Vol. 6, 1920 
PHYSIOLOGY: H. SHAPLEY 
207 
pendent mainly on the velocity of chemical reactions, this result should 
be expected. 
The quantities in the sixth column of table I are the observed mean 
time required for an ant to travel 30 centimeters, at the corresponding 
temperature, and the probable error of that mean; uncertainties of timing, 
frequently contribute nearly a tenth of a second to the probable error. 
The seventh column contains v, the average speed, reduced to centimeters 
per second. 
Temperature is plotted against speed in figure 1 . Observations made at 
station A are plotted as dots, and those made at station Y as crosses. 
Two observations of low weight (Nos. L2A and L27A) are plotted as open 
circles; the recorded temperatures are undoubtedly high, because the 
timing of ants was begun too soon after introducing the cardboard shade 
Speed in centimeters per second. 
1234567 
40^ I ^ ^ 1 ^ ^ ^ , 
Fig. 1. The temperature-speed curve for Liometopum apiculatum. 
and the artificial lowering of temperature is responded to much more 
quickly by the ants than by the mercury thermometer. (Observations 
made in January 1920 at stations F and G confirm the supposition that 
the two measures in question were erroneous and should be rejected in a 
mathematical discussion of the curve.) 
The quantities in the eighth and ninth columns of table I are, respec- 
tively, the percentage difference in speed for inbound and outbound ants, 
and the percentage difference in speed for large and small workers. These 
observations of systematic differences were designed to give evidence of 
the effect on speed of fatigue and mass. A positive value of {i — o)/O.Olv 
