1 76 Mr. Poiul on the Chmiges in the 



jiearod the greatest, passed over almost the same divisions 

 witli others, in which no such discordance could be perceived. 

 Moreover, in examining these discordances in different points 

 of view (that is, both with respect to their right ascensions and 

 l^olar distances), I fancied I perceived something like a general 

 law, that was quite incompatible with any possible hypothesis 

 of error in the instruuient. 



On a point of this importance, I clearly saw the necessity 

 of devising some new method of observation which might de- 

 cide with certainty that which othei'wise would become an 

 endless subject of doubt and conjecture, 



I had often attempted to observe the altitudes of stars by 

 means of an artificial horizon of quicksilver, or other fluid, 

 but had abandoned the attempt from the difficulty of protecting 

 it from the wind, and from the number of observations I lost 

 in fruitless experiments. To this method I had again re- 

 course ; and by means of wooden boxes of different sizes and 

 figures, according to the different altitudes of the stars, I have 

 sufficiently accomplished my purpose. A very few observa- 

 tions were sufficient to convince me that the instrument was 

 in every respect perfect, and that I might repose the greatest 

 confidence in every result it gave. 



Several stars, and particularly those most discordant, I have 

 observed by this new method, and find their places, without 

 any exception, to agree within a fraction of a second with 

 those determined by direct measurement from the })ole. 



Presuming that the observations* which accomjiany this 

 paper will remove every sliadow 'of a doubt as to the accuracy 

 of the instrument, I shall now proceed to state, in as few words 

 as possible, the nature of the changes which appear to me to 

 have taken place since the year 1812. 



If Bradley's catalogue of stars for the 3'ear 1756 be com- 

 pared with the Greenwich catalogue for 1813, it will be pos- 

 sible to deduce the annual variation for each star for the mean 

 period, or for the year 1784, on the supposition of uniformity 

 in the proper motion of each star ; then allowing for the change 

 of precession for each star, a catalogue may be computed for 

 any distant period; as for example, the present year 1822. 

 Suppose such a catalogue computed, which I have named a 

 predicted catalogue; then, if this be compared with the ob- 

 served catalogue for the same year, the following differences 

 will be found to subsist between them. 



The general tendency of all the stars will be to appear to 

 the south of their predicted places, and this teiidency seems 

 to be greater in southern than in northern stars : if any star 

 * Vide Ajipcndix, p. 178. 



be 



