686 



NATURE 



[July 28, 192 1 



that the fringes can be well observed even under 

 conditions of poor atmospheric definition, vvhen 

 the use of a filar micrometer would be impossible. 

 The method has the further advantage that as 

 the distance apart of the slits is varied, the 

 separation which causes the fringes to vanish 

 can be very precisely determined, so that the 

 error of observation is greatly reduced. With 

 small, faint objects, on the other hand, the loss 

 of light arising from the use of narrow slits is 

 serious. At the Paris Observatory an attempt 

 was made to determine by this method the angular 

 diameter of the major satellites of Jupiter, but the 

 light was not sufficient to render the fringes 

 visible. Hamy, therefore, extended the theory 

 to the case in which the slits are of a width which 

 is comparable with their distance apart. If the 

 slits are rectangular, of width a and distance 

 between their centres Z, the formula obtained by 

 Michelson for the distance corresponding to the 

 vanishing of the fringes must be replaced by 



/=r22 X/«{i+o765(a//)«}. 



Michelson and Hamy used the method for the 

 measurement of the angular diameters of the 

 major satellites of Jupiter. Michelson, in i8gi, 

 observing with the 12-in. equatorial at the Lick 

 Observatory, used adjustable narrow slits. Hamy, 

 in 1899, used the large equatorial coudd of the 

 Paris Observatory, and prepared a series of 

 screens of such dimensions that their width was 

 one-third of their distance apart (a = Ji), the 

 widths, being calculated so that the angular dia- 

 meters, a, deduced from the above formula de- 

 creased by o"-i with successive screens. The 

 screens for which the fringes became least distinct 

 were found, and by interpolation the angular dia- 

 meters of the satellites were estimated to o"oi. 

 The angular diameters so obtained, reduced to 

 a distance of five units for Jupiter, were as 

 follows : — 



I. II. III. IV. 



Hamy ... 

 Michelson 



098 

 I 02 



087 

 094 



128 

 1-37 



1-31 

 1-31 



The agreement between the two series is very 

 much better than would be obtained with micro- 

 meter observations. 



The method does not appear to have been 

 further employed until the past year, when, at 

 Michelson 's suggestion, it was tried with the 

 loo-in. Hooker telescope at Mount Wilson. In 

 view of the advantages of the method, this seems 

 somewhat surprising ; possiblv it is due to an 

 exaggerated idea of the diflficultv of the observa- 

 tion. Besides the application to satellites and 

 asteroids, the method might be emploved for the 

 measurement of the oblateness of such bodies as 

 Mercury, which have no satellites from a study of 

 the motion of which the oblateness might be theo- 

 retically deduced, and for which micrometrical ob- 

 servations are not sufficiently accurate. It can 

 easily be shown that by rotating the slits into 

 different orientations the corresponding angular 

 diameters are determined. 



NO. 2700, VOL. 107] 



At Mount Wilson the method has been applied 

 to the measurement of the angular diameters of 

 stars. Theoretical considerations have indicated 

 that the stars of largest angular diameter are to 

 be sought amongst the giant red, or M-type, 

 stars, such as Betelgeuse, Arcturus, etc., but that 

 for no star is the diameter likely to exceed o"-05, 

 a quantity scarcely within reach even of the loo-in. 



Fig. I. — Stellar interferometer attached to end of tube of looin. telescope 

 at th« Mount Wilson Observatory. 



reflector at Mount Wilson. Michelson, in 1890, 

 had, however, indicated the possibility of employ- 

 ing the method in conjunction with an interfero- 

 meter, thereby enabling the original separation of 

 the two beams to be increased very considerably. 

 The arrangement used at Mount Wilson is shown 

 in Fig. I, and diagrammatically in Fig. 2. A 

 steel girder, LL, 20 ft. in length is fixed across 

 the upper end of the tube of the loo-in. 

 telescope. Two adjustable plane mirrors, AA, 



reflect the light from a star along the girder to 

 two other mirrors, BB, 4 ft. apart, which in turn 

 reflect the fight down the telescope tube to the 

 mirror, the two pencils finally uniting, as shown 

 in Fig. 3, and producing interference fringes in 

 the focus of the eyepiece. To obtain the equality 

 in the path of the two beams (which, for inter- 

 ference in white light to be observed, must be of 

 an accuracy of 1/10,000 in.), an adjustable double 

 wedge of glass is placed in the path of one of 



