264 PROCEEDINGS OF SECTION A. 
On comparing two-paired pictures we can hardly find two 
corresponding points which can be satisfactorily marked with the 
fine point of a compass. 
You often think you can do that on some better defined parts ; 
but when the mark is to be made, you know it is only a question 
of rough approximation. 
Consequently it is unnecessary to attempt greater accuracy in 
the measures than that obtainable with a scale of millimeters. 
The photographic part of the work gave more trouble than the 
instrumental questions. It was intended principally to deal with 
the higher clouds, some classes of which are sometimes barely 
visible against the deep blue sky. 
Yellow screens of different densities were used, with stops no 
greater than f/32 and exposures from 1/12 to 1/25*. 
Many experiments were made to determine the best average 
conditions, with a view to give the work as much uniformity as 
possible. If it could have been carried on by the same two 
operators throughout, much could have been left to their 
experience and discretion, to suit the variable photographic 
factors to the variable circumstances. But there were four 
different operators at the far end station, away from the Observa- 
tory, who took up weekly duty in turn, and with very little time 
at their disposal. 
It was essential also to avoid any interference with the camera 
in order to preserve its adjustments, and to facilitate the opera- 
tions as much as possible. 
It would have been undesirable to vary the stop, the time of ex- 
posure, and the screen in each case, and the desired effect had to be 
obtained by an average constant combination of these quantities, 
subject to alterations, only in very exceptional conditions. 
Every possible combination was tried, and at last we found one 
which seemed to act satisfactorily. 
This was 1/25%, for the time of exposure f/25 stop, and a screen 
of medium density call screen No. 2. 
It has been possible hitherto to meet almost every exceptional 
case by sometimes varying the exposure, very rarely using a 
different screen, and never altering the stop. 
In the earlier stages we used Ilford isochromatic plates, which, 
however, gave continuous trouble and poor results. 
The later plates employed were Ilford chromatic, of medium 
rapidity, and these have given every satisfaction, and are in use 
at present. 
The negatives are slowly developed by the Metol-hydro- 
quinone. The finer cirrus clouds sometimes taking forty minutes 
or longer to develop. 
The two stations are connected by telephone, and are also pro- 
vided with a telegraph key and sounder, through which time- 
