2 Mr. C. Barus on tlw Fusion 



2. Remarks on the present Paper. — The following high- 

 temperature work was done nearly five years ago. Having 

 the present occasion to corroborate the former results, I was 

 gratified to find that in the lapse of time my original constants 

 cannot have varied more than 4° or 5° at 1000°, while even 

 this index of the time-error of the method can be largely 

 explained away. It seems justifiable therefore to give the 

 chief features of my method of work * a larger publicity than 

 it will reach in a geological bulletin. 



Apart from this, since M. H. Le Chatelierf has thrown 

 discredit on the platinum-iridioplatinum thermocouple, it 

 becomes incumbent on me to endeavour to restore the prestige 

 of the older couple, lest the same slur should also come over 

 my work. Let me remark in passing, that though I have no 

 doubt of the trustworthiness of the platiDum-rhodium couple, 

 nobody has as yet taken the pains to put it to the severe test 

 of a direct air-thermometer comparison, through which I 

 have put the platinum-iridium couple. I have long hoped to 

 find the time to do this myself. 



The present paper, therefore, though its main object is a 

 study of the thermoelectric or thermal behaviour of the 

 platinum-iridium couple, enters somewhat minutely into high- 

 temperature air-thermometry in general. 



3. Literature. — A consideration of the earlier work, for 

 which there is no room here, will be found in the Bulletin 

 cited, pp. 23-55. 



Apparatus for Air-Thermometry. 



4. Preliminary Remark as to Methods. — Constant-volume 

 air-thermometry is scarcely suitable for high temperature 

 work, for it is not safe to bring pressure to bear on the white- 

 hot porcelain bulb. This objection applies in a measure to 

 Deville and Troost's Sprengel-pump method, which has an 

 additional incovenience in not being continuous, so that the 

 fiducial volume is lost. On the other hand, the constant- 

 pressure method is free from these objections, and my best 

 comparisons were obtained by means of it. Such an air- 

 thermometer, as I shall endeavour to show, is complete in 

 itself, admitting of a redetermination of the fiducial constants 

 (volume of bulb, coefficient of expansion, fiducial zero) with- 

 out taking the instrument apart (§§ 13, 14). 



* 



Of. Annual Report of the Director U. S. G. S., 1881-82, p. 9 ; ibid. 

 1883-84, p. 62 ; ibid. 1886-87, p. 192. The full account is to be found in 

 Bulletin U.S. Geological Survey, No. liv. pp. 56 to 238 (1889). 

 t C.R. cii. p. 819 (1886) ; Bull Soo. Chim. xlv. p. 482 (1886). 



