High Temperatures. 203 



Repeated heating to high temperatures may also under cer- 

 tain circumstances render the wires less homogeneous. 



Attention has been called to this before, but now that the 

 accuracy of thermo-electric measurements has been so materially 

 increased it becomes necessary to go into the discussion of this 

 matter in considerable detail. 



The changes are such as Holborn and Wien observed in 

 platinum and palladium due to the action of combustion gases 

 or hydrogen in the presence of silicium, and which affect the 

 resistance even more than the electromotive force. Even with 

 electric heating where combustion gases should no longer be 

 feared, the walls of an oven which has not been previously heated 

 to a higher temperature than that to be measured, and kept 

 there until burned through and through, develop gases which 

 pass through platinum. 



We have called attention to this in the discussion of aplatin- 

 iridium bulb for the gas thermometer in our earlier communi- 

 cation. The thermo-elements are also affected by exposure to 

 these gases but may usually be protected by the use of porce- 

 lain, which does not itself develop such gases and is very 

 impervious to them even in its unglazed form. For tempera- 

 tures below 1100° still further protection is afforded by glazing 

 of course. Tubes of more refractory material intended for 

 temperatures where porcelain can no longer be used and neces- 

 sarily imperfectly burned we have found to be porous and 

 when first heated to give off gases themselves at high tempera- 

 tures which in the presence of silicium affect platinum strongly. 



If the wires of a thermo-element which have been exposed 

 in this way be afterward used for temperature measurement, 

 there is often little or no change in the electromotive force 

 from the normal values so long as the temperature distribution 

 along the wire remains unchanged, but considerable variations 

 always appear as soon as it is varied. 



We made observations for example with such an element C 

 which had been exposed in an imperfectly burned oven, the 

 hot junction being combined with that of a second element A 

 which had not been so exposed. At 550° the following differ- 

 ences (C — A) in microvolts were observed in an electric oven 

 40 om long in which the common junction of the two elements 

 could be moved from the center toward either end as described 

 elsewhere : 



Hot Junction. I. II. 



At the center (elements symmetrical) V 7 



6 cm to the right j 2 wires heated 14 om ) — 73 -f- 3 



6 cm " left j " ", 26 cm f+126 -(-36 



In column I the elements were crossed while in II the two 

 wires of each element passed out of the oven at the same end. 



Am. Jour. Sci. — Fourth Series, Vol. X, No. 57. — September, 1900. 

 14 



