i8 ANGIOSPERMAE—DICOTYLEDONES 



broaden from apex to base, the latter being hairy. The flower mechanism is hke 

 that of Campanula, but only the anthers wither after the pollen has been shed on the 

 style, the filaments persisting as nectar-covers. Crossing is secured by insect-visits, 

 but automatic self-pollination is possible should these fail, for the stigmatic branches 

 ultimately bend so far back that they touch the pollen clinging to their own style. 



Visitors.— Willis and Burkill observed 2 Muscids, of which one was sufficiently 

 large to remove pollen from the style : also Thrips, and a bug which crept into the 

 flowers. 



1736. W. tenuifolia A. DC. (=Hedraeanthus tenuifolius A. DC). (Kirchner, 

 Jahreshefte Ver. Natk., Stuttgart, liii, 1897, p. 217.) — Kirchner has examined plants 

 of this species in the Hohenheim Botanic Garden, and says that their protandrous 

 mechanism agrees essentially with that of Campanula. The bright blue flowers are 

 arranged in large terminal heads. The style is as long as the corolla, and its end 

 divides into two stigmatic branches which become recurved, though not sufficiently 

 to touch the style with their tips, so that automatic self-pollination is apparently 

 excluded. 



Visitors. — Kirchner observed the honey-bee. 



516. Phyteuma L. 



Literature. — Sprengel, ' Entd. Geh.,' pp. 113-15; Herm. Miiller, ' Alpen- 

 blumen,' pp. 406-9. 



Flowers protandrous, and belonging to class S. 



In this genus (and Jasione, g. v.) the end of the style is at first covered with 

 closely set erect hairs, as in Campanula, and these receive the pollen which is 

 dehisced in the bud. As in Compositae this is swept out of a tube by the elongating 

 style, but in this case the tube is not made up of the anthers but of the long strap- 

 shaped corolla-lobes, which are at first closely apposed. After dehiscence the stamens 

 contract into a crumpled mass. The lower free parts of the corolla-lobes bend 

 somewhat outwards, so that their upper parts can be drawn down. As meanwhile 

 the style elongates, the pollen is not merely pushed up by the stylar brush, but 

 entirely swept out of the tube which surrounds it. 



When the growing tip of the style has reached the upper end of the tube made 

 up by the ends of the corolla-lobes, the three until now closely apposed stylar 

 branches begin to separate, and rupture the tube (already split below), so that this 

 glides down the style. The branches then quickly diverge until their papillose inner 

 surfaces occupy the place where the pollen masses were heaped up during the first 

 stage of anthesis. As the insect visitors (bees and humble-bees) creep over the 

 inflorescences from below upwards, they not only, like all other visitors, regularly 

 cross the older flowers with the pollen of the younger, but also, owing to the fact 

 that anthesis progresses from below upwards, constantly eff'ect crossing between 

 different stocks. 



Should insect-visits be prevented by unfavourable weather. Kerne r says that the 

 stylar branches roll back until they touch the pollen still clinging to the stylar hairs, 

 thus effecting automatic self-pollination. 



