REPORT OF COMMITTEE ON OBSERVATORIES 113 



a great part of the probable error above indicated springs from 

 the inexactitude with which the positions of the comparison stars, 

 chosen for the calculation of the constants of the plates, are affected. 



From this one may conclude that if the photographic observations 

 should be repeated twenty years from now under the same condi- 

 tions of precision we should be able to determine with accuracy the 

 proper motions of about 100,000 stars, as I have indicated on a 

 former occasion. 



Besides, in that wdiich concerns the construction of the Astro- 

 graphic Chart, which includes the exact positions of more than 

 3,000,000 stars, twelve observatories out of the eighteen in this line 

 of work have not only completed their observations, but have also 

 concluded the task of measuring the rectilinear coordinates that cor- 

 respond to them. 



One of the essential requirements under which the degree of ap- 

 proximation mentioned above could be equaled, and even surpassed, 

 lies in the determination of the plate-constants with all possible accu- 

 racy. In investigating this question, the majority of the observato- 

 ries associated in this work have decided that, if the positions of the 

 comparison stars should be taken from the catalogue of the Astro- 

 nomische Gesellschaft, their accuracy would be insufficient and 

 would not correspond with that which obtains in the measurement 

 of the rectilinear coordinates. It is for this reason that out of the 

 twelve observatories above mentioned seven proceeded to the direct 

 determination of their comparison stars with the help of precise me- 

 ridian observations. 



It seems to me, then, that the general catalogue which you are 

 proposing to observe between — 32 ° and — 90 would have a much 

 higher value if, aside from all other applications, it could supply the 

 comparison stars necessary for the reduction of the plates of the 

 Astrographic Chart included between — 32 ° and the south pole, en- 

 trusted to the observatories at the Cape, at Melbourne, Sydney, and 

 Perth (western Australia). 



Influenced by these considerations, Sir David Gill has already de- 

 termined the positions of 8,556 comparison stars of his zone by very 

 careful meridian observation — about twelve for a plate and three 

 meridian observations at least for each star. 



To fulfil the requirements for the chart of the heavens relative to 

 this region, it would be necessary to obtain the accurate positions of 

 about 30,000 stars, which would give about twelve comparison stars 

 for each plate of four square degrees. 



