1 68 D\VIGHT ELMER MIXXICH. 



death, each animal being fed once daily with iM saccharose 

 solution. When the solution failed to effect a proboscis extension 

 through tarsal stimulation, it was found that a response could 

 always be obtained by touching the oral lobes of the proboscis 

 tip to the solution. It was, therefore, possible to force every 

 animal to bring the extended proboscis in contact with the 

 solution even though it drank little or none of it. Two of the 

 eleven flies died within twenty-four hours. The remaining nine, 

 however, lived many days. Exactly, their longevities were 9, 

 10, ii, 28, 30, 33, 35, 44 and 54 days. Six of the animals, three 

 males and three females, thus lived a month or more, the greatest 

 longevity extending to almost two months. It is doubtful 

 whether these six animals would have lived any longer, if as 

 long, in a state of nature. (Compare Glaser '23.). 



The use of proboscis extension as an index of tarsal stimulation 

 entails a certain amount of care to avoid misinterpretation. 

 Extensions of the proboscis are sometimes called forth by causes 

 other than tarsal stimulation, e.g., cleaning operations, distance 

 chemical stimulation not under the control of the experimenter, 

 or hypersensitivity to mechanical stimulation induced by in- 

 anition. In the last respect flies are like butterflies in that they 

 sometimes become so sensitive during inanition that slight 

 mechanical stimuli effect proboscis extensions, the proboscis 

 even remaining permanently extended in some cases. Care was 

 taken to avoid, in so far as possible, such sources of error, the 

 animals being discarded if there was any regularly recurring 

 question of interpretation. Occasionally, particularly with 

 Lucilia, very slight movements of the oral lobes of the proboscis 

 tip were noted which were apparently not coupled with any 

 extensor movement. Such movements were counted as "no 

 responses." Similarly, slight movements of the palp, noted in 

 a few instances, were disregarded. 



The methods used in this paper are similar to those used in 

 previous papers dealing with similar problems. Each trial lasted 

 30 seconds, at the end of which period "no response" was 

 recorded if no extension of the proboscis had been observed. 

 Minimum intervals of ten minutes were allowed between trials. 

 Extensions of the proboscis judged as less than one half of com- 



