8 2 D O C U M E N T 7 4 S E P T E M B E R 1 9 2 5 74. From Auguste Piccard Brussels, 14 Avenue Ernestine, 21 September 1925 Dear Professor, Permit me to submit the following deliberation to you. Miller’s experimental result seems doubtful to me for various reasons. If one does not find any Michelson effect at ground level, only a very small effect at best ought to be detectable on Mount Wilson, since the relative shortage of earth around the peak differs only slightly from the suboptimal conditions at ground level. In my view, in order to have a definitely error-free result, the following condi- tions should be fulfilled: 1) The laboratory must be as light a construction as possible, and it must be able to ascend some thousands of meters above the earth within a short time. 2) It must be possible for the whole laboratory to be set in quite rapid rotation to prevent any source of error that could influence one side of the apparatus by radi- ation or some other lopsided effect. 3) The entire laboratory must be installed perfectly free of tremors. 4) The observation must be made by means of a photographic plate. I believe these conditions can be met at relatively low cost. The apparatus is mounted in a balloon. A small, peripherally suspended electro- motor with a propeller sets the whole balloon into steady rotation (about 1 rotation every 100 seconds). Observation takes place at night when there is no one-sided solar radiation. The interference fringes as well as the position of a magnetic needle are recorded on film. In order to avoid trouble obtaining the interferences, mono- chromatic mercury arc light is used. The entire effect comes to about 1/50 to 1/25 of a fringe this is, of course, not much. If, after the trip, one takes the positions of about 10 lines from one second to the next with the dividing machine, and the ex- periment has lasted through about 100 balloon rotations, it ought to be easily pos- sible to have an experimental series that would yield the necessary precision, by recording the mean of all the readings as a function of the azimuth. For the Earth’s proper motion to be horizontal, the experiment must be conducted around midnight. It must be carried out during a particular season of the year in order for the solar system’s motion of some 20 km/sec to be added to the Earth’s motion around the Sun. If I am not mistaken, it should therefore be during spring. As far as the balloon is concerned, I think no complications will arise. The accumulator battery that feeds the Hg lamp and drives the motor constitutes the main weight of the instrument. As far as I can estimate the weight, a 2,200 m3 bal- loon would be serviceable, such as the one owned by the Swiss Aero Club, filled with coal gas and operated by one man, who would have to serve simultaneously as physicist and pilot.