SHORT MEMOIRS ON METEOROLOGICAL SUBJECTS. 441 



iuliaence of friction tbat only a very small part of that force (due to 

 the gradient) is applied to the preservation of the movement of the air 

 between the inner and outer portions of the cyclone. 



If the friction is proportional to the first power of the velocily, then 

 it results from the equation for F\ that the rotatory movement of the 

 cyclone, and consequently the value of v, is proportional to the sine of 

 the latitude, whence the entire difference in pressure between the inner 

 and any definite exterior portions of the cyclones, or lAB, must be pro- 

 portional to the square of sin (p. Therefore, the mean monthly range of 

 the barometer must increase with the latitude, and nearly in the ratio 

 of sin^ (f. This also corresponds very well with observation. If we put 

 the mean monthly range at the poles at 1.0'°''^ = 41™°>, we obtain the fol- 

 lowing comparison of the observed values (according to Loomis's Mete- 

 orology) and the computed values : 



^ = 0° 30O 450 650 780 

 Computed range 0™"^ 10""" 20°^™ 33"'°^ 35"™ 

 Observed range 3 10 25 34 31 



A more complete agreement is not to be expected, since other matters 

 come into consideration besides the influence of latitude. The violence 

 of the rotation in a cyclone, and consequently also the extent of the 

 central barometric depression, depends to a very large extent upon the 

 amount of aqueous vapor that is conveyed into the cyclone, and this 

 diminishes from the equator toward the poles. On the other hand, the 

 great cyclones that advance from the lower to the higher latitudes 

 increase steadily in diameter, and hence the amount of depression at 

 the center and the amount of the barometric oscillation at any one place 

 are always greater as we approach the Poles. These two effects mutually 

 enfeeble each other to a certain extent. In the slight variation of 3 

 millimeters at the equator is included the diurnal period, which has 

 exactly this amount. In fact, after eliminating this diurnal period, the 

 irregular variation becomes equal to zero, corresponding to the theory. 

 We have here a complete explanation, which has hitherto been wanting, 

 of the [irregular] barometric variations, and their dependence upon geo- 

 graphical latitude. 



These are the most important consequences in reference to cyclones 

 that Ferrel (Amer. Jour. Sci., Nov., 1874), in his above mentioned 

 memoir, deduces from his formula for the gradient. He, however, further 

 applies the same equation to explain also the general distribution of 

 atmospheric pressure under different latitudes on the earth's surface. 

 Wherever in the atmosphere, from any cause whatever, a rotatory motion 

 iis set up over a large portion of the earth's surface, this motion furnishes 

 a value of w and of v in the expression previously given for AB. The 

 term that depends upon u is, in such a case, very small in comparison 

 with that which depends upon 2 n sin ^, and can be neglected. Equally 

 can i be put =0, and therefore, also, v be put = r u. A positive v cor- 



