85 



J. G. HoRSFALL, Esq. inquired what difference was found 

 in the different bars of the best iron ? 



Dr. Scoresby said they were extremely uniform in the 

 best iron ; not one specimen of that kind came anything near 

 the inferiority of the other. The difference did not exceed 

 in any case 33 minutes of a degree ; not the third part of 

 the distance between the two classes. He then went on by 

 the application of another principle to investigate the quality 

 of cast iron. For that purpose he had obtained six or seven 

 sets of the same size of the two extreme qualities. One 

 consisted of the best quality of cast iron, the other was of 

 inferior quality. They were of the qualities usually marked 

 1 and 3. There might be a considerable varfety in No. 1 or 

 3 ; but the difference between 1 and 3 was sufficiently charac- 

 teristic. There was also a considerable difference betwixt 

 the two classes in price, as well as quality. No. 1 had less 

 oxygen and a larger portion of carbon than No. 3, being of 

 a purer description. Without going particularly into the 

 chemical constituents of iron, he might just observe, that as 

 the best quality of iron had in the other case the highest 

 magnetic principle, he expected he should in this case of cast 

 iron find a similar law. He treated it, however, in a different 

 manner. He got the plates cast in the same way in green 

 sand," so that they should be very hard, and might exhibit 

 more of the nature of steel. His plan was to try them by 

 magnetising them and converting them into real magnets, 

 being of opinion that as the best steel produced the best 

 magnets, so the best cast iron would produce the best mag- 

 nets. His first experiment was by magnetising them sepa- 

 rately, and then carefully trying their powers by the compass. 

 One kind. No. 1, had the power of causing the compass to 

 diverge, on an average, 13 deg. 41 min., while No. 3 only 



caused it to diverge 10 deg. 7 min the difference between 



the two being as 136 to 100. Thus it appeared that the best 



