166 ANNALS NEW YORK ACADEMY OF SCIENCES 



of observations. But the curve of St. Lucia is a transitional form be- 

 tween those of Fort-de-France and Barbados, and Para, in Brazil, 1° 

 27' S. lat., shows, between 1902 and 1905, exactly the same abnormal 

 depression as Barbados. A very accentuated depression between 1903 and 

 1904 is also characteristic for Arequipa and Mauritius^® as well as St. 

 Helena. ^^ The temperature curve of Apia, Samoa Island, displays the 

 same very pronounced antipleionian depression, corresponding perfectly 

 with that of Barbados. It is difficult, therefore, to avoid the conclusion 

 that the curve of Barbados must be correct and that if it differs very 

 greatly from the curves of the other West Indian stations this fact may 

 simply be considered as an interesting subject for special investigation. 



For m}^ present purpose the Barbados curve is a very valuable argu- 

 ment against a possible supposition in favor of the volcanic dust hypoth- 

 esis. One could imagine that if the Fort-de-France temperature records 

 as well as those of some other stations of the West Indies have not been 

 very much affected by the eruption of Mt. Pele and the other volcanoes, 

 it may be because the dust veils have been carried into the southern 

 hemisphere, and that there the curves of the stations belonging to the 

 belt of the southern trade-winds show the depression just as much, if not 

 more so, than in the case of the Kral^atoa eruption. 



The fact that Cayenne, which lies between Para and Barbados, has a 

 curve resembling those of St. Croix and Port-au-Prince and that these 

 stations belong to the northern trade winds, and the fact that the curve 

 of Barbados is similar to those of the southern stations, show that this 

 explanation of the observed anomaly is inadequate. 



: Temperatures Observed in Alaska 



Assuming that! the volcanic haze produced by the Katmai eruption of 

 June 6, 1912, must have had the greatest effect on the temperatures re- 

 corded in Alaska and in Canada, I will now compare the curves of seven 

 stations in Alaska with the curves of Victoria and Edmonton, Mauritius 

 and Arequipa. 



^ The temperature data for the Alaskan stations were obtained from the 

 United States Weather Bureau, the Arequipa daily observations were 

 sent to me by Professor Edw. C. Pickering and the Mauritius data by 

 Dr. H. E. Mill and by the director of the Mauritius observatory. The 

 observations considered extend over the five years 1909-1913. 



Since Mt. Katmai could not have affected the temperature conditions 

 of Arequipa and Mauritius, it is safe to take the curves of these stations 



s6 Arctowski : Op. cit. Bull. Am. Geog. Soc, vol. 44, p. 598. 

 "Arctowski: Oft. cit. Annals N. Y. Acad. Sci., vol. 24, p. 109. 1014. 



