On the "Rotational Coefficients " of various Metals. 341 



The thermostat is next placed in position in the tube C 

 (fig. 1), the thermometer placed beside it, and the bath filled 

 up with water ; the air current is then turned on, and the gas- 

 supply so arranged that it enters the thermostat by E and 

 passes to the Bunsen by the side tube of D. A very slow 

 stream of water is allowed to flow through the spiral, and the 

 Bunsen is lighted. The stopcock B, at first closed, is opened 

 from time to time, until, when the desired temperature is 

 reached, the column of mercury in C just touches the end of 

 E; the flame of the Bunsen is thus lowered, and cold water 

 flows through the spiral cooling the bath. As the temperature 

 of the bath falls, the mercury falls below the end of E, and the 

 water is heated before entering the bath; after a few oscilla- 

 tions of temperature a state of equilibrium is reached, in which 

 the size of the flame is so adjusted to the quantity of water 

 flowing through the spiral that the temperature of the bath 

 remains constant. 



I have found that the extreme variation of temperature does 

 not exceed o, 05 for a temperature of 20° C, and that on 

 lighting the gas in the morning the constant temperature was 

 attained in less than an hour, and never varied from day to day. 



XLIX. " Rotational Coefficients " of various Metals. 

 By Edwin H. Hall*. 



THE experiments described below were made at the labo- 

 ratory of Harvard College during the summer of 1882; 

 and most of the results obtained were given at the Montreal 

 meeting of the American Association. 



At the York meeting of the British Association, September 

 1881, I gave a list of certain metals with an approximate 

 value of the " rotational coefficient "f for each, as determined 

 by my experiments. This list was published in the Beport of 

 the Association. Several of these metals had, however, been 

 examined in an extremely inaccurate manner, as was stated 

 at the time, and the numbers assigned them were marked as 

 doubtful. Thus a part of the list ran : — 



Name of metal. National coefficient, 

 arbitrary scale. 



Zinc . . . 



. . +15? 



Aluminium . 



. . -50? 



Magnesium . 



. -50? 



Copper . . 



. -10? 



Brass . . . 



. - 1-3? 



Lead . . . 



. . No effect discovered 



* From Silliman's American Journal for March 1883. 

 t Phil. Mag. Sept. 1881, p. 162. 



