f 102 j 
XXII, On the Depression of Mercury in the Tule of a Ba- 
rometer, depending on its capillary Action. By M. La-~ 
PLACE*. 
Ii is necessary to determine the magnitude of this depression 
in order to render barometers comparable with each other. For 
this purpose we find, in the Philosophical Transactions for 1776, 
a table of corrections formed from experiment by Lord Charles 
Cavendish. At this period the theory of capillary action was 
unknown [?]: but this theory having been since discovered, 
and reduced to the fundamental principle of chemical affinities, 
that of a mutual action of the molecules of matter, decreasing 
with extreme rapidity, so as to become insensible at the smallest 
perceptible distances, it becomes proper to derive from this 
principle the table of the depression of mercury, and to borrow 
from observation only the necessary elements, as is usual in 
astronomy. In this manner we have the advantage of obtaining 
those elements with all possible precision, by comparing the 
whole of the phenomena depending on them with the results of 
theory; and we avoid the small irregularities which the errors 
of observations introduce into a table formed from: experiment 
only. In this case, the elements required are the angle which 
the surface of the mercury makes with the side of the tube with 
which it is in contact, and the depression of the mercury below 
the general level, in a very narrow tube of glass. The more or 
less perfect dryness of the tubes may affect these elements. We 
know that their interior surface is lined with a stratum of water 
which it is very difficult to remove. It is within this stratum 
that the mercury of the barometer rises and falls: its thickness 
is sufficient to render the action of the glass on the mercury in- 
sensible ; and the depression of this liquid in the barometer is 
determined by the reciprocal action of water and mercury. The 
boiling of the mercury in the tube diminishes more and more 
the thickness of the aqueous covering ; but it appears that in 
the best barometers it still remains sufficient to render the action 
of the glass insensible. We find in the excellent siphon baro- 
meter of the observatory, that the convexity of the drop which 
terminates the two columns of the liquid, is not sensibly different 
in the two branches; and I have concluded from the experi-. 
ments of M. Gay-Lussac, that this convexity is the same that 
takes place in a tube of glass completely moistened, provided 
that the water do not cover any part of the surface of the drop ; 
“ for if the whole surface is covered, I have shown that this sur- 
* From the Connoissance des Tems, 1812, p, 315. 
faee 
