ON THE INFLUENCE OF SILICON ON THE PKOPERTIES OF STEEL. 71 



A general summary of the results is given in Table A., wLicli contains 

 both the chemical analysis, the works tests, and the mean of the mechanical 

 tests of the specimens prepared as before described. 



Specimen No. 1 was prepared from the metal taken from the ladle, it 

 being poured into a red-hot crucible, and afterwards treated in exactly 

 the same way as in the ten succeeding experiments. In this case, how- 

 ever, no silicious iron was added, and the values given are intended 

 merely for comparison with those which follow. The analysis of this 

 specimen was not made fi-om any of the samples actually tested, but is 

 what may be considered an average composition of such metal. The 

 mechanical values are the mean of four experiments with metal from two 

 separate charges. 



The silicon pig used had the following composition : — Carbon, 1-96 j 

 silicon, lO'SO ; sulphur, 0U2 ; manganese, 1"90 ; phosphorus, 017. 



On examining the results in detail, the following observations will 

 be made : — 



1. Chemical Composition. — The silicon gradually increases from 

 specimen No. 2, through the series to No. 11. The other constituents, 

 though tolerably uniform, still show sufficient variation to influence the 

 properties of the product, and allowance nmst be made for these 

 differences in composition in drawing any conclusion as to the influence 

 exerted by silicon. 



2. Works Tests. — It will be noticed that all the specimens examined 

 rolled well, and that, with one slight exception, all behaved satisfactorily 

 Tinder the hot test. This is in very marked contrast to what was observed 

 in the previous experiments, when silicod produced distinct red-shortness ; 

 the difference in the present case is doubtless due to the presence of man- 

 ganese in the ingot metal. The cold or bending test was also satisfactory 

 in all cases, with the single exception of No. 11, which contained 0504 

 per cent, of silicon and 0121 per cent, of phosnhorns. In this sample it 

 is not certain that the metal would have behaved in the same way if the 

 phosphorus had been as low as in the other cases. In the welding tests 

 the metal behaved well in every instance, showing that the presence of 

 silicon has no perceptible influence on the welding property. In the cold 

 and welding tests the i-esults are the same as was noticed in the previous 

 series of experiments. 



Mechanical Tests. — In the original metal both the limit of elasticity 

 and the breaking load, which are given in tons per square inch, ai'e rather 

 higher than usual in this class of metal, while the extension and reduction 

 of area ai'e rather lower than is common. These diS'erences are, however, 

 not great, and may be accounted for by the comparatively small scale on 

 which the experiments were performed. 



Limit of Elasticity. — This varies in the first six specimens (Nos. 2 to 

 7) over a maximum I'ange of l'G5 tons, and these small variations are of 

 such a kind as may be explained by differences of composition other than 

 those of silicon. In the specimens with more silicon, however, there is a 

 distinct increase in the elastic limit due to silicon. 



Brealdnrj Load. — This varies in a manner which closely resembles 

 that observed with the limit of elasticity. In the first six specimens the 

 variations are irregular, the maximum range being 41 tons, and this 

 variation can be accounted for apart from any influence due to the silicon 

 present. With more silicon, however, there is a distinct increase of the 

 breaking load, and this is doubtless due to silicon. 



