May 4, 1882] 



NATURE 



works and in the Zcitschrift fitr Meteorologte (Band x.), 

 we need not notice it here except to point out that, 

 assuming the correctness of the formula, the gradient, 

 cateris paribus, should vary (i) directly with the latitude, 

 (2) inversely with the distance from the centre, (3) in- 

 versely as the temperature, (4) directly with the amount 

 of inclination. 



The foregoing results have all been obtained without 

 considering the term depending on temperature and 

 humidity, and which expresses the effect of the disturb- 

 ing force necessary to start and maintain the interchang- 

 ing motions between the interior and exterior portions of 

 the air over a given area. That such a disturbing function 

 is necessary, is evident both from preliminary considera- 

 tions, and also from the form of the general equations of 

 motion, since they would otherwise be satisfied by the 

 conditions for a state of test. The author enunciates this 

 principle in Part I. Chap. III. where he says : " There can 

 be no winds then -without a disturbance of the static equili- 

 brium by means of a difference of temperature or of aqueous 

 vapour in different parts of the atmosphere." And it is 

 important to bear it in mind, if only because we are too 

 often apt to overlook it in the multitude of secondary 

 causes brought to light by a study of atmospheric me- 

 chanics. A consideration of this term, in which tem- 

 perature and humidity are treated jointly, and the former 

 is assumed to vary with the distance from the centre, 

 leads to the remarkable conclusion that there are two 

 species of cyclones, one with relatively warm centres, the 

 more common case, and the other with relatively cold 

 ones. 



These cyclones differ specifically from each other 

 chiefly in the way in which the pressure is distributed 

 and the gyrations directed at different altitudes. 



In a cyclone with a relatively warm centre the air at 

 the earth's surface moves in a cyclonic spiral round and 

 towards the centre, but as we ascend the gyratory velocity 

 diminishes with the altitude, and the annulus of high 

 pressure approaches the centre, until at a very high 

 elevation the highest pressure of that stratum might even 

 be at the centre, and the air gyrate anticyclonically from 

 it over the whole area at that level. In brief, the cyclonic 

 area becomes smaller, and the anticyclonic larger, as we 

 ascend. 



In a cyclone with a cold centre the reverse occurs. 

 At the surface of the earth the initial tendency of the 

 air is to move outwards, and this may be so strong near 

 the surface that there may be only anticyclonic gyrations 

 at this level, with the maximum pressure of the lowest 

 stratum at the centre. As we ascend, however, the 

 gyrations round the centre become more and more 

 cyclonic, while the annulus of maximum pressure gra- 

 dually retreats further and further from it. 



There is, besides, according to the theory, an ascending 

 motion of the air in the interior part of a warm-centred 

 cyclone, and a descending motion in the exterior part, 

 both generally small in comparison with the horizontal 

 motions toward and from the centre. In the case of a 

 cold-centred cyclone these motions are reversed. 



Now as a barometer at the earth* s surface records 

 simply the integrated effect of what happens in all the 

 strata up to the top of the atmosphere, this might 

 obviously vary in the same way for both kinds of cyclones, 

 and so tell us absolutely nothing of such remarkably di- 

 verse conditions prevailing at higher altitudes. The 

 behaviour of the air in the warm centred cyclone is what 

 we are accustomed to obserre in the case of most cyclones, 

 and as they are as often found with relatively cold centres 

 as with warm ones, the former occurring more frequently 

 in summer and the latter in winter, it is difficult to under- 

 stand why the characteristics of the cold-centred cyclone 

 have never yet been found to prevail, at least in moving 

 cyclones. The author indeed offers an explanation of 

 this circumstance, and endeavours also to account for the 



absence of stationary cold-centred cyclones in regions like 

 Central and Eastern Asia and North America in winter, 

 where the temperature gradient would be remarkably 

 favourable to their production. The fact, however, that in 

 the centres of these regions at this season there is not 

 only no cyclonic tendency of the winds or depression of 

 the barometer, but, on the contrary, a pressure greatly 

 above the normal, seems strangely at variance with what we 

 should expect according to the theory of the cold-centred 

 cyclone, and is hardly satisfactorily explained away as the 

 result of the irregularity and size of the area, combined 

 with the excessive cold, which latter is supposed to increase 

 the density and piessure more than the cyclonic tendency 

 diminishes them. The only two cases of cyclones with 

 cold centres which the author seems able to find are the 

 two general wind systems of the northern and southern 

 hemispheres respectively, which " are simply two great 

 cyclonic systems with a cold centre, having the cold poles of 

 the earth for their centres. The motion of the air eastward 

 around and toward the poles in the middle latitudes, giving 

 rise in those latitudes to the noriml south-west winds in 

 the northern, and north-west winds in the southern hemi- 

 sphere, form the cyclones, and the trade-wind region the 

 corresponding anticyclones, with the equatorial calm-belt 

 for the common limit of the two systems. The tropical 

 calmbeltand corresponding maxima of barometric pressure 

 near the parallels of 30', correspond to the similar calm 

 and dividing line between the cyclone and anticyclone in 

 the ordinary and smaller cyclonic systems." 



The primary cause of cyclones, according to Ferrel, 

 is a horizon la/ temperature gradient, so that if a portion 

 of the atmosphere is heated or cooled more than the 

 surrounding parts, and the isotherms are approximately 

 circular, we have the initial conditions for a cyclone ; but 

 after the disturbances due to such primary causes have 

 set in, secondary causes depending on loss of heat by 

 expansion in ascent, and gain of heat by compression in 

 descent, as well as retardation of cooling where aqueous 

 vapour is being condensed, come into play, which on the 

 whole tend to counteract the initial motions. 



The condition of the atmosphere vertically with respect 

 to temperature and humidity, is thus of great importance 

 in regard to the duration of a cyclone when it has once 

 started. 



The author investigates this point at some length, and 

 works out the conditions for cyclone generation in quite 

 a novel manner, from a consideration of both vertical 

 and horizontal temperature gradients. Generally speak- 

 ing, the condition most favourable to the maintenance of 

 an ordinary cyclone is that the vertical temperature de- 

 crement in the interior should be less than in the sur- 

 rounding regions. This condition is found to be more 

 easily sustained where the air is charged with aqueous 

 vapour, since under these circumstances it cools less 

 rapidly in ascending than when dry. He further points 

 out that where the decrement of temperature in the inte- 

 rior is less than outside, especially when this condition 

 occurs throughout the entire atmosphere, a cyclone may 

 arise without any horizontal temperature gradient (pro- 

 vided only a small instant uieoas impulse be given), and 

 that such a state of unstable equilibrium more readily 

 occurs when the air is warm and saturated with vapour. 

 While, however, he thus admits the important role played 

 by vapour in maintaining cyclonic action when once 

 started, he distinctly denies its claim to be considered 

 " either a primary or principal cause of cyclones'' 



As the^e islands in all probability seldom, if ever, form 

 the birthplace of a cyclone, but we are rather accustomed 

 to experience them either fully developed or else in the 

 condition of being "filled up," the circumstances which 

 attend their generation do not practically very much 

 concern us. Still it must not be overlooked that con- 

 ditions which would tend to create and maintain a cyclone 

 in our midst, must of necessity tend to augment the 



