STAR CHARTING. 213 



the focus of the star camera and the rays from it pass out from the 

 object-glass parallel, and falling on the silvered reseau are stopped in 

 aU places except where it is cut through by lines and figures and 

 there it passes on to the sensitive plate and marks it. A number of 

 plates are so treated one after the other and stored ready for use in 

 dark boxes. They are all carefully numbered on the glass and taken 

 out in order and exposed in the star camera on fine nights. The 

 plates are 6 \ by 6 \ inches, and the part actually exposed 6 by 6 

 inches; of this space 4-7 by 4-7 inches is the part which is finally 

 used, that is 2- by 2\ The margin, rather more than \ in., is to be 

 for overlap on the plates, and the stars on this can be compared for 

 verification in each adjoining pair of plates. When the plate is 

 developed after exposure the lines (or grating), as well as the stars, 

 appear. So far everything is simple and mechanical ; but the 

 resolutions of the conference require that one set of plates shall have 

 on them all stars to the 1 1 th magnitude, and the other set all stars 

 to the 14th magnitude, and the difficulty is in an ever-changing 

 atmosphere, and with plates variable in sensitivenesss to give the 

 exposure necessary to secure these results. The Astronomer Roj-al 

 for England, as chairman of the committee appointed to deal with 

 these and other kindred questions, has been making experiments on 

 a fairly good night in London, and has come to the conclusion that 

 two minutes will be enough in such weather for stars of 11th 

 magnitude, and 30 minutes enough for stars of 14th magnitude, and 

 that these times must be modified by the weather — that is, increased 

 if the weather is bad. I am able to show you three plates exposed 

 30 seconds, two minutes, and 30 minutes respectively on the well- 

 known star cluster Kappa Crucis. You will see that 30 seconds is 

 enough to get images of stars to the ninth magnitude, and that two 

 minutes gives images of stars to the 11th magnitude, and takes in a 

 number of 12th and one of 13th magnitudes ; but the plate exposed 

 for 30 minutes is not so satisfactory. It should, according to the 

 rule, show with defined discs stars of the 14th magnitude of 

 Argelander's scale ; in Herschel's monograph of this cluster he has 

 11 stars of the 14th magnitude, and four of the 15th ; of these, eight 

 are invisible, six are visible, but not measureable, and only one is as 

 it ought to, "measureable"; and there are some stars of 12th and 

 13th magnitude that are not measureable. The plates were exposed 

 one after the other on a night that seemed to be uniform, and when 

 the two-minutes' plate was a success, the 30 minutes' ought also to 

 have been. I give the result of this experiment to show the difficulty. 

 At first sight this looks like a failure of the method ; but I find on 

 further investigation that these faint stars in Herschel's list are 

 either much fainter than he took them for, or they are coloured 

 stars. The matter was tested by taking a photograph of the same 

 object and giving three hours exposure ; even then most of the stars 

 referred to above are far too faint to measure, although they can be 

 be seen plainly enough in nearly every instance, and the photograph, 

 hurriedly examined to see if the faint stars were on it, is found to 

 contain at least 20 more faint stars which Herschel did not see. 

 This example will serve to show you better than any statement the 

 difficulty to be met, according to the adopted rule, viz., if two 



