32 MR. JOHN GRAHAM ON CONSUMPTION 
power was not numerically determined, but an improve- 
ment in the power of the boiler was most evident. It was 
now possible to go on with it at least, and the boiler was 
supposed at the time to have recovered about 20 per cent. 
in power. 
This result, however, being still unsatisfactory, I had 
the boiler altered as shewn in Fig. 20. 
fig. 20. 
The direct absorbing surface of the boiler was now re- 
duced to about 7 feet across, the fire remaining the same as 
before. Here again an improvement took place equal to 
at least an additional 8 or 10 per cent. as practically esti- 
mated. 
The boiler was employed in this last condition for some 
time, doing the full work previously allotted to it as a 
plain cylindrical boiler. The uniform impression of those 
who attended it was, that the boiler gave a slightly better 
result in this condition than as a plain cylindrical boiler. 
There were, however, practical objections to the boiler in 
this form, the brickwork over the tubes was constantly 
shaking loose, and the tubes themselves became scaled and 
