Meperîments on the highest audible tone. 
107 
mean variation for a set of results seldom exceeded 5 per cent, it may 
well be supposed that the constant of precision of the measurements 
was in this case determined by technical errors and not by psycho¬ 
logical variations. In future experiments it will be necessary to main¬ 
tain the room at an even temperature during each set of experiments ; 
and to calculate the pitch with the appropriate value of v for each 
change, possibly also to be on guard against sudden barometric 
changes. 
It may seem remarkable that we should have neglected to record 
the temperature of the room. We give the following as reasons : 1. 
we did not expect after the elimination of the error of air-pressure 
to find the psychological sources of error smaller than the technical 
ones ; 2. the temperature of the air has not been regarded in pre¬ 
vious experiments ; 3. it is not the custom of psychological labora¬ 
tories to pay attention to the psychological and instrumental errors 
due to changes in temperature. 
Several means of blowing the whistle have been employed. 
Galton1 used a single rubber bulb. Zwaardemaker2 used a fun¬ 
nel with a rubber membrane stretched across the large opening, the 
smaller end being connected with the whistle by a rubber tube ; the 
funnel was depressed through a constant distance. These methods 
are none of them strictly reliable. The pressure is intermittent and 
cannot be accurately regulated. The only possible method of obtain¬ 
ing a current of air of constant intensity, seems to be by use of a 
rotary-fan blower. This method has been previously described and 
tested.3 The source of power used by us was an electric motor run 
by the city-current ; there were no perceptible fluctuations in 
speed. The fan-wheel of the blower made from 13 000 to 15 000 
revolutions a minute. The blast was carried by a rubber hose into 
a room in another part of the laboratory ; thus all noise from the 
machinery was avoided. The hose led to a rubber tube, in which 
was a stop-cock. The rubber tube ended in a glass T-coupling with 
a rubber tube on each end of the cross-arm. One of these led to the 
whistle, the other to a water-manometer. The manometer scale was 
graduated to millimeters ; the height of the column of water gave the 
1 G Alton, Whistles for audibility of shrill notes; Inquiries into Human Faculty, 
38, New York 1883. 
2 Zwaardemaker, Der Umfang des Gehörs in den verschiedenen Lebensjahren, 
Zt. f. Psych, u. Phys. Sinn., 1894 VII10. 
8 Scripture, A constant blast for acoustical purposes, Am. Jour. Psych., 1892 
IV 582.