122 THE PHYSICAL GEOGRAPHY OF THE SEA. 



Islands, is remarked upon with emphasis hy the microscopist. It 

 is worthy of remark, because, in connection with the investigations 

 at the Observatory, it is significant. 



289. The latitudinal limits of the northern edge of the north- 

 east trade-winds are variable. In the spring they are nearest to 

 the equator, extending sometimes at this season not farther from 

 the equator than the parallel of 15° north. 



290. The breadth of the calms of Cancer is also variable ; so 

 also are their limits. The extreme vibration of this zone is be- 

 tween the parallels of 17° and 38° north, according to the season 

 of the year. 



291. According to the hypothesis (§ 130) suggested by my re- 

 searches, this is the zone in which the upper currents of atmos- 

 phere that ascended in the equatorial calms, and flowed off to the 

 northward and eastward, are supposed to descend. This, there- 

 fore, is the zone in which the atmosphere that bears the "rain 

 dust," or "African sand," descends to the surface ; and this, there- 

 fore, is the zone, it might be supposed, which would be the most 

 liable to showers of this " dust." This is the zone in which the 

 Cape Verd Islands are situated ; they are in the direction which 

 theory gives to the upper current of air from the Oronoco and Am- 

 azon with its " rain dust," and they are in the region of the most 

 frequent showers of "rain dust," all of which, though they do not 

 absolutely prove, are nevertheless strikingly in conformity with, 

 this theory as to the circulation of the atmosphere. 



292. It is true that, in the present state of our information, we 

 can not tell why this "rain dust" should not be gradually precip- 

 itated from this upper current, and descend into the stratum of 

 trade-winds, as it passes from the equator to higher northern lat- 

 itudes ; neither can we tell why the vapor which the same winds 

 carry along should not, in like manner, be precipitated on the way ; 

 nor why we should have a thunder-storm, a gale of wind, or the 

 display of any other atmospherical phenomenon to-morrow, and 

 not to-day : all that we can say is, that the conditions of to-day 

 are not such as the phenomenon requires for its own development. 



293. Therefore, though we can not tell why the " sea dust" 

 should not fall always in the same place, we may nevertheless sup- 



