PHYSICS OF THE GLOBE. 179 



make an angle of 90 with the bearing of the centre, but that 

 there is an indraft of from 25 to 31, beins; on the average 

 29, or 2-| points. It is probable that there is really a great 

 rano-e in the amount of indraft, so that this mean is a rouoh 

 approximation to the truth. The best rule for navigators is, 

 he thinks, about as follows: 



In the northern hemisphere with the wind 



North, the hurricane centre probably bears E.S.E., or more southerlv. 

 East, " " " " " S.S.W., " westerly. 



South, " " " " " W.N.W., " northerly. 



West, " " " " " N.N.E., " easterly. 



The only modification of the ordinary instructions for han- 

 dling ships in hurricanes which these facts suggest is, that 

 when the circular theory states that a ship ought to run be- 

 fore the wind, these facts show that, if possible, she ought in 

 the northern hemisphere, to keep the wind well on the star- 

 board quarter. He also shows that the indraft is greater in 

 one quarter of a hurricane than another, and greater near the 

 centre than farther from it. The mean of the three charts 

 just mentioned gives an average wind force of 10 on Beau- 

 fort's scale, at a distance of 90 miles from the centre, and a 

 force of 6 at a distance of 425 miles from the centre. 



In the chapter on the normal circulation of air during Au- 

 gust, 1873, lie introduces a very suggestive letter from Clem- 

 ent Ley, with reference to the relations between the upper 

 and lower currents of the atmosphere circulating round areas 

 of barometric depression. Mr. Ley has evidently established 

 inductively, from the consideration of a great number of ob- 

 servations, certain laws of atmospheric circulation which 

 have been theoretically and more or less vaguely guessed 

 at by several meteorologists of the past generation. His 

 diagram represents two low areas, With a high between, 

 all advancing due eastward. The lows are preceded by 

 southwest to southeast and northeast surface winds, which 

 are also ascending winds, and are accompanied by clouds, 

 either cumulus, stratus, or cirrus, above which he represents 

 descending currents from the northwest and southwest, which 

 "feed the preceding area of low pressure, and produce an area 

 of clear skv. The advancing: edcre of a bank of cirro-stratus 

 cloud is very well marked, and he states that he has in his 

 diagram given its mean position in relation to the isobar and 



