THE DISTRIBUTION OF VARIANTS IN NATURE 79 



it is obviously thrown back on a relatively small number 

 of experimentally tested cases and the great bulk of the data 

 on local divergence (often associated with valuable ecological 

 and bionomic data) is worthless ! 



We have given in Chapter II a general survey of the facts 

 concerning fluctuations ; but it is desirable here to define 

 how far the deficiency in experimental evidence may be 

 remedied by other means. The following means of inferring 

 whether we are dealing with fluctuations seem to be available. 



A. Certain characters such as size and colour are some- 

 times determined by the amount and type of food available 

 and, though the non-heritability of such variation is only 

 very rarely demonstrated, it is a fair inference that they are 

 not inherited. 



(a) Size. — The adult size of insects obviously depends on 

 the food available for the larvae. In forms with a fluctuating 

 food-supply, such as carrion-feeding flies, adaptability in this 

 respect is very marked (cf. Salt, 1932). Mickel (1924, pp. 15-16) 

 has given a summary of a number of cases, in addition 

 to his own definite evidence that in the wasp Dasymutilla 

 bioculata adult size is dependent on the quantity of food available 

 for the larva. Especially significant is the experiment of 

 Wodsedalek (191 7), who was able to vary the size of the larvae 

 of a Dermestid (Trogodenna tarsale) from large to small by 

 starving them and from small to large by feeding them again. 

 Amongst molluscs, Hecht (1896) records that Elysia viridis 

 grows to a much larger size when its diet is changed from 

 Codium to Cladophora. 



(b) Colour.- — Pelseneer (1920, p. 485) gives a long list of 

 colour-changes in molluscs wrought by differences in diet. 

 In insects which feed on different plants the colour likewise 

 varies with the food. Thus Waters (1928) notes that the 

 moorland form of the moth Coleophora caespititiella, which 

 feeds on Juncus squarrosus, can be distinguished fairly easily 

 by its darker colour from the specimens bred from J. communis. 

 Eisentraut (1929a) attributes the darker colour of certain 

 littoral forms of the gecko Hemidactylus to their feeding on 

 Halophyta. In general it may be noted that there is a tra- 

 ditional suspicion among taxonomists that colour is an unsafe 

 systematic index. This is partly because it is extremely plastic. 

 In some instances, however, experiment is against this view. 



