194 The Rev. Dr. Robinson on the Constant of Refraction. 



The declinations of those refraction stars which are in the Nautical Almanac 

 were compared with its places, as long as they were given to the second place of 

 decimals. Afterwards, they were reduced by the constants of Baily's catalogue, 

 and compared with its mean places for the year, corrected when necessary for 

 proper motion. The others were taken from that catalogue, and reduced by its 

 precession, corrected for Bessel's last value of n, and for secular variation (com- 

 puted from its value compared with the precessions given in the Fundamenta). 

 When any of them have been observed at Greenwich, by Airy, the proper 

 motion has been deduced from his results by the formula, 



_ A — cat 4- f (p — b) — 1".053 X cos a 



'^ - '- WVt ' 



where p — b Is the number found in the last column of the Fundamenta, t the 

 time in years from 1830, and 1.053 the correction for the error in the constant 

 of precession used in that work. When Airy had not observed the star, I use 

 my own declination changed for Bessel's refraction. 



The low stars are often neat spectra (that of aLyrae, I have found 22" long); 

 sometimes the blue and violet disappear for several seconds, and sometimes, 

 though less frequently, the red, the rest remaining unabsorbed. When the 

 colours are distinctly separated, I take the yellow where it borders on green, which 

 I think a tolerable average for the mean of the spectrum. The star should be 

 carefully watched during its whole transit, for the undulations that produce 

 irregular refraction are often of long duration ; and sometimes a star, which is 

 apparently well bisected for several seconds, will leave the wire altogether. 



The temperature is observed by a thermometer of Troughton which I found 

 here. I have verified its freezing and boiling points to assure myself that it had 

 not undergone the change said to have occurred in some thermometers. I have 

 also compared it at several points with a standard instrument made for me by 

 Troughton and Simms, in 1834<, and think it of equal excellence. It is established 

 at a north window of the eastern tower, about four feet above the centre of the 

 circle, and twelve distant in a horizontal direction. In a semicylinder of polished 

 copper, and an interior one of tin, arranged so as to permit a free circulation of 

 air, but excluding all external radiation. In summer, when the rays of the sun 

 reach the northern side of the tower, a second thermometer Is used at a southern 



