828 Proceedings of the Royal Society 
of the steam was entirely absorbed by the water going to supply the 
boiler, and as this beat was sufficient to raise at least eight times the 
weight of water from the atmospheric temperature to that of ebul- 
lition, it will he seen that we are justified in assuming that the solar 
heat which penetrated the walls of the boiler was wholly utilised in 
making steam, and further, that loss by radiation was reduced to a 
minimum. 
Immediately on presenting the mirror to the sun ebullition com- 
menced, and it was kept going all the forenoon and the greater part 
of the afternoon, the position of the reflector being continually 
adjusted by hand. The distillate was collected in a graduated tube, 
and the time taken as every portion of 5 cc. was collected. The 
locality for experiment was that chosen by the eclipse observers, — 
on the banks of the Nile, close to the Egyptian town of Sohag, in 
latitude 26° 30' N. The instrument was very freely exposed to the 
sky, and was directly shone on by the sun from sunrise till 5 p.m., 
when it was obscured by a wood. As originally constructed the steam 
space was of metal, and was much too small, and, consequently, the 
results obtained in the first few days were vitiated by priming. 
When this had been thoroughly ascertained, the metal covering was 
removed and replaced by an inverted test tube, fitted tight with cork 
packing, the internal steam tube was also prolonged, so as to be well 
clear of the boiling water surface. The effect of this alteration was, 
by increasing the volume of the steam space, to eliminate altogether 
the danger of priming, and, by substituting glass for metal, to enable 
the operation to be constantly watched. The results obtained before 
this alteration have been rejected; those made on the 16th, 17th, 
and 18th May 1882, were subsequent to it, and were quite satisfac- 
tory. The results of these observations are condensed in the follow- 
ing table, which gives the amount of water distilled in an hour at 
different times : — 
