480 Mr. Nixon’s Table of observed Terrestrial Refractions: 
Table of observed Refractions. 
11 arcs from | 
\16' 36" to 2)’ 48” 
fi 
ose 29 
— fi ps6 
Perse 530 
| 
50 ares fi | 
Sum, 162 ares = 17°58! 59"; Refr, 3425” = EES 
The class A. forms itself naturally from the total deficiency 
of observations on arcs from 14! 1" to 16! 36". That marked 
B. consists of ares still large enough to mask those local irre- 
gularities of refraction, generally of a negative character, 
which begin to form an occasional but slight feature of the 
class C. The arcs marked D, subdivided at first into two 
classes of about the same number of arcs, were ranged together 
on account of the anomaly of the smaller arcs presenting a 
refraction of ;1,, whilst those of the other class, abounding in 
negative refractions, averaged no more than ;+3. Some of 
the more marked deviations from the mean value may, no 
doubt, be attributed in inconsiderable distances to the want 
of sufficient accuracy in the measurement of the height of the 
eye and pointing the telescope at the base of the signal, yet 
numerous recent observations have clearly indicated a modi- 
fication of the average refraction peculiar to the locality. As 
the ratio of the refraction to the contained are appears to 
increase with the arc, it is more than probable that the con- 
stant error of the sector, considered as — 20", has been esti- 
mated in defect. 
Let a,c be two arcs, of which c is (considerably) greater 
than a, and let a third are & equal their difference c—a. 
aie 1 : 
Admitting =, to denote the constant proportion of the refrac- 
, Abia : 
tion to the arc, then will = be the refraction for the arc a, and 
c A . . 
= that for the are c; that of the intermediate arc } being 
1 Cc 
equa to ~ 
a . : ° 
a Supposing the instrument to give the ele- 
