7 
of measuring Time at Sea. 199 
of vibrations in a given time without increasing the opera- 
tions of the escapement also, and without the freedom of 
the balance suffering some diminution. 
Observation V. 
On the compensation for the effects of heat and cold. 
According to the Gazette du Commerce, and the report 
signed Ludiam, sent to the Academy, to remedy the irregu- 
larities produced in marine watches by heat and cold, Mr. 
Harrison uses a bar composed of two thin pieces of brass* and 
steel, two inches in length, riveted together in several places, 
Jixed at one end, and having at the other two pins across, 
through which passes the balance spring. If this bar remains 
straight in temperate heat (as brass receives more impression 
from heat than steel), the side where the brass is becomes 
convex by heat, and ihe steci side becomes so by cold. Thus 
the pins, one after another, fix the parts of the spring ac- 
cording to the different degrees of heat, and lengthen or 
shorten it; whence follows the compensation for the effects 
of heat and cold. . 
If I had known this ingenious method before I thought 
of my thermometers, probably I should not have hesitated 
to have made use of it in my machine. 
I considered some time whether I should not give it the 
preference. I even made some attempts with this view. I 
shall speak of them presently; but, after having thought of 
them seriously, and after haying put aside, as much as I 
could, that prejudice which we have in favour of our own 
productions, my thermometers appeared to me preferable. 
The following are the reasons which induced me to judge 
$0: 
The first, which would have prevented me from making 
use of it without some considerable change, was, that by 
Mr. Harrison’s method the regulating spring does not re- 
* The author makes use of the word cuivre, which is commonly used to 
express copper ; but in Mr. Harrison’s pamphlet it is lrass; we have there- 
fore given itso. From this it appears probable that the author in other places 
may mean braés, although that is commonly distinguished from copper by 
cuivre jaune—T. 8. E. 
N4 main 
