136 ANNUAL REPORT SMITHSONIAN INSTITUTION, 19 3 5 



earth are controlled from outside there will be close parallelism 

 between these daily and weekly pressures. It was found that be- 

 tween 31 daily contemporary pressures at Honolulu and Batavia 

 the coefficient was —0.12, which is negligible; between 39 weekly 

 ones it was +0.10; between 47 monthly June pressures it was —0.12; 

 and between the pressures of 47 three-monthly seasons of June to 

 August it was —0.46. Between Samoa and Batavia December 

 pressures the coefficient was —0,38, and for the season December to 

 February it was —0.60. Thus it is between the characteristics that 

 persist over months, not over days or weeks, that relationships exist. 



Being forced off short-lived phenomena we search for an expla- 

 nation in terms of slowly changing features, such as ocean tempera- 

 tures; and the big variations from year to year in the amount of 

 pack ice in the antarctic seas forces itself on our attention. But 

 here the reports of 12 years from the South Orkneys yield a rela- 

 tionship of only 0.32 with the southern fluctuation, instead of about 

 0.9, as we should want in a prime cause; and the variations at the 

 South Orkneys come after rather than before those of the southern 

 oscillation. The biggest ocean region is the Pacific, and as an index 

 of its seasonal water temperature we may use the corresponding 

 air temperature of Samoa, which shows a greater persistence than 

 any factor in the world as yet examined; the relationship between 

 its summer and autumn values is as large as 0.94. But unluckily the 

 correlation coefficients show clearly that it is mainly the southern 

 fluctuation in winter that controls the Samoa temperature. Thus, 

 a short-cut to the explanation of our fundamental problems seems 

 as far away as ever. Our three big fluctuations each form a system 

 of changes which are apparently held together by meteorological 

 links ; and there is, in my opinion, as yet no satisfactory proof of any 

 free periods associated with them. 



Let us now consider in what direction new developments seem 

 likely. A moment's reflection will convince us that in view of the 

 variations of rainfall over large areas, such as Brazil and Central 

 Africa, which are scarcely affected by the three big fluctuations, 

 there must be others, some of which are probably on a big scale. For 

 example, we should, on the analogy of the northern oceans, expect a 

 fluctuation of pressure between the antarctic low-pressure belt and 

 the high-pressure belt of 30° S. We are at once reminded of the 

 marked opposition which Simpson found during the short period of 

 4 years for which data were available between pressure at McMurdo 

 Sound and that in a belt round the earth extending from about 25° 

 S. to about 50° S. All students of this subject have found it natural 

 to regard the fluctuations in the amount of pack ice in the antarctic 

 seas as likely to control sea and therefore air temperatures over 



