10 



SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 'J'J 



December, 1924. It is evident that by calc-ulating the con-elation for 

 each week independently, all seasonal influences v^ill be eliminated 

 from the results. 



Fig. 3. — The dotted lines show predicted temperature and rain in the Weekly 

 Weather Forecast for six weeks from May 29 to July 9, 1924, and the lines in 

 full show the observed temperature and rain for the same period. 



July 1924 

 10 11 12 13 14 15 16 17 13 19 20 21 22 23 24 25 26 27 2S 29 30 



Predicted teanperaturo and rain 

 Observed tccporaturo and rain 



Fig 4. — The dotted lines show predicted temperature and rain in the Weekly 

 Weather Forecast for six weeks from July lo to August 20, 1924, and the lines 

 in full show the observed temperature and rain for the same period. 



Figures 3 and 4 each give six weekly forecasts, from May 29 

 to August 20, 1924. The dotted lines show predicted temperature 

 and the full lines the observed temperature. Underneath the tem- 



