E. Loornis — Areas of low pressure. 3 



of the cyclones and continuing for at least 24 hours, we find 

 that the direction of a cyclone's progress accords more nearly 

 with the direction of the principal wind which prevails at the 

 time of the cyclone. 



The average course of the cyclones originating near the China 

 Sea and Bay of Bengal, as enumerated in my fourteenth paper, 

 was 38 degrees north of west, as long as they were moving 

 westerly. The average course of the Asiatic cyclones indicated 

 by the maps accompanying the International Bulletin (oearly 

 all of which originated in the China Sea), was 27-| degrees 

 north of west. Since the average direction of the wind in this 

 region changes nearly 180 degrees from summer to winter, in 

 order to make a satisfactory comparison between the average 

 direction of the wind and that of the progress of storms, we 

 must make a separate comparison for the different seasons of 

 the year. Since nearly all the cyclones, whose tracks are 

 shown on the International charts, originated in the China Sea, 

 I will make the comparison for this region ; and since nearly 

 all of these storms occurred from July to November, I will 

 make the comparison for these months. 



According to Maury's Pilot Chart of the China Sea, the 

 average direction of the winds in this sea for the month of 

 July is south 22° west ; for August it is south 39° west, and 

 for September it is south 39° west. The average for these 

 three months is south 33° west. The average direction of the 

 wind for October is north 53° east, and for November it is 

 north 41° east. The average for these two months is north 

 47° east. For the first three months, the average direction of 

 progress of storms, is 35° north of west, and for the last two 

 months it is 25J° north of west; that is during the first period 

 the average course of storms differs 88 degrees from the average 

 direction of the wind, and during the last period the difference 

 is 68-|- degrees. We also perceive that a change of 166 degrees 

 in the average direction of the wind is accompanied by a 

 change of only 9^ degrees in the average direction of the 

 progress of storms. This fact clearly indicates that the direc- 

 tion in which storms advance is mainly determined by some 

 other cause than the mean direction of the wind. If, however, 

 we make a comparison with the winds immediately succeeding 

 each of the cyclones, and continuing for at least 24 hours, we 

 shall find that the direction of progress of a cyclone corresponds 

 more nearly to that of the principal wind prevailing at the 

 time of the cyclone. It is not, however, claimed that there 

 is an exact agreement between these two directions. 



An examination of the accompanying plate shows that in 

 the middle latitudes of the Northern hemisphere there is a re- 

 markable correspondence between the average direction of the 



