444 
MR. BAILY ON THE CORRECTION OF 
Now although there is enough on the face of the above experiments, to con- 
firm the leading principles we are in search of, yet for the reasons already 
mentioned I should not select them as the most proper for the deduction of 
any very minute results, when compared with others made under more favour- 
able circumstances. 
If we examine the values, denoting the weight of adhesive air dragged by 
the compound pendulums, formed of the spheres attached to the ends of the 
several rods, they will be found to exhibit some apparent anomalies ; more 
especially in the case of No. 45 and 48, where the weight of adhesive air seems 
to be less when the spheres are applied, than with the plain rod. But, it must 
be borne in mind that the deduced weight of adhesive air for each pendulum 
is in each case supposed to be applied to the centre of gyration (which is a dif- 
ferent point of the rod, in each pendulum), and therefore requires correction. 
The collateral column, showing the weight due to the sphere alone (agreeably 
to the formula in page 440) will exhibit more accordance in the results ; and 
denotes more distinctly the quantity we are in search of. 
With a view of obtaining greater accuracy on the points in question, I 
resolved to try the effect of placing the spheres at, or near to, the centre 
of oscillation of the rods : whereby the above-mentioned inconve- 
nient change in the intervals of the coincidences would be avoided, 
and the results rendered more trust-worthy. For this purpose I 
divided the brass and steel rods into two unequal parts at, or near 
to, the centre of oscillation : so that by screwing the longest of the 
two parts into the upper portion of the spheres, and the shortest 
into the lower portion, I might accomplish this object. But, as the 
whole length of the pendulum (from end to end) would, in such case, 
be longer than the rods, by the diameter of the inserted sphere, I 
cut off one inch from each part, in order that the length of the pen- 
dulum, from the knife edge to its extreme end, might, when thus 
used with the different spheres, be more nearly the length of the rods 
prior to the alteration. The two parts therefore of the rods, thus 
reduced, were 3G - 4 and 18‘0 inches respectively. The copper rod 
was the property of Mr. Troughton, and could not be thus divided. 
The following are the results with the spheres thus placed. 
