MR, H. TOMLINSON ON THE COEFFICIENT OF VISCOSITY OF AIR. 
777 
as the square root of the pressure; the amount to be added to the uncorrected 
logarithmic decrement, owing to the above causes, will therefore be 
where p l and p. 2 are the pressures with and without the paper cylinders respectively. 
Further, when the cylinders were screwed on to the suspenders, about 4 mms. of 
the latter entered the former, so that the observed logarithmic decrement was less 
than it should be by an amount which would be nearly equal to the logarithmic 
decrement due to the resistance of the air on two vertical cylinders 4 mms. in 
length and 0'336(3 cm. in diameter; this could be calculated to within a sufficient 
degree of approximation by using the approximate value of p. The amount in this 
particular case was '0000037. 
Lastly, the temperature of the wire was not the same with and without the paper 
cylinders, but, as the effect of change of temperature had been determined previously, 
this difference could be allowed for. 
No correction is required for any variation in the internal friction of the wire itself, 
arising from difference in the vibration-periods with and without the paper cylinders ; 
for I had previously satisfied myself that the diminution of amplitude resulting from 
internal friction is nearly independent of the time of vibration. 
Accordingly we have the following amounts to be added to the uncorrected 
ithmic decrement:— 
Correction. 
For difference of time of vibration with and without paper cylinders 
For difference of pressure of air. 
For difference of temperature of the wire. 
For portions of suspenders which enter the cylinders. 
+ -0000008 
--0000005 
-•0000002 
+-0000037 
Total. 
+ -0000038 
Corrected logarithmic decrement. 
•0027411 
In calculating p, the density of the air, I have assumed that the latter is half 
saturated with moisture, and that the mass of a cubic centimetre of dry air at 0° C., 
and under a pressure of 29*9217 inches of mercury, is '0012930 gramme; thus, in 
the present instance, 
P = 
29'872—-3-x -206 
29-9217 
X; 
Wo 
273 + 12-02 
X '001293='0012334. 
The distance from each other of the axes of the two paper cylinders was 
20"80 centims., and this distance was maintained in all the experiments which 
follow, except the last, where it was 20"78 centims. 
5 G 
MDCCCLXXXVI. 
