With regard to the smaller flat flames, which are the most general for ordinary lighting, the selection of glass globes is a very important matter. It may be said at once that all the old-fashioned style of glasses, with holes in the bottom about 2i in. diam., for fitting into the brass galleries of the older pattern pendants and brackets, are objectionable. The reasons for this condemnation are few and simple. It seems never to have occurred to the makers of these things that the gas flames inside the globes are always wider than the openings beneath them, through which the air required for combustion passes; and that, as a rule, the light of the flame is required to be cast downward. Gas flames always flicker in these old-fashioned glasses, because the sharp current of entering air blows them about. And the light cannot come downward because of the metal ring and its arms, and the glass, which is always thicker and generally dingier at this part of the globe. Perfectly plain and clean glass absorbs at least 1/10 of the light that passes through it; ground glass absorbs 1/3; and the ordinary opal obstructs at least 1/2, and generally more.

Only those globes should be chosen therefore which have a very large opening at the bottom, at least 4 in. wide, through which the air can pass without disturbing the flame. The glass then fulfils its proper duty, screening the flame from side draughts, and not causing mischief by a perpetual up-current of its own. Good opal or figured globes of this pattern may be used without disadvantage, because the light is reflected down through the bottom opening more brightly than if there were no globe, while the flame is shaded and the light diffused over other parts of the room.

The degree to which the luminosity of gas is utilized depends very largely upon the burner, people too often setting down as the fault of the gas, defects which should really be ascribed to the burner. In 1871, the Commission appointed by the Board of Trade to watch over the London gas supply, and whose prescriptions in these matters are more or less recognized by the whole country, made an examination of a collection of gas burners from a large number of sources, and including those in general use. The greater portion of these gave only 1/2, some even only 184, of the light that the gas was actually capable of affording. Two points very often neglected are: (1) that the size of the burner should be proportionate to the quantity of gas required to be consumed by it, and (2) that the gas should issue at a very low velocity. In good argands, the pressure at the point of ignition is almost nil; and in flat flame burners, the pressure should be only just sufficient to blow the flame out into the form of a fan.

It is very necessary that the body of the chamber below the point of ignition should be of material with low heat-conducting power, so that the gas may undergo no increase in volume which would occasion a proportionate increase of velocity, and that the heat may not be conducted away from the flame. To establish this. Evans had 2 argand burners made, differing only in that one had the combustion chamber of brass, and the other of steatite. The latter gave more light than the former in the proportion of 15 to 13 for the same quantity of gas. As another example a No. 8 metal flat flame burner, consuming 5 cub. ft. of gas per hour, gave a light equal to 11.5 candles, while a steatite burner of corresponding size, with non-conducting combustion chamber, gave. 14.6 candles. Another metal burner of a description somewhat generally used, gave about 3/8 of the light that the gas was capable of yielding. Worn-out metal burners generally give the best results, as the velocity of the issuing gas is lower than when the burners are new. A much better result is obtained by burning, say 20 cub. ft of gas from one burner, than by using 5 burners, each of which consumes 4 cub. ft.

This is the reason why the modern argands give so much more light than the older ones, which were drilled with a very large number of holes, and were more suitable for boiling water than for illuminating. If the air which is to support the combustion be heated, before it reaches the flame, especially in the case of flat flame burners, better results are produced, as was pointed out by Prof. Frankland more than 10 years ago, and this principle is now being carried out by some Continental burner makers. Of modern argands there are many excellent varieties, which can evolve 15-30 per cent. more light for the same quantity of gas than the best flat flame burners. One kind consisting of 3 concentric rings of flame with steatite gas chambers was first used in the public lighting of Waterloo Road in 1879. In another the products of combustion are brought down in a flue fastened round the burner, so as to heat the air which supports the combustion as it passes in pipes through the flue above-mentioned to the flame; while a third kind has an arrangement for admitting separate currents of cold air to keep the chimney cool. There seems little doubt that the argand lamp will play a leading part in the gas lighting of the future.

An important point connected with the use of gas is that the heat generated by combustion, may be made to do the work of ventilation, as in the fish-gill ventilator invented by the late Goldsworthy Gurney. In this strips of calico are nailed, by the 2 upper corners, across an opening in the wall, in such a way that each strip laps over the strip next below it. This contrivance, opening and closing like the gills of a fish, is self-acting, as the heated air passes away through the porous material, and cold air is admitted without draught.