are sex-linked. In the little banana fly, over one hundred of these 
differential heredity units have been studied and instead of find- 
ing that each is inherited independently of all the others, the dis- 
covery has been made that they are inherited as though more or 
less loosely linked together in four groups. In plants similar 
discoveries have been made though not on as large a scale. In 
peas themselves, colored flowers are found to be partially linked 
or coupled with the materials responsible for late flowering, while 
the white flowers are partially associated with the materials respon- 
sible for early flowering time. One pea variety with no tendrils 
and having wrinkled seed, crossed with a smooth-seeded, tendriiled 
variety, gives smooth-seeded, tendriiled offspring in the first 
generation, while in the second, where four types of plants are 
expected in the proportion of 9st: 3wt: 3 s non-T: 1 w non-T, 
approximately only two appear, the two central groups being 
represented by very few individuals. The two outside groups of 
course are the same combinations as found in the original 
parents of the cross, and though the two characters naturally 
would not be expected to be associated in any way, experimental 
work has shown us that they are almost completely coupled. 
There are numerous other modifications of Mendel’s ideas, but 
all of these, though important, are to be regarded as super- 
structures built by the methods and on the foundation stones 
laid down by Mendel. 
Through the light thrown upon the nature and inheritance 
of characters by Mendelism, experimenters soon came to a radi- 
cally different conception regarding the role selection played in 
the creation of new varieties. Wheat, beans, and other plants, 
once freed from impurities by constant inbreeding, practically 
bred true, provided the)’ were always subjected to the same en- 
vironmental conditions. Certain varieties of wheat thus treated 
by the Vilmorin family of France have remained stable for forty 
years. Some varieties of peas are said to have bred true for a 
still longer period. Sugar percent in sugar beets has not been 
appreciably increased over that obtained during the first years 
when the selection for a high percentage of sugar began, although 
more stable varieties have been obtained. On the basis of this 
kind of evidence, together with that obtained from more recent 
experimental studies, selection is found not to to be the creative 
agent Darwin and most of the nineteenth century breeders and 
scientists supposed. Selection, so many experimenters believe, 
cannot alter the nature of the hereditary units, nor can it cause 
them to come into existence. Chance, or fortuitous variation, 
brought about no one knows how, seems to be the only method by 
which new units arise. To obtain ablue rose from a red-flowered 
variety, one must wait, so most scientific breeders believe, until 
a blue-flowered variation occurs — for selection is of no avail ex- 
cept as a method of isolating desired variations that are already 
