766 TRUEBLOOD. 



writer has learned in conversation of a thermometer made of the best 

 platinum obtainable, the 5 of which was 1.68; and in a paper on the 

 mechanical equivalent of heat,^ Roebuck describes two thermometers 

 made from the same sample of platinum, for which the 5's were 

 respectively 1.748 and 1.567, with fundamental coefficients of nearly 

 equal magnitudes (0.0035127, 0.0035387). 



h. Calihration of individual thermometers. 



The temperature scale in all the work here described depends on 

 calibrations in ice, steam and naphthalin vapor. The naphthalin 

 used was Kahlbaum's Reagent, its boiling point being taken as 



tn = 218.0 + 0.058 (// - 760) 



in which H is the pressure in millimeters of mercury under standard 

 conditions.® Ordinary drug-store naphthalin boils at almost the 

 same temperature. The apparatus used for the boiling point of 

 naphthalin was thaj: which has become standard for the sulphur point.^ 

 Drip-cones of asbestos paper and aluminum radiation shields were 

 regularly used, but little effect was noted when either or both were 

 omitted. 



Some care was found necessary at the ice point. A thermos bottle 

 was used in the earlier work, but the apparatus finally found most 

 reliable consisted of a 3-inch glass cylinder, lagged with asbestos, and 

 provided with a funnel-shaped bottom, to which a rubber tube with 

 stop-cock could be attached. The cylinder was long enough to permit 

 immersion to the head of the thermometer. It was found expedient 

 to cool the thermometer bulb and stem in a separate ice-bath before 

 immersion in this ice-point apparatus. Experiments conducted to 

 determine whether there was any difference in the freezing points as 

 given by commercial (natural) ice and by ice made from distilled water 

 showed no difference to the order of the accuracy of the resistance 

 measurements, and the commercial ice was thereafter regularly used, 

 because of the lack of facilities for making the other. 



Distilled water was always used at the steam point. 



5 Phys. Rev. (2), 2, 79 (1913). 



6 See Waidner and Burgess, Bull. Bur. St.. 7, 1-9 (1911). 



7 See Waidner and Burgess, Bull. Bur. St., 6, 149-230 (1909-10). 



