February 28, 1890.] 



SCIENCE. 



143 



air, are no longer level : they are deformed into slanting posi- 

 tions, and the slant or gradient is directed toward the warm 

 region in the lower atmosphere, and toward the cold region in 

 the upper atmosphere. Thus far every one is willing to go : and, 

 if it be desired to ti-y the experiment on a class of intelligent 

 scholars, some live interest in the question may be aroused by 

 asking how far they are individually ready to assert that this 

 simple abstract theory is applicable to the case of the earth ; the 

 warm region being the equator, where the mean annual pressure 

 must therefore be low, and the cold region being either pole, 

 ■where the pressure must be correspondingly high. No more 

 salutory lesson can be given in the danger of the purely deductive 

 method in the hands of others than the masters of a subject, for 

 the high pressure that is confidently expected at the poles does 

 not exist. The pressure there is lower than at the equator. The 

 contradiction of theoretical deduction by well-ascertained fact is 

 of the flattest kind, and the scholar may fairly be excused if for 

 a time he loses faith in a theory that has led him into so blunder- 

 ing an expectation. But when he looks further, and fuids that 

 there is a belt of lower pressure at the equator than at the tropics, 

 and that this belt migrates with the seasonal shifts of the heat 

 equator, and that the continents unload their shai-e of atmosphere 

 somewhat in their summer season, it becomes apparent that the 

 theory must be wrong chiefly by omission ; and it may be readily 

 shown that the omitted consideration is the effect of the earth's 

 rotation. Thei-e ai'e very few men in the world who have for 

 themselves avoided this omission; and of these few. Professor 

 Ferrel is the only one who has given the complete theory the full 

 consideration that it deserves. 



The fact that the interchanging convectional circulation of the 

 atmosphere between the equator and the poles takes place upon 

 an earth that rotates on its axis, requires the development of 

 great eastward spiral polar whirls, and the centi'ifugal force of 

 these whirls greatly deforms the simple arrangement of the iso- 

 baric surfaces that would be produced by differences of tempera- 

 ture alone ; so greatly, indeed, that the theoretical high pressure 

 of the poles is reversed to actual low pressure. In consequence of 

 this, the gradients of nearly all the atmosphere are directed pole- 

 wards, the only gradients that lead to the equator being in the 

 lower atmosphere within the tropics, where we have the trade- 

 winds. This may appear more clearly in Fig. 1, which repre- 

 sents a vertical meridional section of the atmosphere, greatly 

 magnified vertically, from pole across the equator to pole ; the 

 meridian line being, for simplicity, straightened out from its true 

 semicircular curve. The pressures at the surface are known by 

 observation, being, on the average, about 29.9 at the equator, 

 80.1 a little outside of the tropics, and perhaps 29.0 at the poles. 

 Remembering that the successive isobaric surfaces diverge from 

 the cold polar regions towai-ds the warm equatorial belt, a num- 

 ber of higher and higher surfaces may be drawn in section, and 

 the prevailing poleward slope of the gradient is then made ap- 

 parent. 



Now, the question asked by Supan and Teisserenc de Bort is 

 practically this: "How does the air, which flows toward the 

 poles on the steep gradient of the upper current, manage to re- 

 turn to the equator against the poleward gradients of the lower 

 levels?" Tliis is as if they asked, "How does the ocean stand 

 thirteen miles higher (i.e., farther from the earth's centi-e) at 

 the equator than at the poles, instead of at once inishing tumultu- 

 ously poleward ?' ' 



The low jjressure at the poles is the indirect pi'oduct of the 

 initial meridional convectional circulation between poles and 

 equator, and the deformation of the simple convectional giadi- 

 ents thus introduced can never go so far as to stop the convec- 

 tional motion by preventing the return of the lower current to 

 the equator. The gi-eat velocity and consequent great centi'ifugal 

 force attained by the upper cuiTent, as it swings around the pole 

 on the steep upper gradients, enable it to ran obliquely against 

 the weaker lower gi-adients as soon as it encounters them in the 

 descending portion of its convectional circuit. That is the 

 essence of the whole affair, though it may be stated in different 

 ways, from words to fomiulse. Perhaps a simpler way of put- 

 ting it is this. The difficulty comes from thinking that the lower 



isobaric surfaces slope toward the pole. But it must be remem- 

 bered that slopes and levels are determined by the local direction 

 of gravity, not by distance from the earth' s centre ; that the local 

 direction of gravity is determined by the local value of the centrif- 

 ugal force arising from axial rotation, and the velocity of axial 

 rotation depends on whether the body that is under discussion 

 goes around the axis once in twenty-four hours, as we do who 

 live on the earth's surface, or in a decidedly less time, as the 

 eastward winds do. If the earth had no rotation, its present 

 level surfaces would be called poleward slopes. The winds 

 which move eastward must regard the sea-level as an equator- 

 ward slope; and the fast winds of the gi-eat eastward whirls 

 around the poles must regard even the lower gxadients of the 

 atmosphere as slopes dii'ected toward the equator, and not toward 

 the pole. It is only the lower winds, whose velocity is weakened 

 by surface friction, that have the same opinion of the lower 

 gradients as we have, and obey them by moving obliquely toward 

 the pole. This is not a matter that needs mathematical state- 

 ment for its demonstration. The rational conception of the pro- 

 cess, on which the validity of any mathematical treatment must 

 be based, is suflicient to demonstrate that the isobaric surfaces, 

 whose an-angement is determined simply by differences of tem- 

 perature, cannot agTce in position with those which are, as it 

 were, defoi-med by the introduction of the deflective forces that 

 arise from the earth's rotation; and to demonstrate, further, that 

 the deformation thus introduced can never go so far as absolutely 

 to stop, although it may greatly retard, the meridional or con- 

 vectional components of motion, on whose persistence all the 

 other motions depend. The reader of the "Popular Treatise on 

 the Winds" can come to no other conclusion than that the 

 essential nature of the circulation of the winds is such as is here 

 outlined ; and the doubts raised by Supan and others will then 

 not be regarded as objections to Fen-el's theory. 



The actual circulation of the winds over continepts and oceans 

 is greatly complicated by seasonal and topographic influences, as 

 well as by the presence of numerous cyclonic storms, marching 

 in continuous procession around either pole. But the ideal 

 planetary circulation is relatively simple; and, as the graphic 

 illusti'ation of its course is seldom given in more than highly 

 diagrammatic forms, we venture to introduce here a more care- 

 fully drawn view of it, the upper winds being exhibited in the 

 northern hemisphere, and the hypothetical return current of 

 middle elevation being drawn on the southern, while the siirface 

 winds are in dotted lines beneath. There is much that is hypo- 

 thetical in this ; but it is as a whole well borne out by actual 

 observation. One of the questions that is still open is the latitude 

 at which the upper poleward overflow from the equator has a 

 directly poleward motion. The latitude certainly varies with 

 the altitude, but it does not appear to be more than ten degi-ees 

 north or south of the equator : for on a poleward gradient, and 

 with a right-hand deflection, both of which are undoubted, the 

 upper overflow cannot long maintain the westward component of 

 motion that it possesses above the equator ; and, as a matter of 

 fact, the oblique pole-eastward motion of the overflow has often 

 been observed in the di-ifting of clouds and in the wind on moun- 

 tain-tops in the so-called "anti-trade." 



The reader must not imagine that all of Professor Ferrel' s book 

 is occupied with theoretical discussions. The citation of appro- 

 priate facts is plentiful and well selected ; quotations are made 

 at length from various sources ; and although the winds are, by 

 the title of the work, its main theme, one needs bvit small ac- 

 quaintance with meteorology to know that nearly all of the 

 science may be fairly presented under this heading. It is most 

 natural that a course in meteorolo,gy should begin and end with 

 a discussion of the circulation of tlie winds ; for pretty much 

 every thing meteorological is, like the deformation of the polar 

 gradients, more or less closely a sequence of the motion of the 

 atmosphere. When the educational value of the study of me- 

 teorology is more widely appreciated, as it must be when more 

 of our teachers are familiar with such works- as this one of 

 FeiTel's, it may come to be true, as an eminent Scottish meteor- 

 ologist some twenty years ago imagined it was already at that 

 time, that "in the schools of the United States of America, 



