approached it from the side of the larger hurricanes. There is a complete gradation from the little dust eddies up through larger whirlwinds and tornadoes to the awful typhoons and cyclones of China and the West Indies; and it has long been my opinion that if meteorologists devoted their attention to the smaller eddies that can be looked at from the outside, and their commencement, continuance, and completion watched and chronicled, they could not fail to obtain a large amount of information to guide them in the study of cyclonic movements of the atmosphere. Unless the smaller whirlwinds are quite distinct from the larger ones in their origin, the theories advanced by meteorologists to account for the latter are certainly untenable. According to the celebrated M. Dove, cyclones owe their origin to the intrusion of the upper counter trade-wind into the lower trade-wind current.* More lately, Prof. T. B. Maury has stated that "the origin of cyclones is found in the tendency of the southeast trade-winds to invade the territory of the north-east trades by sweeping over the equator into our hemisphere, the lateral conflict of the currents giving an initial impulse to bodies of air by which they begin to rotate." Cyclones having thus originated, Prof. Maury considers that they are continued and intensified by the vapour condensed in their vortex forming a vacuum.† Humboldt had long ago ascribed whirlwinds to the meeting of opposing currents of air. There is this dynamical objection to the theory. The movements of * "Law of Storms," p. 246. + Quarterly Journal of Science, 1872, p. 418. the air in whirlwinds are much more rapid than in any known straight current, such as the trade winds; and it is impossible that two opposing currents should generate between them one of much greater force and rapidity than either. If force A joins with force B, surely force C, the product, must have the power of both A and B. But even if this fundamental objection to the theory could be set aside, the small whirlwinds could not thus arise, as they are most frequent when the air is nearly or quite motionless. Then, again, when we turn to Prof. Maury's theory that the cyclones, having been initiated by the conflict of contrary currents, are continued and intensified by the condensation of vapour in their vortex forming a vacuum, we find it negatived by the fact that in the smaller whirlwinds the air is dry, and there is consequently no condensation of vapour; yet, in comparison with their size, they are of as great violence as the fiercest typhoon. Tylor describes the numerous dust whirlwinds he saw on the plains of Mexico,* Clarke those on the steppes of Russia, and Bruce those on the deserts of Africa, and nowhere is there mention made of any condensation of vapour. I have seen scores of whirlwinds in Australia, many rising to a height of over one hundred feet; yet there was never any perceptible condensation of vapour, though some of them were of sufficient force to tear off limbs of trees, and carry up the tents of gold-diggers into the air. Franklin describes a whirlwind of greater violence than any of these. It commenced in Maryland by taking up the dust over a road in the form of an inverted sugar-loaf, and * "Anahuac," by E. B. Tylor, p. 21. Franklin soon increased greatly in size and violence. followed it on horseback, and saw it enter a wood, where it twisted and turned round large trees: leaves and boughs were carried up so high that they appeared to the eye like flies. Again there was no condensation of vapour. We thus see that whirlwinds of great violence occur when the air is dry, and there can be no condensation. When, however, they are formed at sea, and occasionally on land, the air next the surface is saturated with moisture; and this moisture is condensed when it is carried to a great height, forming clouds, or falling in showers of rain and hail. This condensation of vapour is an effect, and not a cause, and takes place, not in the centre, but at the top or at the sides of the ascending column. This is well shown in an account, by an eye-witness, of a whirlwind that did great damage near the shore of Lough Neagh, in Ireland, in August 1872.* It was about thirty yards in diameter. It destroyed several haystacks, and carried the hay up into the air out of sight. It partially unroofed houses, and tore off the branches of trees. The railway station at Randalstown was much injured; great numbers of slates, and two and a half hundredweight of lead were torn from the roof. When passing over a portion of the lake, it presented the appearance of a water-spout. On land everything that it lapped up was whirled round and round, and carried upwards in the centre, whilst dense clouds surrounded the outside and came down near to the earth. As above mentioned, I had in Australia many opportunities of studying the dust whirlwinds; and as I looked upon them as the initial form of a cyclone, I * "Nature," vol. vi. p. 541. paid much attention to them. On a small plain, near to Maryborough, in the province of Victoria, they were of frequent occurrence in the hot season. This plain was about two miles across, and was nearly surrounded by trees. In calm, sultry weather, during the heat of the day, there were often two at once in action in different parts of it. They were only a few yards in diameter, but reached to a height of over one hundred feet, and were often, in their higher part, bent out of their perpendicular by upper aërial currents. The dust and leaves they carried up rendered their upward spiral movement very conspicuous. No one who studied these whirlwinds could for a moment believe that they were caused by conflicting currents of air. They occurred most frequently when there was least wind; and this particular plain seemed to be peculiarly suitable for their formation, because it was nearly surrounded by trees, and currents of air were prevented. They lasted several minutes, slowly moving across the plain, like great pillars of smoke.* When attentively watched from a short distance, it was seen that as soon as one was formed, the air immediately next the heated soil, which was before motionless, or quivering as over a furnace, was moving in all directions towards the apex of the dust-column. As these currents approached the whirlwind, they quickened and carried with them loose dust and leaves * A friend of mine tells me that he saw a similar whirlwind rise at noon one still summer day, and traverse the dusty road on the Chesil Bank between Portland and Weymouth. It travelled fully half a mile, about as fast as he could walk; and the point where it met the ground was not thicker than his walking stick. By-and-by it swept out to sea, where the dust gradually fell. into the spiral whirl. The movement was similar to that which occurs when a small opening is made at the bottom of a wide shallow vessel of water: all the liquid moves towards it, and assumes a spiral movement as it is drawn off. The conclusion I arrived at, and which has since been confirmed by further study of the question, was, that the particles of air next the surface did not always rise immediately they were heated, but that they often remained and formed a stratum of rarefied air next the surface, which was in a state of unstable equilibrium. This continued until the heated stratum was able, at some point where the ground favoured a comparatively greater accumulation of heat, to break through the overlying strata of air, and force its way upwards. An opening once made, the whole of the heated air moved towards it and was drained off, the heavier layers sinking down and pressing it out. Sir George Airey has suggested to me that the reason of the particles of air not rising as they are heated, when there is no wind blowing, may be due to their viscosity: and this suggestion is correct. That air does not always rise when heated, appears from the hot winds of Australia, which blow from the heated interior towards the cooler south, instead of rising directly upwards. Sultry, close weather, that sometimes lasts for several days, would also be impossible on the assumption that air rises as soon as it is heated. This explanation supplies us with the force that is necessary to drive the air with the great velocity with which it moves in whirlstorms. The upper, colder, and heavier air is pressing upon the heated stratum, and the U |