a "dwarf" pollen grain. Hence x will give rise
to tall plants carrying the recessive dwarf character, while x will produce
plants from which the tall character has been eliminated, _i.e._ to pure
recessive dwarfs. Consequently from the 4x ovules of the self-fertilised
hybrid we ought to obtain 3x tall and x dwarf plants. And of the 3x talls x
should breed true to tallness, while the remaining 2x, having been formed
like the original hybrid by the union of a "tall" and a "dwarf" gamete,
ought to behave like it when bred from and give talls and dwarfs in the
ratio 3 : 1. Now this is precisely the result actually obtained by
experiment (cf. p. 17), and the close accord of the experimental results
with those deduced on the assumption of the purity of the gametes as
enunciated by Mendel affords the strongest of arguments for regarding the
nature of the gametes and their relation to the characters of the zygotes
in the way that he has done.

It is possible to put the theory to a further test. The explanation of the
3 : 1 ratio of dominants and recessives in the F_2 generation is regarded
as due to the F_1 individuals producing equal numbers of gametes bearing
the {24} dominant and recessive elements respectively. If now the F_1 plant
be crossed with the pure recessive, we are bringing together a series of
gametes consisting of equal numbers of dominants and recessives with a
series consisting solely of recessives. We ought from such a cross to
obtain equal numbers of dominant and recessive individuals, and further,
the dominants so produced ought all to give both dominants and recessives
in the ratio 3 : 1 when they themselves are bred from. Both of these
expectations were amply confirmed by experiment, and crossing with the
recessive is now a recognised way of testing whether a plant or animal
bearing a dominant character is a pure dominant, or an impure dominant
which is carrying the recessive character. In the former case the offspring
will be all of the dominant form, while in the latter they will consist on
the average of equal numbers of dominants and recessives.

So far we have been concerned with the results obtained when two
individuals differing in a single pair of characters are crossed together
and with the interpretation of those results. But Mendel also used plants
which differed in more than a single pair of differentiating characters. In
such cases he found that each pair of characters followed the same defini