578 THE GENESIS OF NEW SPECIES. 



and conspicuous that they were accepted as the results of crosses between two 

 species even by those amongst the earlier Botanists who were most disinclined to 

 recognize the existence of plant-hybrids. Now, these Mullein hybrids do not for 

 the most part mature any seeds. The pistil itself is usually incompletely developed, 

 and even if one or other of the fruit-capsules does develop, the ovules in it are 

 abortive and infertile. Nevertheless it would be erroneous to say that no Mullein 

 hybrid has ever produced seeds capable of germination. Two such hybrids were 

 artificially generated in my garden, viz.: Verbascum rubiginosum, by crossing 

 Verbascum Austriacum with the pollen of Verbascum phosniceum, and Verbascum 

 pseudophoeniceum, by crossing Verbascum Blattaria with V. phoeniceum. The 

 former of these hybrids, it is true, never produced seeds capable of germination; 

 but in the case of the second, although most of the capsules were empty and 

 abortive, a few containing fertile seeds ripened; so that even the hybrids of 

 Mulleins are not invariably sterile. 



Anyone who will look beyond the limited range of that particular genus will be 

 convinced that in respect of their capacity for sexual reproduction hybrids do not 

 differ essentially from plants which have been admitted to be ' true ", permanent 

 species by Botanists of all periods. For the case of a few of these true species, as, 

 for instance, Cochlearia Armoracia, Grambe tataria, Lilium bulbiferum, Lysimachia 

 Nummularia, Rubus odorus and R. Nutlcaensis, it has long been known that if the 

 stigmas are dusted with pollen from the stamens in the same flowers very few seeds, 

 if any, are set, whilst pollen from other flowers is obviously preferred by them. On 

 the other hand, there are true species whose flowers are pseudo-hermaphrodite, i.e. 

 they have the appearance of being bisexual, but are really unisexual (cf. p. 294). 

 In one individual we find that the ovaries are fully developed, as also the stamens, 

 but that no pollen capable of fertilizing ovules is produced in the anthers; in 

 another plant the ovaries are imperfectly developed, whilst the anthers are filled 

 with effective pollen. For seeds to be set in such circumstances two individuals at 

 least are requisite, and pollen from a plant bearing pseudo-hermaphrodite male 

 flowers must be transferred to the stigmas of the pseudo-hermaphrodite female 

 flowers. Now hybrids with pseudo-hermaphrodite flowers also exist, and in their 

 case, as in that of true species, two kinds of individual are requisite to produce 

 seeds capable of germination. Supposing, however, in such a case that the two 

 kinds of plant necessary for reproduction do not grow close together or do not 

 flower simultaneously, or that one of them is altogether absent a contingency which 

 must often occur fertilization cannot be effected, and consequently no seeds can be 

 formed. It is scarcely necessary to amplify the proposition that dioecious hybrids 

 behave in this respect in the same manner as true species, and that pollination and 

 fruit-formation may in them be impeded likewise by dichogamy or by hetero- 

 stylism. In many hybrids, again, as in true species, the relative positions of 

 stamens and pistil, the height of the stigma, the length of the filaments, and other 

 conditions of the kind are not conducive to autogamy, and consequently no trans- 

 ference of pollen from the anthers to the stigmas in the same flower can take place 



