40 'M. Bessel on the Declination of the principal 



meter ; in which Bradley's rule jierceptibly deviates from the 

 truth. The regularity hereby obtained is of about the same 

 magnitude as the probable errors would lead us to expect. — 

 The reason of the greater deviation at 2a Capricorni is un- 

 known to me ; that of a. Piscis austr. may arise from the centre 

 of the long coloured spectrum (which this lowly culminating 

 star presents) having been differently determined by two dif- 

 ferent observers. 



The constant difference of both series of observations has 

 been so greatly diminished by this calculation, that it is no 

 longer so striking as before. But, whilst the casual errors of the 

 sino^le observations are reduced to a trifle, by the increase of 

 their number, the goodness of the instruments, and the care 

 in their use, the imperfections of the elements of reduction, 

 hitherto unnoticed, distinctly appear. If, after the increased 

 certainty of Mr. Pond's observations (by the application of 

 the method of observing by reflection), the difference were still 

 as great as in the standard catalogue, yet we should be led to 

 believe that there are other unknown sources of error. Now, 

 however, I think I am able to point out some circumstances 

 which might account for differences of this magnitude. 



If, for instance, we were to increase the logarithms of the 

 Greenwich refractions by about 0"0035 (which for 45° of 

 zenith distance amounts to 0",47), the differences would be- 

 come partly positive, partly negative. Such a change would 

 lead us to suppose that either the barometer stand:i 0*24 inches 

 too low, or the thermometer 4° Fahrenheit too high. Or, dis- 

 tributing the error on both, that the first is 0*12 inch too low, 

 and the second 2^ too high. We might add that the sense in 

 which both errors must be assumed, is agreeable to former 

 experience : for the barometers often stand too low, owing to 

 air having been introduced by long use, or to the bore of the 

 tube not being corrected by the scale ; respecting which I find 

 no notice with reference to the barometer of Greenwich. The 

 height of my barometer, for instance, would be 1 line = 0'09 

 inch (English) too short, if I were to use them without cor- 

 rection : the freezing point also of the thermometer moves up- 

 wards in time ; although I do not think that changes of this 

 kind, exceeding 1^ Fahr., have ever been observed. If Mr. 

 Pond vvould be so kind as to give some account of the condi- 

 tion of his meteorological instruments, we should be able to 

 decide at once, whether any difference, and how much thereof, 

 is to be ascribed to this cause. I might readily have deter- 

 mined the constant of refraction, from Mr. Pond's observations 

 (which would have removed every doubt concerning the me- 

 teorological instruments) had there been among the lower 



culminations, 



