The Cytoplasm as Specific Substrate 211 



the Michaelis school themselves. They found, when crossing a Jena 

 race of Epilobium hirsutum with 390 other races of widely different 

 origin, and introducing (by continued backcrossing ) the nucleus of 

 22 of these into Jena plasm, that all hybrids, if reciprocally different, 

 show deviations of the same kind. The activity of peroxidase and 

 respiratory enzvmes is increased, but to a different degree in different 

 combinations, with the result that positive or negative heterosis is 

 found in some, cripples and dwarfs in others, and lethality in still 

 others. "It is possible to transform the lethals into dwarfs and the 

 dwarfs into vigorous plants by inactivation of the oxidizing enzymes 

 at low temperatures or by augmenting the carbohydrates" (Mi- 

 chaelis). This should show conclusively that the interpretation pro- 

 posed above in terms of generalized substrates is correct, and perhaps 

 also in terms of mitochondria. 



It is hardly necessary to go into the details of other cases studied 

 in the plant kingdom, which have been ably reviewed by Caspari 

 (1948). At one extreme are found the facts relating to male sterility 

 in flax, studied by the Bateson school (see Gairdner, 1929), where a 

 Mendelian locus in homozygous condition produces pollen sterility 

 only if it is transferred to the cytoplasm of another species. In 

 other examples the results resemble those in Epilobium, so far as 

 modification of many morphological traits, in the direction of the 

 type present in the form which supplies the cytoplasm of the hybrid, re- 

 mains constant over generations; or remains so though being changed 

 somewhat in the course of generations. In a general way this is the 

 situation both in Wettstein's (1924, 1928a) moss crosses and in 

 crosses of Oenothera studied by Schwemmle et al. (1938). Apart 

 from specific features concerning the plastids (which will be treated 

 separately), the foregoing discussion covers the essentials. The main 

 point in each instance is that the action of known Mendelian loci 

 or of a considerable part of the genome, though not severally refer- 

 able to single MendeHan units, differs in different cytoplasm. In 

 some this means an impaired function leading to generahzed ab- 

 normalities, including certain kinds of sterility and extending to 

 complete lethality. In others, where growth and similar quantitatively 

 varying processes are involved, it means a shfft of gene-controlled 

 reactions and processes toward the type of form which supplied the 

 cytoplasm, that is, more or less of an influence upon developmental 

 kinetics. 



Renner and Kupper (1921) introduced, for this type of action, 

 the term "plasmon-sensitive genes," meaning that the action of some 



