12 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 68 



rise of temperature the pressure falls in the tropics attaining a 

 minimum between the second and third day after the solar change. 

 On the succeeding day the pressure attains a maximum in the temper- 

 ate region and the temperature a minimum. Succeeding this maxi- 

 mum one day, that is, four to five days after the solar changes, is a 

 minimum of pressure in the Arctic circle near the 6oth degree of 

 latitude in both hemispheres, and a maximum of temperature in the 

 oceanic centers of low pressure like that near Iceland. 



This succession of results suggests the flow of air from the tropics 

 to the Arctic circle and brings to the attention of meteorologists a 

 new kind of weather change not heretofore considered in daily 

 forecasting. 



This class of change is entirely different and distinct from the 

 ordinary class of weather changes used in weather forecasting. 



These latter changes usually move from west to east with an 

 equatorial component of motion and are of a very complex nature 

 analogous to a series of waves of different lengths mo\ing with 

 different velocities. (See U. S. Weather Review, Vol. 35, April, 

 19075 Quar. Journal Royal Meteorological Society, Vol. 41, pp. 201- 

 207, July, 191 5 ; Historia y Organizacion del Servicio Meteorologico 

 Argentino, 19 14.) 



A plot of the correlation factors found for the different stations 

 is made on a map of the world in chart 2. In this map the belts of 

 high and low pressure of the earth are shown by colors, the white 

 areas indicating areas of low pressure and the red tints, belts of high 

 pressure. This tinted area is taken from the charts of Buchan for 

 October (Climatological Atlas by Alexander Buchan, Report of the 

 Scientific Results of the Voyage of H. M. S. Challenger) this month 

 being near the middle of the period covered by the observations. 



In making this map there was a choice of plotting the maximum 

 positive or negative correlation at each station or else of plotting the 

 correlation on a given day following the solar changes. Both methods 

 were tried, and do not differ greatly as regards the distribution of the 

 l^ositive and negative values. In chart i the correlation values for 

 temperature are given for the same day at all the stations, namely 

 for the third day following the solar changes. This left the chart 

 free from any personal bias in selection when there were nearly equal 

 positive and negative values. 



Lines were then drawn separating the positive and negative 

 regions. No data were accessible from the central Atlantic Ocean 

 or from central Asia and the lines in these regions are drawn to 

 accord with results from similar regions. 



