t)owLiNG — Steady and Turbulent Motion m Gases. 387 



(2) wbile a bbist of steiun was )iassiiig- steadily around the glass tubes. Tbe 

 arrangements, save for the steam-jackets, were again as represented in fig. 1 ; 

 and the gasometer-method was, of course, employed to draw and measure the 

 air-curreut. 



Tlie method of measuring the air-current for the hot air was exactly 

 the same as for the uuheated air. No correction had to be employed for 

 the temperature of the air in the drum V (fig. 2) ; for, after filling this 

 drum with air through the heated pipes, it was found that the air was not 

 appreciably higher in temperature. In its passage through the pipe P 

 inside the gasometer, it parted completely with its heat. Thus the rate of 

 rise of the drum gave the volume if room, feiiipertiliire flowing in per second. 



The equation we are to verify may be written : — 



J..p- -.V. 



or, in words, that the product V^ p should bo proportional to the viscosity. 



Now Vp is proportional to the mass of gas carried past a point in the 

 tube when the gas has a density p. 



If suffixes denote temperatures, for a given rate of rise of the gusometer, 

 the gas being at room-temperature (10" 0.), we should have: — 



V\o jOio = ^8 pe. 

 i.e., Flu is the velocity the gas would have if the tubes were at 10° C 

 Fio, however, is given by the formula (3) above, viz. : — 



We will call Fjo the "apparent" velocity; and it is easy to see tliat this 

 " apparent" velocity is proportional to the product Vg pe. Consequently, it 

 will be sufficient to show that the "apparent " critical velocity at any known 

 tube-temperature (0) bears the same ratio to the critical velocity at 10° as 

 the viscosity at 0° C. bears to the viscosity at 10° C. 



The temperature of the gas in the tube cannot be assumed to be that 

 of the walls of the tube, for the gas will take some time to become heated. 

 To determine this temperature, as closely as possible, a separate experiment 

 was carried out. Into the tube, AB, was inserted a fine platinum wire 

 stretched along the axis, and the ends of this were connected by thick 

 copper leads to a Post-office box. Its resistance was then measured in 

 the usual waj'. Several determinations were made of the resistance ; first 

 at room-temperature; then when the steam-jackets were heated, but while 

 no air was being drawn down the tubes GAB ; and finally while the air was 

 being drawn down at various speeds, both above and below the critical veloeitj'. 



