i32 An Account of a General Survei/ q/'ihe Heavens 



seen in the annexed Catalogue : but I shall, for the future, pur- 

 sue the advice of my respected friend Tralles ; and in some 

 well-known region of the heavens (perhaps in that of the 

 Pleiades) shall give a number of stars, which I shall class 

 under the different magnitudes. It would be very gratifying 

 to me, and to many others, if Dr. Olbers, who is so perfectly 

 acquainted with the firmament, would publish his opinion con- 

 cerning these differences. 



Those astronomers for whom this account of the commence- 

 ment of a series of observations possesses any interest, will re- 

 joice also to see the result of an undertaking which, notwith- 

 standing the unavoidable rapidity of the objects, has attained 

 to exactness. The first is so arranged that, in regions abound- 

 ing in stars, three stars in a minute, and no more, may, on an 

 average, be observed. If it is desired to observe the right 

 ascension by two wires, (which can only happen with a view to 

 greater accuracy, or when there is some doubt of the first 

 wire,) it becomes scarcely possible to accomplish more than 

 two stars in a minute. 



It may now be inquired, what are the results of the observa- 

 tions made with one wire. This question can only be an- 

 swered by repeated comparisons ; since the state of the atmo- 

 sphere is an important object of consideration. If the stars 

 are indistinct and twinkle much, as is unfortunately very fre- 

 quently the case, the uncertainty of an observation by a single 

 wire is strikingly increased, as I have found by the compari- 

 son of observations made under favourable and unfavourable 

 circumstances. 



As, however, in an undertaking of such great extent, com- 

 menced under the 55th degree of latitude, we must not be too 

 solicitous about the choice of circumstances, but must take ad- 

 vantage of clear weather when it comes, nothing remains but, 

 by a comparison of very numerous observations, to determine 

 a medium value of probable errors. In order to leave no 

 possibility of the existence of arbitrary decisions, and to point 

 out the probable errors which may occur under actual circum- 

 stances, and not under those alone which are pecuharly fa- 

 vourable, I have reduced all the stars (which were observed 

 twice or oftener in the year 1821, since the adoption of the 



greatest magnifying power) to the beginning of 1825; and 

 ave annexed the catalogue of them to this paper: from 

 which it will appear that the probable error of an observa- 

 tion is in 



iR = + 0,"1548, Deck = + 1",013. 

 Bradley's observations give for the ibrmcr 0",1426, for the 

 latter 0",98, both of which are less. I am, however, well satis- 

 fied 



