32. Further Remarks on Dr. Bradley’s Theorem, €&c. 
“In the observations of zenith distances of terrestrial ob 
jects, I have repeatedly noticed, that at the setting of the sun 
the refraction increased from 2 minutes to 2! minutes; so 
that objects which were hid during the whole of the day, became 
visible in the evening. (See Base du Syst. Metr. Decimal, t. i. 
pag. 157, 159, and 165). I have never seen that the state of 
the hygrometer had any sensible influence on the terrestrial re- 
fraction; (ibid. page 166). Messrs. Laplace, Gay-Lussac and 
Biot have proved, that it does not produce any change in the 
astronomic refraction.’’—Delambre’s Asir. tom. i. p. 318. 
It is therefore evident, that to complete the solution of this 
problem much yet remains to be done. From the present view 
of the subject, it would appear as if the best mode would be, to 
take successive altitudes of objects with an accurate altitude 
and azimuth instrument, from the times of their rising in the 
eastern part of the horizon, till they set in the western part, 
giving the preference to those that pass the meridian nearest to 
the horizon below the pole: and to repeat the observations at 
various seasons of the year under a great variety of densities and 
temperatures of the atmosphere. Each single observation should 
be recorded, and not the mean of a number, as is usual for other 
purposes in astronomy. ‘They should be taken in as quick suc- 
cession near the horizon as accuracy will admit; but higher up 
a smaller number will be sufficient. The meridional altitude 
should also be taken with great care above the pole when the 
state of the weather is favourable, as that will materially assist in 
the calculation. 
By this means, the latitude of the place being known, together 
with the declination and azimuth of the object, or the time 
when the observation was made, we may, by computation, ob- 
tain its true altitude; and the difference between this and the 
observed altitude will be the refraction at that point. In 
this manner, from a number of observations, we shall obtain 
the refraction at every degree of altitude, from the horizon to 
the zenith ; and taking them under different densities and tem- 
peratures, we may by equations of condition, or the method 
of the least squares, develop the whole, aud assign, with the ut- 
most accuracy, to each unknown quantity in the formula, its 
due magnitude in all its variety of circumstances. 
The method is, it must be acknowledged, very laborious, and 
will require much time, both in making the observations and 
in computing them; but the advantages which astronomy will 
derive from an exact determination of this equation, are pro- 
portionally great. Did the duties of my situation afford suffi- 
cient leisure, and the smoke of London permit the observations 
to be made, I should feel the utmost gratification in performing 
this labour, tedious as it appears to be. 
