80 STORMS AND WEATHER FORECASTS 



The writer visited St Louis the day after the storm, and was 

 especially impressed with the fact that hundreds of buildings 

 were burst outward at their upper stories, indicating that they 

 were at the time of their destruction near the center of the rotat- 

 ing mass of air, where centrifugal force instantly had reduced 

 the air pressure on the outside to such an extent that the expan- 

 sion of the air in the upper stories of the houses whose windows 

 and doors Avere closed had produced an explosion of the build- 

 ing. In one case all the four walls of the upper story of a house 

 were thrown outward, leaving the lower story intact and the roof 

 resting in proper position one story lower than in the original 

 building. Again, great structures seemed to have been crushed 

 over or taken up bodily and scattered in all directions. 



The fact that this tornado traveled with destructive force 

 through several miles of brick buildings and yet left the city 

 Avith greater force than it possessed on entering it illustrates the 

 futility of planting forests to the southAvest of a city for the pur- 

 pose of protection, as some have advocated. It is probable that 

 the strongest trees would offer but little more resistance to this 

 terrific force than Avould so many blades of grass. 



Whenever the forecast contains the statement that conditions 

 are favorable for severe local storms, it is well for the residents 

 of a city receiving such forecast to observe carefully the forma- 

 tion of portentous clouds and be ready to seek places of safety 

 in the cellars of frame buildings. We have no record of any 

 person having been killed in the cellar of a frame building. 



Chart XVI shows a West India hurricane just making its ad- 

 vent on the Florida coast. A number of stations in the West 

 Indies report to Washington by cable Avhenever hurricanes pass 

 over their region. Sometimes a hurricane composed of a rap- 

 idly revolving eddy of air of only two or three hundred miles in 

 diameter passes between the observation stations on the islands 

 of the West Indies without getting near enough to affect their 

 instruments. Then, if it move rapidly northwest toAvard our 

 Gulf coast, it may reach our seaboard unannounced. Fortu- 

 nately such cases are rare, and in case the storm does reach any 

 ports unexpectedly danger signals Avill be displayed in advance 

 of its coming throughout the remainder of its course until it 

 leaves our shores. At times hurricanes remain several days in 

 the Gulf of Mexico, and the only indication we have of their 

 proximity is a strong suction drawing the air briskly OA'er some 

 of our coast stations toward the center of the Gulf. Again, a 



