_' is pb YSIOLOQICAL GENETICS 



we saw how they demonstrated the existence of a determination 

 stream, comparable to the one discovered in the Lymantria wing. 

 All these facts now lead to the decisive question: How can the 

 action of the genes be linked with these different aspects of 

 pattern localization? In his detailed analysis of the problem, 

 Goldschmidl (1927c) stated thus the general facts and viewpoints 

 that have to be taken into account whenever this problem is 

 being attacked: 



1. The pattern of a lepidopteran wing is formed by the definite 

 arrangement of areas, which contain scales different in color and 

 partly in structure. These areas may be chemically different, 

 e.g., contain melanin, guanin, carotin, etc. They may be differ- 

 ent only in regard to quantity, e.g., contain different concentra- 

 tions or grades of oxidation (or reduction) of melanin. They 

 may be different physically in regard to the form of scales or their 

 surface structure (optical colors). 



2. Viewed as a whole, the patterns result from combination of 

 rather few elements: specifically colored or structured scales; 

 zigzag bands; broad or narrow bands which are transversal, 

 oblique, or parallel to the veins; points, specks, spots in definite 

 relations to veins or wing margin; systems of wavy lines, con- 

 centric lines (eyespots); areas without scales. 



3. Within single families, e.g., Sphingids, Papilionids, Arctiids, 

 the majority of species show patterns that consist of numerous 

 permutations of the same general plan of arrangement. (These 

 plans have been worked out for different groups by Sueffert, 

 Scfrwanwitsch, and Henke). 



4. Patterns of very similar type may be different in regard to 

 chemical or physical structure; the same element may be char- 

 acterized in one species by chemical, in another by optical colors, 

 etc. 



These rather general facts already offer some points of attack 

 for an analysis. As the pattern is generally composed of rela- 

 tively few elements, differing in different species mostly in a 

 quantitative way, it is to be supposed that the processes of 

 pattern formation are also of a relatively simple type, which, 

 however, permits of many permutations and quantitative shifts. 

 Furthermore, since all patterns show definite spatial relations 

 to the areas and general form of the organ (wing base, edge, etc.), 

 the formation of pattern must be connected in some way with the 



