October 7, 1922] 



NA TURE 



479 



Now let this earth be warmed by a source of heat 

 equivalent to the sun, but in the form of a distant 

 ring surrounding it in the plane of the equator. Let 

 the atmosphere be transparent to radiation and take 

 its heat only from the floor of the cell which contains it. 



In the course of time the contents of each cell will 

 reach the temperature of the floor, which will be a 

 maximum at the equator, and will vary as the cosine 

 of the latitude to absolute zero at the poles. 



The barometric pressure in each cell will be the 

 same : were all the cells removed the atmosphere 

 would be in equilibrium. The equilibrium, however, 

 would be unstable, and the least departure from the 

 original stratification of density would cause ulti- 

 mately a circulation to be set up, in which, in the 

 absence of turbulence, warm air would flow from the 

 equator towards the poles at high levels, while cooled 

 air would travel in the opposite direction near the 

 surface of the earth. A steady distribution of tem- 

 perature would be reached when each element of the 

 surface lost by radiation as much heat as it received 

 from the source plus that supplied by the circulation, 

 and this distribution probably would not differ much 

 from that which now exists, though the fact that the 

 real atmosphere is more or less opaque to long waves 

 would introduce a sort of " green-house " effect, and 

 raise the mean temperature above that appropriate 

 to perfect transparency. Again if the imaginary 

 earth were completely covered by a deep ocean, a 

 separate circulation would be set up in the latter, and 

 the temperature distribution would be somewhat 

 modified in the direction of greater uniformity. 



Since the energy of the circulation is derived from 

 the source of heat, there will be no change of pressure 

 due to the velocity, and supposing for the moment 

 that the air is incompressible, then in the nearly 

 horizontal path which constitutes the greater part of 

 each stream line circuit, the cross-section velocity and 

 dynamic head for each will be constant, though not 

 necessarily the same for different streams. The cross 

 section of the ascending and descending parts of the 

 streams will bear to the cross section of the horizontal 

 part the ratio of the length of the earth's quadrant to 

 the height of the homogeneous atmosphere, and thus 

 in the neighbourhood of the poles and the equator 

 there will be a small increase of pressure. The form 

 of the stream lines due to temperature circulation in 

 a spherical shell is indicated diagrammatically in 

 Fig. 2. 



Latitude 

 'ig. 2. — Stream lines of the circulation in a meridional element of a spherical 

 shell, the density of the fluid being supposed constant. 



As regards the distribution of temperature, the 

 results would be much the same whether the earth 

 were stationary or rotating, but the direction and 

 velocity of the" wind referred to a fixed point on the 

 solid surface would be very different in the two cases. 

 If, in the absence of surface friction, the earth were 

 given its present angular velocity the apparent wind 

 would have an easterly component of about iooo 

 miles per hour at the equator while at the poles there 

 would be a calm. If, on the other hand, when the 

 rotation was started, the air was given the same 

 velocitv as the surface under it, the apparent wind 

 would var}- in direction and force in a period equal 

 to that of "the circulation. 



In the real atmosphere, the effects of turbulence, 



NO. 2762, VOL. I IO] 



viscosity, and surface friction will ensure that the 

 average velocity of the apparent wind shall in no 

 place exceed 30 or 40 miles per hour. If unresisted 

 air passes from lat. X to X + AX the change of the linear 

 speed of the ground under it, i.e. the change in the 

 E. or W. component of the apparent wind, is 

 RAX(i-sinX) linear velocity in longitude, and if the 

 apparent wind remains constant, it shows that surface 

 friction is sufficient to accelerate or retard the atmo- 

 sphere by this amount in the time taken in covering 

 the distance RAX. In the case of the earth, this 

 would imply that if the circulating velocity (i.e. theN. 

 or S. component) is 15 m.p.h., surface friction suffices 

 to change the speed of the apparent wind by about 

 15 m.p.h. per hour near the poles while in lat. 30° the 

 corresponding change would be somewhat less than 

 2 m.p.h. per hour. 



On the imaginary seasonless earth, the average 

 wind would everywhere be a definite function of the 

 latitude and coefficient of friction, provided that the 

 going and returning parts of the circulation did not 

 mix on the journey, and in low latitudes this would be 

 true even when the effects of turbulence were taken 

 into account. Farther north or south, however, the 

 hot and cold streams would become interwoven in 

 eddies the forms of which are incalculable, though the 

 average winds would always be either from N. and E. 

 or S. and W. Thus it might be expected that there 



Fig. 3. — AA, Circular conducting plate and tank. BB, Annular hot water 

 trough. C, Axis and cold-water tank. 



would be calms at the equator, moderate and regular 

 trade winds for some distance on either side, and 

 beyond these, irregular winds, the intensity of which 

 increased with the latitude. The barometric pressure 

 would be nearly constant excepf in the eddies, and 

 there the variation of pressure would depend, not on 

 the actual velocity of the apparent wind, but on its 

 difference from the average for the latitude. 



Such a description with modification depending on 

 the seasons, the presence of moisture in the air, and 

 on the distribution of land and water agrees with the 

 average conditions on the real earth. Dampier's 

 maps show that coastal influence may be sensible 

 through io° of longitude or more, and it may be 

 guessed that the direction of the , monsoons is in some 

 way influenced by the great area of land lying to the 

 north of the parts where they blow. 



There is not much information available concerning 

 the wind structure of the atmosphere on the borders 

 of the Trades, and a proper investigation of this 

 subject would form an important addition to meteoro- 

 logical science ; but such an investigation would 

 require more than one Challenger expedition devoted 

 to the exploration of the upper air instead of the deep 

 sea. 



Expeditions of this kind are not likely to be under- 

 taken at the present time, but some notion of the 

 manner in which the Trades break up might be 

 gained by an experiment such as is indicated in 

 Fig. 3, where a thick circular metal plate, provided 

 with descending flanges at the circumference and a 

 thick central axis, carries a shallow circular tank 

 containing fluid. The flanges dip into a circular 

 trough of warm water while the axis is kept cold. 

 If the apparatus is stationary, a circulation is set up 

 in the tank of the type shown in Fig. 2, but if it has 

 an appropriate angular speed about the axis, the 



