May 23, 1890.] 



SCIENCE. 



313 



trary to watch-hands, was plainly visible, together with a 

 rolling motion inward and upward, giving the appearance 

 •of dense volumes of black smoke ascending from a tar-kiln. 

 Coruscations were observe! shooting out on the right or 

 south side of the cloud. A light funnel-shaped cloud, look- 

 ing much like white steam, immediately preceded the black 

 cloud." 



Summary. 

 I have read over about a thousand pages of tornado litera- 

 ture in making these quotations. They are extremely dis- 

 jointed, as must be necessarily the case in the scope of this 

 paper. I have had no theory to support, and in conse- 

 •quence the quotations are without bias. It is hardly proba- 

 ble, however, that any one person can read such a mass of 

 matter and make the best selections to give the more promi. 

 nent appearances; and it is much to be hoped that some one 

 will go over this ground and make independent selections of 

 the salient points. I think the omissions in these appear- 

 ances are oftentimes more suggestive than the positive state- 

 ments. 



The Loud Roar. 



The well-nigh universal testimony is, that there was an 

 indescribable roaring in connection with the tornado. It is 

 probable that this was heard in the earlier cases, but was 

 not regarded of enough importance to note. It is hardly 

 probable that the loudest of this roar could be heard more 

 than a few minutes before the outburst. Little weight can 

 be attached to the observations of a few, that it was not 

 heard after the passage. It is entirely improbable that this 

 can be caused by the wind, or by the tornado whirling in 

 the air. We very much need more careful observations. 

 Attention should be directed especially to a comparison of 

 the sound with a continuous rumble of thunder. 



The ascending current has also been largely commented 

 -on. The attempt to show an updraught by the fact that 

 some persons were let down gently is more than counter- 

 balanced by the fact that pieces of board and timber were 

 driven into the ground many inches. The evidence on this 

 point of an ascending current is very contradictory, and it 

 is highly probable that a fierce blast from two du-ections, 

 together with the assistance of the topography, will account 

 for most of the phenomena. There is almost overwhelming 

 evidence that air rushing into a partial vacuum does not 

 produce this updraught. The evidence shows that the direc- 

 tion of the path of the tornado is pre eminently toward the 

 north-east. 



Whirling of Tornado-Cloud. 



Perhaps the least satisfactory testimony is regarding this 

 appearance. We may set down at once the uselessness of 

 any one trying to determine this whirling if he is more than a 

 thousand feet away, and probably the limit should be five 

 hundred feet. Attention should be directed to the ground, 

 and most careful observations made of the whirling frag- 

 ments. The distribution of the debris is markedly against 

 any whirling. It is impossible to see how a whirl of a hun- 

 dred feet diameter could throw down one tree to the north- 

 east, and its neighbor to the south-east on top of the other. 

 Some one has well suggested, that, if the cloud is whirling, 

 the trees on the edges of the tornado should lie parallel to 



the track, while those in the centre should lie at right an- 

 gles; but we know that precisely the contrary is the appear- 

 ance. The writer once made an observation which may 

 help to elucidate this problem. In the streets of Washing- 

 ton, during the laying of the cable-road, there were em- 

 ployed furnaces for heating tar, and in these furnaces it was 

 customary to consume the remains of the tar-barrels for fuel. 

 The smoke from these furnaces was most dense and black. 

 The draught was so strong that the smoke issued from the 

 chimney at a velocity fully equal to the wind that was blow- 

 ing. The top of the chimney was not more than three feet 

 from the eye, thus giving a most excellent opportunity for 

 noting the slightest movement. When it was nearly calm, 

 the smoke ascended perfectly straight, and with no whirling 

 motion. The moment the wind blew, an extraordinary phe- 

 nomenon was seen. Taking a vertical plane in the direction 

 of the wind, and looking with the wind at the back, it was 

 found that on the right of the plane the smoke whirled from 

 right to left, while on the left it whirled from left to right. 

 The appearance of this smoke thus doubling upon itself was 

 most interesting, and invariably occurred when the wind 

 blew. It seems as though most of the contradictions in the 

 testimony would disappear if some such action as this took 

 place. We may be certain that there is no uniform whirl 

 in tornadoes in either direction. 



Stripping Feathers from Fowls. 



This is undoubtedly a true phenomenon. The attempts to 

 prove that this could be caused by the expansion of air in 

 the quills, due to the passage of a vacuum, have signally 

 failed. Fowls under an air-pump could not be deplumed by 

 exhausting the air. We must also regard Professor Loomis's 

 experiment of shooting a fowl out of a cannon as an entire 

 failure, I mean for elucidating this phenomenon. He 

 thought, that, if the fowl could have been fired at a hundred 

 miles per hour instead of three hundred and forty, the result 

 would have been very diilerent; but this certainly is very 

 doubtful. A wind of a hundred miles per hour would have 

 carried along the fowl, feathers and all. It would be an 

 interesting experiment to fire a fowl at a hundred miles per 

 hour; but it is entirely probable that no fowl could live 

 under such a shock as that, and the feathers would not be 

 driven ofl: until the velocity became enough to dismember 

 the fowl. It would also be interesting to hold a fowl before 

 a blast, and determine, if possible, the velocity needed for 

 depluming. The most singular fact is, that the fowl lives 

 under the depluming process. In some cases roosters have 

 been seen walking around, days after the tornado, crowing, 

 and without a feather on their backs. The appearance can 

 be readily accounted for on the supposition that an electric 

 charge threw off the feathers, and this seems the only way 

 of explaining the stripping of clothes from a person. 



The conclusion seems forced upon us that we need, more 

 than all else, much more accurate observations by persons 

 accustomed to note physical phenomena. It will be seen 

 that in later days the appearance of clouds from the north- 

 west and south-west is attracting great attention. That this 

 fact was not emphasized before, may be due, in part, to the 

 fact that it was not regarded as of any special importance in 

 accounting for the phenomena. It seems that these appear- 

 ances certainly attend the tornado, and are seen all along its 



