188 ROBERT H. BOWEN. 



extraordinary size. The nuclei themselves become more or less 

 lobulated during the growth period and so are difficult to separate 

 and count, but the grouping of the late prophase chromosomes 

 shows that each nucleus retains its identity and that the fusion 

 process involves only the cytoplasms of the uniting cells. The 

 remarkable fact about these giant cells is that they seem to possess 

 only the single set of centrioles characteristic of a normal cell. 

 Often these are not diametrically opposite each other (as is cus- 

 tomary in Hemiptera), and as the spindle is formed for the first 

 maturation division it tends to develop toward one side of the 

 cell, becoming later on symmetrically located. How the centrioles 

 of each of the cells represented in one of these giant spermatocytes 

 are reduced to a normal number is not known ; possibly the loss 

 occurs before the fusion, which might itself be dependent on such 

 an abnormal condition. Further, the centrioles which do take 

 part in the formation of the spindle, are not unusually large, as 

 one might anticipate, but are in fact no larger than the normal 

 ones, at least not conspicuously so. Spindles are normally de- 

 veloped, and the metaphase chromosome plates are formed in a 

 characteristic manner ( though slightly irregular in cells with 

 large numbers of chromosomes). I have been able to count cases 

 with 16 and about 32 chromosomes, the higher numbers being 

 less satisfactory for counting. I have, however, counted con- 

 siderably more than 60 chromosomes in a single plate. Second 

 maturation spindles seem to be normally formed (except, of 

 course, with respect to chromosome number) and spermaticls of 

 a variety of sizes result, each spermatid being formed on a large 

 scale but in an otherwise normal fashion. Whether the giant 

 sperms resulting from the multinucleate spermatocytes reach 

 maturity was not definitely ascertained, but at least some of them 

 seemed to be undergoing the earlier stages of normal differentia- 

 tion. 



The origin of this abnormality is obscure. However, the ap- 

 parent fusion (?) of cells to form multinucleate masses is not 

 altogether unknown, as such aggregates have been noticed by 

 Wodsedalek 1 in a number of mammals, particularly the ground 



i Only a very brief abstract of this author's results has been available; 

 see Abstracts of Papers, American Society of Zoologists, Nineteenth Annual 

 Meeting, December, 1921. 



