chlorine would be absorbed by the water, during its passage through it, and thus lead to variable results. As chlorine is absorbable by water, a slight absorption takes place during the continuance of the experiment. Before proceeding to any trials, it is therefore necessary to ascertain the amount of this, and then to deduct it from the condensation occasioned by the action on the coal-gas. In the tube which I have used, I found the absorption to be exactly 1 degree for every five minutes, and which continues in the same ratio, after the action of the chlorine on the hydro-carbon is over. I have, therefore, always deducted 1 degree for each five minutes, from the total loss, as indicated by the rise of the water in the tube. As, however, the action is over in five minutes, I have seldom continued the trial beyond that time, of course deducting 1 degree from the loss sustained. As chlorine and the condensible matter act on each other in equal volumes, a condensation of 10, when 50 of each are used, indicates ten per cent. of loss by the coal-gas.

Should this method of ascertaining the illuminating power of gases be ultimately found to be correct, another important result may follow its introduction into practice. If we can, by it, fix the illuminating power of one gas compared with that of another gas, the quality of which has been previously determined, and which is consumed with a burner that is known to burn it advantageously, and if the gas which we are subjecting to trial by the shadow test does not show such a high illuminating power as we are led to expect, from the known condensation by chlorine, the probability is, that the burners are not adapted for consuming the gas advantageously, and hence the necessity of altering the apertures, till the power by the shadow is what it ought to be, according to the chlorine test.

There is still another advantage attending the introduction of the chlorine test in addition to those mentioned; it is the facility of comparing different gases with one another, when they cannot be brought together, so as to try them by the shadow. The illuminating power may be considered just as the condensation by chlorine, and thus, then, we may state it numerically. Thus taking a coal-gas having only 1 per cent. of matter condensible by chlorine, its illuminating power may be considered as unity, and all others would be as the per centage of condensation. Hence, also, the illuminating power of gases may be ascertained as compared with other sources of light.

It is evident from what has been said, that, in finding the value of a gas as compared with other sources of light, attention must

be paid to the quality of the gas; a circumstance which by many has been totally disregarded, and hence the very discordant results which have been obtained. In comparing the gas by the shadow given by other lights, we must in fact not only attend to the different circumstances affecting the combustion; we must also at each trial ascertain the amount of condensation by chlorine; for the quality of a gas manufactured on different days, at the same place will be found to vary considerably. In the trials I am now to

state, made with the view of finding the comparative expense of light as got from candles, oil, &c., I have uniformly kept this in view; and by doing so, we are enabled to judge of the expense, not only in this town, but also in other places, provided, of course, we know the illuminating power of the gas by the chlorine test.

The first series of experiments were those with candles, of which ten different kinds were tried. Tallow single wick, tallow double wick, cocoa, palm, composite, margerine, diaphane, composition, spermaceti, wax,all short sixes.

Tallow.-Very different statements have been given of the illuminating power of coal-gas as compared with that from tallow candles, and which has been accounted for by the difficulty of getting the light from the candle to be uniform; the chief cause of the discordance is, however, more probably the difference in the quality of gases manufactured at different places. In conducting my trials, I have paid due attention to the former; trying the candles at different times, so as to have a wick of various lengths. The standard gas light, in all the trials, was a jet burning under a uniform pressure, with a flame of five inches, and consuming exactly one foot per hour.

From numerous trials, I found that the tallow (single wick, short-six), when compared with the gas, and taking the average of all the trials, was as 1 to 3.75. A shortsix will be found, when properly snuffed, to last for six hours, or very nearly so; and supposing candles to be 7 per pound, then the cost for each candle is 5 farthings. Suppose the gas to cost 8s. 4d. per 1000 feet,* then 6 feet will cost 2 farthings, or very nearly so; accordingly, for half the expense, 3.75 times the amount of light is obtained; in other words, for the same light, the expense of tallow candles is 7 times that of gas. The gas I employed in these trials contained, on an average, 12 per cent. of condensible matter. Should the gas contain

I have taken this as being easy for calculation. It is not far from the price of gas in Edinburgh, and in other towns in the neighbourhood of the coal districts.

more or less, then the comparative expense would be greater or less, just according to the quantity. In Edinburgh I have found the chlorine test to indicate from 11 to 14, and 15, very rarely is it up to the latter; of late, I have rarely found it to go beyond 13. Considering the foregoing calculation, as applying to the gas now supplied to Edinburgh, and presuming it to contain 12 per cent. of matter condensible by chlorine, then the expense of tallow candles is 7 times greater for the same light, than that of gas consumed by jet burners.

In England, where the gas is generally manufactured from English caking-coal, the illuminating power is inferior to that of gas got from parrot-coal, or from a mixture of it and common Scotch coal. Now, suppose the price of the gas the same, and that the condensation by chlorine amounts to 6, then the comparative expense of candles and of gas for the same light would be 3.75 to 1.

Similar trials were made with the other candles mentioned.

Double-wicked tallow, 1s. per pound.This candle burns for 5 hours, at a cost of 8 farthings; the light, compared to that of the jet is as 1 to 21, making the expense as 71 to 1. This candle has the advantage of not requiring to be snuffed.

Cocoa candle, 11d. per pound, will burn for nine hours, at a cost of 7.3 farthings; the light, compared to the jet, is as 1 to 3.6, or the same as that of the common tallow candle; thus, making the expense as 7.3 to 1. Palm candle, 1s. 2d. per pound, will burn for 6.6 hours, expense 9.3 farthings, light 1 to 3, expense as 10.5 to 1.

Composite, 1s. 1d. per pound, lasts for eight hours, expense 8.6 farthings, light 1 to 3, expense 8 to 1.

Diaphane (French), 1s. 8d., will last 6.6 hours, at a cost of 13.3 farthings, light 1 to 3, expense 15.1 to 1.

Margerine, nearly in every respect, the same as diaphane.

Spermaceti, 2s. 6d. per pound, will burn for eight hours, cost 20 farthings, light 1 to 2.6, expense as 16.2 to 1.

Composition candle the same.

Wax, 2s. 6d., burns nine hours, cost 20 farthings, light as 1 to 2·6, expense, therefore, as 14.4 to 1.

Thus the tallows, with the exception of the palm, are nearly of the same comparative expense, light for light; the composition is a very little more expensive, the others are more than double the expense.

In the foregoing calculations, I have supposed the gas to be consumed by jets; but I have already shown in the paper read before the Society, and published in its Transactions for 1840, that this is the least profit

able method of burning it. For equal consumpts, the light given by other burners is much greater; thus, taking the jet as 100, that from a fish-tail is 140, from the batwing 160, and from a properly constructed argand 180. Accordingly, by consuming the gas with these, the comparative expense will be still farther reduced. The following table gives the comparative light and expense, according to the kind of burner used.

In conducting experiments with the view of ascertaining the illuminating power of oil, compared with that of gas, I used argand oil-lamps of the common construction, and also others with contrivances adapted to them, which have been lately recommended for increasing the light. The first trials were made with sperm oil, the cost of which, at the time the trials were made, was 98. 8d. per gallon, that is, 18. 24d. per pint. It was burned in a common argand, consuming the oil under the most favourable circumstances. In endeavouring to fix the illumi nating power, I contrasted it with an argand gas burner, having forty-two holes, and consuming 3 feet per hour. I found, however, considerable difficulty in coming to accurate results, partly from the variation in the flame of the oil, partly also from the dif ference in the appearance of the shadow. Six trials were made at different times, and with the lights at different distances. These varied from 2 to 2:4, taking the oil as 1. The average of the different trials gave 2·35. A pint of oil was found to burn 14 hours, at a cost of 144d.; the consumpt of gas for the same time (3 x 14) was 42 feet, at an expense of 44d., but the light was as 2.25 to 1. The comparative expense, therefore, light for light, would be as 144d. × 2.25 to 4d.; that is, as 8 to 1, or very nearly so.

Rectified whale oil was next tried, the cost of which was 48, 8d. per gallon. A pint, when consumed under the most favourable circumstances, was found to burn 12 hours; and contrasted with the gas argand as before, the light was as 1 to 2-54. The cost of oil was 7d., that of gas for the same time was 3 d., but the light was as 1 to 2.54; the expense was, therefore, for the same light, as 7d. × 2.54 to 3d.; that is, 5 to 1.

In the preceding trials the oil was consumed in a common argand, due attention being paid to the different circumstances affecting the consumpt, such as the kind of wick, the height of flame, &c. The next trial was made with the lamp lately introduced under the name of solar lamp. In this a cylinder surrounds that containing the wick, with the upper part bent inwards, so that the aperture being contracted, the current of air that passes up between the one

In the above calculations, no allowance is made for outlay in gas-fittings, &c., &c.

cylinder and the other, striking against the horizontal part of the outer one, causes a contraction and lengthening of the flame; a longer and narrower glass chimney is at the same time required. The advantages said to attend the use of this construction of burner, are, that an oil of inferior quality may be used, while at the same time the light is greatly increased.

The solar lamp, containing solar oil, with a flame as high as could be got to be steady, and without smoke, was contrasted with the gas argand as before, burning three feet per hour. On comparing the lights, and taking the average of numerous trials, conducted at different distances, and when the wick was in different conditions, they were as 0.98 to 1; so very nearly equal that we may consider them as so. The oil, per gallon, costs 38. 8d.; a pint was found to burn eight hours, or very nearly so, at a cost of 5d. The gas required for the same time is 24, or say 25 feet, which would cost 24d.; accordingly, the expense is rather more than twice, o say twice, that of the gas.

To ascertain whether or not there is any saving by using the apparatus adapted to the solar lamp, the solar oil was consumed with a solar wick in the same argand with which the trials with the sperm and whale oils were made, and the light, as before, was contrasted with the argand, burning three feet per hour. The light and the consumpt of oil were found to be the same as with the other oils. The cost of the solar oil per pint is 5d., that of the whale oil 7d.; accordingly the expense is as the cost of the oils. It has been already stated, that by using the solar apparatus, the oil gave a light equal to that from an argand consuming three feet per hour, and that the pint of oil will last for eight hours; the expense is therefore as 2 d. to 5 d., or say ld. to 2d. Now when the solar oil was burned in the common argand, and contrasted with the gas argand, the light was as 1 to 2:54. As the oil lasted for twelve

hours, the cost of gas for that time would be 3 d., or very nearly so. The comparative expense was therefore as 5d. x 2.54 to 3d.; that is, as 3.98 to 1; whereas, by the solar lamp, it was only as 2 to 1; thus making a saving by the use of the solar lamp of nearly one-half of the expense. This peculiar construction of lamp is therefore a very great improvement; for not only is there a saving in expense in the outlay for oil, but for the lighting of large apart. ments a smaller number of lamps is required than when common argands are employed.

Naphtha. This article has been lately recommended as an economical source of light. Though naphtha gives a beautiful and steady light, yet it emits an offensive smell, and unless cautiously burned, is very liable to smoke; the slightest blast against the flame causing dense black smoke instantly to appear. The appearance of the shadow is so different from that from coal-gas, that it is not easy to fix their illuminating power and consequent comparative expense. In the experiments I have performed, I used the gas-argand as before, consuming 4 feet per hour. The naphtha-lamp had a wick of 4 inches in breadth, and burned with a flame of about half an inch in height. In one trial I made the illuminating power of the flames, as naphtha 1 to gas 4-233; in another, they were as 1 to 4:239;-giving an aver age of 1 to 4.236. The consumpt of naphtha was a pint in 24 hours, at a cost of 38. 6d. per gallon, that is 54d. per pint. The gas for the same time would be 24, or say 25 x 4 = 100-that is 10d.; but the light was as 4.236 to 1-therefore the comparative expense comes to be as 2.2 to 1, or very nearly so. Suppose that I have overrated the illuminating power of the gas as com pared with that of the naphtha, say, that instead of 4.236, it was about 4, this would reduce the cost of the latter, and thus make the comparative expense as about 2 to 1.