I 
?17 
" But this is not all. There is a ftill rnorer extraordi¬ 
nary, and in its confequences more wonderful, circum- 
ftance which remains to be noticed. When water is 
cooled to within eight or nine degrees of the freezing 
point, it not only ceafes to be farther condenfed, but is 
actually expanded by farther diminutions of its heat ; and 
this expanfion goes on as the heat is diminifhed, as long 
as the water can be kept fluid ; and when it is changed to 
ice, it expands even ftill more, and the ice floats on the 
Surface of the uncongealed part of the fluid.” 
Count .Rum-ford next (hows how very powerfully this 
wonderful contrivance tends to retard the cooling of wa¬ 
ter when it is expofed in a cold atmofphere. 
“It is well known that there is no communication of 
heat between two bodies as long as they are both at the 
fame temperature; and it is likewile known that the ten¬ 
dency of heat to pafs from a hot body into one which is 
colder, with which it is in contact, is greater, as the dif¬ 
ference is greater in the temperatures'of the two bodies. 
“ Suppofe now'that a mais of very cold air repofes on 
the quiet furface of a large lake of frefli water, at the tem¬ 
perature of 55° of Fahrenheit’s thermometer. The par¬ 
ticles of water at the furface on giving off" a part of their 
heat to the cold air with which they are in contart, and 
in confequence of this lofs of heat becoming fpecifically 
heavier than thofe hotter particles on which they repofe, 
mull of courfe defcend. This defcent of the particles 
which have been cooled neceflarily forces other hotter 
particles to the furface, and thefe being cooled in their 
turns bend their courfe downwards ; and the whole maf's 
of water is put into motion, and continues in motion as 
long as the procefs of cooling goes on. 
“ As ft. on as the water in cooling has arrived at the 
temperature of about 40°, as at that temperature it ceafes 
to be farther condenfed, its internal motion ceafes, and 
thofe of its particles which happen to be at its furface re¬ 
main there ; and, after being cooled down to the freezing 
point, they give off their latent heat, and ice begins to be 
formed. 
“As foon as the furface of the water is covered with 
ice, the communication of heat from the water to the at¬ 
mofphere is rendered extremely flow and difflcult; for 
ice, being a bad conduflor of heat, forms a very warm cover¬ 
ing to the water; and moreover, it prevents the water 
from being agitated by the wind. Farther, as the tem¬ 
perature of the ice at its lower furface is always very nearly 
the fame as that of the particles of liquid water with 
which it is in contact, (the warmer particles of this fluid, 
in confequence of their greater lpecific gravity, taking 
their places below,) the communication of heat between 
the water and the ice is neceflarily very flow on that ac¬ 
count. 
“As foon as the upper furface of the ice is covered 
with fnow, (which commonly happens foon after the ice 
is formed,) this is an additional and very powerful ob- 
ftncle to prevent the efcape of the heat out of the water; 
and, though the inoft intenfe cold may reign in the at¬ 
mofphere, the increafe of the thicknefs of the ice will be 
very flow.” 
Mr. Dalton, in his New Syflem of Chemical Philofo- 
phy, Hates the principal facts of congelation as follows: 
“ ill. The fpecific gravity of ice is lei's than than of wa¬ 
ter, in the ratio of 92 to 100. ad. When water is ex¬ 
pofed in a la'fge fufpended jar to cool in ftill air of 20° or 
30 0 , it may be cooled i° or 3 0 below freezing ; but, if 
any tremulous motion takes place, there appear inftantly 
a multitude of lhinmg hexangular fpicuke floating, and 
flowly attending in the water. 3d. It is obferved, that 
the Ihoots or ramifications of ice at the commencement, 
and in the early ftage of congelation, are always at an an¬ 
gle of 6o° or 1-2-0°. 4th. Heat is given out during con¬ 
gelation, as much as would raife the temperature of water 
140° or 1 50°. The fame quantity is again taken in when 
the ice is melted. This quantity may be.^th of the 
whole heat-which water of .32° contains. 5th. Water is 
• Vol. X. No. 7.06. 
C E. 
detifeft at 36° of the old icale, or 38° of the new; from 
that point it gradually expands by cooling or by heating 
alike, according to the law fo often mentioned, that of 
the fquare of the temperature. 6th. If water be expofed 
to the air and to agitation, it cannot be cooled below 32° ; 
the application of Cold freezes a part of the water, and 
the mixture of ice and water acquires the temperature of 
32°. 7th. If the water be kept Itill, and the cold be not 
levere, it may be cooled in large quantities to 25° or be¬ 
low, without freezing; if the water be confined in the 
bulb of the thermometer, it is very difficult to freeze it 
by any cold mixture above 13° of the old fcale, (Fahren¬ 
heit’s ;) but it is equally difficult to cool the water much 
below that temperature without its freezing. I have obi 
tamed it as low as 7 0 or 8°, and gradually heated it again 
without any part of it being frozen. 8th. In the laft cafe 
of what may be called forced cooling, the law of expan¬ 
fion is ftill obferved as given above. 9th. When water is 
cooled to 15 0 or below in a bulb, it retains the moft per¬ 
fect tranfparency; but, if it accidentally freeze, the conge¬ 
lation is inftantaneous, the bulb becoming in a moment 
opake, and white like fnow, and the water is projected up 
the Item. 10th. When water is cooled below freezing, and 
congelation fuddenly takes place, the temperature riles 
inftantly to 32°,” 
He then proceeds as follows: “In order to explain 
thefe phenomena, let it be conceived that the ultimate or' 
final-left dements of water are all globular, and exaCtly of, 
the fame fize; let the arrangement of thefe atoms be in 
fquares, fo that each particle touches four others in the 
fame horizontal plane. Conceive a feCond ftratum of 
particles placed upon thefe in like order of fquares, but’ 
fo that each globule falls into the concavity of four others 
on the firft ftratum, and confequently refts upon four 
points, elevated 45° above the centres of the globules. 
A perpendicular feCtion of fuch globule, refting upon' 
two diagonal globules of the fquare. Conceive a third 
ftratum placed in like manner upon the fecond, &c. the' 
whole being fimilar to a fquare pile of ftiot. The above 
conftitution is conceived to reprefent that of water at the 
temperature of greateft denfity. 
“ Now the angle of a rhombus is 6o°, and its fupple- 
ment 120°; if any particular angles are manifefted in the 
aCt of congelation, therefore, we ought to expert thefe, 
agreeable to the 2d and 3d phenomena. Whenever any 
remarkable change in the internal conftitution of any body 
takes place, whether by the accefiion and junrtion of new 
particles, or by new arrangements of thofe already exift- 
ing in it,' fome modifications in the atmol'pheres of heat 
muft evidently be required ; though it may be difficult to 
eftimate the quantity, and fometimes even the kind of 
change fo produced, as in the prefent cafe. So far there¬ 
fore the theory propofed agrees with the 4th phenomenon. 
“ In order to explain the other phenomena, it will be 
requilite to confider more particularly the mode by which 
bodies are expanded by heat. Is the expanfion occafion- 
ed Amply by the enlargement of the individual atmof- 
pheres of the component particles ? This is the cafe with 
elaftic fluids, and perhaps with folids, but certainly not 
with liquids. How is it poffible that water Ihould be ex-’ 
panded a portion reprefented by 1 upon the addition of 
a certain quantity of heat at one temperature, and by 340 
upon the addition of a like quantity at another tempera¬ 
ture, when both temperatures are remote from the abfo- 
lute zero, the one perhaps 6ooo° and the other 6170? 
The fart -cannot be accounted for on any other fuppofi- 
tion than that of a change of arrangement in the compo¬ 
nent particles ; and a gradual change from the fquare to, 
the rhomboidal arrangement, is-in all probability efterted 
both by the addition and abftrartion of heat. It is to be 
fuppoled then, that water of the greateft poffible denfity 
has its particles arranged in the fquare form; but, if a 
given quantity of heat be added to, or taken from it, the 
particles commence their approach to the rhomboidal 
form, and confequently the whole is expanded, and that 
8 U the- 
