The h was to take the daily differences between the maximum and 
lain!! temperatures, for daily range. Next to deduce from each of 
ege jroximate true mean daily values, by the application of 
rrecti is as calculated from my Tables of Diurnal Range. 
The 1 erections to be applied to the mean of the three daily observa- 
,ns v e as follows (dependent on the time of the year, the times of 
serv; m, and the daily range of temperature) :— 
o 
O 
[n h try 
when the daily range 
was 7 
the correction 
was 0*1 
subtracts 
it 
30 
a 
0-7 
,, 
„ F( lary 
a 
„ 
7 
a 
0-1 
additive 
11 
it 
36 
a 
0-4 
I M h 
11 
1 
,, 
o-i 
J) 
a 
it 
10 
0-7 
1 
11 
20 
)9 
1-4 
if 
a 
30 
a 
2-0 
,, 
„ A . 
1 
,, 
o-i 
a 
10 
11 
0-7 
,, 
! 
y j 
>9 
20 
) t 
1-3 
j 
J9 
30 
11 
1-9 
,, 
40 
2-5 
Tm 
a 
a 
2 
0-1 
H 1 
- y 
it 
20 
,, 
0-6 
1 
30 
») 
0-8 
, J; 
if 
a 
10 
y > 
01 
,, 
„ 
i i 
a 
40 
0-4 
„ J, 
11 
a 
6 
• ,, 
o-i 
» | 
a 
a 
40 
06 
,, 
„ A ast 
a 
1 
, , 
o-i 
„ 1 
a 
10 
0-4 
” 1 
a 
20 
a 
0-7 
,, 
y9 
11 
30 
it 
11 
,, 
a 
a 
40 
a 
1-5 
i. {jhember 
a 
2 
99 
01 
,, 
» it 
tr 
10 
0-5 
M } ft 
a 
a 
20 
1-1 
,, 
„ 1 „ 
a 
a 
30 
1-7 
„ 
i, 1 „ 
a 
t i 
40 
2-2 
,, 
i> ^ber 
j j 
a 
1 
a 
0-1 
,, 
, 
a 
a 
40 
t i 
0-6 
'it lr. and Dec. „ 
11 
1 
i t 
0-1 
subtractiv* 
” 1 
. 
a 
30 
» 
0-4 
a 
lies< 
iser 
le? 
'as r 
lade 
feqtij 
corrections were applied daily to the mean of the three daily 
tions to determine the true mean daily temperature throughout 
)le times, excepting the three years when the centigrade scale 
nrtunately used. I say unfortunately, for the observations were 
o the whole degree only, and thus the error of reading was 
% as large as 1° Fah. A mean correction belonging to each 
