18 Journal of the Asiatic Society of Bengal. [January, 1908. 
factor, which is the arithmetic mean of that of the two parents; and, 
while it would be legitimate to expect that the F, generation of 
such a cross would contain, among others, plants with typical 
“broad ’-, and typical ‘narrow ’-lobed leaves, it is unnecessary to 
depend on assumption ; since, where observations have extended 
to the F, generation, this expectation has been found to be ful- 
filled. 
The conclusion is, therefore, that the plant which gave rise to 
serial No. 19, is a natural cross between a ‘broad’-, and a 
glectum. 
It has already been noted that in the field in the United Pro- 
vinces there normally occur, in addition to the typical ‘broad ’- 
and ‘narrow ’-lobed forms, plants of which the leaf-factor is 
intermediate. It is difficult to escape from the conclusion that 
these plants with an intermediate leaf-factor are crosses, and that 
cross-fertilization must be of common, though not necessarily of 
general, occurrence. All observations hitherto made in the 
course of these experiments—including those on such definite 
characters a the leaf-glands and colour of the petals—indicate 
the rene fac 
bo 
Orae-fertlation of the cottons in nature. —This fact the 
curr 
denied (3 and again in 4) ; and it is desirable to acer it in greater 
detail. Indirect proof exists in abundance. It would be hardly 
conceivable that a plant, whose flowers are visited by ae as it 
has been shown that cottons are visited (5), is invariably self- 
naturally fertilized, that purity is readily obtained by the artificial 
self-fertilization of selected plants, and that the pollen of one 
species is frequently to be observed on the stigma of a second, are 
points that have come under personal observation. It is 
direct evidence exists. Such evidence is to be found in the occur- 
rence of plants—raised from the seed given by flowers naturally 
fertilized—which, when judged by their offspring, behave in all 
details as a cross. Several instances bea: is have come under 
observation. It will be sufficient, however, to record the salient 
-aitirgasy of that case of which the wins “Aetailed record has been 
t. 
From a packet of seed of @. arborewm 14 plants were raised, 
and of this number 12 proved true G. arborewm. The remaining 
two plants, rate similar to each other, differed markedly from 
e 
plants were self-fertilized, but owing to the Regia of he season 
at which this was carried out, only four plants yielded any seed. 
From these four plants 6 have been rais 
e leaf-factor of the two initial plants was not observed 
since this character had not at that time been identified. The 
records, however, are sufficiently detailed to show that the leaf- 
