342 Dr. Brinkley’s observations for investigating 
each bisection of the star and reading off was considered as 
a distinct observation. The great improvement in the uni- 
formity of the results is very apparent when two or four 
observations are made on the same day, by observing before 
and after the object has been on the meridian. 
The greater errors occur according as the star is more 
remote from the zenith. This is doubtless occasioned by the 
irregularity of refraction, which is so very apparent when 
the object is within io° or 15 0 of the horizon. It may be 
traced by my observations to within a few degrees of the 
zenith. On this account, when the object is 40° or 50^ from 
the zenith, and great exactness is required, it will be necessary 
to increase the number of days of observation, rather than 
the number in the same day, that the irregularity may dis- 
appear from the mean. 
I know not of any observations where the irregularity of 
refraction appears so distinctly as in mine. To illustrate this 
more fully, I have, in Table 8, added the observations of « 
Aquarii. This star I observed with a view of ascertaining 
whether it was subject to changes of place similar to what 
appeared in a. Aquilae. The mean results give a much less 
change of place, but the discordances which appear to belong 
to refraction are more fully apparent. 
By observing before the star came to the meridian, and 
then reversing the instrument, using only the bottom micro- 
scope, I was enabled to get several results on the same day. 
In all the other stars three microscopes were used. 
An inspection of Table 8 appears to show clearly the 
effects of this irregular refraction. Thus it is evident, that 
the differences between the results of the observations of 
