422 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 51 



pressure are separated farther from each other than in the older 

 method, that is to say, we perceive that the zone between these 

 two circles of maximum pressure covers more than one-half of 

 the whole surface of the earth, whereas in the temperature curve 

 the zone between the maxima covers only about three-tenths of 

 the earth's surface. 



Proceeding further in our study of fig. 60 we next consider the 

 curves of mean cloudiness and of precipitation. As these two ele- 

 ments diminish with increasing pressure, therefore, as already 

 mentioned, for the sake of comparability of the curves the ordinates 

 are drawn positive downward. 



This method of collating the different elements reveals the con- 

 nection existing between them in truly surprising manner and 

 clearly brings out the great importance of Ferrel's zones of atmos- 

 pheric pressure. 



Indeed I might go so far as to recommend the considerations here 

 established as the starting point for lectures on Climatology, and as 

 a method of passing gradually from the so-called "solar climate" 

 to the actual local conditions by some scheme that keeps the aver- 

 age values of whole latitude circles plainly in view. 



By always choosing the sine of the latitude as the argument, as 

 has been done here, we at once attain correct ideas as to the weight 

 that is to be attached to the individual values in considering the 

 economy of nature, but independent of course of any uncertainty 

 that may attach to them by virtue of the method of determination 

 and which will gradually diminish. 



Before I go further and draw certain conclusions relative to the 

 temperature curve, I first give at the end of this section the table 

 in accordance with which fig. 61 is drawn, to which I will only add that 

 the numbers interpolated from Spitaler and Batchelder are indicated 

 by 5 and B. 



On examining these numbers one is surprised that the averages 

 of precipitation and cloudiness, resting on rather feeble observa- 

 tional bases, when combined into holospherical averages show a 

 remarkably regular progression. 



After these general remarks which refer only to the presentation 

 in general as also to the average values for the whole year, still 

 further conclusions will be drawn from the whole diagram. 



As already remarked, first of all we see the great similarity 

 that exists between the curves of insolation and temperature as 

 drawn on the scale here adopted, and which at once shows that 

 these two quantities can be connected by one empirical formula at 

 least throughout a considerable portion of their courses. 



