With a small wrench he removed one of the old burners, and the several kinds were successively tested in its place. Some gave a better light, but it was objected that they might consume more gas. Whereupon the chemist tore a strip from his well-worn handkerchief, and, having damped it, wound the ribbon several times around the top of the old burner (which had been replaced), leaving the orifice uncovered. The new burner was screwed down over this, making a gas-tight connection. "There," said he, "we have a gauge. The new burner will receive the same amount of gas that the old one consumed--no more, no less--but the current is slightly checked." The burner gave the same amount of light as before, so far as the eye could perceive. "In the combustion of gas for heating purposes," continued the chemist, "seek the burner with free, rapid delivery through small holes. For light you want something different. Suppose you send a current of gas up into this sewing-thimble: it can find an exit only by turning backward. Then suppose it escapes from the thimble only to enter a larger cavity above it, whence it must issue through a burner-tip with an orifice of the usual size. The current, you perceive, is twice completely broken. It will be seen that only the expansive force of the gas, together with its buoyancy, acts upon the jets, instead of a direct current. Now, it will always be found that the burner which best carries out the principles just illustrated--other points being equal--will give more light with a less quantity of gas than any other. This also exhibits the chief principle of most of the governors or regulators.