420 



SMITHSONIAN MISCELLANEOUS COLLECTIONS 



VOL. 51 



/o as a.6 at 02 o 0.2 <> 1 06 a* 10 



that the distribution of the most important meteorological elements, 

 which of course depends primarily on the insolation, is modified by 

 the distribution of atmospheric pressure, and this is the most 

 important matter. 



The curve of average temperatures has the greatest similarity 

 with that of the theoretical sums of insolation, provided the scale 

 of ordinates be properly chosen. 



But whereas the insolation by its' very nature is accurately sym- 

 metrical on both sides of the equator with a maximum at that 



circle, the maximum 

 of the temperature 

 curve is pushed into 

 the northern hemis- 

 phere. At the same 

 time a second sym- 

 metrically located but 

 much feebler maxi- 

 mum in the southern 

 hemisphere is indi- 

 cated by the change 

 in the differences, i. 

 e., in the differential 

 quotients. 



This peculiarity in 

 the curve of average 

 temperatures for 

 whole circles of lati- 

 tude is still more strik- 

 ing when we seek to 

 approximately elimi- 

 nate the influence of 

 the unequal distribu- 

 tion of water and land 

 in the two hemispheres by uniting into one arithmetical mean the 

 two values corresponding to equal north and south latitudes. 



By this process which has already been once used by my ad vice, 

 by E. Sella, 4 we obtain average values that, as already remarked, 

 may be designated as "holospheric" in contradistinction to the 

 ordinary "hemispheric, " which apply only to the latitude circles 

 of either hemisphere alone. 



FIG. 60. HEMISPHERIC AVERAGES. 



* See Met. Zeit. XIII, pp. 161-166, 1876. 



