350 G. H. KNIBBS. 
At the equinoxes the change of declination is nearly 1’ per hour, 
but the second differences are negligible. The parallax, and 
augmentation with altitude, being the same for each observation 
may be entirely ignored. 
It is sometimes inconvenient to occupy the same station for 
each observation, so that it is necessary to consider the case where 
the longitude and latitude are slightly different—a few minutes 
at most—for the two stations. These differences will be denoted 
respectively by dX and dd. It is easy to see that if the second 
station be nearer the pole than the first, i.¢., if df be +, the 
azimuthal angle measured from the elevated pole will be greater 
than it would be for a point of-the same latitude. Hence we can 
correct the observed direction of the sun, so as to obtain that 
result which would have been given had an observation been made 
at a point on the meridian of the second station, on the parallel 
of latitude passing through the first. 
If this correction be dA it will be given by the equation 
dA = — dpsec ¢ cot (7”’ = $X)......... (51) 
the minus sign in the last factor being used if the second station 
be west of the first. The } term however is generally quite 
negligible. Now, if to the corrected direction, the correction for 
the declinational change be applied, the result will be that the 
deduced meridian will be $(A,+A,); ie, the direction of the 
meridian will have been determined for points whose longitude is 
the mean of the longitudes of the observing stations. 
For the computation of the change of declination, the total 
elapsed time will of course be used, but in evaluating q by (49) 
we should strictly take 
A’ + $vand 7" = 3A, 
v being the convergency of the two meridians. The quantities 
4 v and } 2 are usually, however, so small as to be quite negligible 
in this relation. 
In equal altitude observations we may observe either when the 
sun isin the positions marked 1 in Fig. 4, or when it is in the 
