AUTOGAMY BY MOVEMENTS OF STAMENS. 345 



^nd have no further need of pollen, we are bound to infer that this fresh supply is 

 offered for the purpose of effecting by insects' aid the fertilization of other flowers 

 in which there is as yet no available pollen although the stigmas are accessible and 

 mature. Opuntias and all the numerous species of the Rose genus (Rosa) behave in 

 a similar manner. In them also some of the anthers devote their pollen to auto- 

 gamy, whilst the others devote theirs to cross-fertilization. The filiform filaments 

 are curved and of unequal lengths. The anthers borne by the innermost whorl of 

 filaments open first, but their pollen is of no use for autogamy, notwithstanding the 

 proximity of the stigmas, because the anthers are lower than the stigmas and do 

 not spontaneously come into contact with them. Only the filaments of the outer- 

 most whorl are of the right length, and these alone curve over and lean towards the 

 middle of the flower until their anthers rest immediately upon the stigmas. As, 

 however, these anthers are the last to dehisce, autogamy does not ensue till the very 

 last moment of flowering; for the whole of the time up till now the flower is 

 adapted to cross-fertilization only. 



In a number of Ranunculacese, such as Änemo7ie Hepatica, A. Transsylvanica, 

 Ranunculus alpestris, R. acer, and R. montanus, the structure of the flowers 

 resembles to a certain extent that of Roses. A group of pistils with short styles 

 and almost sessile stigmas rises up in the middle of the flower, and is surrounded by 

 a large number of stamens which are arranged in several whorls and are in their 

 turn encompassed by the petaloid sepals. The flowers are protogynous, and at the 

 commencement of their period of bloom can only undergo cross-fertilization through 

 the agency of insects. At a subsequent stage also after the dehiscence of the 

 anthers of the outer whorl and the exposure of their adhesive pollen, the flowers are 

 still adapted to heterogamy; for the distance of these anthers from the stigmas is 

 ■comparatively great, and insects always alight on the group of carpels in the centre 

 and make their way from them over the anthers to the circumference, whence they 

 again take flight in order to visit another flower. But by degrees the stamens of 

 the inner whorl also mature; the filaments, which hitherto have been very short, 

 ^row considerably and curve inwards, laying upon the stigmas their anthers which 

 have meanwhile burst open. The sepals of these Ranunculacese close up when it 

 begins to get dark, and the flowers assume a nodding position owing to a slight 

 inflection of their pedicels. It might be supposed that these movements also are 

 accessory to autogamy, and, as a matter of fact, in many other RanunculaceaB, which 

 will be the subject of discussion later on, such co-operation does occur; but, in the 

 Anemones and Ranunculuses above referred to, no supplementary action of the kind 

 is necessary, and the closing and nodding of the flowers in wet weather and during 

 the night are resorted to merely with the object of protecting the pollen from 

 moisture. 



To this long list of plants with protogynous flowers must now be added a few 

 protandrous species from the genera Gypsophila, Saxifraga, and Cuphea. The 

 flower of Gypsophila repens contains ten stamens, of which five are inserted in front 

 of the sepals and five in front of the petals. In the bud they are all tucked in like 



