1922] RANDOLPH—MAIZE 355 
workers for the study of cytoplasmic structures have given no 
evidence for the assumption that the group of cell elements under 
investigation consists of bodies belonging to more than one class. 
The observed differences between the various types of plants lie 
in the subsequent behavior of the proplastids with reference to the 
development of chloroplasts. From a structural standpoint the 
cells of all the plant types appear to be initially alike. The differ- 
ences which later appear seem rather to be dependent upon the 
relative amount of chlorophyll developed in organs, the primordia 
of which are present in all cells.” It is probable, therefore, that 
the ultimate explanation of unusual types of behavior may involve 
functional rather than structural differences. 
Discussion 
BEARING OF RESULTS ON PLASTID INHERITANCE.—The hypoth- 
eses which have been advanced to explain plastid inheritance 
have been based chiefly on the breeding behavior of the plants. 
The question remains as to how far these hypotheses will be 
substantiated by cytological observations. For the most part 
such hypotheses involve the assumption that plastids are permanent 
cell organs, arising only by a division of others of their kind, either 
while they are in the form of minute primordia or after they have 
reached the mature stage. The fact that proplastid-like bodies 
are present in the early embryonic stages, and the well substantiated 
evidence that mature plastids arise from bodies which cannot be 
distinguished from these elements, however, do not seem to consti- 
tute sufficient proof of the theory that plastids are cell organs 
having an unbroken continuity through all stages of the life cycle. 
Observations suggest that plastids may also differentiate anew at 
certain times in the life of the cell, which is obviousl¥ of the highest 
importance in connection with the problem in hand, and will be 
fully discussed later. 
The frequent occurrence of chlorophyll abnormalities has 
furnished abundant material for a study of plastid inheritance. 
Many cases have been reported, and the breeding behavior carefully 
studied. The inheritance of some of the known cases is clearly 
Mendelian, but other types are clearly non-Mendelian. Thus the 
