76 GERM-CELL CYCLE IN ANIMALS 
parthenogenesis. If, on the other hand, the re- 
productive cells of the gemmule are of multiple 
origin, they may either be looked upon as true germ 
cells which form a group physiologically equivalent 
to the morula stage in the development of an egg, 
or as a collection of regenerative cells capable of 
producing a new individual. 
In this connection should be mentioned the bud- 
ding of Tethya (Désé, 1879-1880) which develops 
from a group of amebocytes (Maas, 1910) and the 
gemmules of Tedania and Esperella (Wilson, 1902) 
and of hexactinellids (Ijima) which become ciliated 
larvee. Wilson has shown “that silicious sponges, 
when kept in confinement under proper conditions 
degenerate in such a manner that while the bulk 
of the sponge dies, the cells in certain regions become 
aggregated to form lumps of undifferentiated tissue. 
Such lumps or plasmodial masses, which may be 
exceedingly abundant, are often of a rounded shape 
resembling gemmules, more especially the simpler 
gemmules of marine sponges (Chalina, e.g.), and 
were shown to possess in at least one form (Stylo- 
tella) full regenerative power. When isolated they 
grow and differentiate, producing perfect sponges ” 
(1907, p. 295). These “lumps of undifferentiated 
tissue’” have also been noted by F. E. Schulze 
(1904) and recognized as probably reproductive; 
they have been named by this author, “ sorites,” and 
have been called by several authors “artificial 
gemmules.”” The process involved in their forma- 
tion is termed ‘regressive differentiation.”” The 
