262 



the metallic factors; others, on the contrary, deny in toto the 

 influence of friction on the thermo-electric phenomena. Thus, 

 it was recently adverted to in a scientific journal, as a highly 

 paradoxical fact, that in a given case friction had caused a 

 change of sign in the thermo-electric declination produced 

 by the contact of two heterogeneous metals ; but, at the same 

 time, this ' unheard-of fact, as it was called, was explained 

 by supposing, gratuitously, that friction had been effected 

 whilst keeping the metal to be rubbed in the naked hand, and 

 in thus producing an accidental change of temperature. This 

 explanation was offered on the assumption that friction, in 

 itself, is not capable of producing any effect. Between the 

 two extremes of tribothermo-electric omnipotence and nullity, I 

 have tried to discover the middle course of truth." 



Professor Erman then delivers his own opinions, and the 

 facts from which he has deduced them : "A bar of bismuth 

 was joined to that branch of the rheophore of this instrument 

 (the galvanometer), where the silver of a voltaic element (silver 

 and zinc) produces an eastern deviation, and a bar of anti- 

 mony to the other branch of the rheophore. Both of these 

 bars were provided with handles, so that they could be em- 

 ployed without undergoing any change of temperature in the 

 manipulation. When, through these being stationed in the 

 same room, the two bars had previously arrived at the tem- 

 perature of the surrounding space, no deviation whatsoever 

 was produced by their contact, but the slightest friction of 

 either of them against the other gave immediately an eastern 

 deviation. This latter extended even to an entire revolution 

 of the needle in the same direction, if the friction proceeded 

 rather more rapidly. By gently raising the temperature of 

 the two bars to 100° or 111° Fahr., their contact in a state of 

 repose always produced a stationary eastern deviation of about 

 30°, which, by rubbing, was further increased to 60°, and 

 there likewise remained invariable as long as friction continued. 

 At length, when I cooled the bars below the temperature of 



