version, as influenced by the position of
the seed in the capsule, evidently acts. The Blue Imperial pea is the
offspring of the Blue Prussian, and has larger seed and broader pods
than its parent. Now Mr. Masters, of Canterbury, a careful observer and
a raiser of new varieties of the pea, states[868] that the Blue
Imperial always has a strong tendency to revert to its parent-stock,
and the reversion "occurs in this manner: the last (or uppermost) pea
in the pod is frequently much smaller than the rest; and if these small
peas are carefully collected and sown separately, very many more, in
proportion, will revert to their origin, than those taken from the
other parts of the pod." Again M. Chate[869] says that in raising
seedling stocks he succeeds in getting eighty per cent. to bear double
flowers, by leaving only a few of the secondary branches to seed; but
in addition to this, "at the time of extracting the seeds, the upper
portion of the pod is separated and {348} placed aside, because it has
been ascertained that the plants coming from the seeds situated in this
portion of the pod, give eighty per cent. of single flowers." Now the
production of single-flowering plants from the seed of double-flowering
plants is clearly a case of reversion. These latter facts, as well as
the connection between a central position and pelorism and
prolification, show in an interesting manner how small a
difference--namely a little greater freedom in the flow of sap towards
one part of the same plant--determines important changes of structure.

* * * * *

_Analogous or Parallel Variation._--By this term I wish to express that
similar characters occasionally make their appearance in the several
varieties or races descended from the same species, and more rarely in the
offspring of widely distinct species. We are here concerned, not as
hitherto with the causes of variation, but with the results; but this
discussion could not have been more conveniently introduced elsewhere. The
cases of analogous variation, as far as their origin is concerned, may be
grouped, disregarding minor subdivisions, under two main heads; firstly,
those due to unknown causes having acted on organic beings with nearly the
same constitution, and which consequently vary in an analogous manner; and
secondly, those due to the reappearance of characters