ftrongly attracted by thefe fubstances, is confined, and forms a barrier which not only prevents the cold winds from approaching the body of the animal, but which oppofes an almost infurmountable obftacle to the escape of the heat of the animal into the atmosphere. In the fame manner does the air in fnow ferve to preferve the heat of the earth in winter. The warmth of all kinds of artificial clothing may be fhown to depend on the fame caufe; and were this circumftance more generally known, and more attended to, very important improvements in the management of heat could, not fail to refult from it. A great part of our lives is spent in guarding onrfelyes against the extremes of heat and of cold, and in operations in which the ufe of fire is indifpenfable; and yet how little progrefs has been made in that most ufeful and most important of the arts, the management of heat!

Double, windows have been in ufe many years in moft of the northern parts of Europe; and their great utility, in rendering the houfes furnifhed with them warm and comfortable in winter, is univerfally acknowledged; but I have never heard that any body has thought of employing them in hot countries to keep their apartments cool in fummer yet how eafy and natural is this application of fo fimple and fo uteful an invention! If a double window can prevent the heat which is in a room from paffing out of it, one would imagine it could require no great effort of genius to difcover that it would be equally efficacious for preventing the heat without from coming in. But na

tural as this conclufion may appear, I believe it has never yet occurred to any body; at least, I am quite certain that I have never feen a double window either in Italy, or in any other hot country I have had occafion to visit.

But the utility of double windows and double walls, in hot as well as in cold countries, is a matter of fo much importance, that I fhall take occafion to treat it more fully in another place. In the mean time I hall only obferve here, that it is the confined air fhut up between the two windows (not the double glafs plates) that renders the paffage of heat through them fo difficult. Were it owing to the increased thickness of the glafs, then a fingle pane twice as thick would anfwer the fame purpofe; but the increased thickness of the glafs of which a window is formed, is not found to have the leaft fenfible effect in rendering a room

warmer.

But air is not only a non-conductor of heat, but its non-conducting power may be greatly increased. To be able to form a juft idea of the manner in which air may be rendered a worse conductor of heat, or, which is the fame thing, a better non-conductor of it than it is in its natural unconfined state, it will be neceffary to confider the manner in which heat paffes through air. Now it appears, from the refult of a number of experiments which I made with a view to the inveftigation of this fubject, and which are published in a paper read before the Royal Society*, that though the particles of air, each particle for itfelf, can receive heat from other bodies, or communicate it to them, yet

there

there is no communication of heat between one particle of air and an other particle of air. And from hence it follows, that though air may, and certainly does, carry off heat, and tranfport it from one place, or from one body to another, yet a mass of air in a quiefcent ftate, or with all its particles at reft, could it remain in this state, would be totally impervious to heat; or fuch a mafs of air would be a perfect nonconductor.

Now if heat paffes in a mafs of air merely in confequence of the motion it occafions in that air, if it is tranfported, not fuffered to pafs, -in that cafe it is clear that whatever can obftruct and impede the internal motion of the air, must tend to diminish its conducting power: and this I have found to be the cafe in fact. I found that a certain quantity of heat which was able to make its way through a wall, or rather a fheet of confined air, half an inch thick in nine minutes and threefifths, required twenty-one minutes and two-fifths to make its way through the fame wall, when the internal motion of this air was impeded by mixing with it one-fiftyfixth part of its bulk of eiderdown, of very fine fur, or of fine filk, as fpun by the worm.

But in mixing bodies with air, in order to impede its internal motion, and render it more fit for confining heat, fuch bodies only must be chofen as are themselves non-conductors of heat, otherwife they will do more harm than good, as I have found by experience. When, inftead of making ufe of eider-down, fur, or fine filk, for impeding the internal motion of the confined air, I ufed an equal volume of exceedingly fine filver-wire flatted (being

the ravellings of gold or filver lace) the paffage of the heat through the barrier, fo far from being impeded, was remarkably facilitated by this addition; the heat paffing through this compound of air and fine threads of metal much fooner than it would have made its way through the air alone.

Another circumftance to be attended to in the choice of a fubftance to be mixed with air, in order to form a covering or barrier for confining heat, is the fineness or fubtility of its parts; for the finer they are, the greater will be their furface in proportion to their folidity, and the more will they impede the motion of the particles of the air. 'Coarse horse-hair would be found to answer much worse for this purpofe than the fine fur of a beaver, though it is not probable that there is any effential difference in the chymical properties of those two kinds of hair.

But it is not only the fineness of the parts of a fubftance, and its be ing a non-conductor, which render it proper to be employed in the formation of covering to confine heat; there is ftill another property, more occult, which feems to have great influence in rendering fome fubftances better fitted for this ufe than others; and this is a certain attraction which fubfifts between certain bodies and air. The obftinacy with which air adheres to the fine fur of beafts and to the feathers of birds, is well known; and it may eafily be proved that this attraction must affift very powerfully in preventing the motion of the air concealed in the interstices of thofe fubftances, and confequently in impeding the paffage of heat through them.

Perhaps there may be another

*

ftill more hidden caufe which renders one fubftance better than onother for confining heat. I have fhown by a direct and unexceptionable experiment, that heat can pafs through the Torricellian vacuum though with rather more difficulty than in air (the conducting power of air being to that of a Torricellian vacuum as 1000 to 604, or as 6 to 10, very nearly); but if heat can pafs where there is no air, it muft in that cafe pafs by a medium more fubtil than air: a medium which moft probably pervades all folid bodies with the greatest facility, and which muft certainly pervade either the glafs or the mercury employed in making a Torricellian

vacuum.

Now, if there exifts a medium more fubtile than air, by which heat may be conducted, is it not poffible that there may exift a certain affinity between that medium and fenfible bodies? a certain attraction or cohesion by means of which bodies in general, or fome kinds of bodies in particular, may, fome how or other, impede this medium in its operations in conducting or tranfporting heat from one place to another?-It appeared from the refult of feveral of my experiments, of which I have given an account in detail in my paper before mentioned, published in the year 1786, in the 76th volume of the Philofophical Transactions, that the conducting power of a Torricellian vacuum is to that of air as 604 to 1000:- but I found by a fubfequent experiment (fee my fecond paper on heat, published in the Philofophical Tranfaétions for the year

1792) that fifty-five parts in bulk of air, with one part of fine raw filk, formed a covering for confining heat, the conducting power of which was to that of air as 576 to 1284; or as 448 to 1000. Now, from the

fult of this laft mentioned experiment, it fhould feem that the introduction into the fpace through which the heat pafted, of so small å quantity of raw filk as one-fiftyfixth part of the volume, or capa. city of that space (which now contained fifty-five parts, of air and one part of filk) more impervious to heat than even a Torricellian vacum. The filk must therefore not only have completely deftroyed the conducting power of the air, but muit alfo at the fame time have very fenfibly impaired that of the etherial fluid which probably occupies the interftices of air, and which ferves to conduct heat through a Torricellian vacuum: for a Torricellian vacuum was a better conductor of heat than this medium, in the proportion of 604 to 448. I forbear to enlarge upon this fubject, being fenfible of the danger of reafoning upon the properties of a fluid whofe exiftence is even doubtful; and feeling that our knowledge of the nature of heat, and of the manner in which it is communicated from one body to another, is much too imperfect and obfcure to enable us to pursue thefe fpeculations with any profpect of fuccefs or advantage.

But

Whatever may be the manner in which heat is communicated from one body to another, I think it has been fufficiently proved that it paffes with great difficulty through

confined