ERRORS OF CHRONOMETERS. 
11 
Although a change of rate occurred with each 
alteration of position in most of the chronometers, 
it should he observed, that the same change of rate 
did not always recur with a repetition of an ex¬ 
periment in all respects apparently the same. In 
some cases, indeed, the changes were altogether 
anomalous. 
One plan of obviating this source of error in 
small chronometers, it occurred to me, might be 
accomplished, by preserving their uniformity of 
position, by means of a floating needle. For this 
purpose, I placed a chronometer in a light case of 
card paper, supported by a long pin or point, in a 
compass-bowl, on a little cross of thin brass, from 
the arms of which was suspended a perforated 
rhomboidal compass-needle. This needle, I found, 
though only of the ordinary magnitude, was fully 
capable of traversing with a weight of from a 
pound to a pound and a half, and with great fa¬ 
cility, when loaded with a full-sized pocket chro¬ 
nometer. It therefore had the property of keep¬ 
ing the chronometer invariably in the same posi¬ 
tion, and, being suspended on gimbles, of preser¬ 
ving it from the bad effects of the motion of the 
ship at sea. The magnetic needle was hung five 
or six inches below the chronometer, so that its 
influence on the instrument was not greater than 
that of the earth; and, being in an opposite direc- 
