23 
in the Saw fly, Nematus ribesii. 
In the oogonial divisions there are also 8, but in the nuclei of 
the ovarian wall in most if not all cases the number is greater ; it 
is generally 16 but perhaps sometimes more. Either therefore 
these cells must be regarded as abnormal, as has been suggested 
by Wilson for similar phenomena in Hemiptera, or the 8 chromo- 
somes of the germ-cells must be compound, as Petrunkewitsch 
has asserted in the case of the Bee. In the later pupal stages no 
divisions take place in the oogonia, but they undergo prolonged 
growth with deposition of yolk. 
When true reduction had been found in the spermatogenesis 
it became necessary to re-examine the polar mitoses to see 
whether it does not sometimes at least occur in the egg also. In 
some eggs 8 chromosomes were found in the second polar mitosis, 
confirming the opinion previously expressed that two equation 
divisions might occur. But in other eggs there are only 4 chromo- 
somes in the second maturation division, each of which is about 
twice as large as the single chromosomes of the eggs in which 
8 are present. The reducing type of division is found both in 
eggs of virgin and of impregnated females, and is therefore not 
conditioned by fertilization, and in one egg at least only 4 were 
visible in the blastoderm mitoses. 
It must therefore be concluded that in some eggs no reduction 
takes place, and that such eggs may develop parthenogenetically, 
yielding males ; in other eggs separation of complete chromosomes 
occurs, so that the chromosome number is halved, and such eggs 
are capable of fertilization and probably yield females. When not 
fertilized they may develop at least as far as the blastoderm stage, 
but possibly the great mortality among eggs and larvae from 
virgin females is due to the dying off of those containing the 
reduced number of chromosomes. 
Probably in species which yield females from virgin eggs and 
are normally not fertilized, no reduction takes place. 
