Tlcgidarity of Shifts of Wmch 185 



M'^est and South, and returned to S.E., after short intervals 

 of calm. We could thus perceive, on referring to the ship's 

 log, that the entire cycle was completed in five or six days; so 

 that it became quite possible, by examining the central 

 direction of the daily variation, to foretell the wind which 

 must be blowing twelve hours later, when, upon taking 

 into consideration the path described by such central direction 

 from day to day, it appeared that the wind described very 

 nearly a parabolic curve. 



Even the aspect of the heavens, and the state of the 

 weather, were only one degree less regular in their alterna- 

 tions than the hurricanes. With the S.E. wind, the sky 

 was bright, but as soon as it began to veer round, towards 

 afternoon, a few white belts of cirrhous clouds began to appear 

 in the western heavens, constituting a well-marked division 

 of the vault of the sky from one side quite to the other. As 

 it drew still further round, and neared the line of centres, 

 the weather grew foul, a driving scud covered the heavens, 

 and a succession of splendid rainbows were seen, till the 

 ship had reached the nearest spot to the storm-centre 

 when there were sharp squalls of wind, accompanied by heavy 

 showers of rain. The lower strata of clouds, mere vapour, 

 drove before the wind, while those above moved in a directly 

 contrary direction, generally that of the forthcoming wind. 

 The atmospheric pressure, which at first would be consider- 

 able, gradually decreased as we approached the central line ; 

 as we drew away from that centre the barometer rose again. 



