chapter II.; also Gluck, "Untersuchungen uber Wasser- und
Sumpfgewachse", Jena, Vols. I.-II. 1905-06.) afford the most striking
examples of modifications: according as they are grown in water, moist
or dry air, the form of the species characteristic of the particular
habitat is produced, since the stems are also modified. To the same
group of phenomena belongs the modification of the forms of leaves and
stems in plants on transplantation from the plains to the mountains
(Bonnier, "Recherches sur l'Anatomie experimentale des Vegetaux",
Corbeil, 1895.) or vice versa. Such variations are by no means isolated
examples. All plants exhibit a definite alteration in form as the result
of prolonged cultivation in moist or dry air, in strong or feeble
light, or in darkness, or in salt solutions of different composition and
strength.

Every individual which is exposed to definite combinations of external
factors exhibits eventually the same type of modification. This is the
type of variation which Darwin termed "definite." It is easy to realise
that indefinite or fluctuating variations belong essentially to the same
class of phenomena; both are reactions to changes in environment. In the
production of individual variations two different influences undoubtedly
cooperate. One set of variations is caused by different external
conditions, during the production, either of sexual cells or of
vegetative primordia; another set is the result of varying external
conditions during the development of the embryo into an adult plant. The
two sets of influences cannot as yet be sharply differentiated. If,
for purposes of vegetative reproduction, we select pieces of the
same parent-plant of a pure species, the second type of variation
predominates. Individual fluctuations depend essentially in such cases
on small variations in environment during development.

These relations must be borne in mind if we wish to understand the
results of statistical methods. Since the work of Quetelet, Galton, and
others the statistical examination of individual differences in animals
and plants has become a special science, which is primarily based on the
consideration that the application of the theory of probability renders
possible mathematical statement and control of the results. The facts
show that any character, size of leaf, length of stem, the number of
members in a flower, etc. do not vary haphazard but in a very regular
manner. In most cases it is foun