CONTRIBUTIONS FROM THE WISLEY LABORATORY. 167 
in the young trees propagated from them, bud selection of a kind is 
frequently advocated and sometimes practised by nurserymen and 
others with the object of improving, or at least maintaining, the 
cropping powers of the variety. This idea is based upon the sup- 
position that, by continually selecting buds or scions from the trees 
that regularly carry good crops, there will arise a " fixed " good- 
bearing type of that variety, whereas, if, on the other hand, trees be 
so propagated by materials secured from trees that regularly crop 
very poorly, the result will be a gradual deterioration of the variety, 
and the loss of the original cropping powers. 
Methods of selection have without a doubt played a very important 
part in the improvement of all kinds of plants, but it must be remem- 
bered that such improvements almost always follow hybridization 
and seed selection. Reasoning suggesting that like improvements will 
follow bud selection, such as propagating only from vigorous and 
heavy-cropping trees of a variety, or even selecting large buds for 
budding, stout shoots for scions, etc., is invalid. 
In one instance, that of hybridization and seed selection, we have 
the combination of characters from two parents, a combination that is 
seldom exactly reproduced in any two seedlings, so that great variations 
and possible improvements are to be expected. In the other we have 
the bud or scion which has inherited all the characters of the parent 
tree, no more and no less, being an actual part of it, which characters 
will be present in the tree resulting, and only suppressed or intensified 
by external causes. Differences between individual trees are therefore 
due not to heredity but are acquired and fluctuating characters 
due to environment and stock, and as such are not likely to be trans- 
missible. No doubt were it possible to grow trees of a variety 
exhibiting such wide differences under precisely the same con- 
ditions as the original tree of the variety, no variation whatsoever 
would be observed. 
The trees in this experiment have been grown under almost exactly 
the same conditions of soil and situation, and it is unlikely that in this 
instance environment has played any part in causing the wide differ- 
ences exhibited between the trees, and it may be correctly assumed 
that the real cause lies in the great variation among the stocks. 
Much diversity of opinion has existed on this matter for many 
years, but comparatively recent research in other countries, reported 
since the commencement of this experiment, has been followed by 
a decided leaning to the opinion that, whilst several kinds of varia- 
tion may be met with, some of which taking the form of " bud sports " 
etc. may be perpetuated when propagated vegetatively, the very great 
majority of variations exhibited among individual trees of a variety 
in the plantation and garden are not transmissible. 
Further evidence in support of this opinion is added by the result 
of this experiment. 
That great care must necessarily be exercised, however, in propa- 
gating only from healthy trees will be obvious, for buds and scions 
