PRESIDENTIAL ADDRESS. 659 



slips leaves it still a crystal. If part of llie group of bricktats moves forward, 

 keeping parallel to themselves and to the others, the formation remains regular, 

 except that a step is formed on the outermost rows ; the orientation of the ele- 

 ments continues the same throughout. Considerations which I shall mention 

 presently lead to some qualification of this statement. I now see reason to believe 

 that in the process of slip there is a disturbance of the elementary portions or 

 brickbats adjoining the plane of slip, which may alter their setting, and thereby 

 introduce to a small extent some local departure from the perfectly homogeneous 

 orientation which is the characteristic of the true crystal. In very severe straining 

 there may even be a wide departure from true crystalline character. We shall 

 recur to this later ; but meanwhile it will suffice to say that substantially the slip 

 which is involved in a plastic strain of moderate amount is a bodily translation, 

 parallel to themselves, of part of the group of elementary brickbats or molecules 

 which build up the grain. If a crystal whose form has been altered, even largely, 

 by such straining is cut and polished and etched it appears, under the microscope, 

 to be to all intents and purposes as regular in the tactical grouping of its elements 

 as any other crystal. 



Further, in the process of straining we have, first, an elastic stage, extending 

 through very small movements, in which there is no dissipation of energy and no 

 permanent set. When this is exceeded, the slip occurs suddenly ; the work done 

 in straining is dissipated ; if the straining force is removed a strain persists, 

 forming a permanent ' set ' ; if it continues to act it goes on (within certain 

 limits) producing augmented strain. In general a large amount of strain may take 

 place without the cohesion between the gliding surfaces being destroyed. Im- 

 mediately after the strain has occurred there is marked fatigue, showing itself in 

 a loss of perfect elasticity ; but this will disappear with the lapse of time, and 

 the piece will then be harder than at first. If, on the other hand, a process ot 

 alternate straining back and forth be many times repeated, the piece breaks. 



These are now familiar facts. Can we attempt to explain them on the basis of 

 a molecular theory which will at the same time offer a clue to the process of 

 crystal-building as" we find it in metals ? I venture to make this Address the 

 occasion of inviting attention to some more or less speculative considerations 

 which may be held to go some little way towards furnishing the material for such 

 an explanation. 



At the Leeds Meeting of this Association, in 1890, it was my privilege to 

 bring forward certain contributions to the molecular theory of magnetism, and to 

 show a model which demonstrated that the rather complex phenomena of magneti- 

 sation were explainable on the very simple assumption that the magnetic 

 molecules are constrained by no other forces than those which they mutually exert 

 on one another in consequence of their polarities.^ From this were found to result 

 all the chief phenomena of permeability and magnetic hysteresis. Let us attempt 

 to-day to apply consideratious of a similar character to another group of physical 

 facts, namely ,"those that are associated with the crystalline structure of metals 

 and with the manner of their yielding under strain. Just as in dealing with 

 magnetic phenomena, I take as starting-point the idea that the stability of the 

 structure is due to mutual forces exerted on one another by its elementary parts 

 or molecules, and that the clue to the phenomena is to be sought in the play of 

 these mutual forces when displacement of the molecules occurs. 



Iron and most of the useful metals crystallise in the cubic system; for 

 simplicity we may limit what has to be said to them. Imagine a molecule 

 possessing polarity equally in three directions, defined by rectangular axes. We 

 need not for the present purpose inquire to what the polarity along the axes is 

 due ; it will suffice to assume that the molecule has six poles, three positive and 

 three negative, and that these repel the like and attract the unlike poles of other 

 molecules.. We may make a model by using three magnetised rods fixed at right 

 angles to one another at their middle points. I imagine, further, that the 



' ' Contributions to the Molecular Theory of Induced Magnetism,' Proc. Roy. Soe. 

 vol.xlviii. June 19,1890, or Phil. J/i/i?., September 1890, 



nu2 



