4 7 2 D O C . 4 7 8 E F F E C T O F E A R T H ’ S M O T I O N The weakest aspect of Miller’s experiments had already become apparent to physicists beforehand, namely, that the considerable size of his apparatus did not permit sufficient constancy in the temperature of the air through which the interfering light rays move [5] local systematic temperature differences of a few hundredths of a degree could simulate the observed positive effect. Kennedy as well as Piccard avoided this drawback by using substantially smaller apparatuses than Miller the required precision was attained by improvements in the optical equipment while employing special means to achieve constancy of temperature. Mr. Kennedy worked with a light path of barely 5 m. The optical paths were ar- ranged inside a sealed metal casing that was filled with helium at atmospheric pres- sure. The result of the experiments was definitely negative, at such an accuracy that the existence of an effect four times smaller than the one found by Miller could be ruled out. Whereas Dr. Kennedy’s experiments were conducted in the laboratory, Piccard and Stahel successfully carried out the bold plan of performing those extraordinarily subtle experiments in a free balloon. The major difficulty of using an apparatus of relatively low weight in a little space was offset by the advantage that the apparatus was rotatable by having the entire balloon set into slow rotation by means of two small ventilators. Most important, however, was that the experi- ment could be performed at various altitudes and that thus the postulated depen- dence on height could be tested. Unfortunately, the attainable temperature constancy in the apparatus was not large enough to eliminate with certainty the existence of a positive effect of the order of magnitude posited by Mr. Miller. But it could be shown that the observed effects did not increase with altitude, as would have been expected given Miller’s results. Without a doubt, it was meritorious of Prof. Miller that with his experiments he began a scrupulous reexamination of Michelson’s important experiment. However, his result must be seen as refuted by Kennedy’s and Piccard’s experiments.