442 
MR. BAILY ON THE CORRECTION OF 
would be considerably increased. But, to what limit this might extend I had 
no means of ascertaining, since the above steel wire was the finest that I could 
operate with : for, on account of its small weight, a pendulum of this kind soon 
comes to rest : and in order to guard against any probable error arising from 
this source, I took the mean of three consecutive sets of experiments, in de- 
termining each separate result. It also appears from these experiments that 
the quantity of adhesive air decreases as the diameter of the rod diminishes. 
For, a rod, about 59 inches long, and whose diameter is about 4 tenths of an 
inch, drags with it nearly 5 grains of air : whilst another rod of nearly the 
same length, and little more than one sixth of the diameter drags with it 
scarcely half a grain. But, although the thicker rod drags more air than the 
smaller one, yet the effect on the latter is much more considerable than on the 
former. For the 4*904 grains of air added to the weight of the copper rod, 
would reduce the specific gravity of the vibrating mass from 8*629 to 2*939 
only : whilst the 0*479 grain of air added to the weight of the steel rod, would 
reduce the specific gravity of the vibrating mass from 7'687 to 1*024. And 
these are the respective specific gravities which if used in the computations for 
the reduction to a vacuum, would cause n to vanish 
Having thus ascertained the fact that the influence of the air is greater 
upon small rods than upon large ones (increasing considerably as the diameter 
of the rod diminishes), I next tried what effect would be produced by affixing 
various bodies to the ends of these rods. For this purpose I made use of the 
two brass spheres No. 3 and No. 6, already described in the preceding part 
same kind of wire, and having precisely the same specific gravity, but about half an inch longer. The 
results differ from each other more than I could have imagined ; although each set is consistent in 
itself: for we have June 14. Aug. 2. 
4'232 3-975 
4-179 3-947 
Mean = 4-206 Mean = 3-961 
Weight of air = 1'536 Weight of air = 1-431 
I have examined all the steps of each experiment, and of the computations connected therewith ; but 
cannot detect any source of error. In fact, it is one of those perplexing anomalies which occasionally 
occur in our researches after such minute quantities. 
' I cannot trace the exact law of the variations in the three values in the column, indicating the 
weight of adhesive air dragged by each rod ; but the nearest approximation is, that the numbers are 
nearly in the ratio of the square root of the cubes of their diameters. 
