but that which proved the most satisfactory was obtained by dropping a metal ball upon a pine board, the height through which the ball dropped being adjusted until the sound was sufficiently intense to be heard through the discs under investigation. An observer in the adjoining room, with the ear-piece, h, in his ear, could, by alternately closing and opening the rubber tubes, e and d, easily ascertain which of the two discs transmitted the loudest sound.

The rigidities of the discs were measured in the following way; an upright index was cemented to the disc under investigation, and a microscope, provided with a micrometer ocular, was focused on a mark on the index. The disc was then subjected to a pressure of air through the tube, c or d, and the displacement of the index read off in the microscope. The excess of pressure of the air on the inner surface of the disc over the atmospheric pressure was measured by a suitable manometer. From the data thus obtained the displacement of the center of the disc, for a pressure of one gram per square centimeter of surface, was calculated. The value of this displacement is of course a measure of the rigidity of the disc. The following are some of the results obtained with the apparatus just described.

I. Ä lead disc 10.5cm in diameter and 012cm thick, and a glass disc of the same dimensions, were clamped in the two unions respectively, and the intensities of the sounds transmitted through the two dises compared. It was found that the lead disc transmitted sound better than the glass one. The displacement of the center of the lead disc, for a pressure of one gram per square centimeter of surface, was 000106cm and of the glass disc 000053cm.

II. A disc of white pine 65cm thick was compared with a disc of leather of the same thickness. Both discs had been treated with paraffine to render them impervious to air. The displacement of the center of the pine dise was 000013cm and of the leather disc 000212cm for a pressure of one gram per square centimeter of surface. It was found that the leather disc transmitted sound very much better than the pine disc.

In both of the above cases the more rigid disc was found to be the poorer conductor of sound, although, in both cases it was composed of a material much better suited to the transmission of an elastic wave than the less rigid disc.

III. A brass disc 015cm thick was braced by soldering to it a few cross strips of brass. This disc was compared with one formed of two thicknesses of cardboard treated with paraffine. The total thickness of the cardboard disc was 44cm. The displacements of the two discs for a pressure of one gram per square centimeter of surface were found to be the same,

viz: 00008cm. It was also found that they transmitted sound equally well, although the cardboard disc was nearly thirty times as thick as the brass disc.

IV. A disc was built up of ten sheets of cardboard treated with paraffine. The total thickness was 70cm and the displacement of the center for a pressure of one gram per square centimeter of surface was 0002cm. This disc was compared with a single disc of cardboard 22cm thick and which gave about the same displacement. It was found that the two discs transmitted sound equally well, although one consisted of many layers while the other was of a single homogeneous material. "V. The braced brass disc used in experiment III, above, was compared with a disc cut from the same piece of brass, and which had a small mass of brass soldered to its center. The total mass of the two discs was thus made the same. The displacement of the center of the loaded brass disc, for a pressure of one gram per square centimeter of surface, was 0022cm while that of the braced disc was 00008cm. The loaded disc transmitted sound very much better than the braced disc. Even the noises from the street which entered the room could be easily heard through the loaded disc, while nothing of this kind could be heard through the braced disc.

An examination of the above cases, I to IV inclusive, makes it evident that the amount of sound transmitted through the discs as an elastic wave in the material, must be negligibly small as compared to that transmitted as a to and fro vibration. of the disc itself; were this not so, then in case IV, where the two discs possess the same rigidity, one would expect the sound heard through the single homogeneous disc to be louder than that heard through the built-up disc with its many reflecting surfaces. In other experiments, where compara tively rigid discs of materials well suited to the transmission of elastic waves, were compared with less rigid dises composed of materials not so well suited to the transmission of such waves, it was found in every case that the less rigid disc transmitted sound better than the more rigid one in spite of its unfavorable composition. The experiments, therefore, show that when sound is transmitted from the air on one side of the disc, through the disc, to the air on the opposite side, the transmission takes place almost entirely as a to and fro vibration of the disc.

The experiment described under V shows that the effect of mass in a wall is of minor importance as compared to rigidity. The lead disc, in I, had nearly six times the mass of the glass disc, but even this great increase in mass was more than compensated for by the fact that the lead disc gave a displace

ment twice as great as the displacement of the glass disc for the same pressure.

In order to investigate the effect of the mass of a wall on its conductivity for sound, two cardboard discs of the same dimensions were used. Both were treated with paraffine, and when clamped in the unions, both gave displacements of 0004cm for a pressure of one gram per square centimeter of surface. The intensity of the sound heard through the two discs was, as far as could be judged, the same. The discs were

22cm thick, and each weighed 17 grams. To the center of one of the discs was now cemented a mass of lead weighing 34 grams, and it was found that this cut down the intensity of the sound transmitted through the disc by a very appreciable amount. The effect due to the addition of a mass of five grams could be readily detected if the mass was cemented to the disc at its center, but not when the mass was cemented about half way between the center and circumference. This experiment shows that, other things being equal, the wall possessing the greatest mass will be the poorest conductor for sound. When the mass is uniformly distributed through the disc, however, a very slight increase in rigidity will more than compensate for a very considerable decrease in mass. For example, a lead disc, weighing 145 grams, was compared with a disc of red cedar, weighing only 17 grams, the lead disc gave a displacement of 00008cm and the red cedar disc a displacement of 00005cm for a pressure of one gram per square centimeter. It was found that the lead disc transmitted very perceptibly better than the red cedar disc, although it contained over nine times the mass.

It is a common practice in the construction of telephone booths to make them of two, and sometimes of four walls, separated by air spaces, and there seems to be an opinion that such a form of construction is better adapted for the exclusion of sound than one in which the same amount of material is put into a single wall. In order to test the relative merits of the two types of construction, six discs were cut out of cardboard and treated with paraffine in order to render them impervious to air. In one of the unions three of the discs were placed, and separated by cardboard washers, so that an air space of two millimeters was left between the discs. In the other union three discs were clamped in contact. It was found that the discs separated by air spaces transmitted sound better than the discs which were placed in contact. The same experiment was repeated using brass discs and with a similar result. The increased rigidity obtained by placing the discs in contact more than balanced any advantage there might be in having the intervening air spaces.

In all of the experiments above described the source of sound, as has been stated, was a single pulse obtained by dropping a metal ball on a pine board. Some of the experiments were repeated using an organ pipe as the source of sound. In this case it was found that the results might be much influenced by the pitch of the note used. If the natural period of vibration of a disc was in unison with the source of sound, while that of another less rigid disc was not, the transmission might be greater through the more rigid disc.

In conclusion, it may be stated that the experiments described above are representative of many others of a similar character. In every case the rigidity of the disc was found to be the main factor in determining the intensity of the sound transmitted from the air on one side of the disc to the air on the opposite side. The only other factor which seemed to have an appreciable influence on the transmission of sound through the disc was its mass. It was found that of two discs having the same rigidity the one possessing the greatest mass was the poorest conductor of sound. The effect of increasing the mass of a disc is, however, many times smaller than the effect of increasing its rigidity.

The above experiments show that the commonly accepted analogy between the transmission of sound and that of light does not hold where the sound is transmitted from the air on one side of a solid medium, through the medium, to the air on the other side. In such cases it has been found that an entirely different principle is involved, and that the transmission takes place as a to and fro vibration of the wall itself, and not as an elastic wave traveling through it.

Physical Laboratory of Columbia University,

New York City.

ART. XXXVIII.-A new Gauge for the Measurement of Small Pressures; by EDWARD W. MORLEY and CHARLES F. BRUSH.

IN 1888 and again in 1889, one of us constructed two gauges of a new form, intended for the direct measurement of small pressures. About three years ago, a third instrument of the same kind was used by us for the measurement of the pressure of the vapor of water; the time required for a measurement was rather long, and we accordingly constructed two more instruments of a somewhat different form, in which

the time needed for a reading is no more than that required by a filar micrometer. We shall now describe both forms; not only the somewhat costly apparatus which economizes time when many measurements are to be made, but also the simple and less expensive form which may well serve where but few measurements are required.

In figure 1, a b is a siphon gauge, consisting of tubes about five centimeters in diameter, connected below by a smaller