7 70 MR. H. TOMLINSON ON THE COEFFICIENT OF VISCOSITY OF AIR. 
cylinders or spheres had been removed. The caps D, D, were now replaced in Y V, 
and the logarithmic decrement, together with the time of vibration, was once more 
carefully determined. Observations such as these, when certain corrections presently 
to be mentioned, had been applied, enabled one to calculate the effect of the resistance 
of the air on the vibrating vertical cylinders or spheres as far as the diminution of the 
amplitude of vibration was concerned. 
The bar V Y with its appendages was protected by a wooden box B of sufficient 
size to permit of vibrations, which, as regards the resistance of the air, were practically 
as free as in the open. # This box was provided with a window, E E, and two side- 
doors, lined with caoutchouc so as to fit air-tight; these side-doors were kept shut, 
except when it was necessary to make fresh adjustments. The torsional vibrations of 
the wire were observed by means of the usual mirror-and-scale arrangement, which is 
sufficiently shown in fig. 1, where M is the light mirror reflecting an illuminated circle 
of light crossed by a vertical, fine, dark line on to a scale bent into an arc of a circle 
of 1 metre radius, and placed at a distance of 1 metre from the mirror. 
My three years’ experience of the internal friction of metals had taught me that 
this last is by no means constant unless the greatest care be taken to prevent slight 
fluctuations of temperature. The above-mentioned fact seems to have escaped the 
notice of Maxwell and Meyer, probably on account of the internal friction of the 
metal having a considerably less damping effect than the resistance of the air in their 
experiments. With me, however, especially in some cases, changes in the internal 
friction of the metal would have rendered it very difficult, nay, impossible, to attain 
the accuracy which I aimed at, and I deemed it advisable to protect the wire still 
further, as follows:—The top of the air-chamber W was well covered with baize, and 
surrounding W, and concentric with it, was a larger air-chamber X, made of tinned 
iron. This air-chamber was Ilf inches in inner diameter, 15 inches in outer diameter, 
and 46 inches in height; the two concentric chambers of which it was composed 
enclosed between them a space 2 inches thick, stuffed witli sawdust, whilst on the top 
of the chamber was placed a double cover A, also packed with sawdust. Passing 
through the outer air-chamber X, and through the walls of W, were two metal tubes 
in which were placed two thermometers T l5 T 3 , with their bulbs near the wire; these 
thermometers were made by indiarubber tubing to slide air-tight in the metal tubes. 
A section of the two chambers X and W passing through one of the thermometers is 
shown in plan in fig. 2. The whole of this part of the apparatus rested on a stout 
wooden table, in which was pierced an aperture of a size just sufficient to allow the 
zinc tube Z, soldered to the air-chamber W, to pass through it and into the box 
beneath A third thermometer T 3 served to give the temperature of the air in the 
* Ixl fig. 1 the cylinders appear to he closer to the sides of the box than they were in reality; the bar 
V V faced the window, but for the sake of showing the arrangement of the cylinders better, it has been 
drawn facing the adjacent side of the box. The centres of the cylinders were at least six inches from 
the sides of the box. 
