the chromosomes.]

Janssens has found at the time of separation evidence in favor of the view
that some such interchange probably takes place.

We find this same process of interchange of characters taking place in each
of the other three groups of Drosophila. An example will show this for the
Group II.

[Illustration: FIG. 65. Scheme to illustrate double crossing over.]

If a black vestigial male is crossed to a gray long-winged female (fig. 66)
the offspring are gray long. If an F_1 female is back-crossed to a black
vestigial male the following kinds of flies are produced:

Black Gray Black Gray
vestigial long long vestigial
----------------- -----------------
| |
83% 17%

The combinations that entered are more common in the F_2 generations than
the cross-over classes, showing that there is linkage of the factors that
entered together.

Another curious fact is brought out if instead of back-crossing the F_1
female we back-cross the F_1 male to a black vestigial female. Their
offspring are now of only two kinds, black vestigial and gray long. This
means that in the male there is no crossing-over or interchange of pieces.
This relation holds not only for the Group II but for all the other groups
as well.

Why interchange takes place in the female of Drosophila and not in the male
we do not know at present. We might surmise that when in the male the
members of a pair come together they do not twist around each other, hence
no crossing-over results.

[Illustration: FIG. 66. Cross between black vestigial and gray long flies.
Two pairs of factors involved in the second group. The F_1 female is back
crossed (to right) to black vestigial male; and the F_1 male is back
crossed to black vestigial female (to left). Crossing over takes place in
the F_1 female but not in the F_1 male.]

Crossing-over took place between white and yellow only once in a hundred
times. Other characters show different values, but the same value under the
same conditions is obtained from the same pair of characters.

[Illustration: FIG. 67. Map of four chromosomes of D. ampelophila locating
those factors in each group that have been most fully studied.]

If we assume that the nearer together the factors lie in the chromosome the
less likely is a twist to occur between them, and conversely the farther
apart they lie the more likely is a twis