782 
Ancient Observatory at Delhi . 
[December, 
edge of another window, signalised by being splayed out on 
that particular side into an exceptional shape, and leaving 
the edge required for observation dressed with the fine 
“ ivory” cement. The N. and S. line only just clears the 
centre pillar. The same line is carried through the southern 
tower, falling on the other side of its centre pillar (I think 
so at least, though I did not record it in my notes made on 
the spot), and its position is again marked by fine edges of 
the “ ivory ” cement, though the windows are not of such 
exceptional shape as in the northern tower. 
All the above lineal measures were made with a folding 
rule of box-wood, by Adie, of Edinburgh, and specially 
tested at the Observatory of that city for accuracy. It is 
25 inches in length. One side of the rule was marked to 
tenths, and the other to half-tenths of an inch. 
The measures on the walls and sectors we considered 
trustworthy; those connecting the walls with the centre 
pillar were carried out with difficulty. A plumb-line was 
dropped from some well-marked position at the edge of a 
seCtor, and then, by means of two boards alternately moved, 
the measure was carried step by step across the filth that 
covered the ground. A tight string gave a straight line, but 
no special means were taken for levelling. DireCt measures 
from wall to pillar were also made in the same manner. 
The general result of our measures were — 
British inches. 
Radius of the interior of the tower 325*0 
,, ,, centre pillar 32*5 
Height^of finely-cemented portion of tower, 
inside above the sectors : mean of several 
measures 293*8 
In this last measure, which theory demands should be 
292*5 inches, it must be remembered the original index edge 
had been removed for the vertical, and we were guided by 
eye alone, taking the centre pillar’s top and the opposite 
side of the tower as our guide. 
I noticed that the index markings for vertical angle were 
not too truly marked ; several had double lines. Those in 
azimuth were not too accurate. The box-wood rule allowed 
of its being bent against the inner circumference of the 
wall, and the width of the divisions in azimuth were care- 
fully recorded. 
Inches. 
8 are each recorded as 
1 is recorded as 
1 ) 3 yt • • • 
Mean 
••• 575 
... 570 
••• 572 
••• 574 
