of measuring Time at Sea. ; 67 
af a fluid, a second cause again renders the accelerated 
semi-vibrations slower than the retarded; it is in the latter 
that the active principle, whatever it may be, has always 
its full effect ; whilst in the others it acts only with the ex- 
cess of velocity which it has on the body when it returns to 
its point of rest. 
Observation II. 
If in the vibrations of a body the difference of time em- 
ployed for each of the parts which we have distinguished in 
it is not sensible, the small foreign resistance which may take 
place will not sensibly alter the time of the whole vibia- 
tions; for the retardation which happens in the accelerated 
gemi-vibrations will then be compensated by the gain which 
they will produce in the tetarded semi-yibrations, and vice 
versa. 
Observation III. 
But when several caiises rerider the accelerated semi-vi- 
brations sensibly slower than the retarded, then tlie whole 
vibrations aie considerably retarded by the new resistances 
which take place; for, 1st, in the accelerated semi-vibra- 
tions, the body having less velocity than in the retarded of 
the same magnitude, the force which it has to overcome 
the new obstacles which are opposed to it is so much the 
less: @dly, We have seen that the resistances produce al- 
ways in a body in motion obstacles proportional to the time 
that it remains exposed to them; these obstacles are there- 
fore more considerable in the accelerated semi-yibrations 
than in the retarded; consequently, the retardation whicli 
they produce in the former is greater than the advancement 
which they occasion in the latter, and this retardation fol- 
lows the ratio of the square of the ‘difference of times em- 
ployed in each of the accelerated and retarded semi-vibra- 
tions. 
Corollary III. 
Hence we see, ist, That foreign resistances necessarily 
tend to destroy some little of the isochronism of the vi- 
brations of a body, and to render them slower at the same 
time that they diminish their extent; this is what the expe- 
E@ riment 
