on Atmospherical Refraction , &c. 
*99 
the majority have adopted the temperature within the obser- 
vatory. 
It is to be regretted, that astronomers have not informed 
us, by what methods they have deduced these corrections of 
the mean refractions ; I shall therefore explain that which I 
have adopted, in order that this hypothesis may be investi- 
gated ; and any error therein, be the more easily detected. 
I have taken all those stars in the table, below the Pole, from 
jG Ursae majoris to e Aurigae, both inclusive, having noticed 
that those observations made in winter, when the correction 
is additive, the assumed correction by Dr. Bradley’s formula 
appears to be too great ; and this difference will be seen by the 
comparison in Tables II. and III. To these I have added Fo- 
malhaut, containing in the whole 210 observations. Having 
determined the mean place of the star at a certain epoch, I find 
the apparent place ; the difference between which and the 
observed place, is the apparent refraction, which call a; put 
b = mean refraction, c = to the correction for the barome- 
ter, and w = to the correction for the thermometer: then 
The following observation of a. Cygni will suffice for an 
example. Observed zenith distance 83° 47' 58", 34 ; apparent 
zenith distance 83° 56' 27", 12 ; apparent refraction S' 28", 78 ; 
mean refraction 8' 7", 83 ; barometer 30,13; correction by 
Bradley -f- ,017 9; thermometer within 41, without 34: then 
for the thermometer within at 41, or without at 34. 
Proceeding in this manner with the above 210 observations, 
I obtained equations for the thermometer within, from 31 to 
a (8'. 2 8", 78) - b (8'. 7 ",8 3 ) 
b (8'. 7 ">8 3 ) 
= +,°430-c(+, 0179)= w(+, 0251) 
D d 2 
