60 JOURNAL OF THE ROYAL HORTICULTURAL SOCIETY. 
descent, there they are now, grouped into species as we know. How 
did this come about ? b 
We all know the accepted view. We start from the fact that, since of 
all forms of life many more are born than can possibly survive, some 
—indeed, nearly all—must perish and leave no descendants. Next we 
observe the fact of Variation—that even the offspring of the same parents 
are never precisely alike, but vary. Now, since all cannot survive, it is 
clear that different individuals have a different chance of survival and of 
being represented by descendants. For each individual this chance will 
depend on the degree to which its structure and aptitudes fit it to bear 
its part in the struggle to which it is exposed. Briefly, on the whole the 
fittest will survive and breed. 
Lastly, as the places in life that the organisms fit are diverse, so the 
forms of the surviving organisms are diverse too. 
Everyone who cares at all for Natural History knows this reasoning, 
and knows also the difficulties by which its application to the facts of 
Nature is beset—how simple the theory seems when thus stated in 
general terms, but how hard it is to apply it in detail to a particular 
case. ‘ 
Of all these difficulties the most serious are two. The first 
is the difficulty which turns on the magnitude of the variations by 
which new forms arise. In all the older work on evolution it is 
assumed, if the assumption is not always expressly stated, that the 
variations by which species are thus built up are small. But if they are 
small, how can they be sufficiently useful to their possessors to give 
those individuals an advantage over their fellows? That is known as the 
difficulty of small or initial variations. 
The second difficulty is somewhat similar. Granting that variations 
occur, and granting too that if they could persist and be perpetuated 
species might be built up of them, how can they be perpetuated ? 
When the varying individuals breed with their non-varying fellows, will 
not these variations be obliterated? This second difficulty is known as 
that of the swamping effect of intercrossing. Now on each of these two 
points the work of the hybridist and the experimental breeder comes in 
exactly. It is he who can see the variations arise, and can note their size 
and find out exactly how large they are—whether they are great or 
small—whether offspring do really differ but little from their parents, 
or whether, in certain cases and in respect of certain characters, the 
differences in variation may not be very great and definite; whether, 
also, the supposed swamping effect is a real one or not, or to what extent 
it is real, and in regard to what characters. 
I need not tell a body of persons, most of whom have themselves 
made experiments of this kind, that in numberless cases both great and 
thoroughly definite variations do occur. This much every practical man 
now recognises. But we are far from knowing which kinds of variations 
may thus be definite and palpable, and which are not. All we know is 
that both large variations and small variations occur, some in one 
character and others in other characters, and that characters which in 
one species may vary greatly and suddenly, in other species vary only 
slowly or hardly at all. All this isa matter which comes daily under 
