MR. H. TOMLINSON ON THE COEFFICIENT OF VISCOSITY OF AIR. 
799 
Experiment VII. 
The copper wire used in the last experiment was about 4^ feet in length and 
01 centim. in diameter. This was now changed for one of the same length, but of 
0'063 centim. diameter, so that the vibration-period became 8'930 seconds. The 
rest of the arrangements were the same as in Experiment YI. The corrected 
logarithmic decrement was ’0027040, and the value of /x deduced as above was found 
to be '00017902 at a temperature of 13°'100 C. 
The mean of the two last experiments is '00017741 at a temperature of 1'2°'663 C. 
This result agrees so well with the mean of those deduced from the previous experi¬ 
ments that it is unnecessary to make any alteration in the formula already given for 
finding the viscosity at any temperature. 
I have entered more into the details of these last experiments, as I think the 
present method can be more advantageously employed than any of the others. 
Indeed, by spending sufficient time over the experiments, whereby the errors likely 
to arise from the somewhat unstable nature of the internal friction of the metal may 
be more perfectly eliminated, it seems likely that very considerable accuracy can be 
attained by it. 
[Note added Dec. 8th, 1886.—A much greater number of observations were after¬ 
wards made with the same cylinders and wires, and resulted as follows :—With the 
wire used in Experiment YI. the value of [x obtained was '00017708 at a temperature 
of 12°'225 C., and with the finer wire of Experiment YII. the value was '00017783 at 
a temperature of 13°'075 C. The mean of these values is '00017746 at 12°'650 C., as 
compared with '00017711 at 11°'79 C., the mean of the other five sets of experiments. 
If we allow for the difference of temperature by using the previously given formula, 
the agreement between these two means is perfect.] 
In conclusion, my warmest thanks are due to Professor Stokes for his valuable 
suggestions and advice throughout the investigation. To myself the experimental 
verification of Professor Stokes’s formulae has been a source of great pleasure. 
