Intelligence and Miscellaneous Articles. 423 



It follows, from the above observations, that — 



1. On the hardening of steel there ensues a diminution of its 

 specific gravity — that is, an increase of its volume. 



2. The thicker the rod the less is it hardened ; that is, the smaller 

 is the increment of its volume, tins being reckoned from the volume 

 of the crude rod. 



3. By tempering, the specific gravity is again increased, conse- 

 quently the volume diminished ; so that when the temperiug-colour 

 is yellow the increment of volume is reduced to one half its previous 

 amount. This law holds rigorously with rods of all thicknesses * 

 the deviations lie quite within the limits of errors of observation. 



4. If the blue colour of oxidation be reached, the volume under- 

 goes a further diminution, in the mean about one fifth of the pre- 

 vious increment. The less accordance in this case between the 

 observations made on different rods of the same thickness probably 

 arises from the greater difficulty of distinguishing between the 

 darker hues of oxidation. 



5. It was only with the disappearance of all colours (in the 

 grey state) that the rods resumed the specific gravity and volume 

 which they possessed before the hardening — that is, in the crude 

 state in which they were procured. That, notwithstanding this, 

 the two states, grey and crude, are absolutely different follows (as 

 is shown he. cit.) from the fact that a grey rod takes up much more 

 permanent magnetism than a crude one (in an observed case more 

 than twice as much). 



6. If thereupon the rods are heated to bright redness and let cool 

 very slowly, the specific gravity again diminishes, or the volume 

 increases. This increment amounted on the average to somewhat 

 more than one sixth of that produced in the crude rod by quenching. 



I wish, in conclusion, to call attention to the interesting ques- 

 tion, How, in a hardened steel rod, does the density vary inward 

 from the periphery to the axis ? The supposition that first presents 

 itself, and which I made loc. cit. p. 165 ad fin., is that the rod may 

 be imagined as consisting of concentric layers whose hardness di- 

 minishes rapidly from the outside inwards, finally enclosing a uni- 

 formly soft core. I have in the meantime endeavoured to test this 

 view by the following observations : — 



The 7-millim.-thick rod 1, which at last (in the grey state) had 

 shown the specific gravity 7*8174, was hardened in cold water. 

 The sp. gr. then obtained was 7*7417. I now placed the rod in 

 dilute hydrochloric acid, and during a long period 1 daily deter- 

 mined its "specific gravity A. 



During the first nine days it sank, in an absolutely regular 

 manner, to A =7* 7130. The weight of the layer that had been 

 eaten away amounted to only 0*704 gram out of 29*74(3 grams, the 

 original weight of the rod. The mean specific gravity B of the 

 removed layer was calculated =10*5. During the following seven 

 days A rose again, just as regularly, to 7*7412 — thus nearly to the 

 initial value. The layer removed by the acid was m= 1*703 gram, 

 and its calculated mean specific gravity 3 = 7*3. 



