6o4 



NA TURE 



[Oct. 2S, 1880 



globules of sulphurous acid and liquid carbonic acid while in a 

 "spheroidal condition." In these cases, notwithstanding the 

 proximity of the hot vessel, the temperatures of the globules of 

 SOj and CO. are respectively as low as - 10" and - 73° C. 



It has long been remarked by physicists that some substances 

 pass dinrt/y from the solid to the gaseous state, without under- 

 gohig liquefaction : that is, when heated, tkey sublime without 

 vielling. Such bodies, under ordinary atmospheric pressures, 

 have their boiling points lo7i'er than their temperatures oi fusion ; 

 hence they volatilise without melting. Moreover it has lon<' 

 been known that such substances may be made to fuse by 

 subjecting them to an abnormal pressure sufficient to raise their 

 boiling points above their points of fusion. Thus the classical 

 experiments of Sir James Hall show that carbonate of lime may 

 be fused when heated under a pressure sufficient to prevent the 

 COo from escaping (Trans. Royal Soc. Edin., vol. vi. pp. 71- 

 1S6, 1805), In like manner metallic arsenic sublimes without 

 melting at iSo° C, under the ordinary pressure of the atmo- 

 sphere ; but the experiments of Landolt in 1859 show that 

 under artificial pressure it melts in globules at a low red heat. 

 It is evident that in these cases the raj>id vaporisation of the 

 .solids under ordinary circumstances prei'cuts the temperature 

 from reaching the point of fusion ; but when subjected to addi- 

 tional pressure the conditions of liquefaction are secured. On 

 the other hand, in the case of ice, it is obvious that iheiuithdrazaal 

 of pressure by lowa-in; its boiling-point places it in the same 

 category with metallic arsenic under ordinary conditions of 

 pressure. 



In relation to the literature of this subject it is proper to add 

 the following quotations from JI. V. RegnauU's " Elements of 

 Chemistry" (American Translation, Philadadelphia, 1865, vol, i. 

 p. 279). In speaking of the fusion of metallic arsenic under pressure 

 he says : — " The distance between the point of fusion and that of 

 ebullition of any body may, however, be increased at pleasure. 

 For the point of ebullition of a body is the temperature at which the 

 tension of its vapour is equal to the pressure exerted tipon it; and 

 hence by increasing the pressure the boiling-point is raised 

 without sensibly affecting the point of fusion." Again, he 

 says : — " Reciprocally it is evident that a volatile solid body may 

 be always subjected to so slight a pressure that it will boil ^t. a 

 temperature inferior to that at which it melts. Thus ice at the 

 temperature of - 1° C. possesses an elastic force represented 

 by 4-27 mm. ; in other words, it boils at a temperature of 

 — 1° C. under the pressure of 4-27 mm. Ice may therefore be 

 entirely volatilised by ebullition under this fee'ile pressure, with- 

 out reaching its point of fusion, which is 0° C.' 



Berkeley, California, September 30 Joil.N LeConte 



Wire Torsion 

 The phenomena described by JIajor Herschel in his letter to 

 Nature, vol. xxii. p. 557> ^'•d about which he asks for informa- 

 tion, are, v^'e think, quite easily explained by what is known of 

 the fluidity of metals. Yielding, or flowing, seems to occur in 

 all metals after a certain limiting stress has been reached ; indeed 

 it probably occurs, although perhaps to an immeasurably small ex- 

 tent, even with .small stresses (see Proc. Roy. Soc. No. 204, p. 41 1, 

 1880) ; but there is generally a limiting stress beyond which the 

 increase of strain due to yielding becomes comparable in magni- 

 tude with the ordinary sirains, which instantaneously disappear 

 on the removal of the load. 1 he bell-smith pulls his copper 

 wire, and makes it much longer before he thinks it fit for use ; 

 in a similar way the telegraph constructor stretches, or kills the 

 iron wire before he erects the line. Up to a certain limit of 

 pulling force, the wire obeys the well-known laws of elasticity; 

 slightly above that limit there is considerable fluid-yielding, there 

 being but very little yielding below that limit ; and at any instant 

 during the lengthening if the man ceases to pull, the wire shortens 

 a little. In fact at any stage the wire obeys the clastic law for 

 small stresses. Eventually the man ceases to pull, knowing that 

 the metal has lost most of its fluid properties, which can only be 

 restored to it by annealing. The same thiiii; occurs in brass, 

 although to a smaller extent than in copper, vvhicli can be expe- 

 rimentally proved in the following way : — Stretch a piece of well- 

 annealed brass wire in the manner described by Major Herschel 

 until it is nearly breaking ; and immediately set the wire vibrating. 

 Now the note given out by the stretched bra s wire, v.hich, as 

 is well known, -lepends on the tensile stress, will bo found 

 rapidly to go down m pitch. If the wire be li ditened up .again 

 sufficiently with the screw, the original note will a„'ain be heard, 



and the pitch will again go down, but not so rapidly as before. 

 Repeat this process until no flattening of the note is heard ; then 

 in this state we think that the experimenter will find the wire to 

 break with much less torsion than before, and to obey Hooke's 

 law more exactly. If it be desired to repeat the yielding or 

 flowing process, the wire must be previously again annealed. 



Mere sudden straining, even nearly up to the breaking stress, 

 is not sufficient to destroy the fluidity of brass ; time is required. 

 The yielding behaviour of a brass beam when loaded has been 

 studied by Prof. Ihurston (Trans. American Soc. of Civil Eng. 

 vol. vi. p. 28), and we may add that we have found that the 

 permanent state is alviays more rapidly reached when the wire is 

 subjected to rapid vibrations. 



It may be because torsion of a wire is more visible than longi- 

 tudinal strains (the twist being inversely proportional to the 

 fourth power of the diameter for a given twisting moment, 

 whereas the longitudinal strain for a given load is inversely pro- 

 portional to the square of the diameter) that fluidity is so much 

 more apparent in torsional experiments ; but we think it probable 

 that fluidity will be found always much more apparent when the 

 volume of the material acted on is unchanged, that is, when the 

 stress is mainly one of shear as it is in torsion. 

 ■ However this may be we can explain why wire which has been 

 "killed" for pulling forces is not " killed " for twisting, and 

 why it is more difficult to kill for twisting than for tensile 

 stresses. It is well known to wire-drawers that in whatever 

 state copper or brass wire may be, whether annealed or not, it 

 may be drawn smaller, although no doubt it requires less care to 

 draw it if it is annealed. We cannot merely pull wire much 

 smaller, it has to undergo a lateral pressure such as the die gives 

 it. Now in twisting a wire it everywhere receives tliis lateral 

 pressure, that is — imagine a right-handed spiral filament being 

 lengthened by the twist — then the other component of the twist 

 gives to the filament a compression at right angles to its length 

 which enables it to extend. It seems that this lateral pressure 

 is needed to overcome some sort of friction in the particles of the 

 metal tending to prevent their moving into the axis of the vnie, 

 and V hich therefore is greater as the section of the wire is larger, 

 and it is probably for this reason that a very thin -wire extends 

 much more, for a given initial length, before it is killed than a 

 thick wire. We have known a length of about fifteen inches of 

 fine copper wire which had just bsen drawn, and which had 

 been well killed, to bear six or seven hundred complete turns in 

 a lathe, one end being fixed, the other end turned, and the wire 

 kept pretty taut before it was accidentally broken, and even 

 afterwards parts of the wire could be considerably lengthened by 

 pulling. The nature of the explanation of this apparent anneal- 

 ing for tensile stresses arising from previous torsion will be 

 gathered from what follows. 



We infer that the three or four turns given to the wire at the 

 beginning in Major Herschel's experiment were not sufficient to 

 produce permanent torsional set ; why then should increasing the 

 tension during the torsion cause torsional set as well as lengthening 

 of the wire ? This is, we think, a more important question than 

 the one presented to us by the observations of fluidity in the 

 latter half of Major Herschel's letter, and which arose from the 

 metal having belonged to what Prof. Thurston calls the " tin 

 class " as distinguished from metals of the "iron class." 



The explanation we think is as follows, and it leads to the 

 conclusion that torsional fluidity is not independent of tensile 

 stress : — 



Suppose right- and left-handed spirals had been imagined in the 

 wire in question, making everywhere angles of 45° with the axis of 

 the wire ; then torsional strain, however set up, would consist in 

 the production of a difference in length of these two sets of spirals. 

 Now a twisting moment produces this effect ; it lengthens, 

 say the right-hand spiral and shortens the left, and we know that 

 up to a certain limit, which is tolerably high, the sante effect is 

 produced whatever be the tensile stress in the wire, which latter 

 simply tends to lengthen both spirals equally. In fact if 

 Hooke's law is true, the torsion is independent of the tension. 

 But above a certain limit of pull in the wire, the strain in the 

 direction of the right-handed spiral being everyw here due to the 

 sum of two tensile stresses, becomes so great that fluidity sets in 

 and permanent set is produced ; whereas in the direction of the 

 left-handed spiral the stress is due to the difference between the 

 teKsile .stress and the compressive part of the torsional shearing 

 stress, and this difference being small, no permanent tensile set 

 is produced, or at all events one much less than in the case of 

 the other spiral. Consequently if all stresses now cease to act 



