1914-] ATMOSPHERIC PRESSURE. 67 



The relation found in our 19CKI paper was that 



P^; = lcd\/v abwatts per cm. (2) 



where Pc is the linear convection from the hot wire in abwatts per 

 cm., the temperature elevation of the wire, in degrees Centigrade, 

 V the wind-velocity or speed of transverse motion of the wire 

 through the air in cm. per second, and k a constant depending, among 

 other things, on the size and surface-condition of the wire. This 

 formula was found to hold well between the wind velocities z/ = 200 

 and ^' = 2000 cm./sec. (7.2 and 72 km/hr. or 4.47 and 44.7 statute 

 miles/hr.) ; but not to hold below z'==200, unless 30 cm. per sec. 

 were added as an empirical correction to all speeds to take free con- 

 vection into approximate account. This empirical correction, apply- 

 ing fairly well, gave : 



Pc = kO\/v-\-Vf^ abwatts per cm. (2a) 



where Vq is a virtual velocity of free convection approximating 30 

 cm. per second. 



The relation indicated in (2) can be presented graphically by 

 straight lines on logarithm-paper, hut Professor Morris has employed 

 the corresponding relation : 



(5) 



- , ,„ /abwatts per cm. \^ ^ ^ 

 t) =^-' ( diiTC ) (3) 



That is he plots the square of the observed linear convection per 

 degree C, against the wind velocity, thus producing a straight line, 

 if either (2) or (2a) applies. The procedure is followed in Figs. 7 

 and 8. Thus, taking Fig. 7, the broken straight line marked 2.04 cor- 

 responds to the results in an air-pressure of 2.04 megabars, and the 

 observations in Table I appear on or near this line as small circles. 

 Nine different series are indicated in Fig. 7 at pressures of 0.44, 0.66, 

 1.02, 1.54, 2.00, 2.04, 2.81, 3.48 and 3.95 megabars respectively, the 

 first two corresponding to vacua, 1.02 to normal atmospheric pressure, 

 and the six others to extra pressure in the tank. 



It will be seen that the two lowest curves — vacua — deviate dis- 

 tinctly from straight lines. The remainder are drawn as straight 

 lines, and the observations conform to them fairly well, except at the 



