THE NATURE OF THE GENE 291 



There is the work of Lilienfeld (1929) on the form of leaves in 

 Malva parviflora. A dominant mutation produces an inconstant 

 laciniate type. It is characterized by an inherited tendency to 

 show highly laciniated leaves in young plants which are gradually 

 replaced later by more and more normal and even hypernormal 

 leaves. This, as well as the behavior of heterozygotes, would 

 appear as a simple problem of morphogenesis which could be 

 explained easily by genie action upon the rate of production of 

 something needed for full development of leaves (c/. the vestigial 

 case). But the types between laciniate and hypernormal 

 reproduce to a certain extent their kind; the highest grade, 

 hypernormal, may be kept constant, and it mendelizes with Wild 

 type. Thus some condition of the gene must be involved. The 

 details are too different from those in variegation to permit such 

 a simple explanation as somatic mutation to Wild type. Lilien- 

 feld herself accepts Correns' (1919) explanation of the sick gene. 

 Apparently, however, the facts are not sufficiently clear yet, 

 though they suggest the possibility of furnishing important 

 information, if analyzed by new methods. This might also be 

 said of a nearly related case, the "rogues" in peas. 



One more case of this type may be mentioned. Anderson- 

 Kottoe (1931) analyzed a very complicated case of variegation in 

 ferns, which looks somewhat as if it involved unstable genes. 

 Visible changes from green to pale tissue occur at the reduction 

 division, in gametophytic and in sporophytic tissue. As the pro- 

 thallium is a simple and regularly formed structure, the moment 

 of these changes may be exactly determined, and the genetic 

 constitution of a cell may be tested by regenerates from it. The 

 results do not agree with the simple assumption of somatic 

 mutations. It is supposed, moreover, that periodic changes 

 of the gene in question occur, which cannot be of the nature of 

 losses or gains of parts. Anderson-Kottoe formulates a very 

 complicated hypothesis. She assumes that a factor necessary 

 for chlorophyll-formation exists in seven different conditions and 

 that each condition may be stable or unstable. The latter means 

 mutation from one state to another at any point in the life cycle 

 of the plant. Each combination of these states is responsible 

 for a certain phenotype, stable or unstable and different for the 

 phases of the life cycle. From the breeding results, a definite 

 combination for each phenotype is derived, and its mutability as 



