388
PAUL S. EPSTEIN
Now this assumption is not quite correct even from the standpoint of the classical theory. As the nuclei are of electrical constitution, we must in the classical theory also take into account a definite relation between the mass and the velocity of the nuclei. Considering the nuclei as rigid spheres, for instance (Abraham), we should find
m = 1/(1 —ft2)
mtr l/(l-fft2) ■
If we use this formula, the torque will be compensated in part, but not completely as in relativity. It can easily be seen from the formula that 20 per cent of the calculated torque would manifest itself as deflection. The minimum velocity which could be observed by Chase would then be 4V5 km/sec., which brings us near to Miller’s value of 10 km/sec. Although interesting, these experiments cannot therefore decide either for or against Miller’s results. On this account it would be of great value if they could be carried out with increased precision.
Now some remarks about the experiment of Mr. A. Piccard at Brussels: Piccard thought that the height above the earth’s surface should be of influence on the effect Mr. Miller has found. (This is, in fact, a misunderstanding, because Mr. Miller does not claim any such effect.) If the ether drift may be supposed to be larger on Mount Wilson than at sea-level, it should be still larger in the free atmosphere. So Piccard tried the experiment in a balloon. His inter-
• _
ferometer had branches with an optical path 2.8 m long. The steady temperature was controlled by a thermostat. The balloon was rotated about a vertical axis by means of a propeller. A self-recording device was used, and ninety-six rotations were registered. The curves were analyzed harmonically, but it appeared that the thermostat had not functioned as expected. For this reason the accidental errors were too large (the probable error corresponded to a velocity of 7 km/sec.). All that Piccard claims, then, is that the drift in the free atmosphere at 2300 m altitude is not larger than on Mount Wilson. No further conclusions can be drawn from this experiment.
[Note added April, IÇ28.—Both C. T. Chase and A. Piccard have continued their work during the year intervening since the foregoing
CONFERENCE ON MICHELSON-MORLEY EXPERIMENT 389
report was presented. Chase,1 working at Harvard University, increased the accuracy of his measurements about three times. Even taking into account the factor 1/5 mentioned above, his new apparatus could have detected an ether-drift velocity of 3 km/sec. Within this accuracy his results were negative, thus giving strong support to the theory of relativity. The most accurate and recent work of Piccard’s was carried out by him, jointly with E. Stahel,2 on the summit of the Rigi in Switzerland (altitude, 1800 m). The same self-recording interferometer with thermostatic temperature control was used. The results were completely negative, the ether drift being only one-fortieth part of that expected according to Miller.]
Discussion
[Dr. Walter S. Adams, director of the Observatory, opened the discussion, expressing his hope that Professor Lorentz and Professor Michelson would give their opinions in regard to the considerations of Righi and Hedrick.]
Professor H. A. Lorentz: I feel somewhat guilty in regard to the work of Righi. It was a long time ago that I read his papers, and I do not remember their contents very well, as I have been busy with quite different things these last years. I should have read them again of course for this meeting. But this good intention could not be materialized because of my being entertained so much by the people of Pasadena. After having heard Mr. Hedrick’s report, I intend, however, to study these questions again very carefully in relation to Mr. Miller’s experiment. Further, the considerations of Brylinski must be taken up again. Offhand, I can only say that the results of Mr. Hedrick are in contradiction with those which I presented yesterday. Until today I felt myself quite satisfied with the considerations which are based on Fermat’s principle. After Mr. Hedrick’s report, however, I shall have to reconsider these questions carefully. According to Mr. Hedrick’s results it appears, indeed, that the result to be expected in the Michelson-Morley experiment may be numerically different from that which we ordinarily expect on the basis of the classical theory. The numerical value of the second-
1 Physical Review, 30, 516, 1927.
2 Die Naturwissenschaften, 16, 25, 1928.