438 



NATURE 



[September 6, 1900 



graph of rail steel magnified 850 diameters, in which 

 the light constituent is ferrite, and the alternate bands 

 of ferrite and cementite together make up the constituent 

 pearlite. Well-developed pearlite with a conspicuous 

 banded structure is characteristic of good rail steel. It 

 is the form of carbide produced by slow cooling. When, 

 however, steel is hardened by " quenching," pearlite is 

 no longer to be found, and the whole mass consists of 

 interlacing crystalline fibres devoid of banded structure, 

 and is called " martensite." 



With regard to this, Sir William Roberts-Austen says : 

 *'The detection of martensite in a rail should at once 

 cause it to be viewed with extreme suspicion, as showing 

 that the rail is too hard locally to be safe in use." An 

 examination of the running edge of the St. Neots rail 

 (Fig. 2) showed that a surface layer of about i/iooth 

 of an inch thick existed, in which the carbide was mainly 

 in the form of martensite. 



This surface layer forms the lighter upper part of the 

 figure, while the darker portion below it shows the usual 

 assemblage of pearlite and ferrite granules. The actual 

 unning edge is shown against the dark space at the top 



shock to spread through the mass. The induced flaws 

 of Sir Francis Marindin might thus be explained. 



On the other hand, a warning note is struck by Prof. 

 Unwin, who points out that minute transverse fissures are 

 common on the rolling surfaces of old rails ; and as only 

 one rail in 2000 or 3000 breaks on the road, the others 

 being discarded as worn out, it must be rare for such 

 fissures to spread into the substance of the rail. It is 

 suggested that if much silicon is present, the spreading 

 of fissures becomes more rapid ; but the St. Neots rail 

 contained only 0*09 per cent, of silicon, an amount well 

 within the limits of composition put forward by Messrs. 

 Windsor Richards and Martin as suitable for rail steel. 



On considering the problem of how martensite can 

 be formed on the rolling surfaces, it is again evident 

 that the St. Neots rail is a remarkable exception. All 

 old rails become " hammer-hardened " on the surface by 

 long use, and their strength and percentage of elongation 

 may be increased by annealing, but no clear case of any 

 production of martensite in this way could be obtained. 

 The effect produced by the "cold-rolling" action of 



Fig. 2. — St. Neots rail, running edge. Pearlite passing into martens! 

 X 140 D. 



Fio. 3. — St. Neots rail, rolling surface. 



of the figure. Martensite was also found in small patches 

 in some other worn and broken rails, but to a far less 

 extent, the St. Neots rail being unique in this respect. 



The questions that naturally presented themselves on 

 this discovery were, first, How far would this structure 

 account for the brittleness of the rail ? and secondly, How 

 had the martensite been produced ? With regard to the 

 first question, the rolling surface of the St. Neots rail 

 was found to be traversed by a number of transverse 

 cracks (Fig. 3), some just passing through the 

 hardened skin, others running into the substance of the 

 rail. The upper surfaces of rails are subject to tension 

 over chairs by the weight of passing trains applied 

 between the chairs, and cracks are formed in this way. 

 To reaUse the importance of these cracks, it is only 

 necessary to turn to Mr. Martin's memorandum. He 

 found that a heavy steel rail nicked with a chisel to a 

 depth of i/64th of an inch broke under the weight of six 

 hundredweight let fall from a height of twelve feet, while 

 the same rail, if not previously nicked, resisted success- 

 fully the fall of a ton weight from a height of twenty 

 feet. The loss of strength due to these minute cracks is 

 therefore amazing, and can only be accounted for on the 

 hypothesis that shallow nicks are readily induced by 



NO. 1610, VOL. 62] 



passing trains on steel is shown in Fig. 4, which is a 

 photograph enlarged 140 diameters of the running edge 

 of a rail taken up after ten years' wear. The direction of 

 the granules is changed in the surface layer, but other- 

 wise the structure is unaltered. Roberts- Austen succeeded 

 in producing a structure like that of the St. Neots rail, 

 but only by local heating with an electric arc. With 

 regard to this, he observes that this experiment " points 

 to the probability that local heating of a rail by skidding, 

 followed as it would be by the rapid abstraction of the 

 heat by the mass of the cold rail, can produce patches of 

 I martensite, though it may be very difficult in the labor- 

 j atory to imitate the actual conditions by mechanical 

 means." He seems to think that " the structure of the 

 St. Neots rail would point to the changes having been 

 effected while the rail was actually in use." 



Although the St. Neots rail has thus baffled the com- 

 mittee, inasmuch as they cannot say definitely whether 

 other rails are likely to be altered in the same way, it is 

 evident that one of the most important results of their 

 labours will prove to be the full realisation of the fact 

 that steel possesses a complex structure which can be 

 studied with the microscope, that this structure varies 

 greatly with the mechanical and thermal treatment to 



