CLIMATOLOGICAL AVERAGES VON BEZOLD 419 



make this also prominent in the tables the more important extreme 

 values are set in heavy-faced type. 



In the diagram also the ordinates of the first three elements Avhich 

 are in their nature positive have been plotted positively upward, 

 while the two others are plotted positively downward [as shown 

 by the scales at the sides of the diagrams]. Finally the scale of 

 ordinates is so chosen that the relationship between the neighbor- 

 ing curves strikes the eye at once. 



No further explanations of table i seem necessary. It need 

 only be added that the insolation D is expressed in multiples of 

 that for an average day at the equator; the barometric pressure 

 (/?) is in millimeters; the temperature of the air (t) in Centigrade 

 degrees; the cloudiness (n) is in percentages and the rainfall or 

 equivalent depth of melted snow (m) is centimeters. 



The sines of the geographic latitude are used as abscissae, as 

 already mentioned in the introduction. The semicircle at the 

 base of the figure can be considered as one-half of an orthogonal 

 projection of a diminutive globe whose axis 5. N. lies horizontally 

 and on which parallels of latitude are drawn for each io°. 



Hence it suffices to prolong any ordinate of this figure down to 

 the periphery of this semicircle in order to at once perceive the 

 latitude to which it belongs. This figure also elucidates the method 

 by which the interpolation was carried out. 



The numbers on the side of the diagram give the values of the 

 ordinates for the individual curves. Their significance is easy 

 to understand by means of the attached letters and symbols, so 

 that the connection between these numbers and the accompanying 

 curves is seen at once. 



Moreover, the portion of the network projecting above the coor- 

 dinates and leading up to the inscribed numerals is drawn in like 

 the corresponding curves, but rather feebler. 



If now we consider these curves we at once derive the very com- 

 forting impression that our knowledge of the distribution of the 

 most important meteorological elements is far more complete than 

 would be suspected from the ordinary method of presentation. 



That portion of the polar regions for which the averages of tem- 

 perature and pressure for whole circles of latitude can only be 

 formed by bold extrapolations amounts to scarcely one-tenth of 

 the whole surface of the earth, and even for cloudiness and precipi- 

 tation this is true to very nearly the same extent, at least as c< mcerns 

 the principal features of this distribution in latitude. 



Moreover, we see from the tables and especially from the diagrams 



