132 A. L. Day and R. B. Bosnian — 



series represents an entirely different curve of temperature 

 variation along the bar. In some, cases the temperatures at 

 the ends were lower than at the middle, in others higher than 

 at the middle, and in one series one end was higher and the 

 other lower. The mean of all, therefore, probably eliminates 

 any error which might arise from variation of temperature 

 along the bar. 



The results are represented within the limits of error by the 

 straight line equation : 



10 6 /3 = 8-79-1- 0-00161*. 



This may be compared here with the expansion coefficients 

 between 300° and 1000° determined by the authors for the 10 

 per cent iridium alloy,* and of Holborn and Dayf for the 20 

 per cent iridium alloy and for pure platinum : 



80 Pt. 20 Ir. 10 6 /3 = 8*20 + 0-00142* 



90 Pt. 10 Ir. 10 6 /? = 8-84 + 0*00131* 



Pt. 10 6 £ = 8-87 + 0-00132*. 



B. Gas Thermometer Data and Fixed Points.— hi Table 

 YII are given the observed gas thermometer data.J In the first 

 column is the date of the measurement. The measurements 

 are numbered chronologically in the second column for con- 

 venience of reference. In the third column is the measured 

 pressure, p' (or jp/) in millimeters of mercury at 0°, corrected 

 as described on pages 107 and 108. The application of the 

 correction for unheated space (see p. 108) gives the pressure 

 p (or j? ) which is found in the fourth column. In the fifth 

 column is the value of the temperature, t, calculated by formula 

 (5) on page 101. In column 6 are given the readings of the 

 standard thermoelements in microvolts, and in column 7 the 

 positions of these elements on the bulb ; for the significance of 

 these figures see fig. 1 and note on page 104. In the last 

 column are given the other elements which were used on the 

 bulb, together with their positions designated in the same way. 

 The italicized letters represent single platinum wires instead of 

 thermoelements. 



A few measurements in which the value of p changed by 

 more than 0*1 per cent have been omitted ; their position is 

 shown by the absence of their corresponding serial numbers. 



* Published in paper of Day and Clement, loc. cit, pp. 425-441. 



f This Journal (4), xi, 374, 1901. 



% For the measurements in the table, seven furnaces were employed, 

 using three supplies of platinum wire of about 400 grams each. One of these 

 furnaces was wound on the outside, the other six on the inside of the tube. 

 It was possible to rewind the wire at least once after the furnace had 

 burned out. Failure always occurred several cms. away from the bulb in the 

 end portions of the furnace, which, in order to secure uniformity of temper- 

 ature over the bulb, had to be considerably superheated. Only one measure- 

 ment was made at the palladium point, as this one rendered the furnace 

 unfit for further use ; the conditions of this measurement, were, however, 

 perfect. 



