- measurement of Snowdon by the Thermomctrical Barometer. 301 
cooling, the mercury was drawn into the tube from the cap 
at the top, and the previous measure from Gregory’s was 
therefore lost. I mention this, to show how the measure 
was afterwards recovered on descending. 
At this point which I call H, the half height, the thermo- 
metrical barometer was set at 792, in the manner directed in 
my former paper ; thermometer 62°. When boiled again on 
the summit it stood at 86, thermometer 66°. On descending 
again to H, it boiled at 785, thermometer 63°. From this 
point, my measurement to Gregory’s downward, was to be 
made for recovering the accidental loss in the ascent ; and 
for that purpose an operation was to be performed on the 
instrument, by no means so easy as that for taking a suc- 
ceeding stage upward ; for the mercury must in this case be 
expanded out of the tube into the cap over the flame of the 
lamp, until on boiling, it stands as near as may be requisite 
to the bottom of the scale. This was done at H, and the 
water boiled at 14,6, thermometer 63°. On carrying it down 
to the Inn it stood at 695, thermometer 65°. The barometer 
observed at Carnarvon about the time of boiling at H, on 
ascending, was at 29,91 ; and at the time of boiling there 
the second time, it stood at 29,88, where I found it on my 
return to Carnarvon. 
Between Gregory’s and H, the difference 695 — 146 = 549, 
mean thermometer 64°, barometer 29,9. Between H and the 
summit, the difference ( from the mean of the two observa- 
tions at H Z2i±zli—J 788,5 — 86 = 702,5, mean thermo- 
meter 64°. The mean thermometer being the same in both 
lengths, they may be calculated together, the total being 
549 + 702,5 = i25i,5, thermometer 64°. The barometer 
Rr 
MDCCCXX. 
