336 Messrs. J. E. H. Gordon and W. Newall on the Effect 



two small tubes for water, and one larger one for a thermometer. 

 The connexions from the water-tubes to the supply and waste- 

 pipes are made by means of india-rubber tubing, so that no jar 

 can be transmitted to the box. On the base, at each end of the 

 ivory plate, is fixed a brass upright with a horizontal set-screw. 

 The ends of these screws press on the glass clips just over the 

 ends of the magnet, and keep the box in its place. The water 

 supply is conveniently arranged so that by turning one tap hot 

 water is made to circulate through the box, and by turning an- 

 other cold water is sent through. 



The thermometer is by Casella, divided on the stem ; and on 

 its scale the division representing 1° C. is about 1 millim. long. 

 In the earlier experiments the readings were taken at 20°, 25°, 

 30°, and to 50° C; afterwards it was found more accurate and less 

 troublesome to take readings alternately at 20° and 50° C. The 

 bars were hardened by heating them to dull redness (i. e. about 

 530° to 560° C. (Daniel)) and plunging them into cold water. 

 They were then magnetized by being placed in a helix belonging 

 to an electromagnet containing 35 lbs. of insulated wire. The 

 helix was excited by from 4 to 6 pint Grove's cells, After being 

 magnetized, each bar was dipped alternately three or four times 

 into basins of hot (about 80°) and cold water. This was to 

 remove any magnetism which would be permanently lost by a 

 rise of 20° to 50° C. The specific gravities were taken in the 

 ordinary way ; the balance, however, was not good enough to 

 make it worth while to correct for the buoyancy of the air. The 

 apparent specific gravities in water at 15 0, 5 C. (60° F.) and at 

 37°-7 C. (100° F.) were taken. They were multiplied by the 

 specific gravity of water at those temperatures. These specific 

 gravities were calculated from data given by Dalton (see Miller's 

 'Elements of Chemistry/ vol. L p. 249, § 133). They were 



at 15°-5 C 0-9940 



„ 37°-7 C 0-9833 



If there is any error in these values it will merely affect the 

 value of our constant, not its constancy. The analysis was con- 

 ducted as follows, the percentage of iron only being determined, 

 and the amount of impurities inferred from it. A standard so- 

 lution of permanganate was first prepared by dissolving about 

 2 grms. of the pure crystallized salt in 1 litre of distilled water, 

 this solution being standardized in the ordinary manner by dis- 

 solving 0*5 grm. of pure iron in dilute sulphuric acid and testing 

 with the permanganate solution, when it was found that 1 cubic 

 centim. of the solution equalled 0*007573 grm. iron. Pieces 

 were chipped off the magnets, carefully cleaned, dissolved in 

 acid, and then tested with the solution, when the following 

 results were obtained. The analysis is as follows : — 



