SCIENCE. 



225 



refracted. Recent investigations of Henry, Tyndall and 

 others have indicated that when sound encounters a 

 change of density of medium, as when passing from clear 

 atmosphere into a wall of fog, there is a reflection of 

 sound. Altogether there seems no doubt but sound acts 

 like light in these respects, that is, on meeting a change 

 of refractive power, it is bath reflected and refracted, as 

 light is at the surface of water or of glass. The reflected 

 light is found to be always more or less polarized, per- 

 fectly so for the so-called polarizing angle. This polar- 

 izing angle of incidence is such that, as discovered by 

 Brewster for light, the reflected and refracted compo- 

 nents of a single ray, as thev strike away from the point 

 of incidence are at right angles. As reflected light is 

 polarized, reflected sound was supposed to be also. 



Fig. 1. 



Applying the laws of Fresnel and Brewster — 1st, that 

 the index of refraction is equal to the ratio of the veloc- 

 ities of the waves in the media ; and 2d, that complete 

 polarization is obtained for the particular case of right 

 angled reflected and refracted component rays, we are 

 guided to the proper conditions. We conclude that any 

 two substances having different velocities of propagation 

 of waves may be selected For instance, two gases, like 

 hydrogen and air, any two liquids, any two solids, a solid 

 and a gas, or, generally, any two media whatever. Con- 

 siderations of convenience would indicate air and illum- 

 inating gas, and these were chosen for the present pur- 

 pose. The" Velocities of propagation in air and coal gas 

 being as 1 125 and 1420, the index of refraction, accord- 

 ing to the first law above, is n = 1-26. The second law 

 gives for the polarizing angle of incidence, tangent i = 

 n =f26, or i = 51^°, the rays or waves being in the 

 gas. To realize this incidence upon a surface of separa- 

 tion between the gas and air, the cool gas was placed in 

 L-shaped tubes, AB, Fig. 1, hr>ving a portion cut away 

 at the angle, as shown at CD. The branches of the L 

 make equal angles of Si^ with the normal to CD. A 

 delicate membrane was gummed to the tube covering 

 the opening at CD, also shown at F, the object of which 

 was to retain the gas and maintain a polarizing surface, 

 CD. The arrow at A indicates a ray which is incident 

 at E, and is then in part refracted outward at E in a di- 

 rection perpendicular to the reflected component EB. 

 Each tube was about one inch in diameter and three 

 inches long. A number of these were made of tjn, each 

 with one end slightly larger than the other, so that they 

 could be joined up, stove-pipe fashion, to any desired 

 extent. Being cylindrical, the plane of one L-piece could 

 be placed at any angle with the plane of the preceding 

 one, according to the desired polarizing test. 



the effect of nine polarizing surfaces, like nine plates of 

 glass in light arranged at the polarizing angle. The nine 

 plates of glass can be used in two parts — one part, 4 for 

 instance, serving as a polarizer and the remaining 5 as 

 analyzer. The ends at A and B were capped with mem- 

 branes and the whole filled with illuminating gas. Thus 

 AC may serve as a polarizer and CB, or CE, or CD as 

 analyzer. When arranged as in ACB or ACE all con- 

 spire to the same effect of polarization ; but when ar- 

 ranged as in ACD, the plane of all the L-pieces in CD 

 being at right angles to that of those in AC, the effect of 

 one part antagonizes that due to the other, and to a 

 maximum degree as regards the angle. Partial effects 

 may be obtained with intermediate angles between o° and 

 90 . Again, we observe that the L-pieces o( Fig. 2 may 

 be alternately crossed, so that no two contiguous ones 

 will be parallel. It is believed that this arrangement 

 will give the greatest possible antagonistic effect ; also 

 while all Ls are in one plane it is not necessary that they 

 be arranged in a zig-zag line, like AC and CE, but may 

 be indiscriminately connected in that plane. The few 

 experiments made with the above-named arrangements 

 gave very marked results. Of course it need not be con- 

 fined to nine or any particular number of the L-pieces. 



Fig. 2. 



Fig. 2 shows the manner of joining the tubes, giving 



bxo. 3. 



It was found, however, wanting in convemence. The 

 apparatus finally adopted is tnat shown in Fig. 3. A dif- 

 ferent number of L-pieces were used at different times. 

 The portion AB is the polarizer and BC the analyzer. 

 The joint at B was kept tight with beeswax ; the ends at 

 A and C were capped square with the same membrane 

 material as were the angles of the Ls, giving, when 

 charged with illuminating gas, a continuous zig-zag 

 column from A to C. The L-pieces of the polarizer 

 enter half Ls at A and B, the latter having a common 

 axis and resting in bearings at A and B in the standards, 

 as shown. The object of this is to enable the experimen- 

 ter to turn the polarizer readily from cross to parallel, 

 etc. This convenient arrangement of the polarizer is 

 due to my assistant, Mr. Wright. Although applied to 

 the polarizer, it is evidently equally applicable to the 

 analyzer instead. The half L angles were not covered 

 with membranes, but left solid, with gradual inside cur- 

 vature. Membranes might have been applied here with 

 partial polarizing effect. The half L solid angles are 

 supposed to have detracted in a measure from the per- 

 centages of polarization obtained ; but this sacrifice is 

 more than compensated for by the greater convenience 

 and constancy of conditions obtained. If this arrange- 

 ment gives decisive results, of course, more perfect ap- 

 paratus would. The illuminating gas was admitted by 

 a nipple and rubber hose at C, the same flowing the 

 length of tubes and issuing in a small jet at H ; my as- 

 sistant kept this ignited, and used the flame length as a 

 pressure indicator, and it served admirably. 



The first trials were made by blowing an organ pipe 

 in front of the membrane A, to agitate the gas column. 



