36 
The Trigonometrical Survey of India. 
[No. 1, 
63. The computation of heights was performed in the usual 
manner, until the estimation of terrestrial refraction was arrived at. 
The process adopted for this purpose may he briefly stated thus. 
61, Estimation of Terrestrial Refraction. —If the contained arc 
r 
be represented by c, and terrestrial refraction by r, then — = /’ 
c 
the factor, or “ decimals of contained arc.” Whereby if fhe given, 
then r — c.f may be computed. From want of a more accurate 
method of determination, it is usual to adopt that mean value of f 
for finding the height of an inaccessible point, which may be forth¬ 
coming from the reciprocal observations at visited stations. For 
instance if A, B, C, D, be points of the last mentioned order, then 
in the ordinary course of computation, 
there will result three values of f at 
A, as many at C, and two values each at 
B and D, The mean value of f at each 
station would therefore be adopted in 
computing the height of an inaccessible 
point H. To take a real case (at random). 
The values of/’at Batwya T. S, (1) are 
+ 0.011, — 0.017, + 0.065 'and + 0.013, 
Wherein the greatest difference is no less 
than .082 of the contained arc. On 
the other hand, the values of f at hill 
jg stations of observation, will always be 
found accordant within far smaller limits. 
65. The conclusion drawn from the foregoing is evidently this. 
That at plain stations, and when the object observed is placed on an 
ordinary tower, the value of f determined from any given ray A B, 
is not necessarily applicable to any other ray A C. Whereas all 
rays of light at hill stations from terrestrial points appear to be 
nearly equally refracted. These phenomena are clearly traceable 
to local causes. 
66. But of the two mean values of f one obtained at a moun¬ 
tain station of observation, and another deduced in the plains, it 
is evident that the former is more trustworthy, and hence it appeared 
desirable, that the latter should be obtained in terms of the former. 
