THE SELF-FERTILISATION OF PLANTS. 
9 
morphic, bat others, such as P. scotica , are not ; and Primula 
veris and others which are usually dimorphic can become homo- 
morphic, as in fig. 47, which is a short-styled form of the Primrose 
(according to the position of the anthers), but self-fertilising in 
consequence of the style having elongated. Glaux maritima 
(fig. 48) has the style often recurved so as to secure pollination. 
Pinguiculaceje. — P. lusitanica is self-fertilising by the 
stigma being recurved and dipping into the two gaping anther- 
cells, as represented in fig. 51. 
PolygonacejE. — Conspicuously flowering species like the 
Buckwheat, which is dimorphic, and P. Bistorta , with pink 
flowers, are attractive to insects ; but Muller has shown that 
P. aviculare, with its minute blossoms, and without honey- 
glands, is self-fertilising, while P. Convolvulus and P. Hydro- 
piper appear to me to be habitually cleistogamous. 
Alismace 2 E. — The stamens of Alisma Plantago in the first 
stage of the flower are spread away as represented by the left- 
hand stamen in fig. 52, but subsequently they bend back ; and, 
although the anthers are extrorse, they lie immediately over or 
among the stigmas, which are also bent towards them, as in the 
right hand of fig. 52. 
Orchid ace^e. — This order, as is so well known, is almost 
entirely dependent upon insect aid for fertilisation, yet the Bee 
Ophrys is self-fertilising, in consequence of the pollinia falling 
out of the anther-cells, and being retained by their glands, swing 
backwards and forwards, and so strike against the stigma (figs. 
53 and 54). The other case mentioned by Mr. Darwin is Cephal- 
anthera (fig. 55), of w T hich he remarks : — u Whilst the flower 
is still in bud, or before it is as fully open as ever it becomes, 
the pollen-grains which rest against the upper sharp edge of the 
stigma (but not those in the upper or lower parts of the mass) 
emit a multitude of tubes, deeply penetrating the stigmatic 
tissue.” * 
Commelynacejs. — I have figured a case of self-fertilisation 
in Tradescantia erecta , which I found at Kew (fig. 56). It 
was late in the autumn of 1876, and none of the flower-buds 
expanded. The corolla in all cases withered within the bud, 
and several of the stamens were imperfect ; the style was bent 
down beneath the shrivelled corolla, and was pollinated by one 
perfect stamen. The fruits set seed, which I subsequently 
grew. 
G-ramineje. — Hordeum murinum , as far as I am aware, is 
always cleistogamous. The filaments do not seem ever to escape 
from the closed glumes, but remain doubled or twisted back, so 
that the anthers lie in contact with the stigmas (fig. 57). 
* “ Fertilisation of Orchids/' p. 106, 1st ed. 
