GERM CELLS IN THE ARTHROPODA 161 



Polyembryony has been described in an earthworm. 

 Lumhricus trapezoides (Kleinenberg, 1879), in cer- 

 tain Bryozoa (Harmer, 1893; Robertson, 1903), 

 in the armadillo (Patterson, 1913), and in parasitic 

 Hymenoptera (Marchal, 1904 ; Silvestri, 1906, 

 1908) . In every case cleavage is of the indeterminate 

 type, and the cell lineage is unknown. Various 

 theories have been advanced to account for poly- 

 embryony, such as (1) blastotomy or the early 

 separation of blastomeres, each giving rise to a 

 single individual as has been brought about by 

 Driesch (1892) and others by separating the blas- 

 tomeres of the eggs of certain animals ; (2) polyovular 

 follicles may occur in mammals and by some (Rosner, 

 1901) are considered sufficient to account for poly- 

 embryony among the members of this class; and 

 (3) precocious budding has been suggested to account 

 for the production of many individuals from a 

 single egg, most recently by Patterson (1913), 

 who has shown that in the armadillo the blastoderm 

 produces two primary buds from each of which two 

 secondary buds arise, and hence four young develop 

 from each egg. According to the theory of germinal 

 continuity each of the buds must be supplied with 

 germ cells or with germ-plasm which has not yet 

 been segregated into germ cells. Silvestri' s in- 

 vestigations seem to indicate that the former is 

 true for parasitic Hymenoptera, but it is difficult 

 to see how a definite number of germ cells can be 

 supplied to each bud during a process of development 

 which is apparently so indeterminate. If, however, a 



