222 Evaporation from the Surface of the Sea and the Water Budget of the Earth 



1919). The mean values of the gradients used by Wiist are, however, obtained from 

 relatively few observations but appear, as confirmed by later observations, to be of 

 the correct order of magnitude (Shouleikin, 1928, Montgomery, 1936-7/8, Bruch 

 1940). 

 The total reduction factor for a conversion of this kind amounts to 



k = 0-48 ± 008. 



It is seen that the actual evaporation at the surface of the sea is on the average somewhat 

 less than half of the true evaporation measured by an evaporation vessel on board ship. 

 In this way Wiist obtained for the North Atlantic, for example, the following mean 

 evaporation heights (Table 84) for average meteorological conditions. 



Table 84. Evaporation in the Atlantic 



This table also gives some idea of the values measured by an evaporation vessel in 

 different climatic zones and of the meridional distribution of the evaporation amounts 

 over the Atlantic Ocean. The last column on the right gives values obtained by 

 LuTGENS (1911) from his excellent measurements of evaporation; due to unsuitable 

 correction, however, the latitudinal differences are overestimated, especially the evapor- 

 ation amount in the trade regions, relative to that in the doldrums. The total procedure 

 of a direct redaction of the observed evaporation on board a moving ship suggested 

 by Wiist was later again controlled by Cherubim (1931), and he found, after applica- 

 tion of some refined but not very important corrections, a reduction factor of 0-54 

 which, however, he multiplied by 1 -08 in order to account for the influence of the 

 motion of the sea giving the final value 0-583. This latter increase in the size of the 

 correction factor by about 8% for the motion of the sea, for which there was no ob- 

 servational evidence, was regarded by WiJST (1936) as unsuitable since there were other 

 factors, some acting in an opposite direction which had not been taken into account 

 and of which the magnitude was equally unknown. The uncertainties of the direct 

 correction are certainly rather large but if the value obtained by Cherubim is taken 

 as a maximum and that obtained by Wiist as a minimum then a mean of 053 can be 

 taken at the present time as the most probable correction factor. 



