of measuring Time at Sea. 45 
the same interval of time, the number of vibrations of the 
needle of a compass made and suspended on a pivot with 
the greatest: care, has found that it was not always constant. 
I have observed also, by means of an instrument, which 
I call a magnimeter, which by a long index marks on a limb 
the variations of magnetism, first, that this force in a body 
changes according as it is well or ill placed in the direction 
of the magnetic meridian, according as it is more or les 
elevated in the atmosphere, and according to the differen’ 
deerees of heat and cold. I have besides observed that thunder 
produces sensible variations in these forces, and that in the 
aurora borealis there happen also considerable changes, as 
they have observed in Sweden. I only mention these ex- 
periments, the detail of which would draw me too much 
aside from my subject: I hope some day to give an account 
of them to the Academy. 
There now remain only the bodies in vibration by the 
help of the elastic force. Every thing induces us to presume 
that these are the most proper to procure the required mea- 
sure of time. The regularity of certain watches which are 
executed daily, but, above all, the trials that have been made 
with the time-heepers of the celebrated Mr. Harrison, the 
recompense that he merits for them, confirm, in some de- 
gree, what was before only a presumption, and appear to 
demonstrate that the true, and perhaps the only, means of 
measuring time exactly at sea, consists, as we have before 
observed, in the perfected watch. But as watches in generat 
are very distant from the precision requisite in a marine 
watch, we should first search out their different irregulari« 
ties, and the causes, whether physical or mechanical, whence 
they are derived; according to the example of a wise phy- 
sician, who, before he has recourse to a remedy, endeavours 
to understand the disorder well, and what may occasion it. 
This is what I shall do in the following part: in order to 
render the different objects which are treated of more evi- 
dent, I shall make so many separate articles of them. 
As mechanics is here continually intermixed with natural 
philosophy, whatever is only suppotted by reasoning, how- 
ever solid it appears, will always be very uncertain. I have but 
too 
