688 



NATURE 



[July 28, 192 1 



These figures emphasise the remarkable accuracy 

 of the method (the largest residual in distance is 

 only four ten-thousandths of a second of arc !). 

 It is stated that with care both angular separa- 

 tion and position-angle can be measured with an 

 accuracy of about i per cent. This accuracy is 

 much greater than is possible with a filar micro- 

 meter. The method possesses the further advan- 

 tage that the percentage accuracy in both distance 

 and position-angle does not decrease with decreas- 

 ing separation ; with the filar micrometer, on the 

 other hand, the error of observation increases con- 

 siderably with decreasing separation. For the 

 successful use of the interference method good 

 "seeing" is not essential, whereas for the 

 observation of close double stars with a filar 

 micrometer very good seeing is necessary. In 

 explanation of this unexpected result it is sug- 

 gested by Hale that "in bad seeing, when using 

 the whole aperture of the objective, there is an 

 integrated effect of the light-waves meeting in all 

 possible phases, which tends to obliterate the 

 details of the diffraction pattern of the star- 

 image, but that when two light pencils are selected 

 at opposite ends of a diameter the result is not 

 an integration, but a mere displacement of the 

 diffraction-pattern, sufficiently small for the eye 

 to follow." 



If the distance between the slits is greater than 

 the value l = ^\/a, which gives a unique position 

 for which the fringes disappear, there will 

 be four position-angles for the slits in which 

 this occurs, these positions being symmetrical 

 with reference to the line joining the two com- 

 ponents, viz. p±B and p + v + 6, where p is the 

 required position-angle. The most accurate 

 method of observation is to adjust the distance 

 apart of the slits so that 6 is about 30° to 50°, 

 and to measure the four positions in which the 

 fringes vanish, so determining 6 and p ; if the 

 separation of the slits is I, the value to use for 

 the computation of the separation of the com- 

 ponents of the binary is I cos 6, the separation 

 therefore being x/zlcosd. If three different 

 values of I are used and the corresponding values 

 of 6 are determined, the accuracy of the observation 



is increased. We have heretofore supposed that 

 the components are equal in brightness and 

 separated by a distance large compared with their 

 diameters. If they are of unequal brightness the 

 fringes do not completely disappear in any orienta- 

 tion of the slits, but instead the positions of mini- 

 mum visibility (when the fringes of one star fall 

 between those of the other) are observed. The 

 variation in contrast is greater the more nearly 

 equal the components are in brightness. The 

 method is therefore suitable for the observation 

 of close doubles which do not differ too greatly 

 in brightness and are beyond the limit of 

 resolution of, or observable only with difficulty 

 and under the most favourable conditions with 

 the telescope available, in conjunction with a filar 

 micrometer. The time required for a single 

 observation is longer, but this is more than com- 

 pensated by the great increase in accuracy of the 

 observation, by the possibility of observing under 

 poor atmospheric conditions, and by the smaller 

 number of observations required for the deter- 

 mination of an orbit. Prof. Hale expresses the 

 hope that through a co-operative plan of observa- 

 tion, in which several observatories will take part, 

 a large number of close binaries may be 

 measured in this way. 



An attempt is to be made at Mount Wilson to 

 extend the method to the measurement of stars 

 several minutes of arc apart. Until this is tried 

 it cannot be asserted whether or not the difference 

 of atmospheric disturbances along the optical paths 

 of the two stars would prevent the fringes from 

 being observed. If it proves feasible to observe 

 them in this case it may become possible to 

 measure the displacement of a star by the gravita- 

 tional field of Jupiter, and thereby provide a 

 further test of Einstein's theory, scarcely possible 

 in any other manner. The method might then 

 also be employed for the determination of stellar 

 parallaxes and proper moiions (which depend 

 upon the differential displacements of adjacent 

 stars) with a smaller probable error and in a 

 shorter time than by existing methods. The 

 further investigation of these possibilities of the 

 method will be awaited with great interest. 



The Paris Conference of the Museums Association. 



FOR an association which, during the first 

 thirty years of its existence, has confined 

 its meetings to cities in the British Isles, the pro- 

 posal to hold this year's conference in Paris 

 seemed somewhat hazardous. W^hatever objec- 

 tions may have presented themselves to some 

 .TOembers, there can be no doubt that the experi- 

 ment proved a greater success than any antici- 

 pated. During the week July 11-17 the seventy 

 delegates from national, municipal, and semi- 

 private museums, with their president, who, by 

 good fortune, happened to be a man of such dis- 

 tinction as Sir Frederic Kenyon, were received 



NO. 2700, VOL. 107] 



in the most cordial manner by the heads of the 

 State Museums of Art and of Science, by the 

 Conseil Municipal and by the directors of its 

 museums, and by the authorities of Les Invalides, 

 the Biblioth^que Nationale, and similar institu- 

 tions. Receptions at the Louvre, the Mus^e 

 d'Histoire Naturelle, the Hotel de Ville, and the 

 club " Autour du Monde " enabled members to 

 become personally acquainted with many French 

 colleagues ; and visits to the numerous and rich 

 collections of Paris, Versailles, St. Germain, and 

 Malmaison, under the guidance of distinguished 

 authorities, with privileges accorded only to heads 



