316 



Scientific Proceedings, Royal Dublin Society. 



In order now to test what would, be the best shape of magnet to make 

 with tliis same Stubs' steel, four magnets were made, each 10 cms. long, 

 and 0'4 cm. diameter. 



Magnet A was a cylinder witli its end-surfaces at riglit angles to its 

 length ; B had each of its ends rounded off into a hemisphere ; C liad its 

 end-surfaces at riglit angles to its length, but was turned down to 0'3 cm. 

 diameter at eacli end, the middle cross-section being 0'4 cm. diameter ; and 

 magnet D was the same shape as C, with each of its ends turned down to 0"2 cm. 

 diameter, the middle part also being 0"4 cm. diameter. Magnets C and D 

 were approximate elongated ellipsoids. These four magnets were hardened, 

 cleaned, weighed, magnetized, and tested as in the previous cases. The 

 magnetic moment per gramme of each was carefully determined, and also 

 the percentage loss in the magnetism of each, when they were let fall end on 

 (with the true north or south-seeking pole downwards) five times through a 

 height of 100 cms. on to a block of glass. The results obtained are shown 

 in Table III. 



Tablk III. 



Assuming that the magnets are all in the same physical condition, or 

 glass-hard, this shows that the magnet B, with the rounded ends, was the 

 least affected by percussion. 



Section II. 

 At intervals during the five years from 1900 to 1904 communications' 

 were brought before this Society on the magnetic properties of alloys of 

 iron, which were in the form of long rods. Most of the rods liad been 

 returned to the manufacturer for the purpose of being tested mechanically ; 

 there were still available short pieces of nineteen of tliese rods, and eleven 

 more samples had beeu obtained from the makers at the Hecla Steel Works, 



'Barrett, Brown, and Hadfield, Scient. Trans. Roy. Dublin Soe., vol. 

 vol. viii, 1902, p. 1 ; vol. viii., 1904, p. 109. 



vii., 1900, p. 67 ; 



