318 PROCEEDINGS OP THE ACADEMY OF [April. 



that synapsis is brought about by the conjugation of maternal and 

 paternal chromosomes of the same size. This led to the view that 

 homologous chromosomes represent homologous characters and only 

 by the definite association of chromosomes of certain characters can 

 Sutton's theory of the purity of the germ cells be maintained. 



Individuality is further extended by establishing a definite number 

 of chromosomes for each species. The number of chromosomes 

 found in the equatorial plate- of the family Saturnids ranges from 

 thirteen in P. cynthia to thirty-one in A. i<>. and there is no evidence, 

 that P. cynthia is more highly developed than other members of the 

 family. This is in accord with the view of Montgomery (1906), who 

 after tabulating the number of chromosomes in several hundred species 

 of plants and animals, was forced to give up his theory of a correlation 

 between the number of "chromosomes and the evolutionary stage of 

 .the species. McClung (1005) claims that the family AcrididaB is 

 characterized by a fixed number of chromosomes which i- constant 

 for all the genera and species: the genera are characterized by a 

 definite arrangement and association, while the species show the same 

 grouping as the genus but are distinguished by the size difference of 

 the chromosomes and spindles. He consider- that a definite series 

 of chromosome- accompanies a group of somatic characters used by 

 systematic for classification. That the chromosomes may be of 

 classificatory significance has already been mentioned for cynthia 

 and promethea, where the external difference of the pupa was often 

 so slight that only by reference to the number of chromosomes could 

 the species be definitely determined; this was, however, by the marked 

 difference in number, not in arrangement. 



If this individuality is to be maintained, maturation mitoses must 

 show one transverse and one longitudinal division, and while the end 

 result is the same, whether the first division is transverse or longitu- 

 dinal, yet it seems probable, as Montgomery (1003) points out, that 

 the first division will be found in all cases of heterotypic division to 

 be reducing. By a comparison of the forms here described with 

 P. cynthia, where twenty-six pairs of granules can be counted in the 

 spermatogonia and thirteen in the first spermatocyte, I conclude that 

 pseudo-reduction takes place in the family Saturnidae either during 

 the telophase of the last spermatogone division or during synizesis 

 by an end-to-end union of homologous chromosomes, and that the 

 angle in the middle of the rods (fig. 22) and the constriction in the 

 rings (fig. 26) and dumbbells marks this point of union of univalent 

 chromosomes, as well as the point of separation of the first division. 



