390 AUTOGAMY. 



The processes which lead to autogamy in the Water Avens (Geutn rivale), the 

 Raspberry {Rubus Idceus) and some other Rosaceae allied to these are even more 

 complicated than those above described. Thus, for example, the flowers of Geum 

 rivale, on the day that they open, face laterally and have their stalks horizontal; 

 the filaments are short, and the anthers are all closed, while the stigmas which 

 project in a tuft 2 mm. beyond the anthers are already mature. At this stage 

 insects may occasion cross-fertilization, but autogamy is not yet possible. Subse- 

 quently, the filaments lengthen and the anthers of the longest stamens open and 

 come into contact with some of the stigmas at the periphery of the bundle of styles. 

 The pedicel is now curved and the flower nods; consequently, the pollen which falls 

 fi'om the anthers above, when they shrivel, is forthwith received by the outer stigmas 

 of the fascicle of styles, that is to say, by those of the outer stigmas which appertain 

 to the upper half of the flower. The pollen which falls from the anthers of the 

 under half of tlie flower when they dry up, is caught, on the other hand, by the 

 petals on that side of the flower, and is afterwards transferred, by means of an 

 elongation of these petals, to the stigmas of the adjacent reflexed styles. A couple of 

 days later the pedicel is curved into a semicircle, and the flower hangs down with 

 its mouth towards the ground. By this time the anthers of the shorter stamens 

 are open; the whole flower has become loosened, and the fascicle of styles resembles 

 a sheaf of corn. All the styles, including those in the middle, become twisted and 

 reflexed to the extent necessary to bring the stigmas underneath the most recently 

 opened anthers, and when these anthers shrivel and the pollen is forced out, it falls 

 upon the central stigmas, which hitherto have not been furnished with any. Thus, 

 in this case we have (1) the inflection of the pedicels, (2) the elongation of the petals, 

 (3) the elongation of the stamens, and (4) the inflection of the styles — all co-operating 

 towards the same end, namely, that in the event of no insects visiting a flower all 

 the stigmas may receive pollen from the anthers developed in the flower itself. 



The foregoing descriptions, though extremely brief and cursory, give a general 

 idea of the many kinds of contrivances whereby autogamy, as well as heterogamy, 

 is promoted in hermaphrodite flowers. It is evident from them that any mechan- 

 ism which leads to autogamy has full scope for its operation only if ci'oss-pollina- 

 tion has not previously been effected. Again and again we have found that certain 

 processes only take place in the event of a flower being unvisited by insects through 

 whose agency cross-fertilization would have been brought about. In this connection 

 we have also the remarkable phenomenon that many flowers adapted to cross- 

 fertilization by insects do not open at all when there is no chance of their being 

 visited by the agents in question. In the mountainous districts of the temperate 

 zones it often happens that rainy weather sets in just at the time when the flowers 

 are about to open, and that it lasts for weeks. Humble- and hive-bees, butterflies, 

 and flies retire to their hiding-places, and for a considerable time cease to pay any 

 visits to flowers. The growth of the plants is not, however, arrested during this 

 period, and even in the flowers themselves development quietly progresses if the 



