METEOROLOGY. 427 



of tbe opiuion that the organization should go further, and arrange 

 for a regular inspection of all public buildings which had lightning- 

 conductors applied. The necessity for this he demonstrated by ad- 

 ducing a number of striking cases where damage, more or less severe, 

 had occurred to buildings, even though having lightning-conductors 

 attached to them. The cases now cited, he explained, were supple- 

 mentary to those communicated in his paper on a similar subject to 

 the association in 1878. {Nature, xxii, p. 44G.) 



Balfour Stewart, in reference to the subject of terrestrial magnetism, 

 propounds the following working hypothesis: "May there not be a fee- 

 ble magnetic nucleus in the earth around which the great convection 

 currents, the trades and anti-trades, move, as conductors moving across 

 lines of magnetic force, and the tendency of which will be to swell up 

 and sustain the magnetism of the whole earth to the point of saturation ? 

 If this be granted, then the changes in winds and currents must produce 

 corresi)onding changes (diurnal, annual, secular, and nou-i)eriodic) in 

 terrestrial magnetism, so far as we can observe the latter." He adds 

 that, with Mr. Dodgson and Mr. Hiraoka, he has convinced himself of 

 the existence of intimate connection between these phenomena. [The 

 same conclusion has been independentlj- reached by Vines at Havana.] 

 We may look upon the earth as composed of concentric la^'ers: (1) a mag- 

 netic nucleus ; (3) non-conducting primary rock ; (3) the conducting moist 

 earth and ocean 5 (4) the lower dense non-conducting air; (5) the upper 

 thin conducting atmosphere. We observe earth currents in layer 3, and 

 annual discharges or currents in layer 5. The former are apparently 

 secondary currents induced by magnetic changes ; the latter may also 

 have a similar origin. {Nature, xxii, p. 14G,) 



In a lecture on solar physics Professor Stewart stated that he be- 

 lieved one great cause of weather changes to be solar variability, in 

 which we have periods of short length, as well as others extending 

 over many years. 



These terrestrial weather changes, it is sufficiently well known, are 

 l)ropagated from west to east after they have once appeared. 



Again, there are variations in the diurnal declination range which 

 may be said to constitute magnetic weather. 



These are also caused by solar variability, and it is suspected that 

 they are likewise propag.sted from west to east, although more quickly 

 than the well-understood changes of meteorological weather. 



It would thus appear to be at least possible that British magnetical 

 weather of to-day may be followed by corresponding meteorological 

 weather five or six days hence. 



Professor Stewart has made a preliminary trial which induces him 

 to think that this is the case, and that it may ultimately be possible to 

 forecast British meteorological weather by means of magnetic weather 

 some hve or six days previous to it. {Nature, May 5, 1881, xxiv, p. 7.) 



Prof. Stewart, in a lecture on the connection between solar and ter- 



