504 



ANGIOSPERMAEMONOCOTYLEDONES 



the hermaphrodite stage, therefore, cross-pollination will be effected by insect-visit 

 if visitors bring foreign pollen and alight on the stigmas in the middle of the flowei 

 In the second part of this stage both cross- and self-pollination may be brougli 

 about by insect-visits ; autogamy takes place by fall of pollen should such visits fail 

 I could observe the latter in plants which flowered in my laboratory ; the pollen of 

 the three inner stamens fell so plentifully on the carpels that not only were the 

 stigmatic papillae thickly covered with it, but a large portion rolled down the inne^ 

 side of the ovary into the base of the flower. In the first (purely male) stage sue 

 masses of pollen may always be found in the trough-like hollows of the sepals and 

 petals, whence it is removed by gusts of wind. The shallow form of the stigmas, 

 however, renders it scarcely possible for the pollen to be deposited on them by such 

 means. 



Nectar is so abundantly secreted on the carpels that there is always a large drop 

 in the cleft between each two adjacent ovaries. 



Fig. 416. Butoinus umbellatuSy Z.. (from nature). (i) Flower in the first (male) stage: 6 stamens 

 have developed, diverging from the centre and opening their anthers ; the stigmas are still immature. 

 (2) Do., in the second (hermaphrodite) stage : all the anthers have dehisced, but the innermost stamens have 

 only diverged slightly ; the stigmas have become receptive. The pollen in the dehisced anthers is not 

 indicated. j, stigmas. 



I 



Sprengel's account differs from mine in stating that all nine anthers have 

 already lost their pollen when the stigmas mature, so that automatic self-pollination 

 seems to be excluded. Hermann Muller, on the contrary, says that all nine anthers 

 remain well covered with pollen until the stigmas are fully receptive, and come into 

 contact with them partly automatically, autogamy being brought about in this way. 



A. Schulz found flowers near Halle to be usually homogamous or feebly pro- 

 tandrous, more rarely markedly protandrous. There the stigmas are frequently 

 divergent and papillose when the flower opens, but usually they only become fully 

 receptive by the time some of the anthers have dehisced. The stamens are curved at 

 the base and shorter than the pistil; the laterally dehiscing anthers are therefore 

 situated 2-4 mm. below the stigmas, so that in the usually upright flowers automatic 

 self-pollination seems to be excluded, as a rule, and is only possible in occasional 

 cases, when stigmas and anthers are at the same level. 



