202 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 70 



' TEMPERATURE VARIATIONS IN THE UNITED STATES 



We now go on to consider the curves in figure 74 which give the 

 meteorological relations in different regions of the United States, 

 and we find here in the temperature curves as shown in plate 18-L 

 two types. The one type is represented by the curve for the region 

 on the Pacific coast (curve I) and the other by the curves of the 

 eastern states on the coast of the Alantic Ocean (curve IV) as well as 

 those for the Mexican Gulf (curve II). The temperature curves 

 for the interior states form a transition between these two types 

 of curves and have now the one type, now the other. Where both 

 types simultaneously have minimum or maximum these are particu- 

 larly strongly marked in the transition forms, as for example the 

 minimum in the years 1898 and 1899. This agrees also completely 

 with what Hildebrandsson has pointed out, when he indicates an 

 action sphere along the Pacific coast and another in the eastern 

 states. 



Considering now these curves for the time interval after 1900 

 we find that the curve for the Pacific is of a very individual type of 

 its own while the other type characteristic of the easterly stations 

 on the Atlantic coast is the same as that represented by Batavia and 

 the other tropical stations which we have investigated, including 

 Arequipa. After 1900 there appears a dissimilarity between the 

 curve for the Gulf states and the curve for the Atlantic states, while 

 these two curves for the time interval before 1900 had com- 

 plete agreement. This disagreement for the later period of time 

 is of such a nature that the curve for the Gulf states is similar to 

 the curve for Corpus Christi (see fig. 72, XVII) which was indeed 

 to be expected since this lies on the Gulf, but it also is similar to 

 the curve for Bombay and the other similar curves. 



If Hildebrandsson is right in his conception of the action 

 spheres, we should expect that the curves for the eastern United 

 States along the coast of the Atlantic Ocean as well as the curve 

 for the Gulf states would have similarity with the temperature 

 curves for Scandinavia. Placing these American curves with the 

 curve of the coast temperature along the Norwegian coast, the 

 curve for the air temperature for all Norway, and the air tempera- 

 ture in Stockholm, we find a remarkable agreement. This is plainly 

 shown in figure 75 without more particularly describing it. This 

 indicates that Hildebrandsson is right in his view. In the same 

 figure at the top we have given the twelve-monthly smoothed curve 

 for Liepe's station No. i. As the reader will see, this does not 



