344 LOCK : STUDIES IN PLANT BREEDING 
This occurred in the cross already described between the 
native yellow pea and French sugar pea, but was not noticed 
until F s . In this generation a considerable number of 
plants was grown. 
At the end of May, 1904, I observed the time of flowering 
of the offspring (F s ) of 26 plants (F 2 ) which were shown to 
be heterozygotes in respect of floral colour. 
On 25 th May flowers had appeared upon 34 purple-flowered 
and 29 white-flowered plants ; and on 30th May 107 purple- 
and 79 white-flowered plants were in blossom. Thus, on the 
former date, the proportion purple to white was 1*17:1, and on 
the latter 1*35 :1, instead of the 3 :1 of Mendelian expectation. 
Subsequently the proportion of purple-flowered plants 
increased until the final total stood at : 383 purple and 118 
white, or a proportion of 3°25 :1. 
It is clear, therefore, that inthis case earliness was correlated 
with the presence of white flowers, and lateness with the 
presence of purple flowers ; and the original white-flowered 
parent had been observed to come into blossom several days 
before the French sugar pea. 
The case is of interest as showing clear correlation between 
a distinctly variable and indefinite character—time of flower¬ 
ing—and a perfectly definite and constant character—colour 
of the flowers. For I have never observed any flower which 
could be called intermediate in colour between these two 
classes. The purple flowers may show slight differences in 
tint, but these are hardly noticeable. The white flowers 
never, apparently, show any tinge of colour. 
In certain crosses between races of Matthiola, Correns 
(13) found correlation carried to such an extent that the 
hybrids appeared to form only two kinds of gametes, 
containing all the characters of one and of the other parent 
respectively. For this kind of segregation Correns proposes 
the term “ zygolytic ” as opposed to the “seirolytic ” method 
in which the germ plasm is segregated into its individual 
“ Anlagen.” 
