142 THE PHYSIOLOGY OF INSECT SENSES 



chemoreceptor activity by behavioural techniques. It was shown, for 

 example, that a single labellar hair responded to touch, water, certain 

 sugars, and a wide variety of electrolytes and non-carbohydrate 

 organic compounds. In a hungry, thirsty fly bending a hair, applying 

 water, or stimulating with certain sugars eHcited a reflex proboscis 

 extension. The application of such compounds as salts, acids, alcohols, 

 aldehydes, ketones, glycols, ethers, etc., to the same hair prevented 

 extension or elicited retraction if the proboscis was already extended. 



By adapting the hair successively to bending, to water, and to carbo- 

 hydrate, and by noting that removing the tip of the hair prevented 

 response to chemicals but not to bending, Dethier (1955 a) concluded 

 that the neuron whose process terminated at the base of the hair was a 

 mechanoreceptor. He concluded further that of the remaining neurons 

 one was concerned with acceptance (exclusive of mechanoreception) 

 while another was concerned with rejection. In other words, one 

 neuron was conceived of as being sensitive to water and certain 

 sugars while the other was considered to be sensitive to all of those 

 compounds that caused rejection. Later behavioural experiments 

 (Dethier, Evans, and Rhoades, 1956) indicated, however, that re- 

 jection was in reality not a simple modality because it could be initi- 

 ated either by stimulation of the rejection receptor or inhibition of the 

 acceptance receptor. Sugars such as mannose, rhamnose, and sorbose, 

 which of themselves do not cause rejection, are able- to interfere with 

 stimulation by certain other sugars that alone are acceptable. For 

 example, mannose inhibits fructose (but not glucose), rhamnose 

 inhibits glucose (but not fructose or sucrose), sorbose probably in- 

 hibits glucose and fructose. Furthermore, certain rejected salts, 

 notably HgClg and CuClg, render sugar receptors reversibly insensi- 

 tive. 



Still more recently, behavioural investigations into the mechanisms 

 controlling thirst and water satiation suggested that acceptance was 

 not a simple modality either and that there must be a water receptor 

 distinct from the carbohydrate receptor (Evans, 1961 a; Dethier and 

 Evans, 1961). In female flies at certain stages of the reproductive cycle 

 acceptance is further broken down into carbohydrate acceptance and 

 protein acceptance (Dethier, 1961). 



Conclusions drawn from behavioural studies were borne out to a 

 very satisfactory degree when success finally attended attempts to 

 record electrically from chemoreceptive hairs. After many people had 

 tried and failed to record action potentials by standard techniques, 

 Hodgson, Lettvin, and Roeder (1955) and Morita, Doira, Takeda, 



