IN PLANTS HAVING IRREGULAR COROLLAS. 413 
history of which nothing is known, it is unnecessary to invent a 
hypothetical method by which its perfection may have been 
achieved by the compounding of minimal changes; and to pro- 
pose such an hypothesis is to gratuitously invoke difficulty. The 
principle which we have put third, plain as it is upon the facts of 
Variation, is most imperfectly recognized, and indeed, in the 
loose consideration so often given to this subject, the very con- 
trary is frequently assumed. 
We are therefore led to recognize that the forces which 
control the forms of these flowers are such, that they may vary 
greatly, and may, as it were, remake the flower; but the flower 
thus remade may again seem to be a perfect thing in the sense 
that the normal flower is perfect, for at least that semblance of 
perfection which is found in the one is likewise found in the 
other. What those forces are which thus control the form of 
the flower, and how, or why, they thus combine to form symme- 
trical shapes we cannot tell; but the fact that they can do so, 
and that this is one of their attributes, may one day be found to 
be the clue that shall discover to us the nature of those forces. 
Certain reservations must be borne in mind. 
At the outset of the study of Variation, it is at once found 
that argument from analogy from one organ to another, or from 
the case of one organism to that of another, is as yet inadmis- 
sible; for the variation of special organs or of specific forms is 
frequently governed by principles which, so far as we can see, are 
likewise specific. We are therefore conscious that it is by no 
means legitimate to affirm principles like the foregoing as general 
principles of symmetrical variation ; but that these principles are 
obeyed in the special cases now under consideration is sufficiently 
clear. Still the fact that such principles are found operating in 
certain cases should at least suggest the possibility that the 
same principles may have been followed in other cases; and 
especially when a form (e. g. Veronica) is found whose symmetry 
is related to that of its presumptive allies in a way similar to 
that in which the varieties now described are related to their 
respective normals, it may not be unreasonable to suspect that 
variations of this discontinuous character have occurred. 
There is, however, another reservation which is of more im- 
portance. In all the cases now given of a new zygomorphie 
Symmetry arising as a sudden variation, the new form has 
resulted by variation from another zygomorphy already existing, 
