RESULTS OF VARIOUS CROSSINGS. 405 



If ripe pollen from the male flowers of a plant be placed on the mature stigmas 

 of female flowers of the same species, the result may be regarded as certain. This 

 holds good equally for monoecious and dicecious plants, and for such also as have 

 apparently hermaphrodite flowers in which one or other set of sexual organs is more or 

 less abortive, so that the flowers in question are to all intents and purposes unisexual. 

 It has been shown that for plants with true, non-heterostyled, hermaphrodite 

 flowers, a transfer of ripe pollen from the anthers of one flower to the mature 

 stigma of another, belonging to the same species, constantly leads to fertilization. In 

 those cases only in which the stamens are of unequal lengths is pollination attended 

 mth unequal results, according as the pollen has been taken from the longer or 

 shorter stamens. 



The behaviour of plants with hermaphrodite heterostyled flowers is peculiar. 

 Pollen from the anthers of a short-styled flower applied to the stigma of a long- 

 styled flower, or from a long-styled flower applied to the stigma of a short-styled 

 flower, gives the best result. The other possible combinations, i.e. pollen from long- 

 styled or short-styled flowers applied to the stigmas of the same class of flower give 

 indifferent results, and frequently none at all. Experiments have shown in the case 

 of the Loosestrife {Lythrum Salicaria), which, as we have seen (p. 303), possesses 

 long-, mid-, and short-styled flowers, that crossings between stamens and styles of 

 the same length ("legitimate unions") are fruitful, whilst all other crossings ("illegiti- 

 mate unions") are either quite sterile or followed by only a sparing production of seed. 



It has been demonstrated that the pollen-cells of heterostyled flowers vary both 

 in size and colour according to the length of stamen (or height of anther or corolla) 

 producing them. Thus, in the Loosestrife, the dry pollen-cells of the long stamens 

 are green, and 30-38 /j- long and 20-26 /"■ broad; those of stamens of the middle length 

 are yellow and 23-26 /"■ long and 13-16 /^ broad; those of the short stamens are also 

 yellow, but 20-25 /^ long and 11-13 ij- broad (p- = yttw millimetre). In the Cowslip 

 (Primula officinalis [ = veris] ) the pollen-cells, produced by anthers at the mouth of 

 the corolla-tube (and destined for a long style), have a diameter of 30 /i, those arising 

 from anthers low down the tube (and destined for a short style), a diameter of only 

 20 /i. The explanation offered by Delpino of this difference in size — that, other 

 things being equal, pollen-tubes which have to traverse a greater distance to reach 

 the ovules require a more ample supply of reserve-materials than those which have 

 less far to penetrate — sounds plausible, but the problem is probably a rather more 

 complex one than appears at first sight. As a general result of experiments upon 

 heterostyled plants we may say that the most copious production of good seed 

 results from a pollination of stigmas with pollen from stamens of similar height. 



As to the result of autogamy (self-pollination) in ordinary hermaphrodite 

 flowers, it would appear, in all cases where the stamens are of unequal heights, 

 that the pollen has a dissimilar eff'ect according as it comes from a longer or shorter 

 stamen. If pollen, which would not unaided reach the stigma of the same flower, be 

 artificially transferred to that stigma, the product is usually very small. If, however, 

 pollen which would ultimately reach the -stigma of the same flower be artificially 



