166 



JOURNAL OP HOETICDLTURE AND COTTAGE GABDENEB. 



( Fobniorj 19, 1871. 



there is a suction going on tlirougla the whole length, which 

 will draw in the air through any fissure ; whereas, if the exit is 

 contracted and the throat of the furnace large, there is a 

 forcing of ail- into the flue, which seeks its way out wherever it 

 can find an opening. Attention to this principle would cure a 

 large proportion of smoky chimneys. 



But although water neither allows gas to escape nor can 

 become heated above 212°, there is also in its use necessity for 

 ample extent of pipe. The temperature even of 212° is too hot, 

 and causes a certain deterioration of the air. Probably 1.50° is 

 the utmost consistent with salubrity. Further, if the pipes 

 become heated nearly to then- full extent, there is trouble with 

 escape of steam, diminution of water, incrustation of the 

 boiler, and finally a considerable chance that fuel may be 

 wasted, because more is burned than the water can cool. 

 But, subject to attention to this rule, water presents great ad- 

 vantage for the transmission of heat on account of its high 

 specific heat — i.e., the large amount which it absorbs for each 

 degree of rise of sensible heat, and gives off again in cooUng 

 for each degree of loss of sensible heat. Weight for weight, it 

 thus absorbs and gives off fully four times as much heat as 

 air, and bulk for bulk nearly three thousand times. Hence a 

 pipe containing hot water wiU give off three thousand times as 

 much heat as the same pipe containing hot ah', or, in other 

 words, win convey the heat three thousand times as far with 

 the same loss. Brickwork being a much worse conductor of 

 heat than iron (in about the proportion of 1 to 3), and a 

 brick-on-bed being thirty or forty times the thickness of an 

 iron pipe, while a flue is generally four to eight times the area 

 of a pipe, the difference is partly compensated ; but it may 

 still be reckoned that a flue wUl only convey its heat to one- 

 fourth or one-sixth the distance that a hot-water pipe will. 



Before deaUng with the actual amount of fuel which is re- 

 (juired for heating various structures, it is necessary to deter- 

 mine what proportion of the total heat evolved may be fairly 

 considered to be available in practice. The tables on p. 115 show 

 the amount produced by each fuel, and, as already observed, 

 the waste is but trifling when the fuel is burned within the 

 structure to be heated, or when the heat produced is abstracted 

 through a sufficient prolongation of flue. 'When the fire is in 

 a stokehole, it would not be difficiilt in general to make the 

 ventilation of the stokehole pass through close stone pipes led 

 within the house to be heated, and in this way a large amount 

 of waste might be economised. If, then, care is taken to save 

 as much of the heat as is possible, we may reckon that a de- 

 duction of ten per cent, ought to be ample to cover the total 

 waste. The amount requisite to produce a draught in the 

 chimney is too trifling to require notice. I have found by 

 experiment that a temperature of -10° above the au" in a chimney 

 15 feet high occasions a sufficiently strong draught, and a 

 temperature of 70" above the air a fierce draught. Now, for 

 the combustion of each pound of coal about 20 lbs. of aii', 

 equal to 270 cubic feet, aie requisite, allowing, as is found to 

 be the case, that only half the oxygen in the air combines with 

 the fuel. The heat of these 270 feet on passing through the fire 

 would be about 2500° Fahr., and therefore, if the temperature 

 iif the chimney is 60° above the external air, the allowance of 

 heat necessary to ensure a very powerful draught is under one- 

 fortieth part of the total heat generated by the fuel. I pro- 

 pose, therefore, in the following calculations to assume that 

 coal ought to give 12,000 units of available heat to the structure 

 which is to be warmed. And it is, of course, understood that 

 12,000 lbs. of water raised 1° is exactly the same thing as 

 6000 lbs. raised 2°, or 1000 lbs. raised 12°, and so on. Now, 

 12,000 lbs. of water are equal to nearly 192 cubic feet ; but the 

 heat absorbed iu raising water 1° is sufficient to raise 3000 

 times (very nearly) the Ukc bulk of air 1°, consequently the heat 

 yielded by 1 lb. of coal wUl raise 192 x 3000, or 576,000 cubic 

 feet of air 1°. So much for the quantity of ah' that can be 

 heated by 1 lb. of coal. 



Now for the cooling. Let it be kept in mind that air would 

 remain stationary in temperature unless the heat is allowed to 

 pass off, and it wiU be seen that its absolute loss of heat does 

 Dot depend on its bulk, but upon the nature of the media by 

 which it is surrounded : hence the common method of esti- 

 mating heating power by so many feet of pipe required for so 

 many cubic feet of contents of the house is only a rough-and- 

 ready way of approximating to the extent of cooling surface 

 by which the air is smTounded ; but it is more accurate to 

 take the cooUng surfaces lUrectly. The experiments of Hood 

 have shown that each square foot of glass cools about 1.^ cubic 

 foot of air as many degrees per minute as the external tempe- 



rature exceeds the internal. If this excess is V, then 1 square 

 foot of glass cools 1| cubic foot of air 1° per minute ; or -160,800 

 square feet of glass will cool 576,000 cubic feet of air per minute, 

 which, as we have seen, is the quantity of air that 1 lb. of coal 

 will heat V. Therefore 1 lb. of coal burned per minute will 

 supply the heat lost through 460,800 square feet of glass ; or 

 dividing by 60, 1 lb. of coal burned per hour will supply the 

 loss of heat occasioned by 7513 (say, for convenience, 7500) 

 square feet of glass iu the same time, when the external tem- 

 perature is 1° below the internal. For any other difference of 

 temperatures we must divide 7500 by the number of degrees 

 of difference to obtain the surface of glass which 1 lb. of coal 

 per hour will heat. For any different superficies of glass we 

 must divide 7500 by it to obtain the degrees of heat above the 

 external air which 1 lb. of coal will maintain ; and for any 

 additional consumption of coal per hour we have to multiply 

 either the degrees of difference of temperature, or the superficies 

 of glass, found as above, by the number of pounds of coal to be 

 burned per hour to get either the temperature that wiU be 

 maintained in a given structure, or the extent of structure 

 that will be maintained at a given temperature. 



We may express these rules algebraically in the following 

 equations. Let C be the number of pounds of coal to be 

 burned per hour ; D, the difference of external and internal 

 temperature in degrees ; and S, the superficies of glass in 

 square feet. Then 



,, , _ DS _ 7500C imc 



('■■) '•' = 7500 > ■ '' = D . ^ = — g- ' 



From which, when any two of the quantities C, D, and S are 

 fixed, the third can be found. Or, in words : To find the coal 

 requu-ed to keep a given structure at a given difference of tem- 

 perature, multiply the superficies of glass in feet by the pro- 

 posed difference, and divide by 7500. 



To find the superficies of glass which will be maintained at 

 a given difference of temperature by a given quantity of coal, 

 multiply the pounds of coal per hour by 7500, and divide by 

 the intended difference of temperature. 



To find the difference of temperature that wUl be maintained 

 in a given structure by a given quantity of coal, multiply the 

 pounds of coal per hour by 7500, and divide by the superficies 

 of glass in feet. 



These rules are quite independent of the manner in which 

 the heat is conveyed or applied in the structure, whether by 

 Arnott stove, flue, or hot-water pipes, provided only the 

 arrangement be such as to yield twelve pai'ts of the heat of 

 coal, which, when burned, gives thirteen to sixteen parts. If 

 the coal be of very inferior heating power it may be proper to 

 use 7000 instead of 7500 for a constant. The same rules also 

 apply to other species of fuel, substituting a different constant 

 in place of 7500. This constant may be obtained by multiply- 

 ing 7500 by the imits of heat belonging to the fuel in question 

 (less ten per cent, for waste), and dividing the product by 12,000. 



In computing the superficies of glass the woodwork may be in- 

 cluded as a sort of set-off against the extra loss of heat tlirough 

 the laps of the glass. The loss through the back and front 

 waUs of lean-to houses must also be considered, but it is 

 reckoned by Dr. Arnott at only one-twentieth of that of glass ; 

 but, on the other hand, it is important to remember that the 

 effect produced by coal in the foregoing calculation is that for 

 extreme low temperatures in sunless weather, and in compari- 

 son with the temperature of an unheated greenhouse. The 

 night temperatures will always be by fire heat in excess of what 

 has been shown, because as, even in cloudy weather, the day 

 temperature of the air is almost always 10° to 20" higher than 

 the night, there is bottled-up in any glass structure, in the 

 soil, and walls, an amount of heat in the day which is slowly 

 given off at night, and which to that extent diminishes the 

 amount of coal required at night to maintain a given tempera- 

 ture. Gardeners are familiar with the saving of coal effected 

 by early closing in sunny days, and a like saving accrues in a 

 less marked degree even on sunless days. The formulfe given 

 above must therefore be taken as the extremes that can be re- 

 quired by firing alone, in protracted periods of sunless weather, 

 of uniform temperature. In such conditions I have found 

 them hold good in practice. 



The next point is to ascertain how many feet of (say 4-iuch) 

 pipe are required in order to convey and transmit, through the 

 water to the surrounding air, the amount of heat produced by 

 1 lb. of coal per hour. Now, it may bo deduced from Hood's 

 researches on the rate of cooling of iron pipes that 1 foot in 

 length of 4-inch pipe, which contains 150 cubic inches of water, 

 weighing 5-425 lbs., loses -GSV per minute, or 40-86° per hour. 



