ORIGIN OF VARIETIES. 9^5 



is proved most conclusively by the fact that seeds from the same fruit produce 

 diflferent varieties, either exclusively or together v^^ith the hereditary parent-form. 



Although the production of varieties and the form they assume are not the 

 direct results of external influences, yet the continuance of the existence of a variety 

 may be determined by these influences. When a variety is produced, the question 

 arises whether it will thrive best in damp or in dry ground, in sunny or shady places, 

 and so forth ; whether it can reproduce itself under these circumstances, or whether 

 it will die out. The conclusion follows that hereditary varieties arise independently 

 of direct external influences, but that the continuance of their existence depends on 

 external causes. A variety which occurs only in a particular locality is not produced 

 by the conditions of this particular locality ; but it alone furnishes the peculiar con- 

 ditions of life which this particular variety requires, while other varieties which have 

 arisen at the same place disappear. 



It has already been shown in Sect. 34 that hybrids show in general a tendency 

 to the production of varieties. Two diff"erent sets of hereditary characters are com- 

 bined in a hybrid, and there is hence a strong tendency towards the formation of new 

 characters which may be more or less hereditary. Hybridisation is therefore one of 

 the most important means at the command of the horticulturist for disturbing the 

 constancy of inherited characters and producing a number of varieties from two dis- 

 tinct ancestral forms ^. But even the ordinary sexual union of two individuals of a 

 species, as in dioecious, dichogamous, or dimorphic plants, may be considered as a 

 kind of hybridisation ; in these cases also the individuals which unite must cer- 

 tainly be different, since otherwise their cross-fertilisation would be no more pro- 

 ductive than self-fertilisation. In these cases therefore two sets of characters which 

 diff"er, though it may be but slightly, also unite in the descendants ; and if a hybrid 

 from two different species exhibits a strong tendency to variation, the cross-fertil- 

 isation of two diff'erent individuals of one and the same species may at least give 

 rise to a slight tendency in the same direction. It is therefore probable that in the 

 cross-fertilisation of different individuals — towards which there is always a tendency 

 in nature even in hermaphrodite flowers — we have a perpetual cause of variation in 

 plants. But this is by no means the only cause of variation, as is shown by the 

 existence of bud-variation, and by the reflection that the difference between indi- 

 viduals which produce a variable progeny is itself due to slight variation. 



A great number of facts point to the conclusion that almost every plant has a tendency 

 to vary continually and in different directions, while every new character which is not 

 produced directly by external agencies tends at the same time to become hereditary. 

 If notwithstanding this many wild plants and some cultivated ones are very constant and 

 produce no varieties which can be distinguished externally, this is mainly the result of the 

 fact that the newly produced varieties are unable to exist in the conditions by which they 

 are surrounded, or at least soon disappear, a point to which I shall recur more in detail. 

 The hereditary transmissibility of acquired characters exhibits itself in a most peculiar 

 way when it does not affect the whole of the parent-plant, but only a particular branch. 

 A still more remarkable case was observed by Bridgman. He noticed that the spores 

 from the lower inner part of the lamina of the leaves of the varieties Scolopendrium 



^ See also Naudin, Compt. rend. 1864, vol. LIX. p. 837. [Journ. Roy. Hort. Soc. new series, 

 vol, I, p. 1.] 



