.io6 



SCIENCE. 



[Vol. XIII. No. 314 



In a frame structure the safest place is in the cellar, but in a 

 trick or stone structure it is the most dangerous. In the former 

 case, if the building is destroyed, it is invariably carried away from 

 the foundation. In the latter case the cellar is filled with debt-is. 



The safest building to construct is one made entirely of wood, 

 with a " barn frame," and not exceeding one story and a half in 

 'height. Where several buildings are connected together in a row, 

 the height may be increased one or more stories with safety. No 

 structure that rises above the earth, however made, can resist the 

 violence of the tornado, and therefore no building is safe as prop- 

 erty, or as a resort to protect life. Under all circumstances, 

 whether in a building or in a cellar, refrain from taking a position 

 in a north-east room, in a north-east corner, in an east room, or 

 against an east wall. 



The tornado-cave offers absolute security to life and limb, and 

 •nothing can replace it for that purpose. This retreat may be con- 

 structed as a cellar-cave or as a " dug-out." In the former case 

 an excavation is made in the west wall of the cellar, on a level with 

 the floor of it, and carried under ground until a sufficient distance 

 is reached to provide comfortable quarters for those who propose 

 to occupy the cave. The overhangmg roof must be supported by 

 heavy timbers, and then arched over with masonry of brick or 

 stone. This extra precaution concerning the roof is necessary to 

 provide against any serious damage to it by falling timbers or heavy 

 masses of debris. The excavation is made into the west wall, be- 

 cause the storm, always approaching from the west, will carry the 

 debris away from that side. But even this favorable position does 

 not preclude the probability of debris being thrown upon the cave 

 by the whirling currents of the vortex. The extra care bestowed 

 upon the roof is money and time well spent. Careful attention 

 should be given to ventilation and drainage, and to making the 

 retreat in every way as convenient and comfortable as possible. 

 The extent to which this suggestion can be complied with will de- 

 pend upon the pecuniary ability of the person concerned. 



The " dug-out " is a tornado-cave, not necessarily connected 

 with any building. The results to be secured are the same as those 

 derived from the use of the " cellar-cave." The cost of a properly 

 constructed tornado-cave, including material and labor, will range 

 according to the quality, character, and strength of the material 

 with which it is built, together with the price of labor, from a 

 hundred and fifty to three hundred and fifty dollars. Such a cave 

 will comfortably accommodate ten persons. 



The rush of air into the tornado's vortex, and therefore the vio- 

 lence of the wind, depends upon the difference of barometric press- 

 ure between the inside and the outside of the storm. This differ- 

 ence has been observed to be nearly three inches, and may be very 

 much more, for observations have never been made in the centre of 

 the vortex. A gradient of three inches, however, will give a velo- 

 city of 323 miles per hour, which will exert a pressure of about 260 

 pounds per square foot against a surface exposed at right angles to 

 the direction of the wind. 



The explosive force of confined air in a tornado is enormously 

 great ; and frequently it is to this energy, rather than to the direct 

 force of the wind, that the destruction of buildings is due. As a 

 tornado-cloud passes over a building, if the air within is confined 

 by closed doors and windows, and cannot readily escape, the ex- 

 plosive force, due to a very great difference in tension between the 

 air inside and outside of the building, bursts asunder the walls, and 

 ■throws the roof upward to a considerable distance. Eye-witnesses 

 state that under such circumstances roofs have been uplifted a 

 <iistance of five hundred feet. Cellar-doors have been blown away 

 from their fastenings in the face of a strong wind coming directly 

 against them, and corks have been blown out of empty bottles by 

 the sudden expansion of the air within them. Many almost in- 

 credible instances of extraordinary violence by the explosive force 

 of tornadoes could be given, but want of space forbids more than 

 this general reference to such manifestations of the tornado's 

 power. 



There is no fact or record to show that an electrical discharge, 

 or any manifestation of atmospheric electricity, ever directly and 

 -entirely destroyed a large stone or frame building ; ever lifted a 

 locomotive from its track ; ever carried an iron bridge from its 

 foundation, and twisted the framework into a shapeless mass ; ever 



rolled a bowlder from its bed in the ground ; ever embedded one 

 piece of timber into another, after having carried the former several 

 hundred yards in the air ; ever carried bedding and clothing in the 

 air for iniles ; ever elevated to considerable heights in the air, 

 columns of water from ponds, lakes, and rivers; ever lifted animals 

 from the earth and carried them over buildings ; ever drew the 

 water from a cistern ; ever twisted a tree from its stump ; ever 

 turned a building upside down, or end for end, without otherwise 

 injuring it. 



Any method of reasoning which assigns tornado-development to 

 planetary influence is, equally with the electrical theory of their 

 origin, without foundation. We have but to realize that in the for- 

 mation of the tornado, and other local storms of a similar charac- 

 ter, the entire action of all the forces involved, even in the energy 

 of the sun's heat, is embraced in that portion of the atmosphere 

 within from two to three miles of the earth's surface. 



Any influence emanating from the movements, conjunctions, or 

 other periodical mutations of, the heavenly bodies, distant hundreds 

 of thousands and millions of miles, can only reach an infinitesimal 

 amount, and is entirely inappreciable in its effects upon the atmos- 

 phere to produce local or general disturbances, especially near the 

 earth. 



It has been asserted that the conditions which give rise to the 

 formation of the tornado-cloud result from the effect upon the at- 

 mosphere of the mere revolution of the planets in their orbits; that 

 the circular movements in the atmosphere are propagated and con- 

 tinued by such influences. The effect is likened to that which 

 would result from the whirling in different directions, in a large 

 vessel of water, of several globes attached to the same spindle. 

 Upon withdrawing the globes, after a number of revolutions, the 

 surface of the water will be found covered with a network of ed- 

 dies. The inherent fault of this simile is the fact, that, while the 

 illustration provides for the circular movement of the bodies within 

 the medium which is set in motion to give the characteristic whirls 

 or eddies, the subject of illustration, the planets, perform their rev- 

 olutions, not in the atmosphere (the medium to be set in motion), 

 but millions of miles away from it, in another medium, concerning 

 which little is known. The failure properly to apply the method of 

 reasoning by analogy often leads the novice into making the most 

 ridiculous assumptions. It would be more reasonable to assume 

 that the revolution of the planets gave rise to the great disturb- 

 ances of the atmosphere, embracing extended regions of country, 

 which are known on the weather-map as " highs " and " lows ; " 

 but even here the same difficulties operate, although not so ex- 

 travagantly as in the case of the tornado, with the narrow path 

 of a hundred yards or more. 



To forecast successfully the time and place of any atmospheric 

 phenomenon is a difficult matter, largely in proportion to the area 

 of country brought under the influence of the particular disturb- 

 ance. There are, of course, other elements which enter into the 

 calculation to render the problem, except under the most favorable 

 circumstances, an extremely complicated one, where the accuracy 

 of results is demanded. It is well known that the tornado has the 

 most circumscribed area of all storms, while its violence has no 

 equal in the entire range of meteorological phenomena. 



A large amount of field-work, and instrumental and general ob- 

 servations extending over many years, relating to the conditions of 

 formation of this peculiar class of storms in every part of the coun- 

 try where they were found to occur, was necessary as a foundation 

 upon which to base investigations as to the origin, mode of develop- 

 ment, and means of prognostication. I began this work, in addi- 

 tion to other duties, in 1879, under directions of the chief signal 

 officer. The first field-work was done in that year, the second in 

 1882, and the third in 1885. Various reports were prepared and 

 published concerning the development of particular storms, the 

 origin and general characteristics of tornadoes, and the relation of 

 tornado regions to areas of low pressure. 



It was found that tornadoes generally occurred at a certain time 

 of day ; generally moved in a certain direction ; were generally 

 preceded by certain conditions of wind direction and velocity, and 

 by a certain gradient of temperature ; generally occurred in con- 

 nection with a well-defined area of low pressure, and with a " low " 

 of certain form and trend ; and generally occurred in a certain 



