Probably the most accurate determinations were made dur ing the war in connection with gun sound-ranging. Two sensi tive resonant hot-wire microphones, at a known distance apart, received the gun-sounds which were automatically recorded by means of an Einthoven string galvanometer, the time intervals being measured to 0•00i second. A large number of such obser vations, under different atmospheric conditions (wind, tempera ture and humidity), gave a value 337.16 m/sec. at le C. in dry air, and 337.6 m/sec. at Io°C. in air of average humidity, the effect of humidity, producing a slight change of density, being therefore very slight.
Velocity in Tubes or Small Volumes of Gas. Kundt's Dust Tube Method. When a gas is only obtainable in small quantities the method of velocity measurement is entirely changed. The best method available in these circumstances is one which is due to the work of Kundt. The gas is confined to a cylindrical tube containing a sprinkling of fine dry powder (lycopodium seed or cork dust) and is set into resonant vibration by any con venient means, e.g., by the longitudinal vibrations of a rod one end of which is inserted in the tube, or by the vibrations of the glass tube which contains the gas. The dust is heaped up in a repeated pattern indicating the nodes and loops in the vibrating gas column. Rayleigh and Ramsey used this method to deter mine the ratio of specific heats y of the rare gases argon and helium. The velocity c in a gas is given by where p is the pressure and p the density. The value of y found for the rare monotonic gases was 1-66. Kundt finally used a double-tube apparatus, in which stationary waves were produced simul taneously in two different gases by means of the same vibrating rod. The ratio of velocities in gas (I), rod, and gas (2), being equal to the ratio Xi : X, : X2, the values of Xi and X. are obtained from twice the nodal separation (indicated by the dust heaps), and the value of X, is equal to twice the length 1 of the rod when sound ing its fundamental longitudinal vibration. When the frequency N is determined (by monochord and tuning fork, or by a siren) the velocities c in the gas and the rod are known in absolute meas ure, for = and cr.(' = = 2N1. Alternatively the vibrations may be communicated to the gas by means of a steel diaphragm (a telephone earpiece) one end closed and excited electromagnetically, the frequency being varied until resonance is set up in the gas contained in the tube. This method has obvious advantages over the original " excitation. Using a short steel bar and exciting the vibrations by striking one end, Lang (Proc. Roy. Soc., Canada, 1922) has observed the nodes and antinodes in a Kundt's tube up to a frequency of 50,000 cycles/sec. The mean velocity of sound in the air contained in the tube (1.4 cms. dia.) was found to be metres/sec. at C., and in the steel bar 5120 m/sec. approximately. Kundt's tube method of measuring sound-velocities, and indirectly, the ratio of specific heats of gases, has been applied by Dixon, Partington and others to vapours and gases at various tempera tures, and as a means of measuring temperature coefficients of sound-velocity.
The value found by Colladon and Sturm (1826) in Lake Geneva was 1435 m./sec. at a temperature of C.
Velocity of Sound in the Sea.—The velocity of sound, particu larly explosion-waves, has been measured in the sea by a num ber of observers with considerable accuracy. Marti (Comptes. Rendus, Aug. 1919) used three under-water microphones (hydro phones) in a straight line in Cherbourg roadstead, the total base line being 1800 metres. The positions were accurately determined by theodolites. The passage of the sound-wave due to an explosion under water in line with the hydrophones, was regis tered by an electric chronograph with smoked paper and tuning fork time-trace. At a temperature of C. the velocity was found to be metres/sec. in sea water of density gram/cm.' More recently a careful series of observations has been made by Wood, Browne and Cochrane near Dover with four hydrophones covering a base-line of twelve miles in the sea (Roy. Soc. Proc., May 1923). Accurate temperature and salinity observations were made at points along the base line, and a new method (multiple charge method) devised to obviate errors arising through firing the charge at a point not quite in line with the base of hydrophones. The time intervals of pass age of the explosion wave between pairs of hydrophones was recorded on four strings of a six stringed Einthoven galvano meter (photographic), the ticks of an accurate chronometer on the fifth, and a wireless signal sent from a destroyer at the instant of firing the charge was recorded on the sixth string. Thus the record showed to an accuracy of ±.00I second the various time differences and the total time of travel of the explosion-wave to each receiver. The results obtained in summer and winter were 1510.4 m/sec. at and 1477.3 m/sec. at respectively, the salinity being 35%o in both cases. (The theoretical values calculated by D. J. Mathews [see Tables for Velocity of Sound in Fresh and Sea Office Publication H.D. 282 1927] are 1510.4 m/sec. and 1476.1m/sec. respectively.) The experimental results are expressed by the relation, at 35%o salinity (or= 4626+13.8 t 0+3.73 s where s is the salinity in %o at the temperature C). The ratio of specific heats (isothermal and adiabatic) for sea water, deduced from these measurements, is 1.0094 as compared with 1.0090 obtained from thermodynamic data. The adiabatic compress ibility at C and 35%o salinity is X 10" per dyne as compared with Ekmann's isothermal value 42.744 per dyne.