352 
LOCK : STUDIES IN PLANT BREEDING 
offspring of the cross. In other cases it seems necessary to 
suppose that some of the parental characters become latent in 
certain of the cross-bred offspring,’ owing it may be, to the 
inability of certain characters to appear, except in combina¬ 
tion with certain others. On further crossing such characters, 
may once more reappear. In such cases segregation may, and 
probably in many cases does, take place according to Mendel’s 
Law. 
In certain instances the phenomenon of correlation appears. 
In such cases certain groups of allelomorphs show a tendency 
to remain together in the formation of the gametes : such 
correlation may be complete or incomplete. 
Examples occur in which each of a series of characters 
appears to dominate over the next with a definite system of 
valency (of Bateson ;=Werthigkeit of Tschermak). 
In the cases of unisexual crossing described by de Vries, 
the hybrid shows in a reduced state a character .present in 
one of its parents but absent in the other, and in this case 
segregation is said not to take place. In de Vries’ Mutation 
crosses, finally, the first hybrid generation is polymorphic, 
and the different forms subsequently breed true, except for 
fresh mutations. 
Other cases of greater complexity also without doubt 
occur, but the above include the majority of forms of cross¬ 
ing which have so far been definitely worked out. 
Illustrations from Actual Experiments . 
(1) Mendel’s observations were confirmed for the cate¬ 
gories smooth : wrinkled and green : yellow using Telephone 
and Telegraph crossed with a smooth yellow native pea. 
A doubtful exception appeared in the case of one cross. 
{Examples III.) 
(2) Also for the allelomorphs coloured and white flowers, 
on a large scale in two cases ; and for presence and absence 
of parchment layer in one case. {Examples VI.. VII») 
