GRAMINEAE 519 



978. Panicum L. 



2987. P. sangtdnale L. (= Digitaria sanguinalis Scop). (Hildebrand, 

 ' Bestaubungsverh. d. Gramineen/ p. 757.) In this species self-pollination only is 

 possible at first, in consequence of the simultaneous protrusion of stigma and anthers, 

 but crossing may be effected when the anthers have fallen, as thej stigmas are 

 persistent. 



2988. P. Crus-galli L. Hildebrand (op. cit.) says that this species agrees 

 with the preceding one. 



2989. P. miliaceum L. (Kirchner, 'Flora v. Stuttgart/ p. 1 19.) The flowers 

 of this species are homogamous. Stigmas and anthers protrude simultaneously from 

 their fairly wide entrance. The anthers dehisce along their entire length. Although 

 the filaments are rather thin, the stamens are not pendulous, but approach the 

 stigmas when the glumes close. Crossing is therefore favoured at first, and automatic 

 self-pollination possible later. 



979. Setaria Beauv. 



2990. S. italica Beauv. This species, according to Kirchner (' Flora v. Stutt- 

 gart,' p. 119), possesses the same flower mechanism as Panicum miliaceum. 



980. Phalaris L. 



2991. P. arundinacea L. (= Digraphis arundinacea Trin.). (Hildebrand, 

 op. cit,, p. 756.) In this species the stigmas, which project a little from the glumes, 

 can at first be dusted by the pollen of older flowers only, then by their own, and 

 finally by that of younger ones. Warnstorf (Verb. bot. Ver., Berlin, xxxviii, 1896) 

 says that the flowers are feebly protogynous to homogamous ; the anthers are dirty- 

 reddish in colour, and their filaments still stiff when they dehisce ; pollen is scattered 

 during the morning. The pollen-grains are white in colour, irregular, resembling 

 a blunt, usually five-sided pyramid with a rounded base, up to 43 /a long and 25-31 ^ 

 broad. 



2992. P. canadensis L. (Hildebrand, op. cit. ; Kornicke, op. cit. ; Kirchner, 

 'Flora v. Stuttgart,' p. 121.) When the flowers of this species open the outer glume 

 spreads out, but the inner ones only open widely enough for anthers and stigmas to 

 force their way between them. In doing so the anthers often protrude on the side 

 turned away from the axis, but the stigmas on that facing it, so that self-pollination 

 is then prevented. The filaments either remain erect or tip over, and in the latter 

 case the anthers partly dehisce before and partly after this upsetting. Crossing is 

 therefore favoured at some times and self-pollination at others, while the latter is 

 sometimes inevitable. It may also happen that the anthers do not project between 

 the glumes at all, but that their dehisced tips protrude, while the rest of them remains 

 entirely concealed. Kornicke states that the flowers open between 12 to 4, but 

 Hildebrand only observed this to take place towards evening. It never seems to 

 occur in the morning. 



981. Pennisetum Rich. 



2993. P. spicatum Korn. Kornicke describes the pollination of the species 

 as follows: the inflorescence matures from the middle outwards (not above the 



