ERRORS OF CHRONOMETERS. 9 
tion. Twelve or fourteen chronometers, some of 
London, but principally of Liverpool manufac¬ 
ture, were put under trial for nearly two months, 
and the position of eacli was changed about ouce 
a-week. In about one-half of the number, a sen¬ 
sible alteration of rate occurred with every altera¬ 
tion of position. I have been favoured with the 
rate of all these, taken by Mr Davies, with a 
transit instrument by Trougliton; or, in the 
absence of the sun, by the mean of two good 
“ regulators.” 
In a pocket chronometer by Allen and Caith¬ 
ness, the rate was very uniform in two posi¬ 
tions (namely, with the 12 o’clock mark to¬ 
wards the NE. and SW.) ; but, on shifting 
it from NE. to SE., a change of 1".5 took 
place in its rate. In another chronometer by 
Hatton, there appeared to be a difference of 
rate of about a second in two opposite positions. 
In a one-day chronometer by Litherland and 
Davies, there was scarcely any perceptible va¬ 
riation in three positions, namely, N., S. and E.; 
but, on the 12 o’clock mark being directed to¬ 
wards the W., a change of one second per day 
occurred. In another instrument by the same 
makers, between the NW. and SE. positions, 
there was a difference, by the first experiment, of 
2".3 per day, and by the second experiment, of 
