1267 



the mercur} a inucli grealer oppuiliiiütj lliuii in ordiiitirj ca«es to 

 foiiii fsiu'li a fombinauon. Thus llie iiierciiry Je( would Ijave got 

 diffei-eiil coinposilioii tor tlie tirsl niiiiute of the experiment and a 

 changed defleclion wonUI resull. As we did not tind any change in 

 the detlection we cannot ascribe the observed effect to ctieniical 

 contamination of the mercury. 



Our effect may however be explained, as to magnitude and direc- 

 tion, by the tiiermo-electric force between mercury under pressure 

 and mercury without pressure. This force was fii'st observed by 

 DES CouuKKs^), and afterwards measured also by Wagnkk and Hökig'-'). 

 The mercury in our glass apparatus had the pressure of tlie com- 

 pressed air, the mercury jets consisted of mercury at atmospheric 

 pressure and theii' points of contact with the mercury at high 

 pressure were at different temperatures. 



We proved by a separate experiment that this explanation is the 

 right one. We connected the tops of the tubes B by a short close- 

 fitting glass knee, which was thus filled up l)y the mercury from 

 the jets, a small aperture in the upper part serving as an overflow 

 for the mercury. One of the tubes B now being heated, we obtained 

 deflections of the same order as before, although there were no tree 

 mercury jets and therefore no sudden transition of temperature. 



There is, however, one detail which remains unexplained in this 

 way: we always found a greater effect for superficial contact of the 

 jets than for a full contact. The difference increased with the differ- 

 ence in temperature and amounted to 15 mm. in maximo. 



This phenomenon might be due to the increase of the temperature 

 gradient at a superficial contact and then would prove the reality of 

 Benedicks' so called fundamental effect. Of course one should not 

 forget that this thermo-electric force amounts only to 3.5 X 10~^ 

 Volt for the extremely steep ïaW of temperature of 300° over a 

 distance less than 0.1 mu). Therefore this small force may be wholly 

 neglected in all practical cases where it appears together with ordi- 

 nary thermo-electric currents. We devised yet another experiment 

 to demonstrate the minuteness — perhaps even the non-existence — 

 of this effect of temperature fall, and to get rid of the disturbing 

 pressure effect. A thin-walled glass tube was prepared, diameter 

 outside 1.00 mm. inside 0.80 mm, and drawn down at the middle 

 part to an outside diameter of 0.45 mm, inside diameter 0.30 mm. 

 This tube was attached by short rubber tubes to the tubes B, from 



1) Des Goudres. Wied. Ann. 43, p. 673, 1891. 

 -) E. Wagner. Ann. d. Physik. 27, p. 955, 1908. 

 H. HöRiG, Ann. d. Physik 28. p. 871 1909. 



