis minutissima.
My plants produced both perfect and cleistogamic flowers; but I did not examine the latter. Some of the former were crossed with pollen from a distinct plant, and six capsules thus obtained yielded on an average 3.66 seeds, with a maximum of 5 in one. Twelve perfect flowers were marked and allowed to fertilise themselves spontaneously under a net, and they yielded eight capsules, containing on an average 2.38 seeds, with a maximum of 3 in one. Fifty-three capsules produced by the cleistogamic flowers contained on an average 4.1 seeds, so that these were the most productive of all; and the seeds themselves looked finer even than those from the crossed perfect flowers. According to Mr. Bentham O. parviflora likewise bears cleistogamic flowers; and he informs me that these flowers are produced by all three species early in the spring; whilst the perfect ones appear afterwards, and therefore in a reversed order compared with those of Viola and Oxalis. Some of the species, for instance Ononis columnae, bear a fresh crop of cleistogamic flowers in the autumn.
Lathyrus nissolia.
This plant apparently offers a case of the first stage in the production of cleistogamic flowers, for on plants growing in a state of nature, many of the flowers never expand and yet produce fine pods. Some of the buds are so large that they seem on the point of expansion; others are much smaller, but none so small as the true cleistogamic flowers of the foregoing species. As I marked these buds with thread and examined them daily, there could be no mistake about their producing fruit without having expanded.
Several other Leguminous genera produce cleistogamic flowers, as may be seen in Table 8.38; but much does not appear to be known about them. Von Mohl says that their petals are commonly rudimentary, that only a few of their anthers are developed, their filaments are not united into a tube and their pistils are hook-shaped. In three of the genera, namely Vicia, Amphicarpaea, and Voandzeia, the cleistogamic flowers are produced on subterranean stems. The perfect flowers of Voandzeia, which is a cultivated plant, are said never to produce fruit (8/13. Correa de Mello 'Journal of the Linnean Society Botany' volume 11 1870 page 254, particularly attended to the flowering and fruiting of this African plant, which is sometimes cultivated in Brazil.); but we should remember how often fertility is affected by cultivation.
Impatiens fulva.
Mr. A.W. Bennett has published an excellent description, with figures, of this plant. (8/14. 'Journal of the Linnean Society Botany' volume 13 1872 page 147.) He shows that the cleistogamic and perfect flowers differ in structure at a very early period of growth, so that the existence of the former cannot be due merely to the arrested development of the latter,--a conclusion which indeed follows from most of the previous descriptions.