secondary characters and instincts by replacing one secretion with
another, he was faced with the following problem: How can a single
secretion be responsible for innumerable changes as to feather length,
form and colouring, as to spurs, comb and almost an endless array of
other details? To suppose that a secretion could be so complicated in
its action as to determine each one of a thousand different items of
structure, colour and behaviour would be preposterous. Besides, we know
that some of these internal secretions are _not_ excessively
complicated--for instance adrenalin (the suprarenal secretion) can be
compounded in the laboratory. We may say that it cannot possibly be that
the ovarian or testicular secretion is composed of enough different
chemical substances to produce each different effect.

There remains only the supposition that the female already possesses the
genetic basis for becoming a male, and _vice versa_. This is in accord
with the observed facts. In countless experiments it is shown that the
transformed female becomes like the male of her own strain and brood--to
state it simply, like the male she would have been if she had not been a
female. If we think of this basis as single, then it must _exhibit_
itself in one way in the presence of the male secretions, in another way
under the influence of the female secretions. In this way a very simple
chemical agent in the secretion might account for the whole
difference--merely causing a genetic basis already present to express
itself in the one or the other manner.

This may be illustrated by the familiar case of the crustacea _Artemia
salina_ and _Artemia Milhausenii_. These are so unlike that they were
long supposed to be different species; but it was later discovered that
the genetic basis is exactly the same. One lives in 4 to 8% salt water,
the other in 25% or over. If, however, the fresh-water variety is put in