3 5 2 Capt. Kater’s experiments for determining the variation 
The effect of this error on the daily rate of the clock, is 
lessened in proportion to the number of days comprised 
between the two transits ; for if the rate of the clock be 
deduced from transits observed on two successive days, the 
whole amount of the error arising from any deviation of the 
instrument from the meridian mark, will be included in the 
rate; but for any longer interval, it is divided by the number 
of days constituting such interval. 
In order therefore to obtain a true mean, it appears that 
each result should be multiplied by the product of the number 
of the stars into the interval between the observations, and 
the sum of such final products be divided by the sum of the 
factors. 
Observations of the sun are perhaps less entitled to credit 
than those of the stars, as in consequence of an apparent 
wavering of the meridian mark, some degree of uncertainty 
frequently exists in adjusting the transit instrument; setting 
this aside, a transit of both limbs of the sun may be considered 
equal to the transits of two stars. 
Proceeding in the computation in the manner just described, 
we obtain 86090,77 vibrations of the pendulum in 24 hours, 
by the observations of the stars, and 86090,79 by those of the 
sun. But from what has been said, these results are entitled 
to credit in the ratio of the sums of their factors, that is, as 
50 to 16 ; the final mean is therefore 86090,77 vibrations in 
a mean solar day. 
The force of gravity decreasing as the square of the dis- 
tance from the earth's centre increases, the next step is to 
find the correction on this supposition for the height of the 
station above the level of the sea. As the square of the 
