398 PROCEEDINGS OP THE AMERICAN ACADEMY 



to the smallest jars, and it was difficult to revolve it without pro- 

 ducing distortion of the fringes, and an effect amounting to only ^V of 

 the distance between the fringes might easily be hidden by accidental 

 errors of experiment. In the new experiment the apparatus was 

 placed on a massive rock, which rested on a wooden base, which 

 floated upon mercury. The stone was 1.5 meters square and 0.3 of a 

 meter thick. At each corner four mirrors were placed, by reflection 

 from which the length of path traversed by the light was increased 

 to ten times its former value. The width of the fringes of interfer- 

 ence, which were the subject of observation, measured from 40 to 

 60 divisions of the observing micrometer. The light came from an 

 Argand burner sent through a lens. To prevent jars from stopping 

 and starting, the float was kept constantly in slow circulation, revolv- 

 ing once in six minutes. Sixteen equidistant marks were made on 

 the stationary framework within which the float moved. Observa- 

 tions were taken on the fringes whenever any one of these marks 

 came in the ranee of the micrometer. The observations were made 

 near noon, and at 6 p. m. The noon and evening observations were 

 plotted on separate curves. One division of the micrometer measured 

 5o of a wave-length. Mr. Michelson was confident that there was 

 no displacement of the fringes exceeding T ^ of a wave-length. It 

 should have been from twenty to forty times greater than this. Mr. 

 Michelson concludes that this result is in opposition to Fresnel's 

 theory of aberration. 



As late as 1872, Leverrier thought that a new measurement of the 

 velocity of light by Fizeau very important in the interest of astron- 

 omy; and in 1871 Cornu wrote that the parallax of the sun, and hence 

 the size of the earth's orbit, were not yet known with the desirable 

 precision. In 1875, Villarceau made a communication to the Paris 

 Academy on the theory of aberration. He says that the parallax of 

 the sun by astronomical measurement is 8". 86. Foucault's velocity 

 of light combined with Struve's aberration makes the sun's parallax 

 8".86. Cornu's velocity of light gives the same result only when it 

 is combined with Bradley's aberration, which differs from that of 

 Struve by O'^O. Villarceau thinks that there is an uncertainty 

 about the value of aberration on account of the motion of the solar 

 system. In 1883, M. O. Struve discussed seven series of observa- 

 tions made by his father, Nyren, and others, with various instruments 

 and by different methods, at the Observatory of Pulkowa. He was 

 certain that the mean result for the value of aberration was 20". 492, 

 with a probable error of less than T Jj of a second. This aberration, 



