6 H.WALTER 



West Africa) and temperate regions of which I knew the exact chmate 

 (Stuttgart, etc.), it worked, out as well (Figs. 2 and 3). 



The only exception was the temperate steppe chmate with a summer 

 maximum of rainfall. The diagram of Odessa shows a short arid period in 

 the spring and a httle longer arid period in July and August. This was a 

 contradiction as the whole summer m Odessa is dry, though not as dry as 

 in the Mediterranean region. In these cases it proved to be useful to draw 

 an additional rainfall curve with a scale of io°C equal to 30 mm (i : 3) and 

 to distinguish an extreme drought period and a less extreme dry period 

 (Fig. 4, chmatic diagram of Odessa). 



Comparison between the curve of the monthly mean temperature and 

 monthly mean rainfall is allowable because evaporation is mostly pro- 

 portional to temperature. To get the water balance it is necessary to know 

 the ratio of the evaporation to rainfall. As evaporation values are published 

 only for very few stations, temperature values have to be utihsed instead of 

 the missing evaporation values. This is only possible, however, when the 

 temperatures are above o°C, as below o°C the evaporation is negligible. 



Many authors have used the ratio of precipitation to temperature in order 

 to obtain a picture of the water balance (Lang, de Martonne, Koppen, 

 Angstrom, Church and GuefFroy, Thornthwaite, etc.). The formulae they 

 used are, more or less, comphcated. But the ratio of the temperature to 

 rainfall does not much deviate from 1:2 or 1:3 as used by us. For the 

 drawing of thousands of diagrams the ratio must be a simple one, therefore 

 we employ mostly only the ratio i : 2 and distinguish between an extreme 

 drought period and a less extreme dry period if desirable, by employing 

 the ratio 1:3. 



The horizontal extension of the different areas on the diagrams shows 

 us the duration of the periods, the vertical extension, the intensity of 

 humidity or aridity respectively. 



But it is not enough to characterise unfavourable seasons caused by water 

 shortage. To get an idea of the chmate or the chmatic type, we also have 

 to know the unfavourable season, caused by cold. Therefore the months 

 with a mean daily minimum below o°C are indicated by black blocks. This 

 is the average cold season of the year. The months with only an absolute 

 minimum below o°C are hatched (Figs. 1-4). During this season frost 

 was observed, but frost may not occur every year, sometimes only in 

 exceptionally cold years. Some additional chmatic data are provided by 

 figures. Complete diagrams are shown in Figs. 1-4. In the tropics the 

 monthly rainfall may be extremely high and in order to simphfy the dia- 

 gram, the scale for rainfall above 100 mm per month is reduced ten times. 



