Air 32 dust segments Carbonic acid 40 ” Coal-gas 20 ” Hydrogen 9 ” Calling the velocity in air unity, the following fractions express the ratio of this velocity to those in the other gases: 32 Carbonic acid —— = 0·8 40 32 Coal-gas —— = 1·6 20 32 Hydrogen —— = 3·56 9 [Illustration: FIG. 110.] Referring to a table introduced in our first chapter, we learn that Dulong by a totally different mode of experiment found the velocity in carbonic acid to be 0.86, and in hydrogen 3·8 times the velocity in air. The results of Dulong were deduced from the sounds of organ-pipes filled with the various gases; but here, by a process of the utmost simplicity, we arrive at a close approximation to his results. Dusting the interior surfaces of our tubes, filling them with the proper gases, and sealing their ends, they may be preserved for an indefinite time. By properly shaking one of them at any moment, its inner surface becomes thinly coated with the dust; and afterward a single stroke of the wet cloth produces the division from which the velocity of sound in the gas may be immediately inferred. Savart found that a spiral nodal line is formed round a tube or rod when it vibrates longitudinally, and Seebeck showed that this line was produced, not by longitudinal, but by secondary transverse vibrations. Now this spiral nodal line tends to complicate the division of the dust in our present experiments. It is, therefore, desirable to operate in a manner which shall altogether avoid the formation of this line; M. Kundt has accomplished this, by exciting the longitudinal vibrations in one tube, and producing the division into ventral segments in another, into which the first fits like a piston. Before you is a tube of glass, Fig. 110, seven feet long, and two inches internal diameter. One end of this tube is filled by the movable stopper _b_. The other end, K K, is also stopped by a cork, through the centre of which passes the narrower tube A _a_, which is firmly clasped at its middle by the cork, K K. The end of the interior tube, A _a_, is also closed with a projecting stopper, _a_, almost sufficient to fill the larger tube, but still fitting into it so loosely that the friction of _a_ against the interior surface is too slight to interfere practically with its vibrations. The interior surface between _a_ and _b_ being lightly coated with the lycopodium dust, a wet cloth is passed briskly over A K; instantly the dust between _a_ and _b_ divides into a number of ventral segments. When the length of the column of air, _a b_, is equal to that of the glass tube, A _a_, the number of ventral segments is sixteen. When, as in the figure, _a b_ is only one-half the length of A _a_, the number of ventral segments is eight.