420 



NA TURE 



[August 23, 1906 



cubic or normal piling. The group in that case takes an 

 arrangement which is essentially a repetition of this 

 quartette : — 



Fig. 



Along each row the polarity preserves the same direction, 

 but the polarity of each row is opposite to that ol each 

 contiguous parallel row. This description applies equally 

 to all three axes. The whole group (Fig. 3) consists of 

 the quartettes of Fig. 2 piled alongside of and also on top 

 of one another. In this way we arrive at what I take to 

 be the simplest possible type of cubic crystal. 



In this grouping each molecule has the alignment giving 

 maximum stability, and it seems fair to assume that it 

 will take that alignment when the crystal grain is formed 

 under conditions of complete freedom, as in solidifying from 

 the liquid state. .\s a rule, the actual process of crystal- 

 building goes on dendritically ; branches shoot out, and 

 from them other branches proceed at right angles, leaving 

 interstices to be filled in later. We have, therefore, to 

 conceive of the molecules as piling themselves preferably 

 in rows rather than in blocks, though ultimately the block 

 form is arrived at. In this position of maximum stability 

 each molecule has its six poles touching poles of contrary 

 name. 



Now comes a point of particular importance. Imagine 

 two neighbouring molecules in the same block to be turned 

 round, each through one right angle, in opposite senses. 

 They will now each have five poles touching five poles of 

 contrary name, but the sixth pole will touch a pole of 

 the same name as itself. They are still stablv situated, 

 but much less stably than in the original configuration, 

 and they will revert to that configuration if set swinging 

 through an angle sufficient to exceed the limited range 

 within which they are stable in the new position. 



Similarly we may imagine a group of three, four, or 

 more molecules, each to be turned through a right angle, 

 thereby constituting a small group with more or less 

 stability, but always with less than would be found if the 

 normal configuration had been preserved. The little group 

 in question may be made up of molecules in a row, or it 



^^ + + 



F.a. 3. 



may be a quartette or block, or take such a form as a 

 T or L. K sufficient disturbance tends to resolve it into 

 agreement with the normal tactics of the molecules which 

 build up the rest of the grain. 



It is conjecturally possible that small groups of this 



NO. 192 1, VOL, 74] 



kind, possessing little stability, may be formed during the 

 process of crystallisation, so that here and there in the 

 grain we may have a tiny patch of dissenters keeping one 

 another in countenance, but out of complete harmony with 

 their environment. 



If this happens at all during crystallisation, it would 

 seem less likely to happen in free crystallisation from a 

 liquid state than in the more constrained process that occurs 

 when a metal already in the solid state recrystallises at a 

 temperature far below its melting-point. Though rare or 

 absent in the first case, it might occur frequently in the 

 second. There are differences in the appearance of crysta' 

 grains under the microscope in metal as cast and in meta' 

 as recrystallised in the solid state, of which this may be 

 the explanation. It may also explain a difference pointed 

 out by Rosenhain,' that the slip lines in cast metal are 

 straight and regular, whereas in wrought iron and other 

 metals which have recrystallised in the solid they rarely 

 take a straight course across the crystal, but proceed in 

 jagged, irregular steps. These may be due to the presence 

 here and there of small planes of weakness, resulting from 

 the existence of what I have called dissenting groups. 

 ."Vgain, these groups, possessing, as they do, less stability 

 than their normal neighbours, may be conjectured to differ 

 from the normal parts of the grain in respect of electro- 

 lytic quality, and to be more readily attached by an etch- 

 ing reagent. Hence, perhaps, the conspicuous isolated 

 geometrical pits that appear on etching a polished surface 

 of wrought iron. 



It will help in making clear these points, and others 

 that are to follow, if we study the action of a model formed 

 by grouping a number of polarised " molecules " in one 

 plane, supporting them on fixed centres, about which they 

 are free to turn. In the model before you the centres are 

 uniformly spaced in rectangular rows, and the " mole- 

 cules " are + shaped pieces of hardened steel, strongly 

 magnetised along each of the crossed axes, each having, 

 therefore, two north poles and two south poles. The third 

 axis is omitted in the model, the movement to be studied 

 with the help of the model being movement in one plane. 

 On placing these " molecules " on their centres they readily 

 take up the position already indicated in Fig. 3. Each one 

 within the group has its four poles in close proximity to 

 four poles of contrary name, and is, therefore, highly 

 stable. If disturbed by being turned through a small angle, 

 and let go, it swings back, transmitting a wave of vibra- 

 tion through the group, which is reflected from the edges, 

 and is finally damped out in the model by pivot friction 

 and air friction. We may assume some damping action 

 (say by the induction of eddy-currents) in the actual solid, 

 of which the model may be taken as a very crude repre- 

 sentation. 



By turning two molecules carefully round together, each 

 througTi one right angle in opposite senses, we set up a 

 dissenting pair, the equilibrium of which has feeble stability. 

 A slight displacement, such as might be produced by the 

 transmission of a vibrational wave, breaks them up, and 

 thev swing back to the normal configuration, giving out 

 energy, which is taken up by the rest and is ultimately 

 dissipated. By making the dissenting coterie consist of 

 three or more we can give it additional strength. 



An example is shown in Fig. 4, where the three molecules 

 marked a, h, and c are turned round in this way. 



Notice that the normal molecule d, adjoining a line of 

 such dissenters, is in a peculiar position. His neighbours 

 present to him three N. poles and one .S. pole. He has 

 the choice of conforming to the majority, or of throwing 

 in his lot with the dissenters: and he has a third possible 

 position of equilibrium (very feeble equilibrium) which is 

 reached when his two S. poles are turned until the one 

 neighbouring south pole faces just between them. 1 have 

 laboured these points a little because they seem important 

 when we come to speak of the effects of strain. 



Consider now the straining action, which we may imitate 

 in the model by sliding one part of the group past the 

 other part. For this purpose the centres are cemented to 

 two glass plates which can slide parallel to one of the 

 axes. 



I Rosenhain. "The Plastic Yielding of Iron and Steel, ■'/<!«<. fimi and 

 Stct'I Institute^ No. i for 1904, p. 335. 



