338 Dr. Brinkley’s observations for investigating 
obtained completely, without a reference to the correction 
for collimation ; and we are not obliged to depend for some 
days, perhaps, on the stability of the correction for collima- 
tion. We also are more likely, in this way, to improve our 
theory of refraction, because thus the irregular refractions of 
different days will not be mixed together. These considera- 
tions, independently of the interest of the question of parallax 
and aberration, lead me to dwell more on the discordance of 
the Greenwich observations, and of those made here, than 
otherwise I should be willing to do : and I am induced to 
offer a few brief remarks relative to the circumstances of the 
observations that have been adduced by Mr. Pond, to prove 
the non-existence of a visible parallax. 
1. The observations of the Greenwich mural circle are so 
implicated with each other, and the polar distances, even of 
the high stars, depend so much on the index error obtained 
by observations of those stars in which the uncertainties of 
refraction and of other data produce their effects, that it is 
not very extraordinary that the small quantities which I as- 
cribe to parallax should not distinctly appear from the obser- 
vations of the mural circle. There is indeed one exception 
to this explanation, which, I freely confess, occasions in my 
mind more difficulty than any other. This is in regard to 
y Draconis and a Lyras. According to the observations of 
Mr. Pond, there is no difference between the relative places 
of these stars in summer and winter ; and it is from a relative 
change of place I find in these two stars, that I adduce, what 
appears one of my strongest arguments for the parallax of 
a Lyras. In this instance, the two instruments are completely 
at variance, and one of them must give an erroneous result. 
