442 



KNOWLEDGE. 



November, 1910. 



plate tcmporarih- by tliat amount witlKuit fracture. 



\'arious methods \\ere tried to protliiee the 

 required amount of hendin;; artificially : of these 

 the most successful, as one might expect, is that 

 produced by atmospheric pressure on the outer 

 and film side surface of the plate, caused hv 

 exhaustion of the air in a closed chamber at the 

 back of the plate. 



From a trial made with the 16-inch Metcalf 

 Telescope in 1910, June 29th, ocular e\'idence can be 

 obtained hv the examination of the two sets of star 

 images shown on the accompanying illustrations. The 

 region is that surrounding R.A. 19''' 0'" and + 5° and 

 is enlarged nine times from the origin plate, so that 

 10" equal 1""". In obtaining these two exposures 

 the first and left hand image was taken with the 

 plate temporarily curved and the second image shows 

 the same star with the plate flat : each exposure was 

 of ten minutes duration. In Figure 1 the centre of 

 the plate is substantially in the same focus, the flat 

 plate image being slighth' the better one of the two. 

 When the outer parts of the same plate are compared 

 a marked improvement becomes obvious with the 

 curved plate images, the bright stars showing less 

 diffusion and naturalh- smaller images, there being 

 less scattering of light : and the faint stars — the 

 stars" light being more compact and in focus — are 



distincth' improved ; almost invisible and diffused 

 images become quite \-isible and better formed. 



Further plates have since been taken with that 

 instrument to test the effect on the determination of 

 magnitude and position. From an examination of 

 sixty-one stars, fainter than the 10th magnitude, it 

 was found that the flat plate lost 0"17 magnitude at 

 1 '5 from the centre : in the curved plate the brightness 

 increased by 0'06 magnitude : no s\'stematic error 

 was detected. 



The effect on the actual positions of the stars has 

 not yet been determined : this requires careful and 

 accurate measurement of a number of plates, 

 preferably of the same area. 



For the convenience of use, storage, and measure- 

 ment, it w ill be a distinct gain to be able to have the 

 plates flat rather than have to deal with plates with 

 a permanent curvature. When the temporarily 

 curved plates are measured there are obvious 

 advantages in having them flat in the measuring 

 instrument. 



The plate reproduced here is from the Harvard 

 Circular, No. 161. 



Knowing Dr. Metcalfs skill in practical photo- 

 graphic astronom\- and his power to overcome 

 difficulties, we miw expect still better results from 

 his instruments. 



CORRESPONDENCE. 



.SPECTROSCOPIC DOLiBLE STARS. 



To tlie luiitiirs of" Kxowi.EnGE." 



Sirs, — Sir Robert Ball's work, " In the Hi-li Heavens," has 

 a chapter entitled " The New .Astronomy," in which the anthor 

 discnsses, with all his admirable clearness of exposition, the 

 modern methods of determining the motions and masses of 

 Double Stars by means of the Spectroscope. In illnstiation 

 of this subject he takes the well-known star Mizar, the 

 principal component of which was discovered by Professor 

 Pickering to be a Spectroscopic Double, " a discovery," says 

 Sir Robert Ball, " which will take its place in the history of 

 astronomy as the inauguration of a new process in the .study 

 of things sidereal." 



It will be needless to give the readers of " Knowledge " a 

 detailed account of the method by which the examination of 

 the spectrum of a Double Star is held to atloril iis infonnation 

 as to the velocity, period of revolution, nuitnal distance, and 

 masses of its components. Sir Robert l!all iNplains the 

 whole process in his charmingly lucid manner, and the 

 argument would be absolutely convincing were it not for one 

 most important assumption which seems to be scarceh- 

 justified. This is that the earth's line of sight lies in the plane 

 of revolution of the Double Star, or, in other words, that we are 

 observing the orbit of the Double Star "edgeways." 



The whole reasoning is based on the velocities of the 

 components as revealed by the spectroscope. But the spectro- 

 scope can only reveal to us velocities in thr line nf sight, .and 



as the orbits of Double Stars ma\- lii- in .ill possible planes, it 

 is surely an unwarrantable assumption that the velocities 

 deduced from the spectrum are the true velocities of the 

 components. If, for instance, the earth's line of sight 

 happened to be inclined at an angle of sixty degrees to the 

 plane of the Double Star's orbit the velocities of the 

 components, as deduced from the spectrum, would be only 

 half their real velocities, and the whole argument as to the 

 masses and distances of the components would be \itiated. 



There seems to be one, and only one, condition under which 

 the spectrum may be depended on to give us the true velocities 

 of the components of a Double, and that is the condition that 

 the Double is at the same time a Variable. For if we may 

 assume that v.iriabilitx- is due to periodic eclipse of each 

 component by the other, it follows that in the case of a 

 variable star oiu' line of sight lies in or near the plane of its 

 orbit of revolution. 



But no stich condition is nicnlioned in any discussion of the 

 subject which has come under my notice, and my object in 

 writing this is to inquire whether any of your astronomical 

 correspondents can explain what I in the absence of the 

 condition just stated) seems to me an insuperable difficulty in 

 the theorj' of Spectroscopic Doubles. 



Yours faithfulK'. 



A. E. MADDOCK. 



Banparawela, 

 Ceyi.on. 



