PRESIDENTIAL ADDRESS. G65 



the ueighbouihood where it occurs. This is consistent with what we know of tlio 

 development of cracks through repeated alternations of strain. 



Recourse to the model shows that with pyramidal piling the polar axes point 

 in so random a manner that the aggregate may fairly be called amorphous. To 

 illustrate this a group is shown with centres fixed at the corners of equilateral 

 triangles. 



It is obvious that any pyramidal piling at a surface of slip tends to bar further 

 slip at that particular surface. Hence not only the augmented hardness due to 

 strain, but the tendency in repeated alternations to lateral spreading of the region 

 on which slip occurs. The hardness due to straining is, of course, removed when 

 we raise the metal to such a temperature that complete recrystallisation occurs, 

 normal piling being then restored in the new grains. 



Taking a previously unstrained piece, it is clear that the facility with which 

 slip will occur at any particular surface of slip in any particular grain depends not 

 only on the nature of the metal and on the orientation of the surface in question 

 to the direction of the stress, but also on the amount of support the grain receives 

 from its neighbours in resisting slip there. In other words, for a given orientation 

 of surface the resistance to slip may be said to consist of two parts ; one is inherent 

 in the surface itself, and the other is derived from the position of the grain with 

 reference to other grains. 



To make this point clear, think of a grain (under stress) in which there is a 

 gliding surface oriented in the most favourable direction for slipping. Slip on 

 this surface can take place only when its yielding compels the neighbours (which 

 are also under stress) to yield with it, and the surfaces in these on which slip is 

 compelled to occur are, on the whole, less favourably situated. Hence the original 

 grain cannot yield until the stress is considerably in excess of that which would 

 suffice to make it yield if it stood alone, or had neighbours equally favourably 

 inclined. 



Apply this consideration to the case of steel, where there are two classes of 

 grains : the ferrite, which is simply iron, and the pearlite, which is a harder 

 structure. Slip on any ferrite grain is resisted partly by the strength of the surface 

 itself, and partly by the impossibility of its yielding without forcing slip to take 

 place on neighbouring (harder) grains. Now suppose the structure is a very gross 

 one, such as Mr. Stead has shown may be found in steel that is seriously over- 

 heated. On the large grains of ferrite in overheated steel the resistance to slip 

 will be but little greater than it would be in iron, and, consequently, under an 

 alternating stress fatigue of strength, leading to rupture, may be produced by a 

 very moderate amount of load. Mr. Stead ' has shown how the effects of over- 

 heating can be removed by the simple expedient of raising the steel to a tempera- 

 ture sufficient to cause recrystallisation— a homceopathic remedy that transforms the 

 gross structure of the overheated metal into an ordinarily tine structure, where 

 no ferrite grain can yield without compelling the yielding of many pearlite grains. 

 Hence we find, as Rogers^ has demonstrated by experiment, that steel cured by 

 reheating from the grossness of structure previously produced by overheating, has 

 an immensely increased power to resist the deteriorating effects of often repeated 

 stress. 



I trust you will not feel I have abused the license of the Chair in presenting 

 contributions to molecular theory that are for the most part in the nature of specu- 

 lative suggestions, thrown out in the hope that they may some time lead to fuller 

 and more definite knowledge. Remote as they may seem to be from the concerns 

 of the workaday engineer, they relate to the matter which it is his business to 

 handle, and to the rationale of properties, without which that matter would be 

 useless to serve him. We have attempted to penetrate into its very heart and 

 substance in order the better to comprehend the qualities and functions on which 



' See especially a paper by J. E. Stead and A. W. Richards on ' The .Restoration 

 of Dangerously Crystalline Steel by Heat Treatment,' Joum. of the Iron and Steel 

 Inst., No, 2, 1903. 



- F. Rogers, ' Heat Treatment and Fatigue of Steel,' Joum. of the Iron and Steel 

 Inst., No. 1, 1905. 



