CLOUDS. 455 



Whatever may be the regularity presented by the circulation of the 

 winds, and of the clouds in the tropics, and whatever care, also, is 

 taken in studying it, this circulation is not yet exempt from perturba- 

 tions which mask a little the precise moment of the beginning and end 

 of each rotation. The southeast trade-winds, and the configuration of 

 the ground, are among the number of general disturbing causes ; while 

 the land and sea breezes, the cloudiness of the stratum of cumulus, 

 which is more or less time prolonged, and covers that of cirro-cumulus, 

 or these latter that of cirrus, their inclination in space, and conse- 

 quently their transformation accidental and sudden, constitute the local 

 disturbing causes. 



I surmise, besides, the existence of great annual rotations produced 

 by the earth's motion of translation in its orbit analogous to those pro- 

 duced by the rotation of the earth on its axis, both classes of rotation 

 belonging to each climate of the terrestrial zones, having regard to the 

 distribution of continents and seas, and to their physical constitution. 

 These annual rotations appear to commence and terminate at the north; 

 for the cirrus in October, for the cirro-cumulus in November, for the 

 cumulus in December, and for the wind in January. According to this, 

 the superior current emi)loys a month in accomi^lishing its rotation fi'om 

 stratum to stratum, continually approaching the siu'face of the earth, 

 and three months in reaching it. 



Lieutenant Maury claims that the trade- winds are so constant and 

 uniform that their direction no more changes than the current of the 

 Mississippi. I do not share the opinion of this savant, for the observa- 

 tions at Havana demonstrate, on the contrary, that the north trade- 

 wind varies from northeast, and sometimes north- northeast, up to east- 

 northeast, chiefly from December to May, the time at which the current, 

 from the northern hemisphere appears to be stronger than that from 

 the southern, and consequently it approaches the equator. In the sec- 

 ond part of the year, from June to November, the south polar current, 

 being more intense, drives back the first, and advances to the latitude 

 of Havana, and ijrobably to the parallel of SO'^ north, the trade-wind 

 then varying from east-northeast to southeast ; so the limit of the dis- 

 placement in latitude of the trade-winds depends more particularly on 

 the respective intensity of the i^olar currents of each hemisphere. We 

 see, therefore, that the time of appearance which I have established for 

 fracto-cumulus and cumulus seems to correspond with the displacement 

 of the trade- winds. 



In fine, it is at the moment when the rotations of the wind and of 

 cumulus correspond toward the southwest with that of the equatorial 

 current that storms and great showers have generally taken i)lace, in 

 presence of a compact stratum, and a condensation of superior pallio- 

 cirrus, and another stratum of inferior palUo-cumulus. But as soon as 

 the wind and the cumulus revolve to the west, the storm begins to clear 

 off, and the barometer rises. W^hen these first two rotations terminate 



