the furface of the ice, and prepared myfelf to ob-
ferve at my eafe the refult of this moft intereft-
ing Experiment.
A very few moments were fu (Helen t to {how me
that my expectation with regard to it would not be
difappointed. In the former Experiment a fimilar
cake of ice had been entirely melted in lefs than
three minutes ; but in this, after more than twice that
time had elapfed, the ice did not mow any apparent
figns of even beginning to melt. Its furface remained
fmooth and mining, and the water immediately
In contact with it appeared to be perfectly at reft,
though the internal motions of the hot water above
it, which was giving off its Heat to the fides of
the jar and to the air, were very rapid, as I could
diftinctly perceive by means of fome earthy particles
or
244 Of the Propagation of Heat
or other impurities which this water happened to
contain.
I examined the ice with a very good lens, but
it was a long time before I could perceive any iigns
of its melting. The edges of the cake remained
fharp, and the minute particles of duft, which by
degrees were precipitated by the hot water as it
grew colder, remained motionlefs as foon as they
touched the furface of the ice.
As the hot water had been brought from the
kitchen in a tea-kettle, it was not quite boiling hot
when it was poured into the jar. After it had
been in the jar one minute, I plunged a thermom
eter into it, and found its temperature to be at
180.
After 1 2 minutes had elapfed, its temperature at
the depth of one inch under the furface was 170.
At the depth of feven inches, or one inch above
the furface of the ice, it was at 1 69 ; while at
only -| of an inch lower, or above the furface of
the ice, its temperature was 40.
When 20 minutes had elapfed, the Heat in the
water at different depths was found to be as follows :
Immediately above the furface of the ice 40
At the diftance of an inch above it 46
At i inch - - 130
At 3 inches - 159
At 7 inches - - 160
When 35 minutes had elapfed, the Heat was as
follows :
At the furface of the ice 40
j-.an inch above it - - - 7& 9
i inch
in Fluids. 245
1 inch above it - * 1 10
2 inches ^ 1440
3 inches . 148
5 inches - - . 148^
7 inches * 149
At the end of one hour the Heat was as follows a
At the furface of the ice - 40
1 inch above it - - 80
2 inches - - - 1 1 8
3 inches - 128
4 inches * 130
7 inches * - - 131
After i hour and 15 minutes had elapfed, the
Heat was found to be as follows :
At the furface of the ice - 40
1 inch above it - - 82
2 inches * - 106
3 inches - - 123
The Heat of the water had hitherto been taken
near the fide of the jar ; in the two following
trials it was meafured in the middle or axis of
the jar.
When i hour and 30 minutes (reckoning al
ways from the time when the boiling-hot water
was poured into the jar) had elapfed, the Heat of
the water in the middle of the jar was found to be
as follows :
At the furface of the ice - 40
i inch above it - 84
Cinches - - - 115
3 inches - . n6
7 inches - * 117
VOL. II, I i When
246 Of the Propagation cf Heat
When 2 hours had elapfed, the Heat in the mid
dle of the jar was found to be as follows :
At the furface of the ice 40
1 inch above it - - - 76
2 inches - 94
3 inches - 106
4 inches - 108
6 inches -
7 inches - - . -
An end being now put to the Experiment, the
hot water was poured oft' from the ice, and on
weighing that which remained, it was found that
5 oz. 6 grains Troy (=12406 grains) of ice had
been melted.
Taking the mean temperature of the water at
the end of the Experiment at 106, it appears
that the mafs of hot water (which weighed 73^-
ounces) was cooled 78 degrees, or from the tem
perature of 184 to that of 1 06 during the Ex
periment. Now, as it is known that one ounce of
ice abforbs juft as much Heat in being changed to
water as one ounce of water lofes in being cooled
140 degrees, it is evident that one ounce of water
which is cooled 78 degrees, gives off as much Heat
as would be fufficient to melt T 7 -~ of an ounce of
ice ; confequently the 73^ ounces of hot water,
which in this experiment were cooled 78 degrees,
actually gave off as much Heat as would have been
fufficient to have melted !i2Sl? Ao-A- ounces of
i 4 o
ice.
But the quantity of ice actually melted was
only about 5 ounces $ and hence it appears that
left
in Fluids 247
Iff s than one-eighth part of the Heat loft by the water
was communicated to the ice ; the reft being carried
off by the air.
As the fame quantity of hot water was ufed in
this Experiment, and in that, No. 15, which im
mediately preceded it, and as this water was con
tained by the fame veffel, -(the glafs jar above
defcribed) it appears that ice melts more than
eighty times Jlower at the bottom of a mafs of boil
ing-hot water, than when it is fuffered to fwim on
its furface : For, as in the Experiment No. 15,
lo-g- oz. of ice were melted in 2 minutes and 58
feconds, 5 ounces at leaft muft have been melted
in i minute and 29 feconds ; but in the Experi
ment No. 1 6, 3 hours or 120 minutes were em
ployed in melting 5 ounces.
The ice however was melted, though very flowly,
at the bottom of the hot water, and that circum-
ftance alone would have been fufficient to have
overturned my hypothecs refpedling the manner
in which Heat is propagated in liquids ; had I not
found means to account in a fatisfactory manner
for that faft, without being obliged to abandon my
former opinions.
In about half an hour after the hot water had
been poured into the jar, in the laft Experiment,
examining the furface of the ice I difcovered an
appearance which fixed my attention and excited
all my curiofity ; I perceived that the ice had been
melted and diminifhed at its furface, excepting only
where it had been covered, or as it w&ejhadoived
by
248 Of the Propagation of Heat
by the flat flips of deal by which the cake of ice
was fattened down in its place.
Had the ice been protected and prevented from
being melted by that piece of the wood only, which,
being undermoft of the two, repofed immediately
pn the furface of the ice, I fliould not perhaps have
been much furprifed ; but that part of the furface of
the ice being likewife protected which was iituated
immediately under the other piece of wood, -that
which, lying acrofs the Under piece, and refting on
it, did not touch the ic$ any where except juft at its
edge, that circumftance attracted my attention, and
I could at firft fee no way of accounting for thefe
appearances but by fuppofing that the ice had been
melted by the calorific rays which had been emitted
by the hot water ; and that thofe parts of the ice
which had been Jhadowed by the pieces of deal
receiving none of thefe rays, had of courfe not
been melted.
I was fo much ftruck with thefe appearances*
that I immediately made the following Experi
ments, with a view merely to the elucidation of this
matten
Experiment, No. 17*
Into a cylindrical glafs jar, 6~ inches in diame*
ter, and 8 inches high, I put a circular cake of ice,
as long as could be made to enter the jar, and
about 3^ inches thick ; and on the flat and even
furface of the ice I placed a circular plate of the
thinneft tin I could procure, near 6~ inches in di-
ameter s
in Fluids, 249
ameter, or fufficiently large juft to cover the ice.
This plate of tin (which, to prefer ve its form, or
Jceep it quite flat, was ftrengthened by a ftrong
wire, which went round it at its circumference)
had a circular hole in its centre, juft two inches in
diameter, and it was firmly fixed down on the
upper furface of the cake of ice, by means of fever-
al thin wooden wedges which paffed between its
circumference and the fides of the jar.
A fecond circular plate of tin, with a circular
hole in its centre two inches in diameter, and in
all other refpeds exactly like that already defcribed,
was now placed over the firft, and parallel to it, at
the diftance of juft one inch, and like the firft was
firmly fixed in its place by wooden wedges.
Thefe perforated circular plates being fixed in.
their places, the jar was placed in a room where
Fahrenheit's thermometer ftood at 34; and ice-
cold water was poured into it till the water juft
covered tne upper plate ; and then the jar was
filled to within half an inch of its brim with boil
ing water ; and being covered over with a board,
was fufiered to remain quite two hours.
At the end of this time, the water, which was
ftill warm, was poured off, and the circular plate
being removed, the ice was examined.
A circular excavation, juft as large as the hole
in the tin plate which covered the ice, (namely two
inches in diameter) and correfponding with it,
perfectly well defined, and about -~- of an inch
deep in the centre, had been made in the ice.
This
250 Of the Propagation of Heat
This was what I expecled to find ; but there
was fomething more, which I did not expect, and
which, for fome time, I was quite at a lofs to
account for. Every part of the furface of the ice
which had been covered by the tin plate, appeared
to be perfect, level and fmooth, and fhowed no
figns of its having been melted or diminiflied, ex
cepting only in one place, where a channel, about
an inch wide, and a little more than T *- of an inch
deep, which fliowed evident marks of having been
formed by a ftream of warm water, led from the
excavation juft mentioned, in the centre of the
upper part of the cake of ice, to its circumference.
As the edge, or vertical fide, of the cake of ice
was evidently worn away where this ftream patted,
there could be no doubt with refpecl to its direc
tion. It certainly run out of the circular excava
tion in the middle of the ice ; and though it might
at firft appear difficult to explain the fact, and to
{how how this hot water could arrive at that place,
yet it was quite evident that the immediate caufe
of the motion of this ftream of water could be no
other than its fpecific gravity being greater than
that of the reft of the water at the fame depth :
and that this greater fpecific gravity was at the
fame time accompanied by a higher degree of Heat,
is evident from the deep channel which this ftream
had melted in the ice, while other parts of the
furface of the ice, at the fame level, were not
melted by the water which refted on it. To eluci
date this point, I made the following Experiment 5
Experiment^
m Fluid*. 251
Experiment^ No. 18*
Thinking it probable, that if the circular exca
vation in the ice, which anfwered to the circular
hole in the middle of the tin-plate which covered
the ice, and alfo to that in the fecond plate, which
was placed an inch higher, had been melted by
radiant Heat, (as it has improperly been called)
or by the calorific rays from the hot water ; that,
in that cafe, as forne of thefe rays xnuft probably
have been refle&ed downwards at the furface of
the water, in attempting to pafs upwards into the
air ; I thought that by preventing this part of them
from reaching the ice, which I endeavoured to do
by caufing them to be abforbed by a light black
body, (a circular piece of deal board, covered over
with black lilk) which I caufed to fwim on the
furface of the water, their effects in melting the
ice might perhaps be fenfibly diminifhed. Had
this really been the cafe, it would certainly have
afforded ftrong grounds to fufpecl that thefe rays
were in fact the caufe of the appearances in quef-
tion ; but on making the Experiment with the
greateft care, I could not perceive that the cover
ing of the furface of the hot water with a black
body produced any difference whatever in the
refult of the Experiment as it was firft made,
(Experiment No. 17) or when this black covering
was not ufed.
After fome meditation on the fubjecr, it oc
curred to me that this melting of the ice at its
upper
Of w Propagation of Heat
upper furface could be accounted for, in a manner
which appeared to me to be perfectly fatisfaftory ;
without fuppofing either that water is a conductor
of Heat, or that the effect in queftidn was produced
by calorific rays.
Though it is one of the mofl general laws of na
ture with which we are acquainted, that all bodies,
folids as well as fluids, are condenfed by cold ;
yet, in regard to water, there appears to be a very
remarkable exception to this law. Water, like
all other known bodies, is indeed condenfed by
cold at every degree of temperature which is con-
fiderably higher than that of freezing, but its
condemnation, on parting with Heat, does not go
On till it is changed to ice ; but when in cooling its
temperature has reached to 40 degrees of Fahren
heit's fcale, of eight degrees above freezing, it
ceafes to be farther condenfed ; and on being
cooled ftill farther, // aftually expands, and con
tinues to expand, as it goes on to lofe more of its
Heat, till at laft it freezes ; and at the moment
when it becomes folid, and even after it has be
come foiid, it expands flill more, on growing
colder. This fact, which is noticed by M. DE Luc,
in his excellent treatife on the modifications of the
atmofphere, has iince been farther inveftigated and
put beyond all doubt, by SIR CHARLES BLAGDEN*
See Philofophical Tranfactions, vol. Ixxviii.
Now, as water in contact with melting ice is
always at the temperature of 32% it is evident that
water at that temperature muft be fpecifically
lighter than water which is eight degrees warmer,
or
in Fluids. 253
or at the temperature of 40 ; confequently, if two
parcels of water at thefe two temperatures be
contained in the fame veffel, that which is the
coldeft and lighteft, muft neceffarily give place to
that which is warmer and heavier, and currents
of the warmer water will defcend in that which is
colder.
In the two laft Experiments, as the circular tin-
plate which covered the furface of the ice ferved
to confine the thin meet of water which was be
tween the plate arid the ice, as this water could
not rife upwards, being hindered by the plate,
and as it had no tendency to defcend, it is prob
able that it remained in its place ; and as it was
ice-cold., it was not capable of melting the ice on
which it repofed.
But as the tin*plate had a circular hole in its
centre, the furface of the ice In that part was of
courfe naked, and the ice-cold water in contact
with it being difplaced by the warmer and heavier
water from above, an excavation, in the form of a
fhallow baiin, was formed in the ice by this de-
fcending warm current.
The warm water contained in this bafin over
flowed its banks as foon as the bafin began to be
formed ; and iffuing out on that fide which hap
pened to be the loweft, opened itfelf a paffage
under the tin-plate to the edge of the ice, over
which it was precipitated, and fell down to the
bottom of the jar. The water of this rivulet being
warm, it foon formed for itfelf a deep channel in
the ice ; and at the end of the Experiment it was
VOL. II. k found
2 54 Of the Propagation of Heat
found to be every where deeper than the bottom of
the balin where it took its rife.
This manner of accounting for the appearances
in queftion, feemed to me to be quite fatisfactory ;
and the more I meditated on the fubject, the more
I was confirmed in my fufpicions that all liquids
mufl neceffarily be perfect non-conduttors of Heat.
On thefe principles I was now enabled to ac
count for the melting of the ice at the bottom of
the hot water in the Experiment No. 1 6 ; as alfo
for the flownefs with which that procefs went on ;
and encouraged by this fuccefs, I now proceeded
with confidence to plan and to execute ftill more de-
cifive Experiments ; from the refults of which, I
may venture to fay it, the important facts in quet
tion have been put beyond all poilibility of doubt.
If water be in fact a perfect non-condu&or of
Heat, that is to fay, if there be no communication
whatever of Heat between neighbouring particles
or molecuks of that fluid, (which is what I fuppofe)
then, as Heat cannot be propagated in it but only
in confequence of the motions occafioned in the
fluid by the changes in the fpecific gravity of thofe
particles which are occafioned by the changes of
their temperature, it follows that Heat cannot be
propagated downwards in water, as long as that
fluid continues to be condenfed with cold ; and
that it is only- in that direction? (downwards) that it
can be propagated after the water has arrived at that
temperature, where it begins to be expanded by
cold ; which has been found to be at about the
4oth degree of Fahrenheit's fcale,
Reafoning
in Fluids* 255
Reafoning on thefe principles., we are led to
this remarkable conclufion ; namely, that water
which is only eight degrees above the freezing point,
or at the temperature of 40 0? , miift be able to melt as
much ice in any given time, WHEN STANDING ON
ITS SURFACE, as an equal volume of water at any
higher temperature, EVEN THOUGH IT WERE BOIL
ING HOT.
My philofophica! reader will doubtlefs think
that I muft have had no fmall degree of confidence
in the opinion I had formed on this interefting
fubjecl, to have had the courage to make, even in
private, the Experiments which were neceiTary to
afcertain that fad.
Experiment, No. 19,
Into a cylindrical glafs jar, 4.7 inches in diame
ter, and 13.8 inches high, I put 43.87 cubic inches,
or lib. n -j oz. Troy, in weight, of water, and
placing the jar in a freezing mixture, compofed of
pounded ice and common fea-falt, I caufed the
water to freeze into one compact mafs ; which ad
hered firmly to the bottom and fides of the jar,
and which formed a cylinder of ice juft three inches
high.
Had the bottom of the jar been quite flat, inflead
of being raifed or vaulted, the cylinder of ice would
have been no more than 2.67 inches high.
As foon as the water in the jar was completely
frozen, the jar was removed from the freezing
mixture, and placed in a mixture of pounded ice
and pure water, where it was fuffered to remain
four
Of the Propagation of
four hours, in order that the cake of ice in the
jar might be brought to the temperature of 32.
The jar ftill Handing in a mallow difti in the
pounded ice and water, the furface of which cold
mixture was juft on a level with the furface of the
ice in the jar, I covered the top of the cake of ice
with a circular piece of ftrong paper, and poured
gently into the jar 73^ oz. Troy of boiling-hot
water, which filled to the height of eight inches
above the furface of the ice. (See Plate II.)
I then removed very gently the circular piece
of paper which covered the furface of the ice, and
after leaving the hot water in contact with the ice
a certain number of minutes, I poured it off, a.nd
weighing immediately the jar, and the unmelted
ice which remained in it,^ I afcertained the quantity
cf ice which had been melted by the hot water during
the time it had been fufFered to remain in the jar.
This Experiment I repeated four times the fame
day, (i6th March, 1,797) varying at each repe
tition of it the time the water was permitted to
remain on the ice. The refults of thcfe Experi
ments were as follows ;
Time the
Temperature
Temperature
Number of
the Exper
iment.
hot wa
ter re
mained
on the
of the hot
water when
it was pour
ed on the
of the water
i inch below
its furface at
theendofthe
Quantity
of ice
melted.
ice.
ice.
Experiment.
Minutes.
Grains.
No. 19
I
1 86
Not obferved
1632
No, 20
3i
185
Not obferved
1824
No. 21
15
184
170
1757
No. 22
60
186
140
2573_
From
in fluids* 257
From the refults of thefe Experiments., it was
plain- that a very coniiderable portion of the ice
which was melted, was melted in the very begin
ning of the Experiments, or \vhile the hot water
was adually pouring into the jar ; which operation.
commonly lailed about one minute ; and the ir
regularities in the refults of the Experiments, and
particularly of the three firft, fhewed evidently
that the quantity of ice melted in that opera
tion was different in different Experiments. I had
indeed forefeen that this would be the cafe ; and
o that account it was that I covered the furface
of the ice with a circular piece of ftrong paper,
and always took care to pour the water very gently
into the jar : but I found that all thefe precautions
\vere not fufficient to prevent very confiderable
anomalies in the refultsj of the Experiments ; and
as I found reafon to fufpecfc that the motion in the
mafs of the hot water, w r hich was unavoidably oc-
calioned by removing the circular piece of paper
which covered the ice, was the principal caufe of
thefe inaccuracies, I had recourfe to another, and
a better contrivance.
Having procured a flat, fliallow difh, of light
wood ; half an inch deep, 44 inches in diameter,
(or fomething lefs than the internal diameter of the
jar,) and about -i of an inch thick at its bottom,
I bored a great number of very fmall holes through
its bottom, which gave it the appearance of a
fieve. This perforated wooden diili having been
previoufly made ice-cold, was placed on the furface
of the ice in thejarj and the hot water being
gently
*5 8 Of *h* Propagation of Heal
gently poured into the dim, through a long wooden
tube ; as this perforated dilh floated and remained
conftantly at the furface of the water, and as the
water paffing through fuch a great number (many
hundreds) of fmall holes, was not projected down
wards with force, it is evident that by this fimple
contrivance thofe violent motions in the mafs of
water in the jar, which before took place when
the hot water was poured in the ice, and when the
paper which covered the ice was removed, were,
in a great meafure, prevented.
In order that the water which was poured
through the wooden tube (the bore of which was
about half an inch in diameter) might impinge
perpendicularly and with force againft the bottom
of the dilh, the lower end of the tube was clofed
by a fit cork Hopple, and the water was made to
iffue horizontally through a number of fmall holes
in the fides of this tube, at its lower end.
As foon as the operation of pouring the hot (
water into the jar was finilhed, the perforated dilh
was carefully removed, and the jar was covered
with a circular wooden cover, from the centre of
which a fmall mercurial thermometer was fuf-
pended.
The effecls produced by thisfnew arrangement of
the machinery will appear by comparing the refults
of the two following Experiments with thofe juft
mentioned*
in Fluids,
59
Number
of the Ex
periment.
Time the
hot water
remained
on the ice.
Temperature of the
hot water,
Quantity
of ice
melted.
At the be
ginning.
At the
end.
No. 23
No. 24
Minutes.
I
3
196
190
196
188
Grains.
4 2 3
703
In order ftill more effe&ually to prevent the in
accuracies arifing from the internal motions in the
mafs of hot water which were occafioned in pour
ing the water into the jar (and which could not
fail to affect, more or lefs, the refults of the Ex
periment) I had recourfe to the following con
trivance.
I filled a fmall phial containing 8^ cubic inches
with ice-cold water, and then emptying the phial
in the jar, I covered the furface of the ice with
this ice-cold water to the height of 0.478 of an
inch.
On the furface of this ice-cold water, inftead of
that of the ice, 1 now placed the perforated wooden
dilh previoufly made ice-cold, and poured the hot
water upon it.
The refults of the following Experiments fliow
that this contrivance tended much to diminifh the
apparent irregularities of the Experiments.
The air of the room in which thefe Experiments
were made was of the temperature of 41 ,
260
Of tie Propagation of Heat
No. of the
Experiment.
Time the
hot wa
ter was
on the ice.
Temperature of the
hot water one inch
below its furface,
Quantity
of ice
melted.
At the
beginning
At the
end.
No. 25
No. 26
No. 27
Minutes.
IO
3
180
192
190
190^
182
165
95
Grains.
580
914
3200
From the refults of thefe laft three Experiments,
we can now determine with a very conliderable
degree of certainty how much ice was melted in
the a5i of pouring the water into the jar^ and con-
fequently the rate at which it was melted in the
ordinary courfe of the Experiment ; fuppoiing
equal quantities to be melted in equal times.
As in the 27th Experiment 3200 grains were