164 ANNUAL OF SCIENTIFIC DISCOVERY. 



five acres, and was set on fire at the part most distant from him. 

 " Whirlwinds were now observed in the hottest part of the fire. 

 They did not unite in one column, but were scattered throughout the 

 fire, and several were formed at the same time. The first were on a 

 comparatively small scale. Their height was from thirty to forty 

 feet. To these succeeded others on a larger scale, until they readied 

 the height of more than two hundred feet, and the flame and smoke 

 which formed their columns were perfectly distinct from the general 

 mass which arose from the fire. They appeared rather to increase in 

 size and frequency toward the latter part of ' the burning,' and many 

 were formed on the ashes, after the fire had, to a great extent, gone 

 down. 



" Among the whirlwinds there were several points of difference, by 

 which they might be classed under four heads. The most common 

 one was that which was stationary over a part of the fire which was 

 hotter than the neighbouring portions. A second variety was that 

 which had a progressive motion, and advanced over the burnt track, 

 throwing up ashes and cinders, and thus marking its course through 

 the fire. Some of these emerged from the flames. This was probably 

 the case with quite a number, although, having nothing to mark them 

 after leaving the fire, they became invisible. One, however, passed 

 near enough to us to be observed, and attracted our attention by its 

 rustling sound, and by the leaves which it carried up. This was 

 about fifteen or twenty feet high. At the time this passed us, we had 

 moved from our first station, and were about three hundred yards from 

 the fire. These whirlwinds differed from the others in form, being 

 very wide at the top, and contracting to a point at the bottom, like a 

 top or a spindle, or, more exactly, they were of the form of the upper 

 cone of an hour-glass. An interesting phenomenon which attended 

 some of the whirlwinds might render it proper to arrange them under 

 a third class. In these the flame was violently whirled at the base ; 

 then above succeeded a dark interval, where the flame seemed to be 

 extinguished entirely, but towards the top it broke out anew. It was 

 a mixed whirling of flame and smoke, the smoke occupying the cen- 

 tral portion. The dark interval where the smoke was unconsumed 

 was greater or less, as the flame above approached to, or receded from, 

 that beneath. There were quite a number of this class. The fourth 

 kind were formed of immense columns of smoke, so narrow and lofty 

 that they resembled towers of several hundred feet, or trunks like 

 those of trees in form, extending into the sky. The rotary motion 

 was obvious throughout their entire length. These columns of smoke 

 were generally straight, but sometimes bent at the top by the wind. 

 In connection with the whirlwinds, there were several other facts of 

 interest observed during the burning of the cane. We noticed the 

 direction of the wind was changed. At first it was from the north- 

 east, and continued in that direction in the upper part of the atmos- 

 phere, as was evident from the way in which the columns of smoke 

 were bent. But shortly after the commencement of the burning, the 

 air beneath blew in all directions towards the centre of the fire. The 

 columns of smoke were not bent for more than a hundred yards ; hence, 



