PHYSICS, PROGRESS OF, IN 1891. 



733 



one. Hoy*'* fleet rometer has a disk tlmt weighs of instability is n ached, and thi-ir middm Hwing- 

 f a gramme. \\'. E. A\ rl..n. .1. IVm . and ing into their inking. 



Mini). HIT i London Ki.yal Society, .lunc -1) Their bcha\ i<.r thus explains exactly wh> 



find that the ordinary nnadrant electrometer 

 not even approximately obey the usually 

 ;ed law when the potential of the needle is 

 altered. This is due to electric action between 

 the guard tube and needle and to the tilting of 

 the needle at high potential. They have altered 

 the const ruct ion of the instrument so as to elimi- 

 nate thi>souree of error. 



'iliiimj 1 1 ml Li ijlttn ing-Rods. X. D. C. 

 Hodges (Institute of Electrical Engineers, April 

 '.Ml has noticed that in all recorded eases of 

 lightning stroke the dissipation of a small con- 

 dud or (>ay ^V of tin inch in diameter) has always 

 protected surrounding objects between two Hori- 

 zontal planes passing through its upper and low- 

 er ends. A query inserted in "Science" weekly 

 since .In ne 19 had elicited no exception to this 

 law by .Ian. 1. To apply it practically Mr. 

 Hodges has devised what he names a "lightning 

 dispeller," consisting of a thin, easily volatilized 

 conductor, that is intended toconvert the electric 

 energy at once into a harmless form. In oppo- 

 sition to his views. Prof. Elihu Thompson ("Elec- 

 tric World," Xew York. June 27) claims that 

 this invention rests on a fallacy, since the invent- 

 or seems to suppose that overcoming the resist- 

 ance of the conductor is the only work done by 

 the Mash, whereas this work is only a minute 

 fraction of the whole amount. The practical 

 experience of several jean is necessary full v to 

 decide on the value of the dispeller. Prof. 

 Thompson (ibid.) asserts that not all lightning is 

 oscillatory (see "Annual Cyclopa'dia," 1889. p. 

 699). A discharge that passes over a great 

 length of cloud he says takes some time to do so. 

 Magnetism. Its Nature. Prof. J. A. Ewing, 

 by further advances in the direction noticed last 

 year (" Annual Cyclopaedia," 1890, p. 719), has 

 thrown most important light on the molecular 

 proeoses of magnetization. His theory, which 

 is popularly set forth in a Royal Institution lect- 

 ure (May 22). is an extension of the generally ac- 

 cepted one (Weber's), that magnetization imparts 

 no new qualities to the molecules of a substance, 

 but simply reveals their magnetism (which had 

 been ineffective because their axes had all possi- 

 ble directions) by bringing them more and more 

 into parallelism. Ewing represents the magnetic 

 molecules by small magnetic needles, free to turn 

 on pivots, and studies their behavior when they 

 are subjected to the influence of an electric cur- 

 rent in a surrounding copper coil. When there 

 is no current through the coil, the magnets ar- 

 range themselves in stable groups. A feeble 

 current turns each slightly, but all return to 

 their original positions when it ceases. As the 

 current grows stronger, some of the less stable 

 groups suddenly assume new positionsof equilib- 

 rium, in which their axes are more nearly par- 

 allel to the magnetizing force, and others follow, 

 till finally with a very strong current all the 

 groups break up and new ones are formed. A 

 large proportion of the magnets, in this case, will 

 not go back to their old positions when the cur- 

 rent is stopped. A still stronger current of 

 course alters their direction little, as they are 

 now nearly parallel. The slow movement of 

 the needle* at first, their wavering as the point 



external forei- steadily increases, the ma;.-! 

 tioii of a metal first increase* slowly, thn. 

 rapidly, and then n. wly again, till 



finally a large addit ional force produces iiln 

 change at all. It also explains the phenoi; 

 of retentiveness, which on this thcon 

 to the assumption by the molecules of m-v. 

 manent configurations. ll\ "Annual 



Cyclopa'dia," 1889, p. TOL'j. which Ewing di 

 as "the tendency of the changes of magnetism 

 to lag behind changes in the magnetizing f 

 conies in, on his theory, whatever may be the 

 cause of the change, whenever it involves such 

 deflections of the molecules as t make tin-in 

 unstable. The unstable movements are i 

 versible with respect to the agent that produced 

 them ; that is to say, they can n< t be simply un- 

 done step by step as the agent is removed. The 

 dissipation of energy attendant upon a rcver.-al 

 of the magnetism of iron or steel, or upon any 

 cyclic change in it, is explained by the fact that 

 as a molecule " tumbles," as it "were, from an 

 unstable grouping, it oscillates, till its motion is 

 damped by the electrical eddies that it causes. 

 The well-known effects of vibration, tapping, or 

 strain in facilitating magnetization also follow 

 directly from the new theory. And since the 

 molecules of a piece of iron wire are ma_ 

 though there be no magnetization of the wire as 

 a whole, its physical qualities change during 

 loading and unloading in a manner involving 

 hysteresis. The theory also throws light on the 

 "time lag " of magnetization, which it shows to 

 be due to the way in which one molecular group 

 after another breaks un. The sudden loss of 

 magnetism when a metal becomes red-hot, Ewing 

 suggests is due to the fact that the molecular 

 vibration may be then so violent as to set the 

 molecules spinning, thus masking their polarity. 

 Arthur Hoopes ("Electrical World. "New York, 

 May Hi) has drawn curves represent ing the mag- 

 netic susceptibility of a Ewing's model. and they 

 correspond remarkably with those obtained for 

 actual metals. Prof. S. I'. Thompson (London 

 Physical Society, Nov. ~'s. 1890), in illustrating 

 Swing's theory with small charm com), 

 found that with small openly spaced magi 

 weak external force was comparatively effective, 

 and he thinks this may throw light on the molec- 

 ular groupings of the loadstone, the crystalline 

 variety of which has little or no magnetic sus- 

 ceptibility, while the heterogeneous variety has 

 a great deal. 



jfaonttiam and Strain. Prof. 0. 0. Knott 

 (Edinburgh Royal Society, .lunc 1) has investi- 

 gated some relations between magnetism and 

 twist in iron, nickel, and cobalt. A rectangular 

 rod of nickel 01 cobalt twists left-handedly when 

 a current is passed along it in the direction of 

 magnetization, while iron twists right-handedly, 

 unless strong fields arc used, lie obsetTfd ef- 

 fects, when an apparently demagnetized ire 

 was subjected to twist, that siigg'-ted that a 

 magnet i/ed wire may, in certain circuit)-;, 

 consist of alternate 1 , layers of opposite polarities. 

 From his own experiments and th- -e of others, 

 he concludes that the first effect of a -hearing 

 on molecular groupings is not only to 



