IN THE TROPICS. 
309 
Sixteen plants from yellow seeds gave rise to 542 yellow 
seeds only. 
Sixty-two plants from yellow seeds produced 1,809 seeds, 
of which 1,331 were yellow and 478 green, or 2*78 : 1. 
Each of these plants produced both green and yellow seeds. 
A considerable number of seeds in this generation were 
damaged by insects or fungus. Every perfectly developed 
seed of the series was clearly recognizable as yellow or 
green without removal of the testa. Such seeds only were 
counted. 
The proportion of heterozygotes to extracted dominants 
in F s — 62 : 16, or 3*9 : 1—deviates widely from Mendelian 
expectation (2 : 1). Such a difference illustrates well the 
necessity of growing a considerable number of plants for 
determinations of this kind. 
The following table* gives a diagrammatic representation 
of the course of this experiment : — 
F l 50 y. 
41 
F 2 488 y. 158 g. 
#. (24-6°/ 0 ) 
f --- A -N 
16 (36°/ 0 ) 28 11 
F 3 1013 y. 877 y. 275 g. (4y) 291 g. 
| 78 (23*9°/ 0 ) 
9 16(20-5°/ o ) 62 47 
F 4 224y.(7f) 542 y. 1331 y. 478 g. 1325 g. 
(26-4 °/ 0 ) 
The above case agrees in the closest way possible with one 
of Mendel’s, who himself used a variety in which in the 
seeds of many plants “ the green colour of the albumen 
* This arrangement is used by Correns (11). 
t Showed a slight tinge of green almost certainly due to imperfect 
ripening, the seeds appeared to be fully developed, but were not properly 
hard. 
8(10)04 (23) 
