704 PROCEEDINGS OF THE AMERICAN ACADEMY. 



Several years ago, when I had to have a set of the best magnets I 

 could get for measuring purposes, carefully ground into shape after the 

 steel had been hardened, I experimented * upon hundreds of seasoned 

 magnets made of many kinds of steel ; ordinary tool steels, self-harden- 

 ing tool steels, and special magnet steels. In comparing round steels, 

 it appeared that such " Stub's Polished Drill Rod " as I could get made 

 slightly less desirable magnets than did " Crescent Drill Rod," or the 

 common brand of " Jessop's Black Rod," which last two were magneti- 

 cally indistinguishable. I found nothing better in tool steels than the 

 Crescent Drill Rod, or the Jessop's Black Round Tool Steel, and I have 

 used these as standards in testing my round cast-iron magnets. The sub- 

 joined table shows what may be expected of seasoned magnets made 

 of these steels. 



Most of my experiments on the strengths of round cast-iron bar 

 magnets have been made with pieces 18 centimetres long and either 

 0,95 centimetres or 1.25 centimetres in diameter, of which I have a good 

 many, some new and some cast two years ago. These were usually re- 

 laxed "f after their hardening by being boiled in water for some time ; next 

 they were magnetized to saturation in a solenoid, and then they were 

 again boiled and " aged " in the usual manner. The resulting magnets 

 were finally tested in competition with a large number of seasoned tool 

 steel magnets, of different brands but all of the same dimensions as the 

 castings, with the help of a mirror magnetometer. The cast-iron mag- 

 nets looked, of course, rather rough in comparison with others made of 

 polished rod, but their moments differed among themselves less than 

 those of an equal number of the steel magnets made of any one brand. 

 Just one of the tool steel magnets had a moment sensibly greater 

 (about 4 per cent) than any of the cast-iron magnets, but the average of 

 the moments of the cast-iron magnets was rather greater than those of 

 the steel, even after the records of the two or three steel magnets with 

 abnormally low moments had been rejected. 



Two years ago t I measured the temperature coefficients of a good many 

 bar magnets of cast iron and steel. In every instance, as was to be 

 exi)ected, the rate of loss of moment per degree of rise of temperature 

 was much greater at temperatures near the boiling-point of water than 

 at room temperatures, but if for purposes of comparison we used the 



* Peirce, American Journal of Science, 1898, p. 334. 



t Barns and Stronlial, Bnlletin of the U. S. Geological Survey, No. 14, 1885. 



J Peirce, These Proceedings, Feb., 1903. 



