GEITONOGAMY. 319 



other plants, but not with that which lies on the sweeping hairs below the stigmas. 

 As soon as the strap-shaped corollas begin to wither and shrivel, the two stylar 

 branches diverge strongly, and twist and turn like tiny snakes sideways and 

 downwards. At the same time adjacent styles come nearer to one another, and it 

 is therefore natural that the stylar branches of neighbouring flowers should get 

 entangled. In this way the stigmas of one flower (which are still in a receptive 

 state) necessarily come in contact with the pollen on the sweeping hairs of another, 

 and pollination ensues. 



The same process occurs in the flowers of the Lettuce (Lactuca), the Alpine 

 Sow-thistle (Mulgedium), and in Oiondrilla, only here the heads contain more 

 florets than in the form just described. The stylar branches do not undergo snake- 

 like movements, but they diverge widely and roll back a little, an action altogether 

 suflScient to bring them into contact with the styles of neighbouring flowers and 

 to promote a crossing. It is worth noticing that the corollas of the ray-florets of 

 Prenanthes roll outwards when they begin to fade, while those of the Lettuce and 

 of the other Composites mentioned fold up and form a hood over the stylar 

 branches during crossing. The Salsify (Tragopogon), Hawkweed {Hieracium,), 

 Crepis, Scorzonera, Hawkbit (Leontodon), Dandelion (Taraxacum), and many 

 other Composites, of which these plants may be regarded as typical, contain in eacli 

 head as many as 100 ray-florets ai-ranged in spiral series (of. fig. 222^, p. 112). The 

 strap-shaped corollas separate in the morning and fold together in the evening, and 

 similarly the anther-tubes and styles are inclined somewhat to the circumference 

 of the capitulum in the morning, but come close together and assume an upright 

 position in the evening. This gradual ajjproach ultimately becomes actual contact, 

 and since the development of the protandrous florets proceeds from the circumfer- 

 ence towards the centre of the capitulum, the stigmas of the outer florets are 

 mature at the time when the pollen has only just been swept out of the anther- 

 tubes of the inner florets. The contact of the adjoining flowers, therefore, neces- 

 sarily leads to cross-pollination. The fact that the corollas of the ray-florets in 

 any capitulum are of unequal length (fig. 222*, p. 112) has also a close bearing on 

 this process. If they were all equally long this contact and crossing would be 

 impossible, for division walls would be interposed between the styles of the outer 

 and inner florets. But the inner corollas are just short enough to allow the styles 

 to touch one another. In many of these plants, e.g. in the Salsify (Tragopogon), 

 geitonogamy is also assisted by the arrangement of the flowers in each capitulum, 

 each flower of an outer row being placed exactly between two of the next inner 

 series. When the capitulum closes, the two curved stylar branches of an outer 

 floret, with their exposed stigmatic surfaces, become applied to the pollen-covered 

 styles of the inner flowers immediately to right and left in front of them. 



There are comparatively few species of Composites having exclusively tubular 

 florets in which cross-pollination occurs between the members of the same capitulum. 

 The most remarkable of these species belong to the Hemp Agrimony genus (e.g. 

 Eupatorium, aroTnaticum and cannahinum; see figs. 294 ^ and 294 -). The capitula 



