15S 



AXXAL.i XEW YORE ACADEMY OF SCIEXCE^ 



terral between the crests C and E, of nearly six years or the double of 

 the interval generally obseiTed. cannot be considered as an argument in 

 favor of a correlation between the temperature depression K and the 

 eruptions of Krakaioa and St. Augustin. But the drop of temperature 

 for the mean of Xovember. 1882. -October. 18S3; coincides with the ap- 

 pearance of the Krakatoa veil in the States and also with a possible 

 influence of the St. Augustin eruption, so that a simple chance circum- 

 stance can hardly be admitted. Then the curve starts up. Just as it went 

 up before the consecutive mean ending with October, 1883, and this 

 tendency is maintained for a few months. The small crest d in the 

 depression K seems to be another detail showing clearly that this de- 

 pression is abnormal, that it is not a true antipleionian depression. 



Besides, on the table of monthly departures, we notice the exceptional 

 iigure — 5.1 ^F. for October, 1883, and the succession of negative de- 

 partures for the year 1881. 



In order to show that the depression K was really abnormal and that 

 Tinder normal conditions we ought to have had a pleionian crest (D) m 

 place of this depression, I will consider now the consecutive temperature 

 curve of Port Darwin, a station of Xorth Australia, 12^ 28' lat. S. 



Tempeeatuezs at Poet DaPwWix axd some other Stations 



In Table II the monthly temperatures for the years 1880-89 are given 

 in departures from the means figuring on top of the table. *^ The dia- 

 gram (Fig. 2) represents the succession of consecutive annual means. 



Fig. 2. — Thermopleionian variation of Port Daricin 



A comparison of Figures 1 and 2 shows striking similarities between 

 the two curves. I marked, therefore, the crests of the Port Darwin curve 



<^- Meteor. Observat. made at the Adelaide Observatory. . .during the year ISSO. . . ISSO. 



