HYBRIDISATION VIEWED FROM SYSTEMATIC BOTANY. 199 
considerations will help us to realise their essential distinctness as a 
class. 
A species may be described as a collection of like individuals inhabiting 
a definite geographical area, more or less continuously or discontinuously 
according to circumstances, and over which they have spread by diffusion 
from some original birthplace. Widely diffused species are generally 
variable, owing to modifications effected by change of environment, which 
eventually result in the production of distinct but allied species, mostly 
in distinct geographical stations, though by subsequent diffusion their 
areas may again overlap. Many of these allied species retain their con- 
genital affinity, yielding hybrid offsprmg when intercrossed, and when 
such species grow in proximity to each other the pollen frequently gets 
transferred by natural agencies, and under such conditions natural 
hybrids frequently occur. Hybrids—natural or artificial alike—almost 
invariably show a distinct combination of the characters of the two 
parent species, which may usually be traced on careful examination, 
though sometimes the influence of one parent preponderates to such an 
extent that it becomes difficult to identify the second one. ‘They are 
usually very variable, some individuals derived from the same two species 
being so dissimilar as to have been at first considered essentially distinct 
in their origin. Those which occur wild are usually found intermixed 
with the parent species, though in the case of wind-borne seeds this may 
not always apply. And they are frequently rare, as compared with the 
parents, though some species hybridise with such facility as to form 
notable exceptions. 
These are, of course, only relative peculiarities, and afford no clue to 
the identification of a hybrid as such. The fact is no absolute differential 
character can be pointed out. It was formerly thought that hybrids 
between distinct species were universally sterile, and consequently a so- 
called “fertile hybrid ’’ was taken as evidence that the parents were only 
forms of one; a view which long ago proved altogether untenable. We 
now know that some bond fide hybrids are as completely fertile as their 
parents, and even those that are sterile—i.c. have their reproductive 
organs functionally impotent—frequently appear structurally perfect. 
It is this absence of any reliable character by which a hybrid can be 
recognised that constitutes the real obstacle to progress in our knowledge 
of “natural hybrids.’ It is open to every sceptic to ask for “ proofs,”’ 
and we have already seen how the views of many shrewd observers have 
been treated. It is not always sufficient to point out that such and such 
a plant is rare, is only found where two other species grow together (not 
where they grow separately), that it is intermediate in character, or that 
it combines the characters of the two, and is evidently a natural hybrid 
between them. The test of direct experiment is alone convincing in 
many cases, and believing, as I do, that natural hybrids are more common 
in certain groups than we have yet realised, I welcome any additional 
light that can be thrown on the subject, from whatever source it may 
come; and instead of viewing the work of the hybridist unsympathetically, 
I only wish some of his energy could be diverted into other channels, for 
I believe that nothing would do more to advance our knowledge than a 
few well-directed experiments in those groups in which so much difference 
