24 



SMITHSONIAN MISCELLANEOUS COLLECTIONS 



VOL. 71 



Extending the correlations of the means of 14 minus 28 days to the 

 first half of 191 6, a maximum correlation is found at 36 days. This 

 result could be explained by a movement of the prominences to higher 

 solar latitudes. In order to study the question more in detail the year 



Table VIII. — Correlations Between the Analyzed Temperatures of Buenos Aires for 



Successive Days 



1915 was divided into 14 periods of 26 days each and correlations for 

 one day later, tv\ro days later, etc., were found for each group of 

 values up to 40 days. These are shown in table VIII under the head- 

 ing I to 2, 2 to 3, etc., and are plotted in figure 6 with numbers 

 I, 2, 3, 4, etc. 



