A close observer will notice that the flame from a match, a candle, or a gas-jet, is not uniformly bright. The reader may verify this by lighting a match and observing the flame. There is always a bluish or darker portion near the bottom. In this less luminous part the air is combining with the hydrogen of the hydrocarbon which is being vaporized and disintegrated. Even the flame of a candle or of a burning splinter is a miniature gas-plant, for the solid or liquid hydrocarbons are vaporized before being burned. Owing to the incoming colder air at this point, the flame is not hot enough for complete combustion. The unburned carbon particles rise in its draft and become heated to incandescence, thus accounting for the brighter portion. In cases of complete combustion they are eventually oxidized into carbon dioxide before they are able to escape. If a piece of metal be held in the flame, it immediately becomes covered with soot or carbon, because it has reduced the temperature below the point at which the chemical reaction--the uniting of carbon with oxygen--will continue. An ordinary flat gas-flame of the "bats-wing" type may vary in temperature in its central portion from 300 deg.F. at the bottom to about 3000 deg.F. at the top. The central portion lies between two hotter layers in which the vertical variation is not so great. The brightness of the upper portion is due to incandescent carbon formed in the lower part. When scientists learned by exploring flames that brightness was due to the radiation of light by incandescent solid matter, the way was open for many experiments. In the early days of gas-lighting investigations were made to determine the relation of illuminating value to the chemical constitution of the gas. The results combined with a knowledge of the necessity for solid carbon in the flame led to improvements in the gas for lighting purposes. Gas rich in hydrocarbons which in turn are rich in carbon is high in illuminating value. Heating-effect depends upon heat-units, so the rating of gas in calories or other heat-units per cubic foot is wholly satisfactory only for gas used for heating. The chemical constitution is a better indicator of illuminating value.