108 BULLETIN OP THE 



the late Professor Coffin in his Winds of the Northern Hemi- 

 sphere. But of course this value, depending upon so small a 

 gradient, cannot be actually determined from the observed gradient 

 by the preceding relation. Again, the value of G deduced from 

 the table of barometric pressures given by Buchau in his Mete- 

 orology, for the parallel of 20° north, is about 0.02 of an inch. 

 The wind in the Atlantic Ocean in this latitude being from the 

 N. E., we have the value of i here about 225°. With this value 

 of i and the preceding value of G, the equation gives t;=15 

 miles, v^hich is probably about the velocity of the trade winds on 

 the sea in this latitude. 



From the same table of barometric pressures given by Buchan 

 we deduced, for the parallel of 52° south latitude, the value of 

 G^ = 0.07 of an inch. Supposing i here to be small, that is, that 

 the constant wind blows nearly from the west, the relation above 

 gives u = 21 miles for the velocity of the wind from the west. 

 All accounts represent the west wind in this latitude as blowing 

 very strong all round the globe, and Mr. Laughton speaks of it 

 as frequently amounting to a half gale. 



In an ordinary cyclone the value of ii in compai'ison with 

 2 ?i sin Z cannot in general be neglected, and it is readily seen 

 that near the centre of a cyclone where r is small the value of u 

 may be very large. When the isobars of a cyclone drawn to 

 every tenth of an inch of the barometer are one hundred miles 

 apart, we have (7 = 0.1 of an inch. With this value of G, sup- 

 posing the value of i to be so small that we can put sec i = 1, 

 we get from the expression of G at the distance of four hundred 

 miles from the centre of the cyclone, and on the parallel of 45°, 

 i; = 29 miles for the velocity of the wind, and this would be very 

 nearly the velocity at sea where i is so small. But with the value 

 of z = 45° nearly obtained by Professor Loomis, we get = 21 

 miles nearly. At a distance of one hundred miles only from the 

 centre, all the other circumstances remaining the same as above, 

 we get i; ^ 22 miles in the case in which i is small, as at sea, and 

 in the case in which the value of i is 45° we get = 18 miles 

 nearly. Hence we see that the law gives the velocities for the 

 same gradient considerably greater near the centre than at greater 

 distances. 



The preceding law cannot be tested by comparing the observed 

 velocity of the wind in any individual case with that deduced by 

 means of the law from the observed gradient obtained from the 

 isobars laid down on the weather maps of the Army Signal 

 Service; for these, being laid down from observations made at 

 stations which are in many cases several hundred miles apart, 

 cannot take in the efiFects of the numerous local and compara- 

 tively small disturbances, and these latter may affect very much 

 the velocity of the wind at any station The law can only be 

 tested by comparing the average velocity of a great many in- 



