1446 
came to a phase-difference of about fourteen months, whereas we 
found only two months. 
3. Charts 3 and 4 contain a representation of the synchronous 
mean-pressure distribution and temperature deviations for the months 
December—February, for 1902 and 1910 when it was “too warm” 
on the continent; charts 5 and 6 for 1901 and 1909 when the 
winter was “too cold”. 
Horizontal shading represents positive temperature-deviations, 
negative deviations are represented by vertical shading; the widest 
shading agrees with deviations ranging between O° and 1°.4, the 
narrower respectively with deviations from 1°.5 to 2°.9, 3°.0 to 
4°.4 and from 4°.5 to 6°.0 C. 
Attention is drawn by the following points in the isobaric-charts : 
a. relative low pressure in warm, relative high pressure in cold 
winters ; 
6. the centre of the Iceland-Faréer-low lies in warm winters to 
the east, in cold winters to the west of Iceland ; 
c. a rather sudden bend in the isobars in warm winters, which 
fails in cold ones; 
d. the isobars are less undulated in warm winters than in cold 
ones ; 
e. to the north of the Iceland-Faréer-low the region of high 
pressure appears in warm winters, it fails in cold ones. 
These facts generally govern the temperature-distribution. 
4. Finally the degree of certainty of a prognostication about 
the character of the coming winter has to be examined and also how 
far our knowledge about the numerical value of the standard-devi- 
ation is improved. 
The greatest success of a prognostication may be expected for a 
region where the largest correlation-factor was found; in our case 
the part of Germany containing the stations Berlin, Görlitz, Posen, 
and Ratibor (Table II, chart 2). For the temperature-deviations 
over the period 1899/1900—1914/1915 with regard to the mean 
value over these 16 years and for those of the Trade winds 
over the months June—November the following values were found, 
they are given in hundredths of a degree Celsius and of Beaufort- 
units. 
