AUTOGAilT BY A CO-OPERATIOX OF iI0VEilENT5. 3S1 



of the perianth have expanded the pedicels stand out horizontally from the axis 

 of the inflorescence, and the flowers face sideways. The flowers are protandrous. 

 The anthers of the three stamens, situated in front of the little pits in the ovary 

 in which honey is secreted, dehisce at the same moment as the perianth-leaves 

 expand, and these anthers are placed in such a position as to be touched by 

 insects as they enter the flower. The stigma is still immature at this stacfe. 

 A Uttle later, when the stigmatic tissue has developed the power of retaining poUen. 

 the stamens move out of the way of insects towards the periphery of the flower 

 and thus render it pc«sible for cross-pollination to be efiected by such of these 

 visitors as bring with them poUen from younger flowers. In the third stage of 

 the flower's duration the pedicel bends down until the flower is at last tndy 

 nodding. The stamens have meantime executed a reverse movement towards the 

 middle of the flower, and the stigma is found to be just underneath one of the 

 anthers belonging to a stamen of the shorter class. These anthers always have 

 s-jme poUen left in them, for they do not open till the second stage of the 

 flowers development and cannot have undergone contact with insects. The 

 gradual shrinkage of the anthers now causes this store of poUen to fall out of 

 them on to the adjacent stigma, and thtis autogamy is efiecte-l just before the 

 flower fades. 



The hermaphrodite flowers of certain Rosaceee — Dryas octopetalo., Geurii 

 CKcine'inn, G. wionto.nwm, G. rejAo.v.s, PrAentillOj o.trosangv.ineo., P. repen-'i, and 

 Walditiinia geoides — and those of some Eannnculaceae, viz., Adonis vernoJA-s, 

 Anemone a.lpina, and A. balckmis afibrd particularly instructive examples of 

 autogamy. In aU these plants the flowers are protogynous and are characterized 

 ty having a large number of carpels crowded together in the centre and surrounded 

 by equally numerous stamens, which are disposed in several whorls. In the 

 Eosacea; in question the stamens are tucked down before the bud unfolds, and 

 they do not straighten out until the anthers are nearly ready to open. Dehiscence 

 occTtts first in the anthers which belong to the outermost whorl of stamens, and 

 are furthest away from the stigmas of the bunch of ovaries in the middle. This 

 relative position of the two sets of organs excludes all possibility of autogamy, 

 especially when the flower is erect: on the other hand, cross-pollination is quite 

 Ekely to be effected by insects, which alight on the stigmas, thence proceed 

 towards the circumference of the flower, licking up honey and collecting pollen 

 on the way, and finally take wing from the edge to visit other flowers. By 

 degrees, the stamens of the innermost whorl come to maturity; they straighten 

 out and elongate, and their anthers with pollen exposed upon them are brought 

 to the same level as the stigmas of the central pistils. A transference of pollen 

 to some of these stigmas is now certain to ensue, and is rendered all the more 

 inevitable by the outward inclination and inflection of the styles belonging to the 

 pistils most remote from the centre which now take place, and bring the corre- 

 sponding stigmas into direct contact with the poUen. But if this were all, the 

 ^igmas in the centre might get no share of pollen in the event of an absence 



