HOGG. — VISCOSITY OF AIR. 625 



in this case be about 5.5. On ihis account one would have expected a 

 larger result than that given. 



In the experiment performed by F. G. Reynolds the s{>lierical shell 

 which he used had a diameter of nearly 13 cm. It was made to vibrate 

 about a diumeter, and the period of oscillation was 16 seconds. The 

 maximum angle, measured from the position of equilibrium, was appar- 

 ently about 0.15 radians, so that the maximum velocity of a point on the 

 equator of the sphere must have been approximately 0.19 cm. per second. 

 The square of the quantity is not small com|)ared with the quantity itself, 

 and the quantity Vcipjfx for this case is about 9.4. There may be in 

 this an explanation of the fact that the result obtained by Reynolds is 

 about 3 per cent greater than the mean of the results given in this 

 paper. 



The analogy between the method of the single cylinder and that 

 employed in the present paper is apparent. As has been seen, the 

 result obtained by Tomlinson by this method is almost identical with 

 that obtained in the present investigation ; but Reynolds, by this method, 

 obtained /7.o = 0-0001768. An effort has been made to discover the cause 

 for this discrepancy. 



In order to eliminate the effect of friction on the mirror and wire and 

 the viscosity in the wire itself, Reynolds made use of a telescopic arrange- 

 ment of two hollow cylinders. The decrement was measured, first, with 

 the outer cylinder covering the inner, and then with the inner one pulled 

 out a distance of 25.4 cm. The difference between these decrements he 

 took to be that due to the friction of the air on the inside and outside of a 

 length of 25.4 cm. of the inner cylinder. The inside diameter of the inner 

 cylinder was 4.70 cm., and the inside diameter of the outer was 5.00 cm. 

 The thickness of the metal was 0.04 cm., so that there was a space of 

 0.11 cm. between the cylinders. In obtaining in the manner described 

 the decrement for the extended portion of the iinier cylinder the as- 

 sumption was made that the friction on the inner surface of a length of 

 25.4 cm. of the outer cylinder was the same as the friction on the same 

 length of the inside surface of the inner cylinder together with the friction 

 due to a length of 25.4 cm. of the layer of air between the cylinders. 

 This assumption would seem to require justification. 



On account of the two distinct operations involved in the method used 

 in the present investigation, it is not well suited for the determination of 

 viscosity at low pressures. Another investigation is in progress, one of 

 whose objects is to determine the law relating viscosity to pressure at 

 pressures where slip is appreciable and, when this is done, to measure 



VOL. XL. — 40 



