396 The Mther "Drift " and Rotary Polarization. 



the beams buck to the reference mark. Similar effects were 

 noted late in the afternoon. 



My next attempt was to obtain observations with light 

 which had traversed the liquid the full number of times, — ■ 

 six, which I had provided for. To do this, I took precautions 

 to exclude extraneous light, in order to obtain the full optical 

 sensibility, by immersing screens made of tubes, within which 

 the successive rays should pass. These adjustments were very 

 tedious, and before I succeeded in getting the light from the 

 last mirror into my field, the noon hour had passed. On 

 dismounting 1113' half-shade system, I found that the nicol A 

 was seriously impaired by the presence of a capillary film of 

 oil and the strip of spar had become detached. On removing 

 the latter, I found it broken. In attempts to remount the 

 remainder I destroyed it still more, beyond hope of further 

 use. These strips are very fragile, ground as they are to 

 0*1 mm. thick, so as to give as near a vanishing line as 

 possible. Mr. Halle of Steglitz-Berlin has, however, suc- 

 ceeded in making strips with square ground edges, in lengths 

 of 45 mm., as thin as this. As it was quite doubtful whether 

 I could obtain a proportionately greater sensibility with six 

 as with four passages, even if I should be fortunate enough 

 to secure again so fine a strip, and as it was probable that I 

 would have to renew all the optical pieces as well as redistill 

 the oil, I concluded to discontinue the experiment, although 

 I had not obtained the final sets of systematic observations as 

 originally planned. The likelihood of attaining a second 

 order sensibility in this test seems very remote indeed, although 

 ttase observations have carried the test probably a thousand 

 times beyond the first order and within a factor of ten of the 

 second order. 



Thus assuming white light and a rotation varying as the 

 inverse square of the wave-length, we may take 130° per 

 decimetre as the mean rotation of the oil. Considering the 

 total length to be 410 cm. x 4 = 1640 cm., we have a (com- 

 pensated) rotation of 21300°. As I should have been able to 

 detect a direct change of 1/120° on reversal, or one half this, 

 if the mean readings mentioned be allowed, the total effect 

 of the motion of the earth in space on the rotation of the 

 plane of polarization is certainly less than 1/5000000, and 

 probably less than 1/10000000. 



Physical Laboratory, University of Nebraska, 

 June 7, 1905. 



