226 



SCIENCE. 



A small mirror was attached to the membrane C, reflect- 

 ing a pencil of light upon a screen. The deportment of 

 the image indicated complex and inadmissible vibratory 

 movements of gas column, and besides quantitative in- 

 dication was found preferable to qualitative ; thereupon 

 the quantitative impulse and indicator pendulums were 

 adopted, as shown at E and F respectively, Fig. 3. An 

 ivory ball, % inch > n diameter, suspended by a thread of 

 8 inches length, was used at E, and so placed that when 

 at rest the ball would just touch the membrane at A. 

 The impulse was imparted by bringing the ball back 

 against the stop, shown by means of a spatula held in 

 the hand, and then allowing it to swing free against the 

 membrane, each time with a definite, predetermined arc. 

 So much of the impulse as reaches C knocks the pendulum 

 F through a certain arc, the same being measured on the 

 scale D. This pendulum was a small, hollow glass bead, 

 suspended by a silk fibre and trained delicately against 

 the membrane. The bob carried a pointer for the scale 

 D. 



In the experiments the ball would be dropped against 

 A some five or ten times, at intervals of about ten sec- 

 onds, the corresponding deflections at D being noted 

 and recorded ; then the polarizer would be turned 90 

 and like observations noted. Again, 90 would be 

 turned off, etc., etc.; occasionally the length of impulse 

 arc would be changed, or more or less L-pieces applied, 

 and in each case a large number of observations made. 



In the experiments the initial pulse seemed to be fol- 

 lowed by a series of vibrations in rapidly decreasing 

 amplitudes; but it is believed that the initial pulse is 

 equivalent to a genuine sound wave, or an undulation. 

 Evidence of soundness of this view is found in the fact 

 that the velocity of sound can be satisfactorily deter- 

 mined by similar pulses sent through tubes of 25 or 5c 

 feet length. It was evident that the initial pulse only, 

 was concerned in the first swing of pointer at D. 



Observations were made with as few as one L in the 

 polarizer, and two in the analyzer. But the results were 

 small, averaging about 4 per cent. But during the noting 

 of 455 individual results of observation the number ot 

 Ls in the polarizer ranged from 1 to 3 and in the analyzer 

 2 to 3. The average of all of these 455 observations 

 gave a percentage of 8.87 hundredths of quenching of 

 the polarized beam. That is to say, where the analyzer, 

 in light, entirely quenches the polarized beam in turning 

 through 90 , in the above 455 experiments only 8.87 

 hundreths of the polarized ray was quenched. 



But as a higher percentage was looked for, the instru- 

 ment itself was now examined for possible faults. The 

 membranes were all found under considerable tension, 

 whereas, of course, they should be perfectly free from it. 

 After completely slackening, then, as was supposed, the 

 experiments were continued with the following results : 



POLARIZATION CONTINUED. 

 Coal Gas in L-tubes and Air outside. 

 Individual results. 



Polarizer having 4 Ls and Amlyzer 5 Ls. 





'+ 





+ 





+ 





+ 





6.0 



6.0 



8.0 





65 



5-8 



6.2 



6.0 



6.0 



6.0 



6.0 



7-5 



5-5 



6-5 



5 5 



6.0 



5-2 



6.2 



6.2 



6.0 



7.0 



5-3 



6-3 



5-6 



6-5 



5-2 



7.2 



6.1 



5-8 



7-5 



5-2 



6-5 



5-2 



6.4 



51 



6.2 



6.0 



5.7 



7.0 



50 



6.2 



5-7 



63 



5-5 



6.1 



6.1 



5-5 



7.2 



S-a 



6.6 



6.0 



6-3 



5-i 



6.1 



6.2 



5-8 



6-5 



5° 



6.8 



52 



6.2 



5-2 



6.1 



6.0 





7.0 



4-9 



6.2 



5-2 



6.5 



5-4 



6.5 







7.0 



4.0 



&3 



5-3 



6-3 



5-8 



6.8 



6.07 









6-43 





6-3 



5-2 



6.4 



S83 



7.19 



5-oi 



5-50 



6.30 



5-37 



°-34 



In this series each value given in any column is the 

 number of divisions on the scale D, Fig. 3, of the deflec- 

 tion of the indicator pendulum bob. After obtaining one 

 column of results the polarizer, AB, was turned 90 , and 



the next column obtained. Thus the several columns 

 were obtained. The = and + signify that the polarizer 

 and analyzer are in parallel or perpendicular planes re- 

 spectively. 



On completing this series of observations air was 

 passed into the L tubes, completely displacing the coal 

 gas, so that the membranes were now suspended in mid 

 air. Other conditions remained the same. The indicator 

 pendulum now responded to the same impulses so 

 slightly as to be barely observable, but not measurable, 

 and they were apparently the same for the polarizer in 

 all positions ; that is to say, when air was upon both 

 sides of the reflecting surfaces there was almost no ap- 

 preciable reflection. This evidently should be the case, as 

 the light membrane itself is now the chief cause of reflec- 

 tion. This, compared with the results obtained with 

 gas in the tubes, shows that a considerable reflection is 

 due to surfaces of separation of gases differing in den- 

 sity. 



As regards the polarization, we observe that every 

 mean under " = " is larger than any mean under " + ," a 

 fact which cannot be assumed accidental, nor explained 

 on any other ground than polarization. To find the per- 

 centage of this in a single result, we have 



Means of the above series. 

 = + 

 6.07 5.83 

 7.19 5.01 

 °-43 5 5o 

 6 -3° 5-37 

 6 -34 



Means, 6 47 5.43 

 6.47 — 5.43 



Per cent. = = 16. 1. 



6.47 



Whole number of observations, 80. 



At this point the membranes were again examined and 

 found to have appreciable tension, though supposed to 

 have been entirely slackened at the beginning of the ser- 

 ies of observations just cited. It seemed that the damp- 

 ness in the gum used, expanded the membranes so that 

 they became tight after thoroughly drying as the experi- 

 ments went on. It was thereupon determined to slacken 

 them with the utmost possible care, and continue the ob- 

 servations The following table of individual results was 

 then obtained in the manner explained above, the mem- 

 branes being constantly watched for entire slackness. 



POLARIZATION CONTINUED. 

 Coal Gas in tubes and Air outside of tubes. 

 Polarizer having 4 Ls and Analyzer 5 Ls. 



Individual results. 





+ 





+ 





+ 





+ 





+ 





+ 





+ 



1.2 



09 



2.0 



I.O 





1.0 



1.6 



1.0 



1.8 





'•5 



0.9 



2.0 



1.0 





















1 2 



'•5 



1.0 



2.1 



1.2 







1 8 





T 8 





1 8 













































1.5 



1 . 1 



2.0 



1.0 



1-3 



1.0 



1.9 



1.0 



>-7 



1 . 1 



i-5 



1.1 



2.0 



1.0 





1.0 



2.0 



0.9 



1.2 



1.0 



1.8 



1.0 



%•! 



1.2 



i-5 



1.0 



2.0 



0.9 







2.Z 



0.9 



1-3 

 «-S 



1.0 

 1.0 



1.6 



1.0 

 1.1 

 1 . 1 

 1.1 



1.8 

 i-7 

 '•7 

 1.6 



1.0 

 1 .0 







2.1 



0.9 



1.44 



1. 00 



2.0O 



1 .01 



1.45 



1. 00 



1. 71 



1.07 



1.67 



1. 19 



1.50 



1. 00 



2.00 



1. 01 



Polarizer turned 90 immediately following each col- 

 umn of results. 



For gas displaced by air, other conditions the same ; 

 deflection, 0.00. 



The smallness of these results, compared with those of 

 the previous series, may be explained on the ground of 

 extreme and entire slackness of the membranes ; also the 

 slackness is still further evinced by the fact that, when air 

 displaced the gas no deflection of the indicator pendulum 

 was observable in response to the impulses. Tense mem- 

 branes would have turned the sound waves somewhat, 

 and in the manner of the rebound of a drumstick from its 

 drumhead. 



To obtain the percentage of polarization effect, we 

 have : 



