of the Heat-condacting Power of Mercury, 487 



accurately known coefficients of expansion of mercury for every 

 temperature, the temperatures m could be calculated from the 

 quantities which flowed out. 



Finally, a piece of identical tube shut off by itself, of about 

 20 centims. length, which was likewise arranged as an outflow- 

 thermometer and extended below the main tube in direct conti- 

 nuation with it, was to serve for the determination of the lower 

 temperature t' of the tube, which only in consequence of a cur- 

 rent rose above the surrounding temperature. In order to cause 

 this piece to be heated in like manner by the current, the latter 

 had, immediately after passing through it, to enter a third piece 

 of tube, which possessed an equal width and a like closure of 

 iron as the two others. This lowest tube stood on a long, 

 screwed ivory point ; and all three were held each by a wide ring 

 into which three long points of ivory were screwed as radii — the 

 uppermost, proper experiment-tube at its lower end. Thus no 

 diversion whatever of the current or heat could take place. 



I must yet go briefly into the manner of heating, which had 

 to be effected from above. The upper iron closure of the proper 

 experimental tube had been made of somewhat greater diameter 

 outside of the tube, and possessed an upward-directed rim. The 

 vessel hereby formed was filled with mercury; and into this 

 dipped a large conical-bottomed iron reservoir, which was 

 crammed full of brass refuse from the turning-lathe. A circle 

 of gas-jets was then brought round the middle of the reservoir. 

 With this arrangement, low as well as high temperatures T 

 could be maintained in the tube with extraordinary constancy. 

 In order to prevent downward radiation from the (especially in 

 higher temperatures) strongly heated iron reservoir, underneath 

 the burners a screen of sheet brass closely surrounded the reser- 

 voir, and below this were three of stout pasteboard, all separated 

 from one another by strata of air. Between the lower screens, 

 also, ran several coils of a leaden pipe, through which cold water 

 was incessantly flowing. Lastly, the small projecting part of 

 the iron closure of the tube was itself underlaid with loose 

 asbestos. The effect of all these precautions was entirely satis- 

 factory ; indeed the entire length of the experiment-tube was 

 found to have the same cool environment. That this surround- 

 ing temperature might not be disturbed by the observer himself, 

 the entire apparatus was surrounded, from the screens down- 

 wards, with glass windows. The experiments were performed 

 in a cellar of the laboratory, which was not exposed to rapid 

 vicissitudes of temperature. 



With this apparatus I obtained the results set down in 

 the following Tables — first for heat-conduction alone, without a 

 simultaneous passage of a current, so that /' and 6 were equal. 



