238 
CAPTAIN SABINE ON THE REDUCTION TO A VACUUM 
difference of temperature is always in favour of the tropical stations, the 
error will be of a constant nature, unlike the greater part of the small irre- 
gularities to which pendulum experiments are liable, which may compensate 
themselves by the multiplication of the experiments. 
The stations which I have myself visited with the pendulum embrace a 
greater range and variety of temperature, I believe, than those of any other 
experimentalist. The preceding remarks would therefore apply particularly 
to them, but for a circumstance which has fortunately occasioned, in all cases, 
a compensation of the errors, which would otherwise have arisen from the in- 
fluence of the variations of temperature on the density of the air. This compen- 
sation is a consequence of the “ correction for temperature,” (i. e. the correction 
of the vibrations for the temperature of the pendulum,) having been obtained 
by the peculiarly practical mode, of vibrating the pendulum in London in 
temperatures differing so widely as to include the whole range experienced 
elsewhere, — instead of deriving the correction from the expansion of the metal 
in pyrometric experiments. Had the vibrations at high and low temperatures 
in London, from which the correction for temperature was obtained, been re- 
duced to a vacuum by the true reduction, as it is now known, — instead of by 
the “ correction for buoyancy,” which was then thought to constitute the true 
reduction, — the value of a degTee of Fahr. on the daily vibration would have 
been found 0.43 instead of 0.421. The first of these numbers is the true cor- 
rection for temperature due to the expansion of the metal ; the second is that 
correction diminished by the effect of a degree of temperature on the part of 
the reduction to a vacuum heretofore neglected. It is the second number 
(0.421) which has been used throughout the experiments to which I allude, in 
reducing them to a mean term of comparison ; consequently the corrections 
for temperature so applied are every where too small, if regarded as represent- 
ing only the effect of the expansion of the metal on the vibrations of the 
pendulum ; but they are experimentally correct, when regarded as representing 
the joint effects of the expansion of the metal, and of the difference of tem- 
perature on the part of the reduction to a vacuum not comprehended in the 
correction for buoyancy. 
The effect of differences of barometric height on the results of those experL 
ments is too inconsiderable to require express correction ; not exceeding 0.02 
