ARCT0W8EI, CHANGES IN DISTRIBUTION OF TEMPERATURE 103 



otherwise. Since the pleions displace themselves, the crests of the curves 

 cannot occur simultaneously everywhere, and since it has not been possi- 

 ble to detect important and persistent seesaw centers, it seems a priori 

 very improbable that the direct and inverse types of variation could be 

 characteristic for certain given locations. Therefore, there is no fixed 

 location for the inverse or compensating type. 



An inspection of the consecutive curves shows, however, that in Xew 

 Mexico, Arizona and Southern California the existence of a center, where 

 the variation displays a striking preference to belong to the inverse type, 

 may be suspected, and that, on the contrary, in Pennsylvania and Oregon, 

 the direct type must be predominant. 



Are these locations centers of origin of pleionian variations? Not 

 necessarily. Besides, the records of only ten years of observations are 

 insufficient to give definite results. Even such as it is, the result gained 

 leads to further investigations in regard to the question of the mode of 

 formation of pleions in situ. 



Leaving this question as an unsolved problem, I will pass to another 

 most puzzling subject. 



Since, for certain parts of the United States, the consecutive tempera- 

 ture curves belong to the direct type, — that is to say, are similar and 

 coincide more or less in time with the equatorial curves, — the impulse 

 producing these variations must be the same as that which produces the 

 tropical variations. This impulse is evidently extra-terrestrial. There- 

 fore, where the variation is direct, the departures of temperature vnll not 

 be due to abnormal conditions of atmospheric circulation but will, on the 

 contrary, produce such changes of atmospheric pressure, wind direction 

 and velocity, etc., as may be characteristic for pleions or antipleions. On 

 the maps the pleions do not disappear : they move away. 



Now the question is how — in a direct type of variation — the pleion 

 corresponding to the second crest of the consecutive curve is renewed. 

 Is it the same pleion coming back from the region it was pushed away 

 from by the formation in situ of the direct antipleion, or is it a new 

 pleion, and if so what became of the first one ? 



Let us call the pleionian crests of the Arequipa curve A, B, C and D 

 (Fig. 1). The consecutive maps show that the crest B of New York 

 went northwest over Canada and then southwest towards California. The 

 pleion came back nearly the same way during 1904-1906. The crest C 

 of New York is therefore the same as B ; but, if we try to follow this 

 pendulation on the consecutive curves of individual stations, we do not 

 succeed very well. This is because, as has been shown in Fig. 52, the 

 amplitude of the departures changes independently of the pendulation. 



