4 le Astronomical Society. 



in the door of the case; and this being screwed close, (as described 

 in the Armagh Observations (or 1828,) all free communication with 

 the air is precluded, and the changes of temperature within it are 

 slower and more regular than would be possible in clocks of the 

 ordinary construction. This made observations at other than the 

 stated periods unnecessary. The barometer was observed whenever 

 its movements seemed irregular, and the mean for the day corrected, 

 if necessary, by interpolation." 



Dr. Robinson then shows that the equations, when treated by the 

 method of least squares, give, for the gaining rate of the clock at 

 the standard temperature and pressure, 0%2G3 ; for the gain by an 

 increase of l°in the temperature (his clock was over compensated), 

 0%G69 ; and for the gain by a depression of 1 inch in the barometer, 

 0%24]. As the mere effect of the air's buoyancy on the mercurial 

 pendulum is 0%15 for a change of 1 inch in the barometer, the re- 

 mainder is due to the inertia of the air carried along with the pen- 

 dulum in its oscillations. 



Dr. Robinson finds that the barometrical pressure has no sensible 

 effect upon the arc of vibration, and, consequently, does not affect 

 the clock through this cause. 



With respect to the barometric variation of rate, having once 

 determined its co-efficient, nothing is easier than to allow for it 

 when any sudden change occurs : it may, however, be corrected 

 mechanically with the utmost facility. A rise in the barometer 

 tends to make a clock go slower. Suppose a syphon barometer to 

 be attached to the pendulum, then a rise in the barometer will 

 transfer a portion of the mercury in the syphon to the tube, and 

 the effect will be the same as if a small weight were shifted from the 

 syphon end to the tube end of the barometer. Now if the syphon 

 be so adjusted to the pendulum, as to bore and situation, that this 

 alteration of the mercury in the barometer will accelerate the rate 

 as much as the increase of the air's buoyancy and inertia retard it, 

 the clock will be unaffected by any variation of the atmospheric 

 pressure. 



Dr. Robinson shows how the syphon barometers (he places two, 

 one on each side of the pendulum rod) may be constructed and 

 adjusted for any particular case. 



" I had previously recognised the influence of the barometer on 

 the clock before the mercurial pendulum was applied ; and, without 

 employing the method of minimum squares, satisfied myself that 

 the effect was even greater, being nearly 03 for an inch. In 

 examining this point for any given clock, it is however to be 

 remembered, that it is alwa)^s necessary to make the probable 

 errors of observation considerably less than the quantity sought; 

 otherwise, no conclusive result can be obtained. This may be done, 

 either by taking the rate for intervals of several days, or by observ- 

 ing several stars on each night." 



J List 



