72 NOTICES OF THE MEETINGS [May 9, 
watched through a complete revolution, the time of which was ob- 
served to be 28h. 26m. 
The sources of probable error are numerous and not easy to be 
effectually guarded against. The most formidable, perhaps, is the 
extreme difficulty of causing the pendulum to vibrate truly in oneplane, 
and to prevent its motion in a narrow ellipse. When this takes place, 
and the arc is considerable, the direction of the major axis is con- 
tinually changing, owing to a well known mechanical cause (see Her- 
schel’s Outlines of Astronomy, p. 444); but this deviation is 
always in the same direction as that of the original motion of the 
pendulum, and consequently changes when that direction is changed. 
The true deviation may be distinguished from this, in that it is 
always from E. to W., independently of the direction of the original 
impulse ; and the ball always passes accurately through the centre 
in every oscillation, whereas in the former case it never does. 
For great accuracy, a variety of other precautions are requisite, as" 
to the perfect freedom of suspension, guarding against currents, &c. ; 
It is, however, possible that the elliptic deviation may oppose that due 
to the earth’s rotation, while the latter may manifest itself in spite of 
the former. 
It is extremely probable that many of the public repetitions may 
have been affected bythese causes of error; yet some of those 
referred to have been made by men of so much eminence and experi- 
ence as observers, as to render it highly improbable that they should 
not have been sufficiently guarded against every source of fallacy. 
The accordance of many of the results at different places within fair 
limits of error, is also a strong argument in favour of their accuracy 
and trustworthiness, 
The rates of deviation for one hour as determined at different 
places do not seem to be more discrepant than would accord gene- 
rally with the differences of latitude. The experiment at Paris gave 
about 11° 30’, at Bristol 11° 42’, at Dublin rather more than 12°, at 
York about 13°. 
To apprehend the theoretical principle it is necessary to take 
into account, lst, the simple inclination of two successive posi- 
tions of the meridian of a place to each other after any interval of 
time: 2nd, the independence of the motion of the ball of the pendu- 
lum, of the rotation of the point of support: and, 3rdly, that the ball, 
though free in this sense, is not however wholly free, being continually 
drawn down by gravity in a direction continually changing (relatively 
to the original direction of vibration,) as the earth revolves. Hence, 
though from the second cause the ball would have a tendency always 
to preserve a motion parallel to its original motion, and thus to deviate 
regularly from the meridian, it will (from the third cause) not preserve 
this exact parallelism, but will take an intermediate direction. The 
exact determination of this direction cannot be made on any general 
considerations, but must be the result of detailed mathematical inves- 
tigation. 
ee 
