326 THE CROSSING OF FLOWERS. 



region of the younger flowers, where pollen is Ijeing shed, and their stigmas thus 

 actually come into contact with the pollen. The process is still further promoted 

 in this species of Woodruff by the fact that the last flowers to be produced are 

 staminate. In the Red-berried Elder (Sambucus racemosa), various species of the 

 Cornel and Dogwood genus (Cornus florida, 7)ias, sanguinea), in the Vines (Vitis), 

 which bear true hermaphrodite flowers, in the Tufted Loosestrife (Lysimachia 

 thyrsiflora), and in many Spiraeas (Spircea), the arrangements for geitonogamy 

 resemble those of Siler trilohuTii in that the direction of the style and the position 

 of the stigma remain unaltered, but the filaments of the anthers elongate and bend 

 over so as to deposit the pollen on the stigmas of adjacent flowers. In the Way- 

 faring-tree and Guelder-rose (Viburnum Lantana, V. Opulus) we have yet another 

 contrivance — the pollen which is shed from the bent anthers of one flower falls to 

 the bottom of the cup-shaped corolla of an adjacent one, where the large cushion- 

 like stigma is situated. 



The process of geitonogamy in the Snake-root {Calla palustris) and in Saxi- 

 fraga juniperifolia to some extent resembles the fall of pollen in Compositae. The 

 flowers in these plants are crowded in short spikes or fascicles. They are proto- 

 gynous, the stigma in the lower half of the inflorescence not ripening until the 

 upper flowers are shedding their pollen. Now, when the anthers begin to shrivel 

 and the pollen is thrown out, it necessarily falls on the fertile stigmas below. In 

 those species of Veronica which have spicate inflorescences {Veronica maritima, 

 spicata, spuria, &c.), the method of geitonogamy is slightly difierent, for here the 

 style undergoes peculiar movements during the flowering period. The crowded 

 flowers are all protogynous, and the stigmas of the flowers which first unfold are 

 exposed to the pollen of plants of other species. This continues for two days. 

 Meanwhile the stamens of the lowest flowers on the spike have elongated and 

 pushed their anthers into the place first occupied by their stigmas ; the anthers 

 then dehisce and shed their pollen. But shortly before this the style has bent 

 sharply dow^nwards so that it is impossible for its stigmas to come in contact with 

 this liberated pollen. Not until all the pollen has fallen down by the shrivelling up 

 of the anthers or has been carried away by insects do the styles again straighten 

 and project almost horizontally from the axis of the spike. The upper flowers on 

 the spike undergo the same course of development, but the stages here are two days 

 later. On this account the pollen falls from the anthers of the higher flowers just 

 when the styles of the lower flowers again become straight. The still fresh stigmas 

 at the ends of the styles thus come into line with the falling pollen and are 

 eflSciently pollinated by it. 



A similar process occurs in Eremurus (see fig. 293^, p. 309), but here there is 

 no fall of pollen. The stigmas at the end of the just straightened style are brought 

 by their change of position directly into contact with the orange-yellow pollen still 

 clinging to the withered anthers of the higher fiowers. Many of the styles, of 

 course, brush by the anthers without effecting this contact, and accordingly many 

 stigmas in the racemes of Eremurus remain unpollinated. The transfer of the 



