Studies on Chromosomes 185 



to the spermatid nuclei. All the spermatid nuclei receive the 

 haploid type-group of II chromosomes, half including the small 

 idiochromosomes and half the large; but in addition each may 

 receive one or more supernumeraries. The total number of 

 chromosomes in the sperm nuclei is therefore variable in the same 

 individual. 



7 Both the number of the supernumeraries and their size, indi- 

 vidually considered, are constant in the individual. 



The first question that the foregoing report of results will raise 

 is whether the number and size relations of the chromosomes in 

 each individual are really as constant as I have described them. 

 I have for the most part selected for illustration and description 

 the more typical conditions ; but, granting the accuracy of the figures, 

 does such a selection really give a fair presentation of the actual 

 conditions ? It is almost needless to say that very many cases 

 might have been shown that would seem to give conflicting results. 

 By far the greater number of these discrepancies are, I believe, 

 only apparent. Numerical discrepancies of this kind are very 

 often evidently due to mere accidents of sectioning or to the super- 

 position or close contact of two or more chromosomes. Again, 

 apparent discrepancies in the size relations of the chromosomes, as 

 seen in polar views, very often arise through different degrees of 

 elongation (particularly in the^ maturation divisions). But apart 

 from such apparent variations, real deviations undoubtedly occur 

 in almost all of the relations described. Now and then, for exam- 

 ple, a spermatogonial or ovarian group is found that clearly shows 

 one chromosome too many (as in Fig. 9, m), 9 and the same is true 

 of the first spermatocyte-division, but such cases are very rare. 

 The former case is probably a result of an abnormality in the forma- 

 tion of the chromosomes from the resting nucleus, the latter not 

 improbably to a failure of synapsis. Again, both spermatogonial 

 and spermatocyte-cysts are occasionally found in which the num- 

 ber of chromosomes is doubled or quite irregular. These are 



9 A perfectly clear case of this has been found in the pyrrochorid species Largus cinctus (a particu- 

 larly fine form for study). In this form the normal male number is n, the female 12; but in one 

 female three cells were found each of which shows with all possible clearness 13 chromosomes, very 

 many other cells showing the normal number. 



