222 T. H. Morgan 
metabolism and the temperature determining the result. In 
support of this conclusion he cites some experiments in which the 
animals were starved when the percentage of males was increased. 
This conclusion is open to one serious criticism. If the condition 
of nourishment determines the number of large eggs that can 
develop, the proportion of the two kinds of eggs that mature— 
if two kinds really exist at an earlier stage—will be altered. This 
would mean not that sex is determined by the nutrition of the parent 
but only that a well-nourished parent can produce more female 
eggs than one that ig starved. ‘This is precisely the state of affairs 
shown by the phylloxerans. Poorly nourished females produce 
only one egg, but this is of normal size, while well-nourished 
females may produce eight to ten or more. Such differences 
would seriously affect the sex ratio if the males are somewhat less 
affected, owing to the small size of the egg. We have no means of 
knowing at present whether the large egg of Dinophilus apatris 
formed by the fusion of several ovarial cells is large on account of 
the number of cells that unite to produce it, or whether, being 
already a female egg, it absorbs a larger number of cells than does 
the male. One fact speaks strongly in favor of the view that the 
result is not due simply to the number of cells absorbed but to the 
determination of the egg before or at the time of absorption, 
namely, the great difference in size of male and female eggs. If 
the result were due to the factor of absorption alone, we should 
expect to find a graded series, not a dimorphic condition. ‘The 
evidence, while as yet not decisive on this point, indicates, never- 
theless, that there must be already in the ovary two kinds of eggs 
prior to the absorption period, and that the nature of the egg deter- 
mines how much of the surrounding material it will absorb. If 
this view prove correct, sex in Dinophilus apatris must be deter- 
mined in the egg long before the maturation stage is reached. 
The conclusion is in harmony with the facts here described for 
Phylloxerans. 
THE PROBLEM OF THE THREE CHROMOSOMES 
In those insects where the spermatogonium of the male has 
one accessory, the odgonium of the female has two chromosomes, 
