270 



SCIENCE. 



[Vol, XV. No. 378 



sun's direct rays be hidden, this same oppressive feeling is 

 noted. This heat seems to have a special characteristic en- 

 tirely unlike that from the ordinary bright sun's rays, and 

 it has never been properly accounted for. The same heat is 

 often felt just before a thunder-storm, and we may consider 

 that the same cause is acting in both these cases. 

 Clouds. 

 The clouds have the general motion that they always have 

 in the neighborhood of an extensive storm. This vcould 

 seem an exceedingly important consideration, and v?ould 

 show that the conditions leading up to the tornado are the 

 same as are present at innumerable times when there are no 

 serious outbreaks. A beginning in a study of these condi- 

 tions has already been made {Journal of the FranMin In- 

 stitute, July, 1888, p. 48). In that investigation it was 

 learned that near the centre of the general storm, the clouds 

 moved mostly with the surface winds, from right to left, or 

 counter-clockwise, and if they had any other direction it 

 was siinply a slight turning toward the east in the line of 

 the general flow of the upper current. This fact was also 

 ascertained by Hildebrandsson in Denmark, though the 

 storms there may have slightly different characteristics from 

 the proximity of the ocean. A score of tornado regions 

 have shown this tendency. It will suffice to give here il. 

 lustrations from three tornadoes which have been specially 

 studied ; and a fourth will be described later, in connection 

 with the Louisville tornado. 



Washington Court-House, O. 



This place was visited Sept. 8, 1885. The general storm 

 was central in southern Wisconsin, and the cloud-move- 

 ments were as follows: from south-west at Dubuque, Daven- 

 port, Keokuk, Springfield (III.), Chicago, Indianapolis, 

 Louisville. Nashville, Cincinnati, Columbus, and Sandusky; 

 from south at Cairo, Toledo, Pittsbiirgh, and Erie; from 

 south-east at Detroit, Cleveland, and Buffalo. In all this 

 region there was not a single cloud-movement from the 

 north-west or even west. On April 14, 1886, St. Cloud and 

 Sauk Rapids, Minu., ware injured by a tornado. The gen- 

 eral storm was e«.itral in north west Dakota, and the cloud- 

 motions were as follows: from south-west at North Platte, 

 Des Moines, and Valentine; from south at Dead wood, 

 Omaha, Leavenworth, Springfield (Mo.), Springfield (111 ), 

 Cairo, St. Louis. Dubuque, and La Crosse; from south-east 

 at Bismarck, Moorhead, St. Vincent, and St. Paul. Here 

 not only are there no west or north-west cloud-movements, 

 but also in the immediate tornado region the movement is 

 markedly from the south-east. 



Mount Vernon, 111. 



This tornado occurred Feb. 19, 1888. The general storm 

 was central in eastern Iowa, and clouds were moving as 

 follows: from south-west at Keokuk, Memphis, and Louis- 

 ville; from south at Springfield (111.), Chicago, Indianapolis, 

 Toledo, St. Louis, Cairo, and Chattanooga; from south east 

 at Milwaukee and Davenport. 



This is an exceedingly important fact, as will be seen 

 later on, and seems to be abundantly established. 

 Distribution of Tornadoes. 



Rarely do these violent outbursts occur singly. In any 

 region favorable for tornado development, two hundred to 



four hundred miles to the south-east of the centre of the 

 general storm, we shall find, after the hottest part of the 

 day, a line of tornadoes occurring one after the other, and 

 moving almost invariably to the north-east. An hour or so 

 later, another line will be found parallel to the first, and 

 about fifty miles south-east of the first In some cases, as 

 in the great Louisville tornado, there may be six or even 

 more of these lines. They are entirely independent of each 

 other, and very clear cut, there being no destruction be- 

 tween. The last line may not begin until 8.30 p.m., more 

 than five hours after the hottest part of the day. 



Velocity. 



A marked feature of these outbursts is their most rapid 

 translation across the earth for two hundred and sometimes 

 three hundred miles. The speed is rarely under forty miles 

 per hour, and there have been well-authenticated cases 

 where it has reached over eighty miles per hour, notably in 

 the Louisville tornado. It should be noted that the general 

 storm, while travelling in the same direction, attains a 

 velocity only half as great; for example, at Louisville it 

 was thirty-eight miles per hour, which is very much above 

 the average. 



Thunder-Storms . 



All through this region thunder-storms are very numer- 

 ous, and are an invariable accompaniment. Sometimes 

 vivid displays of lightning occur in the tornado, and un- 

 doubtedly these would be observed much oftener were not 

 the beholder awe struck by the terror-inspiring phenomenon. 

 A careful study of thunder-storms, occurring at any time, 

 has shown that they have a velocity double that of the fos- 

 tering general storm, and are always found within theregion 

 six hundred miles to the south-east. It has come to be gen- 

 erally admitted that tornadoes and thunder-storms are anal- 

 ogous phenomena, and that the former are an intensification 

 of the latter. If this be true, we see at once what an enor- 

 mous advance has been made. We do not need to wait for 

 the full-fledged monster, whose very appearance drives away 

 all thought save flight for safety, but we may study it in its 

 gentler moods. 



Lurid Sky. 



Let us take a slightly closer view. Almost the first warn- 

 ing after the appearance of dark and threatening clouds in 

 the west, as in a thunder-storm, is a peculiar lurid or greenish 

 tinge in the sky from the south to the west. This "tornado- 

 sky" is a characteristic feature, and it is believed that to 

 those who have studied the phenomenon this will always 

 serve as a warning for the more serious results which soon 

 follow. 



Two Clouds. 



Many observers have seen very black clouds, — one to the 

 west or north-west, and the other to the south-west, — which 

 seem to rapidly advance, and, when they meet, to form the 

 tornado. This is believed to be in the nature of a perspec- 

 tive effect, and the clouds to be an immediate accompani- 

 ment of the tornado. A full explanation of this phenome- 

 non will be given later. 



Cloud of Dust. 

 It is a very significant fact that in very many cases the 

 funnel-cloud, which has a sharp outline and should be the 



