6o 
Dr. Brinkley on the 
7 6° or 77 0 , and not less than about 69 0 or 70°. Let us sup- 
pose it in its mean quantity at about 73 0 , and then the value 
of m will be about — n ^ 73 ■ = 1 ,6, and for the stars near the 
Pole, the mean value of n about 0,7. Hence m — n is less 
than unity, and, consequently, the error of the constant of 
refraction is greater than the error of N — C. Now I think 
it will be conceded to me, that it requires the exactest instru- 
ments and exactest observations to determine the quantity 
N — C certain to half a second. A greater number of stars 
can be used for determining C than for N, but then the greater 
zenith distances will probably occasion C to be more inexact 
than N. In C, the irregularities of refraction, and in N, the 
errors of division, have most influence. 
In a series of observations in which I am at present engaged, 
for determining anew the constant of refraction, I use for N 
the Pole star only, and I lessen the effect of the errors of 
division, that may be apprehended, by being enabled to ob- 
serve the Pole star in all parts of its daily course. 
I shall not anticipate here the probable result of these 
observations. 
The object of the above, is not to examine whether the con- 
stant of refraction has been determined with greater exact- 
ness at Dublin or at Konigsberg, but only to endeavour to 
show that the uncertainty, which exists, cannot be considered 
in any manner adverse to the received opinion respecting the 
exactness of modern observations and modern instruments.* 
Before leaving this subject, I may be permitted to make a 
* A catalogue of Polar distances necessarily exhibits, for low stars, the error of 
the constant of refraction as it were considerably magnified. 
