A6b LOCK : STUDIES IN PLANT BREEDING 
of fertility, is shown to be a foreshadowing of the distinction 
between Mendelian and unisexual crossing; and forms 
which differ only by such characters as obey Mendel’s Law 
when crossed are fertile together. Forms, on the other 
hand, which differ by true specific distinctions exhibit 
infertility which increases with the number of such differ¬ 
ences up to complete sterility. 
Correns (22) has criticized these views of de Vries. He 
finds that de Vries’ supposition—Mendel’s Law holds good 
for varietal characters, whereas species-characters on crossing 
yield constant hybrid-characters—does not hold generally 
good. Correns has described cases in Zea Mays in which 
hybrids in respect of varietal characters give constant pro¬ 
geny. On the other hand, certain “species-” characters, 
for example, some of those which Bateson has examined in 
the case of poultry, have been shown to follow Mendel’s 
In opposition to de Vries’ view of the existence of an 
unpan-ed character in the case of unisexual- (species-) 
hybrids, Correns believes that the rudiment of a correspond¬ 
ing character exists in the other parent form ; since the 
“progressive ” character does not really pass over unmodified 
m the unisexual cross, but is reduced to half its former . 
value. Correns believes that a progressive mutation takes 
p ace y the modification of a pangen already existing, not 
by the formation of a completely new one. He would, 
therefore, speak of a pair of « Anlagen ” in this case also. 
De Vries regards a “retrogressive” mutation as taking 
Place simply by a process of becoming latent (*a, inactive), 
o a pangen. Bat Mendelian hybrids not infrequently show 
Character intermediate between those of their parents, as 
barrens has clearly shown; hence in this case the “re- 
.TT * ^lelomorph has a distinct effect, and is not entirely 
, ere agam Correns would speak of a pair of “ An- 
XoXr Bate9 ° n ’ Wh ° — * h ° " f 
