BUD VARIATION. 
255 
that we are unable to say positively whether this was a case 
of reversion or no.” 
It may perhaps be said that seedling variations such as happen 
in the apple or the pear are the necessary outcome of the cross - 
fertilisation to which the plants in question have been 
subjected for ages, just as the bud variations in the case of the 
pelargonium are. This may be true in some cases, but can 
hardly be so in all : for instance, in a bed of seedling conifers, 
such as Lawson’s cypress or deodars, raised from imported seeds 
taken from wild plants, often from the same cone, the number 
of seedling varieties is often large. In the case of cultivated 
plants as of domestic animals, Mr. Darwin . has shown how the 
variations that arise are directly connected with the objects for 
which the particular plant or animal is cultivated. A plant, 
for instance, grown for the sake of its fruit is apt to vary in its 
fruit characters more than in its leaf characters. But although 
this may and no doubt does apply to a considerable extent in 
the case of seminal variations, it seems less applicable in the 
case of bud variations, as will be judged from the illustrations 
before given, as also from the negative evidence afforded by a 
plant like the Jerusalem artichoke, which is propagated exclu- 
sively by its tubers, and indeed never ripens its seed in this 
country, and which has produced no variation by “ sport ” or 
dimorphism, although so largely grown and for so long a period. 
Mr. Darwin attempts to explain the phenomena of bud 
variation, as of inheritance and reproduction generally, by his 
hypothesis of pangenesis. This hypothesis proceeds on the 
assumption that every cell of a living organism gives origin 
to an innumerable host of u gemmules ” in minuteness as 
in number transcending conception. These gemmules divide 
and multiply, or they lie dormant possibly for ages. They 
circulate throughout the organism or they become aggre- 
gated together, and so form embryos or buds, and they are 
transmitted from one generation to another. There is nothing 
improbable in the assumption of the existence of these gem- 
mules ; and, if we take their presence for granted, it is easy to 
see how they afford an explanation of the phenomenon of 
reversion to an ancestral condition, such as bud variation so 
frequently presents. Gfemmules derived from a plant’s remote 
progenitors are, according to the hypothesis, circulating in the 
present generation, and it only requires the occurrence of 
favourable conditions to determine the revivification of these 
now dormant gemmules to reproduce the ancestral form. There 
still remains the difficulty of ascertaining what the favourable 
conditions are which determine this change. The reason for 
the prolonged dormancy of the gemmules is also not obvious. 
But, supposing we admit the gemmule hypothesis as sufficient 
