GRAMINEAE 525 



mature at the beginning of anthesis. These remain receptive within the still divergent 

 glumes when the anthers have fallen. Cross-pollination is therefore possible at the 

 beginning and end of anthesis. Warnstorf says that the anthers dehisce between 

 4 and 7 p.m. for the second time. 



3025. A. sativa L. — This species is homogamous. Godron states that the 

 flowers open between 2 and 4 p.m.; and that the stamens fall over before the anthers 

 have dehisced, so that pollen cannot drop on the stigma of the same flower. 



Although Godron thus considers that crossing predominates, Rimpau seems to 

 think that automatic self-pollination is very certain, as the filaments elongate very 

 slowly and the anthers dehisce close to the stigma. Rimpau observed several times 

 that the anther-walls suddenly contract actively, and part of the pollen is thereby 

 flung immediately upon the stigma. When cross-pollination is excluded the species 

 is entirely fertile. Allogamy seems to occur as an exception in the smaller, upper 

 flowers. (C/! Hackel, Bot. Ztg., Leipzig, xxxviii, 1880.) Hildebrand considers that 

 crossing is favoured rather than self-pollination. He states that in dry weather the 

 flowers open in the afternoon or towards evening. In unfavourable weather the flower 

 remains closed and pseudo-cleistogamous autogamy takes place. Von Liebenberg 

 also describes the .species as self-fertile. 



Appel (in a letter to the author) considers the existence of transition-forms 

 between A. sativa L. and A. fatua L. a proof of the occurrence of cross-pollination, 

 at least in individual species. Haussknecht does not consider these transition-forms 

 to be of hybrid origin, and infers that A. sativa is a cultivated form of A. fatua, 

 produced by centuries of breeding, but Kornicke and Appel (the latter of whom 

 observed these forms in large numbers near Coburg, Wiirzburg, and Schaff"hausen) 

 are more inclined to suppose that they are really hybrids. This opinion is strengthened 

 by the fact that when further cultivated the plant becomes variable, even if cross- 

 pollination is excluded. 



Kirchner ('Flora v. Stuttgart,' p. 133) says that anthesis begins in the afternoon 

 when the weather is favourable, and lasts until evening, proceeding from the apex 

 of the panicle to its base. The lower of the two flowers of each spikelet opens first, 

 and the other shortly after if the weather is favourable ; as the flowers are usually 

 pendulous the anthers sway in front of the stigmas during anthesis. If they dehisce 

 before the stamens have finished elongating, self-pollination is effected freely, if later, 

 this is prevented. Pollination takes place almost regularly within the flowers of the 

 same panicle, and rarely crossing, which occurs only in the small flowers at the top 

 of the spikelet. 



According to Kornicke many varieties, besides opening in the afternoon, do 

 so also, rarely or frequently as the case may be, in the morning, beginning in 

 favourable weather before 8 a.m., and continuing for some hours. The bulk of the 

 flowers, however, always open after midday. 



The time of anthesis is, however, altogether subject to certain external influences ; 

 it may be delayed during hot dry weather by sunshine and the dryness of the soil. 

 A brief shower can then cause a rapid opening of many flowers. Damp warm 

 weather induces earlier anthesis. 



3026. A. orientalis Schreb., and 3027. A. nuda L. — Hildebrand describes 



