368 Astronomical and Nautical Collections. 



He therefore recommends the employment of the French Tables 

 for moderate zenith distances, and remarks that nearer the horizon 

 it is useless to expect minute accuracy in any conclusion from 

 astronomical observations. 



In the Xllth volume of the Irish Transactions, we find a memoir 

 of Dr. Brinkley, read in 1814, (Coll. VI,) on the thermometrical 

 correction of refraction; giving a method of correction " derived 

 from the formula" of Simpson, " obtained in the hypothesis of a 

 densi'ty decreasing uniformly." The author observes, that "at pre- 

 sent we have not sufficient observations to determine, whether the 

 actual variations of refractions at low altitudes are most conform- 

 able to the theory of Mr. Bessel, to that of Dr. Young," or to his 

 own : which, indeed differs less from each other than they do from 

 the corrections employed by the French, by Groombridge, or by 

 Bradley ; and after all, it seems impossible to expect much advan- 

 tage from any theory in applying this correction to the accidental 

 variations of any one climate, though it may very probably be of 

 use for finding the mean refractions in distant latitudes. (See Coll. 

 XIII.) 



It was in the interval between the publication of Dr. Brinkley 's 

 two papers, that Dr. Young annexed a new Table of Refractions 

 to the Nautical Almanac, founded on an approximation of his own, 

 but agreeing almost exactly in the mean refractions with the French 

 Tables, adopting, however, a correction for temperature derived 

 from theory, and greater near the horizon than that which the 

 French have employed. Having observed that the series obtained 

 for expressing the refraction in terms of the density failed at the 

 horizon, because the altitude was a divisor of the coefficients, it 

 occurred to him that this inconvenience might be avoided, by ex- 

 pressing the density in a series of the powers of the refraction; 

 and the formula thus obtained, though not always convenient in 

 extreme cases, is still very useful for obtaining a tolerably accurate 

 result with great facility from any imaginable theory, and is also 

 capable of representing, by a few of its first terms only, with their 

 coefficients empirically modified, the refraction either actually ob- 

 served, or correctly computed, upon any possible hypothesis re- 

 specting the constitution of the atmosphere. 



