THE FLORAL STEM. 741 



duced. Ver}' many composites witli ligulate florets, e.g. the species of Salsify and 

 Hawkwoed {Tragopogon and Hieraciwm), periodically open and close their capitula, 

 i.e. the ligulate portions of their flowers curve for a time outwards, so that the upper 

 side is turned towards the sky; they then again become erect, curve inwards, and 

 at length close tightly together. In this closing of the capitulum the stigmas of 

 the peripheral flowers become pressed against the pollen of the central ones, and in 

 this waj' a crossing is necessarily brought about between neighbouring florets. All 

 these crossings, however, could not occur if the flowers of a plant were developed 

 at great distances from each other and all unfolded at the same time, and there is 

 no doubt that the formation of capitula, umbels, close i-acemes, spikes, and cymes, 

 ranks as an importjint contrivance for accomjjlishing the cross-fertilization of the 

 flowers. 



Another advantage obtained by the close grouping of the flowers consists 

 in the fact that certain portions of one flower serve as temporary resting-places for 

 the pollen falling from an adjoining flowci-, which at the moment of dehiscence is 

 not yet ready for dispersion in the air. In order to clearly explain this contrivance, 

 which may be observed in catkins, I will take the case of the flowers of the Walnut 

 (Juglans regia) figured on the next page. As long as the male inflorescence is 

 immatm-e, the flowers are crowded together in a short, thick spike, the free end of 

 the rachis being directed upwards. Simultaneously with the development of the 

 pollen in the anthers, however, very remarkable changes are brought about in a 

 short time in the whole inflorescence. Within a few days the rachis elongates to 

 three or four times its former length, and becomes limp and pendent; the flowers 

 are in this way somewhat separated and brought into an inverted position, so that 

 now the open side of each flower is directed downwards, and the lower side upwards. 

 When the wind is still, the anthers, hanging on thin short filaments, open, and the 

 pollen rolls out of them as a powdery mass. It does not fall directly into the air, 

 but, first of all, drops on to the under side of a neighbouring flower which previously, 

 in the erect spike, stood above the anthers in question, but now that the spikes have 

 become pendent, is situated below them. This under side is plainly excavated as 

 a depression, and the pollen of the flower above is deposited in it for a time, as 

 shown in the illustration over page (fig. 184 2). The pollen has to reach the stigmas 

 of flowers developed a long distance from the catkins, often on other branches up 

 above. It would be highly disadvantageous, if, after the dehiscence of the anthers, 

 the pollen should fall immediately to the ground; it would then be lost and wasted, 

 and neither favourable winds nor lightly hovering insects would be able to carry 

 it from the earth to the stigmatic flowers on the branches of the tree. But in the 

 depressions on the under sides of the flowers, as if in a waiting-i-oom, it occupies the 

 most favourable position conceivable. While there is no wind, the tassel-like spikes are 

 undisturbed, and the pollen remains quietly in its temporary resting-places; but as 

 soon as a gust of wind comes, the spikes oscillate, swinging to and fro like pen- 

 dulums, and the pollen, emptied and blown out of the pit-like cavities, is carried to 

 the neighbouring branches and whirled round the tree-crown on to the stigmas, in 



