352 



Scientific Proceedings^ Royal Dublin Society. 



would seem as if the effects of the diminished carbon and increased 

 manganese in No. 7 were just balanced by the presence of the 1'9 per 

 cent, of tungsten and the same amount of copper. No. 7, however, is 

 less retentive, as will be seen from the percentage loss being nearly doubled ; 

 and when this material was tested in the form of a thin magnet 10 cm. 

 long, and having a dimension ratio of 60, its magnetic moment per gramme, 

 when glass-hard, was found to be 55"7, and the percentage loss due to falling 

 four times through a height of one metre was 8*1 as compared with 6'4, 

 when tlie dimension ratio was 33. On the whole, therefore, the specimen 

 No. 2 would be the best material to use in making a permanent magnet. 



Dimension -ratio. 



A good length of wire, one millimetre in diameter, was obtained of some- 

 what the same composition as No. 5, but without the copper, viz. (C = 0-77, 

 Si =0'5,Mn = 0'61, andCr = 5'19); and from this wire fotirteen magnets were 

 made, varying in length from 20 em. to 1 em. These magnets were 

 hardened, one at a time, cleaned, weighed, and tested for magnetic moment 

 per gramme, exactly as in the previous oases ; and the results when plotted 

 with the dimension-ratios as abscissae, and the corresponding magnetic 

 moments as ordinates, gave a very smooth curve, as shown in figure. From 

 the curve we see that a magnet 10 cm. long, or of dimension -ratio 100, is 

 the most effective magnet to make of this material, and has a magnetic 

 moment per gramme of 65, an increase of 65 per cent, on the value for No. 5 

 in Table I., which has a dimension-ratio of 33. 



