Sex Determination in Phylloxerans and Aphids 301 
bulges sometimes to one side. That artificial conditions, such as 
handling or osmosis, might break such a delicate connection at 
this time is not at all improbable, and such an artificial result might 
give the impression that the accessory is actually divided. More- 
over if the bridge arches toward or away from the observer, the 
effect may be produced at certain focal levels of discontinuity 
between the ends of the lagging chromosome, when none such 
exists. The arch containing the connecting thread of the acces- 
sory might also be cut off by the knife, and lie in another section, 
where its extreme tenuity might defy discovery. Such considera- 
tions make me sceptical as to whether the lagging chromosome in 
this species of aphid ever normally divides into two. Figures like 
that of Plate I, Fig. 3, in Stevens’ paper in which two nearly equal 
cells contain each one end of the accessory do not seem to me 
to represent accurately the conditions that I have seen, if on no 
other ground than in the absence of a bridge of cytoplasm around 
the connecting thread of the accessory. 
In the willow aphid the rudimentary cell is extremely small and 
contains very little cytoplasm. It does not subsequently divide 
and slowly degenerates. 
The equatorial plate of the second spermatocyte division of the 
large cells shows three chromosomes (Fig. XX, Z). These divide 
equally to produce each two spermatids (Fig. XX, 4’). In all 
essential respects the results agree with those that I have found in 
the phylloxerans. 
Von Baehr’s facts on a willow aphid A. saliceti with the same 
number of chromosomes agree with those given here. He also 
determined that while the female parthenogenetic egg contains six 
chromosomes, the male egg, after extrusion of its polar bodies, 
contains only five. He leaves open, of course, as I had done also, 
the question as to how the number becomes reduced to five. The 
most plausible inference was obviously that one was lost in the 
polar body. Stevens’ earlier statement (1905) that “there is no 
evidence in my material of any difference between the maturation 
of the female parthenogenetic egg and that of the male”’ is probably 
incorrect, and her count of the same number of chromosomes in the 
somatic cells of the male and female embryo must have been an error. 
