of Edinburgh, Session 1884 - 85 . 
161 
at VII., VIII., IX., from Inchkeith ten miles seaward ; and under 
4 all farther to the east. 
Table IV. 
Place. 
1 Jan. 
Feh. 
Mar. 
April. 
May. 
June. 
July. 
Aug. 
Sept. 
Oct. 
Nov. 
Dec. 
1 
37-9 
36-5 
41-0 
44-5 
56-7 
57-4 
517 
46T 
2 
38-7 
40-3 
40-7 
43-0 
49-0 
52-7 
57-8 
55-2 
52-2 
48:1 
40-8 
3 
42 9 
51-0 
55-3 
55-6 
54-9 
52-8 
47-6 
42-3 
4 
42-5 
50-9 
54-3 
54-0 
52-8 
48-8 
43-7 
The general result of the observations is that in the landward part 
of the Firth the range of temperature is greater and the period of 
the maximum earlier than farther seaward; and that as the sea is' 
approached the range becomes less and the date of the maximum is 
retarded. Thus in 2 and 3' the maximum occurs in July, in 4 not 
until the end of August. No doubt observations made at Alloa 
would show a winter minimum of 32°, or nearly so, and the summer 
maximum would probably approach 70°. The extreme probable 
range at Alloa is 40° in the year; at Queensferry a range just half 
of that, 20° from 58° to 38°, has been observed; and at the Isle of 
May it is most likely that the annual range will not much exceed 
10°. The extreme temperatures observed near the Isle of May were 
in fact 55° in August and 43° in December. 
The mean temperature for the year appears not to vary more than 
half a i degree from 47° ’5 in any part of the Firth; but an extended 
series of observations is necessary in order to clear up this point. 
In order to show the continuity of the seasons, the temperature 
curves for surface water at North Berwick and at Granton were 
drawn on a new plan. A circle is divided into 12 sectors, one for 
each month, and the temperatures are measured on the radii, con- 
centric circles being drawn for 30°, 40°, 50°, and 60°. The line 
representing the temperature becomes a closed curve, and the regu- 
larity of its form is very striking. The curves on Plate VI. fig. 2,' 
representing the monthly means in the four divisions of the Firth, 
are less regular, in consequence, probably, of the intermittent nature 
of the observations. In Plate VI. fig. 1 the same curves are shown. 
This plan of plotting curves by means of polar instead of 
rectangular co-ordinates will be found to have many advantages in 
