184 F. A. P. Barnard on the Pendulum. 
Arr. XXI.—On the Pendulum; by F.-A. P. BARNARD, Pres 
dent of the University of Mississippi,—with a description of an 
Electric Clock, constructed by E.'S. Rrrcu1e, of Boston, 
the University of Mississippi, under the direction of President 
Barnard.— With a Plate. 
[From the Proceedings of the American Association at Baltimore, 1858.] 
THE importance of the pendulum as an instrument for the 
measurement of time, is sufficient to justify any amount of effort 
which may be made to secure the regularity of its performance. 
The causes which disturb this regularity exist ‘partly in the 
nature of things, and are partly introduced by the contrivances 
employed to maintain the motion of the instrument. -Amoug 
these, the effect of varying temperature in altering the distance 
between the centre of oscillation and the point of suspension, * 
one which has given occasion to many ingenious invenhons; 
yet, however effectual some of these may have been in removilg 
the irregularity due to this cause, it is probably true that no plan 
of compensation has been found in practice to be entirely saul 
factory. It is an opinion entertained by the writer, though 168 
proposed with some diffidence, that the problem of compensation 
cannot be experimentally studied with results to be perfec Te 
lied upon, so long as the pendulum has any work to do; an he 
must be the case whenever the maintaining power is derived, di 
rectly or indirectly, from a train of wheel-work. ‘The different 
forms of anchor or pallet escapement involve friction upo? the 
pallets, which, however nearly constant it may be, cannot be 
wholly so, and however slight it may be, either absolutely ot 
its variations, cannot be altogether insensible as @ GI5™ this 
cause. Hor small as may be the amount of fluctuation er 
resistance, it is to be considered that all the quantities conce™ ” 
in the question of maintaining the motion of the pendulum 2 
small, and that every minute variation is multiplied thousa” 
sorbed, and is at different times unequally commun! 
ndulum. : , 
pe : 
When we attempt to study, in the actual performance 1 
pendulum, the efficacy of any plan of compensating the 
of temperature, it is impossible entirely to distinguish the’ 
ularities due to one cause from those which may P ante ray 
- another. Cold, for instance, by stiffening the lu ri at 
_cause the clock to gain, and ‘this effect may be re the 
_ ascribed to an under-compensation of the contractio? 
