80 



Scientific Proceedings, Royal Dublin Society. 



Comparing Nos. 5 and 14, which have high carbon and a high percentage 

 of chroniiinn, the former " having also over 3 per cent, of manganese, we see 

 that the percentage loss in No. 5 is less than half that in No. 14, due no doubt 

 to the presence of the manganese. 



Of the four tungsten steels, Nos. 9 to 12, inclusive, the first seems the best 

 from the point of view of retentivity ; and it is known that a steel containing 

 about 7'5 per cent, of tungsten makes the most permanent magnet ;i and 

 No. 9, though the actual magnetic moment is low, is a fairly permanent 

 magnet. No. 15 is a good magnet, though .its moment is low compared with 

 No. 16, which has 1 per cent, more of chromium ; the retentivity of the former 

 is over three times that of the latter. 



In the paper above referred to (April, 1910], in Table IV, page 318, it 

 was shown that the magnetic moments all increased slightly one month after 

 they had had the final five hours' annealing by steam, so that if we were to 

 calculate the total loss from that time, July, 1909, till now, January, 1920, 

 the magnets, with the exception of No. 8, arrange themselves into three 

 natural groups, as shown in Table II, where Column E gives- the total 

 percentage loss in ten and a half years. 



Table II. 



From 2-7, manganese steels, the loss is about 25 per cent. ; from 9-11, 

 tungsten steels, the loss is about 20 per cent. ; and from 12-16, chrome steels, 

 the mean loss is over 25 per cent. 



^Soient. Trans. Roy. Dub. Soc, vol. vii, Plate vi, January, 1900. 



