38 PROCEEDINGS OF THE AMERICAN ACADEMY 



Admirable fundamental catalogues have appeared from the observa- 



O 



tories of -Konigsberg, Abo, Dorpat, Cambridge (Eng.), Greenwich, 

 Edinburgh, and Paris ; and two from Pulkova and one from Leiden 

 are to appear within a few years. 



I wish to call the attention of American observers to this subject, 

 and will give a short sketch of the process, and enumerate the syste- 

 matic errors which must be guarded against, and their sources. 



First, it is assumed that an approximate catalogue of time-stars is at 

 hand ; the first object is to obtain a correct right-ascension of Polaris, 

 peculiar it may be in some degree to observer and instrument, and this 

 is done by ojiposite culminations, at all seasons ; or at two opposite 

 seasons, as spring and autumn. If the values for spring and for 

 autumn differ by any sensible amount, the instrument is ill-mounted, 

 but a constant difference of three seconds of time here did not prevent 

 Bessel from obtaining a value accurate to 0^3. In our Cambrido-e 

 observations of Polai'is, this difference has been much less ; in fact 

 hardly sensible. 



The next step is to derive the right-ascensions of other polar stars 

 from that of Polaris ; and it is here necessary to observe at two 

 opposite seasons of the year, or, which is the same thing, at two oppo- 

 site culminations. 



Now comes the comparison of time stars among themselves, and the 

 best method, and the method which must in all cases form the ultimate 

 test, is to begin by comparing at two opposite seasons the stars Castor, 

 Procyon, and Pollux with the stars y, a, jS Aquilaj ; and, finally, to 

 compare the remaining fundamental stars (Maskelyne's 36, for ex- 

 ample) with both these groups, though oftener with the nearest one. 

 But this is a method which requires a very perfect clock, ' especially 

 with perfect compensation, or else not exposed to changes of tempera- 

 ture. In case it is not practicable, we may content ourselves for a 

 time with tlie more ordinary process of assuming the freedom from 

 constant error of large groups of stars, and thus getting the individual 

 errors of the assumed fundamental catalogue ; and the only objection 

 which I know to it is, that being almost universally employed, and 

 Bessel's Fundamenta being the old Catalogue always used for proper 

 motion, we may in this way accumulate dangerous errors, sensibly 

 identical, in nearly all modern observations. It is, therefore, desirable 

 that this point, too, should be tested oftener than it is. 



At the Observatory of Harvard College, a fundamental catalogue 



