704 



SCIENCE. 



(Vol.. II., No. 4:i. 



reason for httertil motion has now appeared : 

 the gravitative pressure of the upper air is no 

 longer at riglit angles to the convex isobaric 

 surfaces, and consequently there will be a ten- 

 dency for the air to slide down from the centre. 

 In obedience to this impulse, some of the cen- 

 tral expanded air moves lateral!}' or radially 

 outward to the marginal region ; and now there 



is no longer a uniform pressure of oU inches 

 at sea-level. At the centre, whence the upper 

 ail- has rolled away, the pressure will be reduced, 

 let us say, to 29 inches : on the surrounding 

 district, over which the air has advanced, the 

 pressiu'e has increased to 30.25 inches. In this 

 new arrangement of pressures there is cause 

 for still further gravitative motion ; namelj', a 

 rising of the air at the centre, a sinking at 

 the marginal region, and a horizontal motion 

 along the sea-surface, toward the centre of low 

 pressure, in the attempt to restore an equilib- 

 rium. But this will not fully overcome the in- 

 equality of pressures, or correct the sloping of 

 the isobars ; for the existence of an ascending 

 and expanding warm current at the centre re- 

 quires that the isobaric surfaces there shall be 

 separated bj' a greater vertical distance than in 

 the normal cool- 

 er air of fig. (i. 

 Further, the mar- 

 ginal descending 

 current of air, 

 greatly cooled hy 

 radiation in the 

 upper regions, is 

 heavier, volume 

 for volume, than 

 the ascending cur- 

 rent, and hence 

 has its isobaric 



surfaces closei- together than usual. A shorter 

 vertical column of it is needed to balance an 

 inch of mercury in the barometer. Fig. 8 shows 

 this final condition, — the diminished pressure 

 and greater separation of the isobaric lines 

 at the warm centre ; the increased pressure and 

 the approach of the isobaric lines in the cooler 



margin. Now, in virtue of the greater dis- 

 tance between the isobars at the centre, the 

 altitude of some surface, say that of 24- inches, 

 will be as great there as over the marginal 

 region, in spite of the inequalities of pressure 

 and inward slope of the isobars at sea-level ; 

 and at greater altitudes the isobaric surfaces 

 will become convex, and hence slope outwards, 

 instead of inwards, as 

 below. The two direc- 

 tions of slope will be 

 .jr. separated bj- a level or 



~ neutral plane, on which 



Za. there will be no ten- 



dency to motion. Here 



' ^"^ we have excellent illus- 



-— — ^ tration of the convee- 



So tional motion of the wind 



in a storm. It ascends 

 at the centre, where it 

 is lightest ; it then flows outward, down the 

 barometric gradient ; it sinks at the marginal 

 region of higher pressure, and then flows in- 

 ward, down the reversed gradient, back to the 

 centre again. This may be called the vertical 

 circulation of the storm ; and it will be contin- 

 ued as long as the central current is warmed to 

 excess, so as to raise its isobaric surftices. In 

 the desert-whirlwind we have seen that the 

 supply of warm air depends immediately upon 

 contact with the surface-sands heated by direct 

 sunshine. In the cvcloue at sea, the greatest 

 part of the warmth needed is given out by the 

 vapor that condenses at the centre, and falls 

 in the heavv rains, without which a cyclone 

 cannot form. Such a storm may last manj' 

 days. 



The explanations thus far given of the be- 



ginning of a cj'cloue apply strictly' only to tlie 

 hurricanes of tropical latitudes ; for in the tem- 

 perate zones our numerous storms are bj' no 

 meansi always dependent on local warmth and 

 calmness of the air. The most that can now be 

 safely said of the origin of such storms is, that 

 they depend on some immediately preceding 



