512 SUMMARY OF CURRENT RESEARCHES RELATING TO 



of that constituent of steel which in my lecture at the last annual 

 meeting I described as the pearly compound. High powers show that 

 it really has a ctructure closely resembling that of pearl, the surface 

 being marked by fine straight or curved parallel lines, due to the pre- 

 sence of alternating very thin plates of varying hardness. After only 

 a few hours of observation I felt almost certain that these thin plates 

 were iron free from carbon, and the intensely hard substance seen so 

 well in blister steel ; but the facts were so extraordinary and so unlike 

 anything I had ever seen or heard of in any mineral substance, that it 

 was not until after several months devoted to the careful study of all 

 the chief kinds of iron and steel that I felt confidence in the results. 



The chief facts are best seen in the case of an ingot of steel of 

 medium temper. On fracture comparatively large crystals are visible, 

 radiating from the surface to the interior. When a properly pre- 

 pared microscopical section is viewed with a moderate power, it is 

 easy to see that, after having crystallized out from fusion at a high 

 temperature, these large crystals break up on further cooling into 

 much smaller, as described in my lecture. What is now seen with 

 very high powers is that these smaller crystals finally split up into 

 alternating very thin plates. Taking all the facts into consideration, 

 it appears as though a stable compound of iron with a small amount 

 of carbon exists at a high temperature, which at a lower breaks up 

 into iron combined with a larger amount of carbon, and into iron free 

 from it. If these two products had not differed so much in hardness, 

 or if the alternating plates had been considerably thinner, or if 

 definite plates had not been formed, such a compound structure would 

 never have been suspected. It has probably never been specially 

 looked for in other substances, and might exist without being visible, 

 even with the highest and best magnifying powers. In those cases 

 where no subsequent segregation has occurred, these alternating 

 plates are often remarkably regular and uniform in thickness ; and 

 as far as I am able to judge, the softer plates are about double the 

 thickness of the harder. If so, we may say that the thickness of the 

 softer plates is about 1/40,000 in., and of the thinner 1/80,000, thus 

 giving well-marked stri89 1/60,000 in. apart. To define even these 

 requires very careful adjustment of the object-glass j and, considering 

 all the circumstances of the case, it could not be expected that the 

 two bounding edges of the thinner hard plates could always be defined 

 so as to show a flat intermediate surface. We are, in fact, brought 

 face to face with an optical difficulty, depending on the considerable 

 length of waves of light compared with the objects under examination, 

 and are obliged to infer the nature of the very fine structure from 

 what is seen when it is somewhat coarser. In some cases it is easy 

 to trace the gradual passage from these extremely thin plates up to 

 those which are sufficiently thick to show clearly that the structure 

 is due to thin plates of the hard substance between soft iron. No 

 mere cleavage would explain all the facts, though it is extremely 

 probable that the direction of the alternating plates was determined 

 by the previous crystalline structure. In some cases the plates are 

 less well marked, and the structure is more granular. 



