184 Holborn and Day — Gas Thermometer at 



The measurements made at that time were upon wires of 

 pure rhodium, iridium, palladium, gold and silver together 

 with alloys of platinum with palladium and ruthenium, whose 

 thermo-electric force referred to pure platinum was measured 

 up to 1300°. The elements with palladium or palladium 

 alloys differed from the others in that the range within which 

 the parabolic formula applies is smaller. 



The formula for the normal element T 2 derived from the 

 comparisons with the gas thermometer follows, together with 

 certain of the others ; those metals being chosen which are 

 best suited to the measurement of high temperatures and which 

 offer a basis for exterpolation. 



These have been compared with the normal element from 

 50° to 50° approximately and the thermo-electric forces cor- 

 responding to the exact temperature intervals as contained in 

 Table XVI, obtained from the observed data by graphical 

 interpolation. 



Since the publication of the earlier paper some observations 

 above 1300° have been added with the aid of an oven equipped 

 with a coil of platin-iridium wire. 



Normal Element Pt-90 1% 10 Eh.— The thermo-electric 

 force e 2 in microvolts of the platinum — platin-rhodium stand- 

 ard element T 2 is given by the following equation from 250° 

 upwards 



e 2 = — 310 + 8-048^ + 0-00172^ 



when the hot junction is maintained at the temperature t° and 

 the cold junction at 0°. The formula is based upon the obser- 

 vations with the platin-iridium bulb above 500° and with the 

 glass bulb and nitre bath below, as contained in Tables XIII, 

 XIV and XV. In the columns " Obs. — Calcul." of these tables 

 will be found the differences between the observed values and 

 those obtained from the formula. 



It will be recalled that the element T t was compared (first 

 paper, Tables I to VII) above 500° with the gas thermometer 

 equipped with porcelain bulbs. These observations differ 

 among themselves somewhat, due in part to the behavior of 

 the porcelain bulbs and in part to the less perfect conditions of 

 measurement then obtaining. In the normal curve these 

 observations are therefore not included, the curve being based 

 entirely upon the temperatures measured with T 2 and the 

 platin-iridium and glass bulbs. It is nevertheless not without 

 interest to compare the temperatures obtained with T x and the 

 porcelain bulbs, with the later ones. 



This may be done by referring the normal values for T t 

 (Table II former paper) to those of T 2 , by adding the differ- 

 ences T 2 — T 1 (Table XI, former paper) which were obtained by 



