CONTRIBUTIONS TO METEOROLOGY, 



29 



is accoiiipauied by a change of only 9|o in the average direction of the ])rogTess of storms. Tliis 

 fact clearly indicates that the direction in which storms advance is mainly determined by some 

 other canse than the mean direction of the wind. 



41. 1 next endeavored to ascertain wliat was the prevalent direction of the wind wliich ])re- 

 cedcd each of the storms referred to in Table IX, and also the prevalent wind which sncceeded 

 the low center. The observations pnblisiied in the United States International Bulletin include 

 only two stations within the tropics in the neighborhood of the China Sea, via, Manila and 

 Tnguegarao, both of them on the island of Luzon, one of the Philippine Islands (see Plate V), and 

 the observations at the latter station commence witii the year 18SL The following table shows 

 the observations at these two stations for all those storms of 1881 which passed near enough to 

 either of these stations to canse a decided fall of the barometer. The observations include (I) The 

 barometer in English inches reduced to sealevel; (2) the thermometer (Fahrenheit); (3) the rela- 

 tive humidity; (4) the wind's direction; (5) the wind's velocity in miles per h> ur; and (6) the rain- 

 fall, in English inches, during the preceding twenty-four hours : 



Table XI. — Obnervations near the time of cyclones. 



