Oil of Citronella on tvjo species of Dacns. 417 



Another suggestion is that the olfactory sense of flies may 

 be highly developed in certain directions and within 

 certain narrow limits, while outside these limits it is com- 

 paratively inoperative. We should on this hypothesis 

 expect to find instances where the males were very 

 sensitive to the smell of the females or vice versa, the 

 sensitiveness being, however, probably confined to one 

 sex; the smell of the food of the adult fly would attract 

 both sexes if they fed on the same substances, while the 

 food of the larva would, by its smell, direct the female in 

 oviposition. Other smells, unless very strong, would have 

 little effect. 



Regarding the matter as thus crudely put, we might 

 look on each species as tuned to respond to three or four 

 notes on the scale of smell, and we should expect to find 

 the most delicate adjustment and most accurate " tuning " 

 in the direction of the sexual smell, since errors of per- 

 ception would here be most disadvantageous to the species. 

 There would be a correlation between the degree of 

 specialisation of the larva in the matter of diet and the 

 definiteness of the smell which would prompt the female 

 to lay eggs. In many cases the food-smell of the adult 

 fly would be least narrowly adjusted. At all times other 

 senses such as those of sight and touch might play a more 

 or less important part as auxiliaries or controls. 



If we accept for the moment some such view as this, 

 then among those species in which the male finds the 

 female by smell we must regard each one as an assemblage 

 of individuals in which one sex is tuned to respond to a 

 certain definite kind of molecular vibration corresponding 

 to some compound or mixture of compounds emitted by 

 the other sex, and these compounds would thus constitute 

 definite specific characters. We might even perhaps go 

 further and define some of the larger groups by those 

 " generic " smells which characterise certain kinds of 

 chemical substances, such, for instance, as the organic acids, 

 the alcohols, amines, terpenes, etc., and which depend on 

 the presence of certain atoms or of atomic groups of some 

 particular configuration. 



In any case it seems a very remarkable fact that two 

 species such as D. zonatus and D. diversus which live in 

 the same district, and have always been regarded as quite 

 distinct, should have exactly the same sexual smell. There 

 is, of course, the possibility that citronella does not repre- 



