CYTOPLASMIC HEREDITY 417 



divides in somatic mitosis. When the genomeres of a gene are unHke 

 with respect to their effect upon chlorophyll development, their segrega- 

 tion results in variegation. 



The cytological evidence bearing on the foregoing theories is as yet 

 rather meager. Green, yellowish, and colorless plastids have frequently 

 been observed in variegated plants, but too little is known about their 

 origin and behavior to warrant generalizations. In normal plants and in 

 several Mendelian and non-Mendelian abnormal chlorophyll types of 

 Zea Mays, Randolph (1922) found the same kind of visible structure in 

 the meristematic cells: all contain very minute proplastids (plastid 

 primordia) which, so far as the microscope shows, are of one kind. In 

 normal plants the proplastids develop into large green chloroplasts, 

 whereas in pale-green, yellowish, or white plants, or in the various regions 

 of variegated plants, they fail to attain the normal color, the normal size, 

 or both. No evidence was found that green and colorless plastids 

 represent fundamentally distinct types; they appear rather to be end 

 results of different modes of development of one type of initial body. 

 Moreover, they are connected by a series of all conceivable intermediate 

 conditions, particularly in cells on the boundary between green and white 

 areas in variegated leaves. Hence no support is found for the hypothesis 

 that plastids of two initially distinct types undergo a sorting out during 

 the development of the tissues. Why it is that the primordia develop so 

 differently in different cells can only be conjectured, but it appears 

 probable that their peculiar behavior is an indication of some invisible 

 differentiation being carried out in the cytoplasm. In Abutilon, Tsinen 

 (1923, 1924) attributes variegations to alterations occurring in the 

 plastids before, during, or after their development from primordia. 

 Although the green, yellow, and colorless plastids in variegated individ- 

 uals of Hosta japonica are indistinguishable in the primordial stage, 

 Yasui (1929) is inclined to regard them as initially unlike. 



The full evaluation of the various theories of non-Mendelian chloro- 

 phyll inheritance must await a fuller knowledge of the cytoplasm and its 

 many differentiations. When certain variegation patterns are compared 

 with the patterns assumed by the leaf-cells as a result of their lineage 

 (Noack, 1922), it is difficult to regard a simple sorting out of plastids or of 

 genomeres in successive cell-divisions as an adequate explanation of the 

 variegation. It seems more likely, in some cases at least, that the peculiar 

 behavior of the plastids is associated causally with less evident conditions 

 in the cytoplasm. Such conditions arising during the course of tissue 

 differentiation might well affect the growth and greening of plastids, as 

 they do other developmental processes, and so result in the appearance of 

 ''chlorophyll characters." The development of all such characters 

 doubtless involves the interaction of both nuclear and cytoplasmic ele- 

 ments, but in the non-Mendelian category the latter elements exert a more 



