In animals, it would appear, such saltatory variations occur much more rarely than in plants; the case examined in detail by Darwin of the 'black-shouldered peacock' which suddenly appeared in a poultry-yard is an example of this kind. Much more numerous, however, are the instances among plants, and especially among plants which are under cultivation. This indicates that we have here to do with the effect of external conditions, of nutritive influences which cause the slow variation of certain determinants, sometimes abetting and sometimes checking. As soon as a majority of ids varied in this way comes to lie in a seed, a sport springs up suddenly and apparently discontinuously--a plant with differently coloured or shaped petals or leaves, with double flowers, with degenerate stamens, or with some other distinguishing mark, and these new characters persist if the variety is propagated without inter-crossing.
But it happens sometimes, though more rarely, that not the whole plant but individual shoots may exhibit the variation. To this class belong the 'bud-variations' of our forest trees, the copper-beeches, copper-oaks, and copper-hazels, the various fasciated varieties of oak, beech, maple, and birch, and the 'weeping trees'; also the numerous varieties of potato, plantain, and sugar-cane. It seems that only a few of these breed true when reproduced from seed, or in other words, they usually exhibit reversions to the ancestral form: on the other hand, in the weeping oak for instance, nearly all the seedlings exhibit the character of the new variety, though 'in varying degrees.' The records as to the transmissibility of bud-variations through seed are probably not all to be relied upon, and new investigations are much to be desired, but the fact that in many cases they may be propagated not only by means of layers and cuttings but by seed also, is most important in our present discussion, for it proves that here too the varied determinants must be contained in a majority of ids. As it is only a single shoot that exhibits the saltatory variation, only the germ-plasm which was contained in the cells of this one shoot can have varied, and it must have done so in so many ids that the variation prevailed and found expression. But that, in this case also, the variation does not appear in all, but only in a small majority of ids, is proved by the frequent reversion of bud-varieties to the ancestral form. I have already reported a case of this kind shown to me by Professor Strasburger in the Botanic Gardens in Bonn, where a hornbeam with deeply indented 'oak-leaves' had one branch which bore quite normal hornbeam leaves. In my own garden there is an oak shrub of the 'fern-leaved' variety, whose branches bear some leaves of the ordinary form; variegated maples with almost white leaves often exhibit in individual branches a reversion to the fresh green leaves of the ancestral form. We see from this that what is so energetically disputed by many must in reality occur--namely, differential or non-equivalent nuclear division--for otherwise it would be unintelligible how the ids of the new variety, if they once attain a majority in the tree, could give place in an individual branch to a majority of the ancestral ids. Only differential nuclear division, in the manner of a reducing division, can be the cause of this. Of course this implies only a dissimilar or differential distribution of the ids between the two daughter-nuclei, not a splitting up of the individual ids into non-equivalents.