274 French National Institute. 
mometer. Thus this water, being a little warmer, is aise, 
heavier, 
: This experiment rests upon count Rumford’s own theory, 
regarding the manner in which heat is propagated in the Ji- 
quids. He thinks that the latter do not conduct it as all the 
other solid bodies do ; the metals, for instance; and that the. 
contact of a warm body only heats the mass of a liquid) in 
proportion as the heated molecules at first elevate themselves 
in virtue of the lightness they acquire, and allow the cold 
molecules to oceupy their place and be heated in their.turn. 
He has recently given us upon this doctrine an experi- 
ment more delicate and still more precise than all the pre- 
ceding. A portion of watery, heated to 80°, was only sepa- 
rated from a thermometer placed above it by a layer of cold 
water of some lines in thickness: not one of, the heated 
molecules was able to descend, and the thermometer did not 
rise one degree. 
The same chemist has made some experiments upon 3 
question in physics which, nearly concerns the doctrine 
of affinity, or rather the adherence of the molecules of a 
liquid with each other. The following is the manner in 
which he renders it palpable. He places oil upon water, 
and drops into the oil some very small grains of tin, or 
some very small drops of mercury : these small bodies pass 
through the oil quickly enough ; but when they come to the 
water, they stop on the surface of it, although much hea- 
vier than the water. The adherence of the water here forms 
something equivalent to a kind of pellicle which supports 
them ; but if we accumulate them, their mass acquires a 
weight which surmounts that adherence, they tear this pel- 
licle, and are precipitated. The appearance of a similar 
pellicle is also formed at the lower surface; for if we place 
water over mercury, and then drop some globules of the lat- 
ter into the water, they also stop at the bottom of the water 
without mixing with the rest of the mercury, until they have 
been enlarged and become heavy enough. M. Rumford 
adds to these experiments the striking remark, that without 
this adherence the least wind would carry off the water 
i ~ from 
