OoTuuEu 19, 1883.] 



SCIENCE: 



.V2l 



A SYSTEM OF LOCAL WARNINGS 

 AGAINST TORNADOES. 



I HAVE lately exainiucd with some (.'are the 

 excellent compilation by Sergeant Finley of 

 the signal-service, ' Characteristics of six iiiin- 

 dred tornadoes,' with reference to the question 

 of devising a simple apparatus for saving hu- 

 man life. 



Saving property seems to be out of the ques- 

 tion, as no structure can withstand the force 

 of the tornado-wind. Life may be saved by 

 recourse to underground shelters, cellars, etc., 

 such as actually have been built in many places 

 for this end. 



Two facts may Ije quoted from the work 

 named, — 1°. Three hundred and fortj^-seven 

 out of three hundred and ninety-three torna- 

 does (that is eighty-eight per cent) originated 

 between the west and the south-south-west 

 points ; 2°. The average velocity- of progres- 

 sion was about one mile in two minutes. 



From what we already know of the atmos- 

 pheric conditions necessary to the production 

 of tornadoes, it seems probable that in the 

 future it may be practicable for the general 

 weather service in Washington to send out 

 warnings a day in advance to large regions 

 of countrv within which tornadoes arc likely 

 to occur. These warnings would necessarily 

 be of a very general nature. The\- would 

 simpl}' state that the conditions were such on 

 two sides of a large region (like the state of 

 Wisconsin, for example) as to make it probable 

 that tornadoes would occur somewhere inside 

 that region within tweuty-four hours. The 

 local weather services of states like Ohio and 

 Iowa could, perhaps, make these predictions a 

 little more specific ; but there is no prospect 

 whatever that warnings of anj- particular tor- 

 nado can be given in the immediate future. 

 It can be said, that, within a district five hun- 

 dred miles square, tornadoes are likely to occur 

 within twenty-four hours, and such a warn- 

 ing would be of value ; but it does not seem 

 to be probable that it can be said that a par- 

 ticular thirty miles square of this region is 

 in particular danger. Under these circum- 

 stances, it is of interest and importance to in- 

 quire whether some efficient method of local 

 warnings cannot be devised. If five minutes' 

 warning could have been given at anj- of the 

 late tornadoes, many lives might have been 

 saved. If each household could be warned by 

 the continuous ringing of a bell, for example, 

 that a wind of destructive force (say, sevent}' 

 miles per hour and upwards) was approaching, 

 and that five minutes were available in which 



to seek shelter, this w could be well worth do- 

 ing. 



A wind of seventy miles an .hour is suflicient 

 to blow dowu chimneys and to uniouf houses, 

 unless thej' are well built. Ordinary tree.v will 

 not stand under it. Tlie pressure on a sijuirvj'e 

 foot is in the neigliljorhood of fifty i)ounds." 

 There might lie occasions where seventy miles 

 would be the maximum wind-velocity : and the 

 person who had taken refuge in the cellar might 

 be inclined, after the gust had blown over, to 

 find fault with the indicator which had predicted 

 a tornado, when onh" a violent gale occurred. 

 But such storms do not occur as often as once 

 a year ; and it would seem that one could afford 

 to be frightened as frequenth' as this for the 

 sake of immunitj- from an occasional tornado, 

 which might be following in the track of such 

 a \ ioleut gale. 



1 have found that it is practicable to erect, 

 at a moderate expense (less than five hundred 

 dollars) , an apparatus which would give from 

 three to five minutes' warning to all the inhabit- 

 ants of a small town, by the firing of a cannon, 

 for instance ; and in addition, and without any 

 increased expense, this apparatus could ring a 

 bell in every house. The additional expense 

 to each house would be less than ten dollars, 

 the cost of maintenance would be less than a 

 hundred dollars a year, and the work could 

 1)6 <lone by any intelligent person. The sys- 

 tem, for a small town, would be something 

 like the following : suppose a circle described 

 about the town with a radius of from two to 

 two and one-half miles. The only serious 

 danger from tornadoes is to be feared from the 

 part of this circle between the west point and 

 the south-south-west point. Along the cir- 

 cumference of this circle, between the south- 

 south-west and west points, run a line of single 

 telegraph-wire on twentj' posts to the mile, 

 and from the west point bring the wire into the 

 town, letting it end at the telegraph-office. 

 It is grounded at each end of the line, and at 

 the telegraph-office it is connected with a bat- 

 ter}-, which sends a constant current over the 

 line. Within the town, connection is made in 

 various houses with magnets. Each magnet 

 holds a detent, which prevents a bell from be- 

 ing rung bj- the action of a cheap clock-work 

 governed by a coiled spring. If the circuit is 

 broken anywhere in the line, eacii l)ell begins 

 to ring, and continues to sound till its spring 

 is run down ; for four or five minutes, for ex- 

 ample. A cannon could be fireil by a simple 

 device, which would warn persons in the fields, 

 etc., to seek shelter. In a large town the 

 circuit mijjht end in one of the engine-houses 



