SABINE’S EXPERIMENTS. 65 
in my power to have made, in order to have brought their results in compe- 
tition, in point of minute accuracy, with those of the detached pendulums, sup- 
posing even that the two methods were equally unobjectionable in theory: a 
mean between the extremes in the 24 hours registered, by the same Six’s 
thermometer, has been taken as the mean temperature of the interval, in the 
present instances, as also, indeed, at all the equatorial stations ; wherefore the 
registered may be supposed generally to exceed by a small amount, the true 
mean temperature in cases where there has been no interference of artificial 
heat. : 
The experiments by this method were designed rather to exemplify the de- 
pendence which may be inferred on results obtained by means of a pendulum 
attached to a clock, than toafford a corroboration of the results of the de- 
tached pendulums; they may, however, be regarded as a confirmation, to a 
certain extent, when they agree within an amount which may reasonably be 
ascribed to the insufficient registry of the temperature, as is the case in the 
present instance, and has been hitherto, at all the stations at which the two 
methods have been tried in comparison. In the subjoined tables it is shown that — 
the pendulum numbered |, being kept in motion -by the clock, would make 
86334, 698 vibrations, and No. 2, 86444, 878 vibrations in a mean solar day, 
in a vacuum at the level of the sea, and at the temperature of 53 degrees; 
and that the same pendulums in the same clock, in London, would make res- 
pectively 86376, 244, and 86486, 26 vibrations under like circumstances and 
temperature ; the reductions in these tables having been computed as already 
described, except that as the pendulums are of cast instead of plate brass, 
their expansion has been taken at 0, 0220 of an inch per foot for 180 degrees 
of Fahrenheit, making a corresponding correction of 0,44 for each degree of 
temperature to be applied to the number of vibrations in 24 hours. 
The acceleration deducible from the preceding results is as follows; from 
the first half interval 41, 546 seconds per diem by pendulum No. 1; and 
from the second half interval 41, 382 seconds by pendulum No 2; a mean of 
the whole interval and of the two pendulums being 41, 464, which is less than 
the result of the detached pendulums by 0,474 per diem. 
9 3 
’ 
