424 
MR. BAILY ON THE CORRECTION OF 
the included air: and there is little or no difference whether the hollow body 
be hermetically sealed, or whether the ends be loosely closed, and a free com¬ 
munication left between the internal and external air: due regard being had, 
in all these cases, to the correct determination of the vibrating specific gravity 
of the body. When both ends of the cylinder are left open, the effect of the 
air appears to be the least, as in the pendulum 15 ; and it is increased when 
either the top or bottom pieces are replaced, as in pendulums 16 and 17 : which 
seems to show that some slight modification of the results is caused by leaving 
the ends open to the circumambient air. I would observe that, with the ex¬ 
ception of No. 14 and 19, the specific gravities of the cylinders could not be 
practically determined, but were computed only; and from assumptions rela¬ 
tive to the contents of the cylinder, which could not be completely verified. 
But they are probably very near the truth ; and the comparative results cannot 
be materially affected by any error that is likely to have occurred. The repe¬ 
tition of three of the experiments, as stated in the preceding note, after the 
cylinder had been altered, and its contents subjected to a new computation, 
shows the degree of accordance that may be attained in these experiments. 
If the results of the experiments with these hollow cylinders be compared 
with those made by M. Bessel, with a hollow brass cylinder of a somewhat 
similar form, vibrating in air and in water, there will be found a very con¬ 
siderable and remarkable difference; inasmuch as he makes the value of n 
equal to9T00*. But, on examining the steps of the process by which he 
deduces this value, it will be easy to discover the source of this apparent dis¬ 
cordancy. The specific gravity of the brass, of which the cylinder was formed, 
is stated to have been 8’3 ; but by reason of the included air, the specific gra- 
interior of the cylinder, in the experiment with pendulum 19, when the vacuum tube was exhausted, 
I repeated the experiment, and found the following result: 
2-076 
2-160 
Mean = 2'118 
But, here also, from some appearances round the screw of the bottom piece, I had again reason to 
suspect the escape of some of the air from the interior of the cylinder; which may perhaps account 
for the slight discordancies apparent in the partial results. The whole however are very satisfactory. 
* See his work, page 67. He makes the value of k, from two experiments, equal to 7'99 and 8'21 ; 
mean = 8'100 : to which, unity must be added, in order to obtain the value of n. The diameter of 
M. Bessel’s cylinder was 2-84 inches, and its height 3 - 20 inches. 
