5 
on the Rates of Chronometers. 
qfy when the same power passed through nearly the middle qf 
the balance and its spring. 
The effects produced on the rates of the preceding chrono- 
meters, by the direct transmission of the magnetic power through 
the centres of their respective main-springs, having been thus in 
some measure found, an attempt was next made to estimate 
what alterations of rate would result from the partial transmis- 
sion of the attractive power through the same centre. For this 
purpose, chronometer (A) was again employed, and placed in 
the position represented in Fig. 5., so that a radial line proceed- 
ing from the centre of the time-keeper through the middle of the 
main-spring, might form an angle of 27° with the longitudinal axis 
of the magnet. The consequence of this application, was an 
immediate increase of 4- 20".l, its detached rate, to -f 52".3 ; a 
quantity less than the mean of the two applications of the chro- 
nometer in the position denoted by Fig. 1., of -|- 16".6. By 
turning the chronometer a quadrant, in order that the radial 
line before alluded to might form an angle of 117° with the axis 
of the magnet, as shewn in Fig. 6., and which position also 
brought the centre of the balance into the exact direction of the 
longitudinal axis of the magnet, the daily increment declined 
from -f- 52".3 to -f- 29".l ; the magnitude of which rate, when 
contrasted with that observed in the same chronometer, when in 
the position of Fig. 2., being very remarkable. By again turning 
the machine through another quadrant, so as to bring the chrono- 
meter into the situation of Fig. 7., wherein the radial line formed 
with the axis an angle of 153°, and the attractive power of the 
magnet was only transmitted through small segments of the spring 
and balances, the rate was augmented from -j- 29". 1 to -f- 33".7 ; 
being less than the rate determined for the same chronometer, 
when placed as in Fig. 3., by 9".7. Lastly, by turning the 
time-keeper another quadrant, in order that the angle formed 
by the radial line and the axis should be 63°, .as represented in 
Fig. 8., another declension in the rate took place, from -f 33". 7 
to -p 18". 5 ; and when the machine was afterwards restored to 
the situation it occupied in Fig. 5., the rate again increased 
to -p 52". 3, agreeing within 3" of its former rate when in the 
same situation ; — an aberration by no ifneans considerable, when 
the great irregularities here alluded to are considered. These 
