IN PLANTS HAVING IRREGULAR COROLLAS. 419 
ceeded in making it, a symmetrical and perfect thing—perfect, 
that is to say, in the sense that the normal flower is perfect. 
How, then, if the seed of this flower had been saved and arace of 
Streptocarpus having the symmetry of No. 4 had arisen? Such 
a thing may be possible enough. By what means would the mor- 
phologist of the future have traced the descent of this flower by 
comparison, and so forth? Of course we recognize that had it 
been possible for us to have observed the development of this 
flower, some indication of a change of plan might perhaps have 
been found ; but if all trace of such a great and essential change 
as this is at the moment of its occurrence thrust far back into 
development, the prospect of finding traces now of any large 
Proportion of the changes which happened long ago is not 
hopeful. We are therefore disposed to think that the first 
teaching of the facts of Variation is this: that comparison of 
forms is not likely to be a good guide to the history of those 
forms; and that there is no evidence that degrees of apparent 
relationship of form are an indication of degrees of actual rela- 
tionship by descent; aud that nothing short of an actual know- 
ledge of the processes of Variation and a discernment of the 
changes which are possible to living things from those which are 
Impossible to them, can be of any use in the solution of the pro- 
blem of Descent. Until such knowledge shall have been reached, 
any hypothesis of the * atrophy " of parts, of the “fusion” of 
parts, and generally the attempt to reconstruct what is unknown, 
must, of necessity, be unfounded and misleading, and had better 
not be undertaken. 
The fact that there are certain variations which are, as it 
were, integral, and which, if they occur at all, occur in their 
complete form always, or nearly always, is of course perfectly 
well known. Darwin* gives several cases of this as illustrations 
of the phenomenon that certain characters cannot blend. In 
particular he instances the case of peloric Antirrhinum, which is 
closely akin to Linaria. He states that of 137 plants raised as 
the second generation of a cross between normal and peloric 
Antirrhinum, only two were in an intermediate condition, while 
the others were all either normal or peloric. Our object now is 
to show that this principle is widely true of variations which are 
* Anim. & Pl. 1885, ii. p. 71 ; see also Mélicocq, in Bull. Soc. Bot. France, 
1859, vi. p. 716. 
