TRANSACTIONS OP THE SECTIONS. 19 



gatlon, because the error cliauges its sign between these two years ; and his first 

 process is to examine, for all the zeuitli-point groups having a suihcient number of 

 star-observations, the results given by north stars, by south stars, and by observa- 

 tion of the reflected image of the wire. If the means of these separate results were 

 exhibited in the Greenwich observations, it would be easy to see at a glance that 

 the zenith-points given by north stars frequently differ considerably from thos(> 

 given hy south stars, though the mean of these always agrees well with the nadir 

 result. This would show that the circle-readings require a variabh; correction 

 depending on the .«ine of the zenith-distance, but none depending on tlie co.sine, 

 or that the error has for its cause cither an erroneously assigned flexure of telescope, 

 or something that produces a similar error in tlie circle-readings. Now, in any 

 one of these cases, as the assumed zenith-point is the mean of all the results (north 

 stars, south stars, and nadir), it is plain that, while it is applied bade again to the 

 circle-readings for direct and reflexion observations of north and south stars, the 

 zenith-distances deduced from the direct and reflexion observations will be dif- 

 ferent, and that the difference will change signs at the zenith, or that the l! D will 

 have ditt'erent signs for north and south stars. Pursuing this iuvestiu-ation through 

 all the groups, Sir. Main succeeds in reproducing the mean results given in what 

 is usually called the R D Table, which appears in each Greenwich volume. This 

 discordance then is plainly traced to changes (much more frequent than is usually 

 supposed) in the state of the instrument, which render the corrections depending 

 on the sine of zenith-distance variable, and which we may therefore presume to be 

 of the nature of flexure. 



Remarls on the British Association Catahr/ue of Stars. 

 Bij the Rev. R. Main, M.A., F.R.S., F.R.A.S. 



The object of this paper was to show the necessity which exists for the construc- 

 tion of a new general compiled catalogue of stars to replace that which was pub- 

 lished in the year 1845, under the direction and at the expense of the British As- 

 sociation. The author would not have ventured in his ovni person to bring this 

 matter before the Association, if his attention had not been directed to it by the 

 circumstance of his having nearly completed the reobser\ation of all the .stars con- 

 tained in it which are visible at "Oxford, and which require observation. 



By this means a rather singular class of errors was brought to light, which, it 

 was thought, might be interesting to astronomers, and to which it is desirable to 

 give publicity. 



The principal modern catalogues which were used by Mr. Baily for comparison 

 with the more ancient catalogues of Bradley, Lacaille, and Piazzi, were those of 

 Taylor and Jh-isbane, the places of the stars in both of the last-mentioned cata- 

 logues being, from causes which it is not necessary now to mention, very faulty. 



Now, unfortunately, the differences of the right ascensions and declinations given 

 by the ancient and modern catalogues, when reduced to the same epoch, have been 

 attributed not to errors of the observations, but to proper motion, and the resulting 

 proper motion is not only set down in the column appropriated to it, but is used in 

 bringing np the star's place to the epoch 18.50. The results of the observations 

 of Bradley and Piazzi, which of themselves woidd have been very reliable, are 

 thus seriously vitiated ; and any one interested in the inquiry may convince him- 

 self of the truth of the above statement by consulting the notes ap]3ended to the 

 catalogues of stars in the Radclifle Observations, commencing with 1862. 



In addition to this, the author observes that the utility of' the catalogue of the 

 British Association is greatly diminished, not only on account of the badness of 

 the star-places, but on account of the interval wliich has elapsed since the epoch for 

 which the star-constants are calculated, namely 1850. For results of the utmost 

 exactness it is now unsafe to use these constants ; and, on the whole, it appears de- 

 sirable that the Association should shortly consider the propriety of undertaking 

 a new catalogue, which shall embody the plentiful and accurate results of star- 

 observations during the past quarter of a century. 



9* 



