338 ANGIOSPERMAE—DICOTYLEDONES 



Wamstorf (Verb. bot. Ver., Berlin, xxxviii, 1896) says that three forms of the 

 species occur near Ruppin : — 



(i) With larger, homogamous hermaphrodite flowers, in which the stigmas do 

 not project from between the perianth leaves. 



(2) With smaller female flowers, among which are found hermaphrodite ones. 



(3) With very small, purely female flowers, in which the stigmas project far 

 beyond the perianth leaves. 



The pollen-grains are white in colour, smooth, rounded-tetrahedral, 37-5 to 44 /x 

 in diameter. 



2453. R. obtusifolius L. (Tulberg, Bot. Not., Lund, 1868, p. 12; Herm. 

 Miiller, 'Fertilisation,' p. 516, ' Weit. Beob.,' II, p. 222; Kirchner, 'Flora v. Stutt- 

 gart,' p. 210; Schulz, 'Beitrage'; Kerner, 'Nat. Hist. PI.,' Eng. Ed. i, II, 295, 298. 

 315.) — Kirchner says that the flower mechanism of this species agrees with that of 

 the preceding one. Tulberg describes the flowers as markedly protandrous ; the 

 stigma is not accessible until after the falling of the stamens, when it is rendered so by 

 the bending back of the perianth leaves — which have hitherto covered it. Kerner 

 states, on the contrary, that the plant is protogynous. Schulz observed gyno- 

 monoecism and androdioecism. Warnstorf describes the pollen-grains as white in 

 colour, irregularly polyhedral, tuberculate, up to 44 /x in diameter. 



Visitors. — Herm. Miiller observed the bee Halictus cylindricus F. 5 on the 

 anthers. 



2454. R. sanguineus L. (Schulz, ' Beitrage,' I, p. 95.) — Schulz describes 

 this species as feebly protandrous, Kerner as protogynous. Schulz observed 

 andromonoecious and androdioecious as well as hermaphrodite flowers. Automatic 

 self-pollination is not possible in the hermaphrodite ones, as pollen cannot easily 

 reach the stigma. Numerous flowers remain unfertilized because the perianth 

 remains closed and the stigmas are therefore hidden. 



2455. R. conglomeratus Murr. (Schulz, loc. cit.) — The comparatively small 

 stigmas of the hermaphrodite flowers of this species usually mature at the same time 

 as the anthers, sometimes a little later, rarely before them. The relative position 

 of the reproductive organs and their homogamy render automatic self-polhnation 

 inevitable. Schulz observed andromonoecious as well as hermaphrodite flowers. 

 He says that this species is not anemophilous, as the flowers are not pendulous on 

 flexible stalks. 



2456. R. maritimus L. (:== R. aureus ]\Iill.). (Schulz, op. cit., I, p. 94.) — 

 The flowers of this species are homogamous. Each of the inner perianth leaves 

 possesses two tooth-like processes on its lowest third, on which the stigmas lie while the 

 flower is still in the bud. When anthesis begins, the stigmas are situated immediately 

 below the anthers, and as both organs mature simultaneously, automatic self-pollina- 

 tion is inevitable. Pollen is seldom removed from the flower, as the anthers project 

 only a little or not at all beyond the perianth, and the stalks are not pendulous or 

 flexible, but either horizontal or erect. Schulz denies the existence of anemophily. 

 The flowers are sometimes gynomonoecious or androdioecious. 



2457. R. pulcher L. (Schulz, op. cit., II, pp. 153-4.)— The perianth of 

 flowers belonging to this species remains almost closed, the anthers projecting only 



