1900] THe CELL PLATE IN HIGHER FLANIS 161 
such a relation to the central spindle as to continue the first 
process of growth. 
Another question of importance in this connection is, does 
the growth in area of the cell plate depend entirely upon the 
fibers already existing, or are new peripheral fibers formed f So 
far as my observations upon the larch go, I think it probable 
that there is no formation of new fibers, but that the whole growth 
takes place as a result of the changes in the existing fibers; yet 
the evidence on this point was not conclusive. If the hypothesis 
is true, whether the mother cell divides into the four cells which 
form the pollen grains by successive or by simultaneous division 
depends upon the number of spindle fibers existing in connec- 
tion with the first nuclear divisicn. If there are enough fibers 
" form a cell plate completely across the cell, successive divi- 
sion results, while if the cell plate does not reach across the 
= it is absorbed into the rest of the protoplasm, and the final 
division takes place simultaneously after the second nuclear 
division. In the onion the conditions are such as to indicate 
the necessity for the formation of new peripheral fibers. It will 
be remembered that at the time of the formation of the young 
“cil plate there were but few radiating fibers visible, and those 
that could be seen were relatively short, with the exception of 
sr Ay the periphery of the central spindle. In the later 
es Pe ter the fibers had entirely disappeared from the central 
tion ok .. relations of spindle fibers in the peripheral por- 
Viously ste could be distinguished that have bese pre- 
that there Sun ed in the larch (fig. 30). This relation indicates 
ing fibers sage a growth at least of the original radiat- 
There: ig s sa probable production of new ones. 
nother point of importance here. The fact that 
peripheral fibers may be traced to the nucleus 
uous production of new peripheral fibers from 
and the onion ; . The difference, then, between the larch 
'0n is possibly as follows: in the former the extent 
ate depends upon the number of exist- 
latter, where there is necessity for the 
