1447 
FF DE == 
| 
| Deviations (1900, Ol | 02 | 03 | 04; 05 | 06; 07 (08 | 09 |! 10 | LI | 12) 13 | 14 | 15 | 
| 
| 
“Temperature —103 — 2174123 —30 —33 +30 +93 —200 + 7—240 +193 +100|—27 +87 +53 -+167| 
| Tradewind |— 43 — 69-+ 43 —50 +10 +10 +24 — 23-++10—102 + 19 + 23/+20-+418+56 + 56 
| | 
Thus r= + 0.8476 and f= + 0.0475. 
In the following diagram the full line represents the deviations 
of the Trade wind to their full magnitude; the stroke-dot line gives 
those of the temperature, but reduced to */,, of their value. 
In 14 out of 16 cases the deviations show a corresponding sign, 
and generally it would have been possible in seven out of eight 
cases to make a good forecast about the sign of the departure of 
the coming wintertemperature. 
Let M be the average wintertemperature for the 4 stations over 
the period 1900—1915, the standard-temperature deviation 6,, 6, that 
of the Trade wind. 
From the observations the following numerical values result: 
M= + 0°.027 C., o, = + 1°.295 C., 6, = + 0.438 Beaufort units; 
for the regression-equation we find: 
2, = 2.508 a 
1 = 2 
where ., represents the temperature-deviation, 7, the corresponding 
one of the Trade wind. 
The improved probable error in the prognostication of the tem- 
perature, being: 
Win NDR V(1—r?) tea 2-4 56: 
it is evident, that our knowledge about the value of the standard- 
deviation is improved by 
LV (1?) X 100 = 46.93 °/,. 
