THE GENETIC NATURE OF TAXONOMIC DIFFERENCES 273 



characters. One of the most important is the latent period before 

 hatching. The eggs are laid in the early summer and immediately 

 develop into young caterpillars, which, however, do not leave the egg 

 till the following spring. The physiological mechanism determining the 

 length of time spent in the egg is compUcated, but an essential part is 

 played by a reaction which depends on the summation of all the time 

 spent above a certain temperature. The difference between the local 

 races is a difference in the length of this time which is required before 

 emergence takes place, and the various races are exactly adjusted to the 

 meteorological conditions in which they live. Goldschmidt has shown 

 that this eminently adaptive character is at least partly inherited 

 through genes, though it seems that in each race there are several genes 

 which influence the character, so that the F2 segregation is not into 

 sharply defined classes but gives the wide variation characteristic of 

 inheritance by multiple factors. Some cytoplasmic differences may also 

 be involved since there is a slight tendency for offspring to resemble 

 their mothers rather than their fathers, and the results of a cross thus 

 differ according to the way in which it is made. 



A third character studied in detail by Goldschmidt is the pigmenta- 

 tion of the caterpillars. The types can be roughly classified as light, dark, 

 and those which are at first light and darken later. Closer examina- 

 tion shows that these differences depend on the rate of pigment 

 formation (p. 173). The genetic basis seems to be a set of allelomorphic 

 factors, but again there may be a cytoplasmic factor involved which 

 causes a tendency for matroclinous inheritance. 



Each race differs from the other races in all these characters and in 

 many more which have not been fully analysed. Thus the races have 

 characteristic complexes of genes, and the same conclusion must 

 certainly apply to different species. 



(b) Peromyscus. — Sumner^ has made a long series of crosses between 

 geographical races of several species of the American deermouse, 

 Peromyscus. The differences between the races are mostly quantita- 

 tive, affecting such characters as relative sizes of parts, depth of colour 

 and size of coloured area, etc. The crosses nearly always give inter- 

 mediate Fi's and do not show any clear-cut segregation in the F2's 

 and subsequent generations. There is, however, an increase in varia- 

 bility in F2 and later, and it is clear that the races differ in many poly- 

 meric factors each with only a small effect. The segregation found is 

 usually rather slight, presumably because of the large numbers of 

 ^ Sumner 1930, 1932. 



