IN PLANTS HAVING IRREGULAR COROLLAS. 421 
vary also among themselves ; and it is possible enough that the 
comparative frequency of these variations obeys the Law of Error, 
taking the form No. 2 as the mean form. It seems, in fact, 
a cases where changes of symmetry are concerned, that the 
intermediate forms are, as it were, points of unstable equilibrium, 
and that the body therefore assumes these forms rarely, as in 
some instances, or never, as in others. A simple illustration, 
though somewhat loose, may make this more clear. 
The distribution of forms with regard to the normal symmetry, 
on the one hand, and the symmetry of the variety, on the other, 
may be perhaps compared with the spread of agricultural settlers 
into a country divided by a mountain-range. The mass of the 
people will settle in the nearest plain. Some will perhaps settle 
on the hillside, but these will be found fewer and fewer as the range 
is ascended ; but those who get over the top of the range will 
mostly go down into the plain beyond. Those in the hither 
plain are many and are the normal; those in the valley beyond 
are the variety, and the few on the hillsides are the intermediates. 
We wish, then, to insist upon the fact that there are at least 
two classes of variation ; and it is suggested that this is a fact of 
great importance. It may be remarked that if, as may be 
alleged, there is little evidence that species may arise by what 
may be called discontinuous variation—a variation in kind—there 
is still less evidence that new forms can arise by those variations 
in degree which at any given moment are capable of being 
arranged in a curve of Error; and no one as yet has ever indi- 
cated the way by which such variations could lead to the con- 
stitution of new forms, at all events under the sole guidance of 
Natural Selection. Whatever may be hereafter determined as 
to the scope of either of these classes of variations in the consti- 
tution of Species, it is of the first consequence to recognize that 
these two classes of variation exist; and the problem of the 
history of any given form or structure will never be solved until 
it shall have been first determined whether it is the result of the 
one class of Variation or of the other, and whether the changes 
which produced it were continuous or discontinuous. 
Norz.—The observations above recorded were made a year 
ago, and as since that time the case of Veronica has becn much 
more fully investigated, a brief abstract of these observations 
may here be added. 
