ARCT0W8KI, CHANOEH IN DISTRIBUTION OF TEMPERATURE 43 



The punctuated lines show this variation, which is absolutely distinct 

 from the pleionian variation and may have a period of 18, 19 or any 

 other number of years. The length of time makes no difference, the only 

 important fact being that such a long range or macrochronic variation 

 exists. By eliminating the effect of the pleionian variation, we obtain 

 macropleionian crests and macromeionian depressions, but it is evident 

 that for that purpose a long series of meteorological records is necessary. 



Fig. 1. — Diaf/rumn of conseciitiie ainiiial ami consecutive monthly means of temperature 



at Arequipa 



On the other hand, the second curve of the diagram shows the succes- 

 sion of consecutive monthly means for the years 1906 and 1907 and dis- 

 plays brachypleions. In the case of Arequipa, the brachypleionian crests 

 have a mean period of about 55 days. 



If we were absolutely certain that the pleionian variations of the equa- 

 torial regions were due to changes of the energy radiated by the sun, it 



Herbert H. Kimball: "Solar radiation, atmospheric absorption, and siiy polarization, 

 at Washington, D. C." (Bull. Mount Weather Observatory, Vol. 3, p. 69. 1910.) 



: "The effect upon atmospheric transparency of the eruption of Katmai Vol- 

 cano." (Month. Weather Review, Vol. 41, p. 153. 191.3.) 



: "A return to normal atmospheric transparency." (Journ. Washington Acad. 



Sciences, Vol. 4, p. 17. 1914.) 



W. J. Humphreys : "Volcanic dust and other factors in the production of climatic 

 changes, and their possible relation to ice ages." (Bull. Mount Weather Observatory, 

 Vol. 6, p. 1. 191.3.) 



