1932 ] 
The Olfactory Responses of Bloivflies 
147 
tank never varied more than 1/5° C. Each of the units used 
for the control of humidity consisted of two bottles ar- 
ranged so that air was drawn through solutions they con- 
tained; in addition, the “test” set had a vessel containing 
ethyl butyrate shunted between these bottles and the U- 
tube. This vessel was held at a constant temperature of 
30° C. in all tests. 
Relative humidity was controlled by chemicals. A con- 
centrated solution of a given compound gives the air with 
which it comes in contact a definite relative humidity. Four 
were used : 
Chemical Relative humidity 
98-100 
H 2 0 
NaCl 
NaOH 
NoS0 4 
73 
30 
0-4 
The general method was to change the humidity, the 
temperature being constant. The tests at the four humidi- 
ties were then repeated for another temperature. Three 
temperatures were used: 20° C., 30° C., and 40° C. 
Flies were collected in the afternoon and were kept with- 
out food for five hours. They had free access to water. 
Sixty-four flies were used in each test. The method of 
procedure was as follows : a fly was secured on the end of a 
galvanized iron wire by wrapping its wings and the end of 
the wire with adhesive tape. The opposite end of the wire 
was fixed in the lower surface of the rubber stopper of the 
test chamber. This last, containing the stopper with the 
attached fly and tubes was placed in the constant tempera- 
ture bath for one minute. One inlet tube was attached, by 
means of a rubber tube, to the “check” apparatus. The 
other inlet tube was similarly attached to a water-pump. 
The fly was observed for thirty seconds. If it did not re- 
spond (which was generally the case) the tube attached to 
the “check” apparatus was transferred to the “test” ap- 
paratus. Again the fly was observed for thirty seconds. 
If it extended the proboscis within this time, it was con- 
sidered “positive.” Sexes were used in about equal num- 
bers. Percentage of response was based upon the number 
