1440 
When considering the table below the largest correlation for the 
whole of Europe — '/, (Germany + Iceland) — will be found between 
fluctuations of the Trade-wind over May—October and those in the 
temperature over December—February. 
When we consider the combined correlationfactors for Germany 
and Iceland — */, (b + c) — with regard to the following periods, 
during which the Trade-wind was taken into account, we get a 
clear insight into the march of the correlation. 
fs ; ; El 
March —Aug. April Sept. May—Oct. | June -Nov. | July - Dec. 
| 
0.4192 | 0.5847 | 0.6673 | 0.6213 0.2884 | 
| 
With the collaboration of the Directions of the steamship companies 
and of our staff of volunteer-observers, the observations about the 
strength of the Trade winds over May—October may be at our 
disposal about medio November and so it is possible to make in time 
a prognostication about the winter to come. From a closer exami- 
nation of the correlation factors in Table 1 we learn that the Trade 
wind observations of April 
September provide an approximation, 
which may be improved when the observations of October are at 
our disposal; for Germany we have to take the Trade-wind obser- 
vations of May or June—October, for Iceland those of May—October. 
For Holland the observations over the periods March-—August, 
April—September or June—November give a good result, for Europe 
we can make use of the mean value of the observations *, (April— 
September + May—October + June—November). 
When taking into account the Trade wind of March or of December, 
the value of the correlation-factor decreases rather fast. 
Table II] contains the correlationfactors between departures in the 
strength of the Northeast Trade over three periods and those of the 
temperatures during the following winter in Europe for 135 stations 
or distriets. They apply to the period 1899/1900—1913/1914. 
In charts 1 and 2 the iso-correlates have been traced, they are 
regular in their progress and they follow more or less the line of 
separation between the low-lying plain and the highlands as in the 
neighbourhood of the Alps, the Carpathian Mountains and the Apen- 
nines; chart 1 applies to the Trade winds of May—October, chart 
2 to those of June—November. 
Positive departures in the strength of the Northeast Trade are 
