82 Scientific Proceedings^ Roi/al Dublin Society. 



By comparing Nos. 1 and 2 in the table, it looks as if the high silicon in 

 No. 2 had decreased its retentivity ; the same appears to apply to No. 4. 

 Taking Nos. 1 and 7, with nearly the same amount of chromium, the 

 smaller carbon and larger manganese in No. 7 seem to counteract each 

 other ; also, the equal quantities of tungsten and copper neutralize to make 

 the retentivity of the magnet poor. 



The magnets 3, 5, and 6 are good, and confirm an observation made by the 

 writer, and given in the paper referred to above, that the best proportion of 

 chromium for a magnet was about 2'5 per cent. ; the results for No. 3 magnet 

 more nearly illustrate the fact. The actual moment per gramme of No. 5 is 

 low, though its retentivity is high, due possibly to the presence of copper 

 neutralizing the chromium, as well as its fairly higli carbon. On the whole 

 it is evident, as far as permanence is concerned, that the best magnets would 

 be Nos. 3, 5, and 6. 



Column D in Table III gives the total percentage loss in the magnetic 

 moment per gramme due to the magnets being allowed to fall, end on, from 

 the height of one metre /owr times on to a block of glass ; in every case with 

 the true north or soioth pointing pole downwards. 



