ARCT0W8KI, CHANGED IN DISTRIBUTION OF TEMPERATURE 41 



of the curves for many other places are reproduced in tin's memoir. Two 

 facts of some importance are clearly demonstrated by all these curves: 



1) Although there are some most interesting agreements with tlie curve 

 expressing tlie sun spot cycle, this cycle cannot be considered as being the 

 main factor producing the pleionian variations, simply because the pleionian 

 crests and depressions of the temperature curves repeat themselves three to 

 four times more often than the maxima and minima of the solar curve. 



2) The temperature curves for distant stations, belonging to absolutely 

 different climates, present in certain cases such striking coincidences that it 

 is impossible to ascribe them to simple chance circumstances. I may add that 

 the consecutive curves of rainfall, of sunshine records and of atmospheric 

 pressure display also pleionian variations and demonstrate the fact that we 

 have to deal with more or less periodical alterations of the atmospheric circu- 

 lation. 



In order to make comparisons, a standard curve was necessary. The 

 records of the exceptionally undisturbed climate of Arequipa, in Peru, 

 gave this necessary standard. ° 



The consecutive temperature curve of Arequipa, for the years 1900- 

 1910, shows four characteristic crests and four depressions. The curve 

 of Bulawayo, in Ehodesia, is absolutely similar to the Arequipa curve. 

 The same may be said about the curve obtained from the Mauritius 

 observatory' tigures. and for Tananarive, Madagascar. Batavia, Java, dis- 

 plays also an indisputable resemblance with the Arequipa curve. Xorth 

 of the equator, Havana gives a similar curve, but the data of San Juan, 

 Porto Eico, give a slightly retarded curve, and this is a most interesting 

 fact. Indeed, the pleionian crests of Porto Rico could not be retarded if 

 these temperature anomalies did not have a tendency to persist combined 

 with a tendency of displacement. 



The question, therefore, was whether all pleionian variations observed 

 all over the world were not in immediate correlation with the Arequipa 

 variation. Together with this question, it was necessarv to solve the 

 problem of the displacement of pleions. 



Long-range Variations of Temperature 



The waves expressing the changes of temperature are of different 

 amplitude and different length. If we take monthly means of tempera- 

 ture, we do not take into consideration the groups of cold and hot waves 

 which characterize the changes of weather, and we eliminate also the 

 short diurnal waves of the more or less regular daily variation. If we 



*Heneyk Aectowski: The "Solar Constant" and the Variations of Atmospheric Tem- 

 perature at Arequipa and Some Other Stations. Am. Geog. Soc. Bull., Vol. 44, p. 598. 



