explosion, or rapidity of transmission of flame, being in proportion to the fire-damp in the air, and the proportion of gases, other than light carburetted hydrogen, which exist in the former. It has hitherto been accepted that a mixture of air and fire-damp, to be explosive, must contain the latter in some proportion not lower than 7 parts by volume in 100 parts of air; this having been arrived at, in former days, by experiments with mixtures of gas contained in small narrow tubes; but the late Commission found, during the experiments upon a comparatively large scale carried on with safety-lamps, that flame could be transmitted by a mixture of air and marsh-gas containing little more than 4 per cent. of the latter. It was experimentally demonstrated that the facility with which flame travels, through a mixture of air with comparatively small propor tions of marsh-gas or pure fire-damp, is dependent upon the diameter of the channel containing it. Thus, a mixture of marsh-gas and air, containing 5.5 per cent. by volume of marsh-gas, could not be ignited in an ordinary eudiometrical tube by an electric spark, nor in a tube 1.75 inch in diameter by a lighted taper; but it was ignited, and burned round the flame applied, in a tube 3 inches in diameter. In this tube a mixture of air containing 6 per cent. of marsh-gas propagated the combustion readily; in some of the Commission's lamp-experiments at Llwynypia, in South Wales, a violent explosion was produced by the immersion of a Clanny lamp in a current of air containing about 6 per cent. of fire-damp, travelling through a channel, which had a section of about 13.5 inches by 10 inches; and there is no doubt that, in the more spacious passage of a mine, the power of mixtures of air with comparatively small volumes of fire-damp is still further developed. These observations of the Commission have received confirmation from the results of somewhat analogous experiments made in Germany by Professors Kreischer and Winkler.

The plan, hitherto pursued, for ascertaining whether the air in any part of a mine-way or working is contaminated to a dangerous extent with fire-damp, is to observe the effect of the air upon the carefully-trimmed and properly-adjusted

flame of a safety-lamp, by preference a Davy, Clanny or Mueseler lamp. The first effect of a small proportion of fire-damp in air is to produce an elongation of the lampfire, which increases as the proportion of fire-damp rises, until a very feebly luminous bluish flame, due to the com. bustion of the gas in the mixture, makes its appearance outside the elongated flame. This so-called cap, or aureole, extends as the fire-damp increases in amount (presenting appearances which differ with different kinds of lamps), until, when the atmosphere becomes actually explosive, the gas-mixture burns at the air-inlet of the lamp; or if, as is the case with the Davy lamp, the air enters from all sides, the gauze cylinder becomes filled with the bluish flame, and the lamp-flame is extinguished with more or less rapidity.

Mr. William Galloway, formerly one of H.M. Inspectors of Mines, was the first to carry out a series of experiments, with the object of estimating the proportion of fire-damp in the air of a mine by the behavior of a safety-lamp flame, employing definite mixtures of air with a natural fire-damp, almost entirely marsh-gas, which was supplied by a "blower" in Llwynypia Colliery. More extensive experiments with the same object have since been made by Kreischer and Winkler, who employed accurately-prepared mixtures of pure marsh-gas and air, and made use of five different kinds of safety-lamps, two of which burned rape oil, two benzine, and one a mixture of the two.

Galloway's results led him to the conclusion that a practised eye could just detect 2 per cent. by volume of pure fire-damp in air, "an exceedingly faint cap, 1⁄2 inch high," being observed by him on the most carefully-adjusted flame of a Clanny lamp when immersed in air containing that proportion of marsh-gas. The results of Kreischer and Winkler were confirmatory of the broad result of Galloway's experiments; but they found that the readiness with which firedamp could be detected, when present in air in small proportions, differed with different lamps and illuminating materials. An amount of 1 per cent. of gas was not at all indicated by a safety-lamp burning rape oil, while only the

faintest traces of a cap were exhibited when benzene, or a mixture of benzine and rape oil, was used. With 2 per cent. of marsh-gas, only an indistinct cap was detectable on the flames of lamps burning rape oil, while the flames furnished by benzine and by the mixed illuminants showed distinct caps. They found that the Davy and the Stephenson lamps furnished less marked results than the lamps (of the Clanny and Boty types) used by them, so that their conclusions as to the minimum amount of fire-damp which can be detected, by means of a safety-lamp burning the ordinary illuminant, agree very well with Galloway's earlier results. Other authorities on these subjects have also instituted experiments in this direction, especially with the view of ascertaining which form of safety-lamp is preferable as a fire-damp indicator; thus Mr. A. R. Sawyer has been led, by an extensive series of experiments with the Davy, Mueseler, and Marsaut lamps, to the conclusion that the second of these is distinctly superior to the others for testing purposes; and Messrs. Mallard and Le Chatelier have also been led by experiment to prefer, as a gas indicator, the Mueseler to the Davy lamp, especially when certain modifications are made in it, whereby the flame is rendered more sensitive.

The results of Galloway and of Kreischer and Winkler, and those of the Commission which relate to the minimum proportion of marsh-gas needed to impart to the atmos phere in a mine explosive characters, or the power of propagating flame, appear to leave no doubt that a contamination of the air by fire-damp to an extent little, if at all, more than double the minimum amount which can be detected by a fairly expert observer by the only means hitherto employed (i.e., by means of a Davy, Clanny, or Mueseler lamp), suffices to impart to it explosive properties. Moreover, as will be demonstrated later on, the existence of fire-damp in the air of a mine, to an extent below the minimum proportion which an intelligent observer of average experience can detect by means of the lamp-flame, is now known to become a source of great danger under conditions which are prevalent, though in different degrees, in a

very large number of coal-mines. It is therefore a matter of great importance that some more delicate indicator of fire-damp than the flame of the miner's lamp should be provided, which, while being thoroughly trustworthy, shall not be complicated in action nor difficult of employment underground, by a properly instructed person. Much ingenuity and scientific skill have been applied to the production of sensitive and reliable fire-damp indicators, and instruments and appliances of great variety have been elaborated, which, however, have only in a few instances afforded promise of successful practical application. The great importance of this subject must be the Author's excuse for examining, somewhat in detail, the most prominent results of work done in this direction.

One of the most recent suggestions, which has received considerable attention in Germany, and may lay claim to great simplicity, has been to employ alcohol in place of oil as the fuel, in a very tall form of Davy lamp, having an Argand wick, the air-supply to which passes up through a properly-protected vertical channel in the centre of the vessel containing the fuel. There is a short conical chimney round the flame, and the height of the latter is so regulated as not to be seen above the top of the chimney. The author of this mode of testing, Mr. Pieler, states that so little as 0-25 of fire-damp is indicated by the production of a cap 1.25 inch in length; in experiments made for the Commissioners in the return-air of a mine, where other experiments had afforded indications of the existence of about 1 per cent. of gas, the cap shown on the flame was 3 inches in height. This form of lamp is a most sensitive gas-indicator, but, at any rate in its present form, its use is attended with danger and uncertainty; it is at once extinguished by a very moderate current of air, and, if introduced into an explosive fire-damp mixture, it will bring about an explosion in a few seconds.

The late Mr. Ansell was one of the earliest and most persevering workers in this field, having patented ingenious and apparently promising fire-damp indicators, more than twenty years ago, which were based upon Graham's clas

sical researches on the diffusion of gases. If a vessel, partly composed of a porous body, such as gutskin, sheet india-rubber, or unglazed earthenware, and filled with atmospheric air in its normal condition, is surrounded by air containing an admixture of marsh-gas and at the same pressure as the contents of the vessel, the gas will diffuse through the porous body and pass into the vessel at a more rapid rate within a given time than the air will leave the vessel by diffusion; hence the pressure in the latter will be temporarily increased, and this increase will be proportionate, approximately, to the amount of fire-damp in the surrounding air, so that the percentage present may be inferred by comparing the observed maximum pressure with those pressures known to arise when given percentages of marsh-gas exist in the atmosphere. Conversely, if the contaminating gas in the surrounding air is carbonic acid instead of marsh-gas, the tendency will be to temporarily reduce the pressure in the vessel. To render the increase of pressure apparent Mr. Ansell adopted several ingenious devices, such as enclosing an aneroid barometer in a small case, one side of which was made of thin porous earthenware, and arranging the indicator of the barometer so as to show the percentage of marsh-gas present in the air (on the assumption that the results are not disturbed by the presence of any other gases). In another instrument the distension of a small india-rubber bulb by the pressure, temporarily exerted, serves to close an electric circuit, and thus an alarum is rung, when the pressure exerted corresponds to the percentage of fire-damp to which it has been adjusted. Or, the interior of a vessel, partly composed of porous earthenware, is connected with one limb of a mercury siphon guage, the other limb of which is in communication with the air of the mine. Another form of indicator, on precisely the same principle, was brought out a few years ago by Mr. Libin. Indicators of this class only operate when first, or suddenly, exposed to the air to be tested, being obviously no longer available as indicators of a given proportion when the air within the vessel has once become contaminated with marsh-gas. To make these