IN THE HALL OF THE ROYAL SOCIETY. 
573 
formation of the mucilaginous-looking matter had opened a permeable com¬ 
munication between the water in the boiler and the atmosphere ; by which not 
only the water was carried off by evaporation, which would account for the 
deficiency, but the air passed in and was absorbed: and we have little doubt 
that if the stratum of oil had been thicker, the change would have been con¬ 
fined to the lower surface, and the water would have been perfectly protected, 
as was the portion set aside in the glass. 
I shall now proceed to notice two or three more circumstances of interest, 
which I remarked during my observation of the water-barometer. 
It is extremely curious to watch its action in windy weather ; the column 
of water appears to be in a perpetual motion, resembling the slow action of 
respiration. During a heavy gale of wind on the 16th of November 1830, I 
made the following observations : 
Time. 
Thermometers. 
Water-. 
Barometer. 
Mercurial 
Barometer. 
Intern. 
Extern. 
h 
m 
Inches. 
Inches. 
2 
30 
56 
55*5 
387-87 
29-092 
2 
45 
387 59 
29-090 
3 
0 
387-44 
29-090 
3 
15 
387*28 
29-090 
4 
0 
387-64 
29-090 
4 
15 
387-85 
29-090 
About half-past two, the maximum range of the oscillations was about 0‘28 
inch ; about half an hour later, one gust of wind caused an oscillation of 0’43 
inch, and the minor oscillations were generally nearer the lower than the 
higher extreme. At four o’clock the movement became sensibly less in extent, 
and the mean point of the oscillations began to rise, and, as I ventured to pre¬ 
dict, the wind very soon began to abate. It became very suddenly calm, and 
the next day was very fine. The time of this change, as indicated by the in¬ 
strument, was certain within five minutes. 
On the subjoined scale (Plate XX.) I have laid down the hourly observa¬ 
tions of Mr. Hudson of the water and mercurial barometers obligingly com¬ 
municated to me by that gentleman. They have not been corrected ; but the 
corrections would be of little importance in the rough comparison which I at 
present design to institute. A very slight examination will show that there 
4 E 
MDCCCXXXII. 
