508 Trans. Acad. Sci. of St. Louis. 
given equal weight, and the azimuths given by the circum- 
polar stars were weighted according to the judgment of the 
computer. 
The values of the collimation deduced by this method from 
the observations of each night are given in the following 
table :— 
RESULTS FOR COLLIMATION DEDUCED FROM THE PRELIMINARY REDUCTION. 
Collimation. 
DATE 1891. 
Columbia St. Louis 
Transit Inst. Transit Inst. 
October 25 +08.407 — 08.218 
Ws Of 0 .329 0 .212 
+e 26 Q .814 0 .264* 
wes 26 0 .412 0 .225 
pacar 2 0 .325 0 .194 
ix BO ser sss 0 .392 0 .209 
ate 29 0 .320 0 .207 
kes DUNNE vb ende Wevos eres 0 .3827 0 .238 
November 2. +0 .356 0 .219 
on 2 —0 .202 
Means +08.354 —0*.216 
These results show, in case of each instrument, little or no 
evidence of change in the collimation constant and in the 
reduction of the observations by the Method of Least Squares 
the mean values of the collimation as found above were 
adopted as the true values. The effect of any small error in 
these adopted collimations is eliminated by the reversal of 
the instruments. This reduces the number of unknown quan- 
tities to two, 47 and a, thus diminishing very much the 
labor of computation. 
n the reduction by Least Squares the stars were give 
weights depending on the declinations. The Coast Survey 
system of weights was used, then an observation equation was 
* This determination is given one-half weight because of poor determina- 
tion of the azimuth. 
