124 CARNEGIE INSTITUTION 



Kapteyn's method should be adopted, and all the photographs 

 taken at different seasons on the same plate should be exposed at 

 the same hour angle. Although this involves some sacrifice of 

 parallax factor it eliminates many possible sources of sj^stematic 

 error. I would advocate an observ^atory (that is to say, a telescope 

 and two photographic observers) devoted exclusively to this work, 

 with an office and staff of measurers and computers located else- 

 where, where the services of students and others could be secured, 

 living is less expensive, and facilities for instrumental construction 

 and repair are more accessible than in the southern hemisphere. 

 To complete the work there must of necessity be a corresponding 

 observatory in the northern hemisphere, and to complete the whole 

 work in any reasonable time there should be several such pairs of 

 observatories. 



To make a beginning, so as to test the accuracy and probable 

 value of the work thoroughly, it would be well to install one observ'- 

 atory of the kind in the most favorable situation and to confine the 

 work, say, to four overlaping areas at each alternate hour of R. A. 

 in each zone of 4° in declination and from declination 60° at each 

 4 hours of R. A. The results of such a series of pictures, taken 

 and discussed, would lead to results of immense general importance, 

 and would give some close approximation to the average parallax of 

 stars of different magnitudes and proper motions, and would be an 

 excellent pioneer program to ascertain the weak points of the 

 original arrangements. It would not too greatly increase the pro- 

 gram if plates having for their centers a number of the stars of 

 more remarkable proper motion were added — indeed, perhaps the 

 program might be best begun with these. 



A, ^. I am hardly disposed to support the plan suggested by you 

 of extending the zones on the plan of the Astronomischc Gesellschaft 

 from declination — 32° to the south pole. 



It is a far more accurate and useful plan to select the stars which 

 are best distributed for determining the constants of photographic 

 plates, as has been done at the Cape for declination — 40° to — 52°, 

 and then to determine from the photographic plates the places of 

 all required stars — say all stars to the eighth or ninth magnitude. 



On the Astronomische Gesellschaft plan you get for this purpose an 

 unnecessary number of stars in some parts of the sky and an insuffi- 

 cient number in others. For the zone — 40° to — 52° I found 8,000 

 stars ample, and they are as uniformly distributed as it is possible 

 to select them. From the coordinates of our plates (and we are 



