3 zo 



THE QUARTERLY REVIEW OF BIOLOGY 



gations are produced on the basis of a single 

 Mendeli an uni t . In appearan ce they show 

 "a continuous series of quantitative varia- 

 tions ranging from apparently deep self 

 colors with only occasional color changes 

 that are apparent through a series of dilute 

 self-colors with all gradations of color 

 from deep red to whitish and with increas- 

 ing numbers of dots, splashes, lines, bands 

 and larger segments of darker and lighter 

 colors; and through a series of variega- 

 tions varying in pattern from very heavy 

 to extremely light. ' ' This description by 

 Eyster (z8) of the pericarp case in maize 

 covers also the more or less similar types 

 in other cases, viz. Antirrhinum, Mirabilis, 

 Capsella, etc. These unstable types are 

 not confined to different individuals but 

 occur within the same individual. It 

 has therefore to be explained why on the 

 basis of an apparently single Mendelian 

 gene a complex mosaic is produced with a 

 number of typical features. The first 

 idea is that there is something unusual 

 the matter with the gene in question, 

 Correns ('19) directly speaking of a sick 

 gene. All explanations which have been 

 tried, though looking at first sight very 

 different, are of much the same order if 

 viewed from some distance. All interpre- 

 tations agree that something happens to 

 the gene in question so that different cells 

 are produced and the stage of development 

 in which this something occurs determines 

 the resulting pattern. The event in 

 question is regarded by Emerson ('13, '17, 

 'zz) as a temporary inactivation of the 

 gene. Baur (14) also does not wish to 

 touch the old conception of the gene as 

 the genetic unit. But realizing that some 

 diversity of the genie basis is needed to 

 explain the facts he assumes that a number 

 of unilocal genes are present which recom- 

 bine in different ways. Correns ('19) 

 was the first to make the revolutionary 

 step of explaining the facts by assump- 



tions regarding the finer structure of the 

 gene. He writes: 



"In order to have at least a model one might 

 assume that the gene consists of a large molecule to 

 which the same sidechain of atoms is attached, say, 

 ten times. This number might be mutable, might 

 undergo changes in the plus or minus direction under 

 unknown conditions 'external' for the gene. To 

 each number of these sidechains would correspond a 

 definite ratio of white and green in the mosaic plant 



The difference in this interpretation 



would be that the state of the gene, the 



number of sidechains, attached to the gene molecule, 

 is not constant, but that new chains could be added or 

 old ones detached, and this during ontogenesis of the 

 individuals." 



Here then we have already a somewhat 

 quantitative conception based on the 

 assumption of a complex structure of a 

 gene which may change quantitatively in 

 regard to definite parts of the whole. A 

 step further leads to the theory of Ander- 

 son and Eyster (^4, 'z 5 , 'z8). The last 

 named author assumes that the facts can 

 only be explained if a quantitative segre- 

 gation of parts of the gene takes place in 

 development. The gene, then, is regarded 

 as composed of a constant number of 

 genomeres or gene elements, which may 

 or may not be chemically identical. The 

 genetic difference between pigmented and 

 pigmentless forms is given in the geno- 

 meres, the genomere C being mutated into 

 c; both genes contain the same number of 

 genomeres, either C or c. If, however, 

 only a certain number of genomeres 

 mutate from C to c an unstable gene with 

 two types of genomeres is produced, the 

 gene of the variegated forms. This gene 

 may be divided into its genomeres during 

 somatic mitosis, giving different combina- 

 tions of genomeres up to a complete 

 separation of the two types C and c. It is 

 clear that all the facts can be represented 

 by this conception, which is very closely 

 akin to that of Correns. 

 If we compare these different concep- 



