of Edinhurgh, Session 1883-84. 
931 
The improved thermometer having been constructed and tested 
in the interval, a set of hourly observations was made at the three 
stations during a period of twenty-five hours, from 9 a.m. on July 3rd 
to 10 A.M. on July 4th. The temperature was taken at the surface 
and the bottom as on previous occasions, and also at a point midway 
between the two. 
The third was a hat clear day, but the sky clouded in the evening, 
and rain fell very heavily during the entire night. A severe 
thunderstorm was experienced from 11 p.m. on the 3rd to 4.30 a.m. 
on the 4th. The continual dazzling produced by the lightning 
flashes made it difficult to read the thermometers with accuracy, but 
there is reason to believe that no serious error was made, and the 
observations were carried on regularly. 
On the third the surface temperature was high, and the interme- 
diate curve remained close to the bottom one, until 5 p.m,, when it 
rose rapidly. During the night and on the forenoon of the 4th, 
the temperature varied but little, and the curves for the surface, 
bottom, and half depth interlace each other curiously. A distinct 
rise of the bottom temperature, and fall of that at the surface, marks 
the inflow of the tide at 8 p.m. The means for this set are — 
Air, 
Surface 
Half-depth 
Bottom 
19 Hours Daylight. 
6 Hours Darkness. 
For 25 Hours. 
57-6 
55-2 
57-0 
Entrance, 
57-95 
57-36 
57-78 
Ark, 
57-73 
.57-35 
57-65 
N.W. corner. 
58-10 
57-41 
57-90 
Entrance, 
57-59 
57-48 
57-56 
Ark, 
57-31 
57-47 
57-35 
KW. corner, 
57-23 
57-43 
57-29 
Entrance, 
57-19 
57-49 
57-27 
Ark, . 
57-20 
57-30 
57-22 
N.W. corner. 
56-93 
56-83 
56-92 
In order to eliminate the tidal effect as much as possible, a mean 
was taken of the hourly results of the first and second series ; the 
tides, being at opposite phases at the same hour, annul each other 
to a certain extent. The surface temperature curve in the 
diagram, representing the result of tidal elimination, follows the air 
curve very closely; the bottom curve also does so, though to a 
much slighter extent, and the phase is clearly retarded. The 
accurate curves embodying all the results have been laid on the 
table. 
