LILIACEAE 



463 



stigma mature simultaneously, but automatic self-pollination is excluded, for these 

 organs do not come into contact, there being about 3 mm. distance between them. 

 So long as I observed the honey-bee only as quite a casual visitor, I was of opinion 

 that regular autogamy was probable as well as pollination, by means of the extremely 

 violent wind prevailing in the islands. The filaments are 3-4 mm. long, and beset 

 with very numerous dense yellow hairs, almost one mm. long and directed obliquely 

 upwards, which not only increase the conspicuousness of the flowers, but also serve 

 as hold-fasts for insects, and as weels for catching pollen. Sometimes this is 

 temporarily deposited here when it has missed the stigma of the same flower until 

 it is carried to it by a fresh gust of wind, when automatic self-pollination is effected, 

 and sometimes the pollen is carried by the wind to a flower of an adjacent plant, 

 when it is again caught first by the filament-hairs and transferred thence to the 

 stigma when occasion offers, crossing thus being brought about. It may be seen 

 that this latter occurrence is not exactly rare from the fact that numerous pollen- 

 masses — sharply distinguished from the yellow filament-hairs by their yellow-red 



Fig. 401. Narthiciiim ossifragitni^ Huds. (from nature, diagrammatic^ (i) Flower, seen from 

 above. (2) Do., from the side, after removal of some stamens and part of tlie perianth (x 2A). 



a, anthers ; p, perianth ; s, stigma. 



colour — are to be found not only on the inner sides of the stamens, but are also 

 often found caught by the hairs on the outer sides of the filaments ; this latter 

 must therefore be foreign pollen. Having at a later date observed various bees 

 and flies as busy and constant visitors of these flowers in the North Frisian Islands, 

 1 am now convinced that the anemophily just described must be considered only as 

 an exception, and that pollen is as a rule transferred by insects. Kerner states 

 that autogamy takes place towards the end of anthesis by fall of pollen. 



Willis and Burkill's description (' Fls. and Insects in Gt. Britain,' I, p. 267) of 

 the flower mechanism for Central Wales agrees with mine for the North Frisian 

 Islands; they, however, frequently observed automatic self-pollination, the flowers 

 opening so late that the anthers had already dehisced and had dusted the stigmas. 

 The reason for this might have been that the plant was almost at the end of the 

 flowering season. The tissue at the base of the filaments is, according to Willis and 

 Burkill, juicy, and is perhaps bored by bees, if ihey visit the flowers. 



Visitors. — The following were recorded by the observers, and for the localities 

 stated. — 



