Forces in Voltaic and Thermoelectric Piles. 271 



Everything being so simple and straightforward from this 

 point of view, we have next to ask, What fault can be found 

 with it, and why is it not readily admitted ? 



The crux is the proposition 2'. It is granted that E.M.F. 

 is work per unit electricity for the whole circuit, i. e. proposi- 

 tion 2 ; but it is not admitted that the same thing is true for 

 every point of a circuit. 



Well, then, supposing this legitimately questioned, there 

 must be something to put in place of it. What is this some- 

 thing ? Dr. Hopkinson answers the question completely and 

 satisfactorily. I want to know if Professors Ayrton and 

 Perry accept his answer. 



The " something " is this. Electricity is supposed to have 

 a specific heat like any ordinary fluid, and this specific heat is 

 supposed to vary with temperature and with the substance in 

 which it happens to be flowing, to the extent even of being posi- 

 tive in some metals, negative in others ; and the variations in 

 this hypothetical specific heat, since nothing is otherwise known 

 about it, can be relied on to perform whatever office is required. 

 Electricity can let heat drop, or pick it up again, at spots 

 where its specific heat is supposed to change, and this without 

 introducing any rigid work- and -energy considerations, nothing 

 but a free and airy " specific-heat " hypothesis. 



Dr. Hopkinson denies that there is anything " confusing ,J 

 in this mode of regarding the matter ; and Ayrton and Perry 

 very naturally object to their view being considered " absurd/'' 

 I should be sorry to bandy adjectives of any but a compli- 

 mentary nature ; and considering the high authority of many 

 who have found themselves for various reasons able to main- 

 tain much the same position as Ayrton and Perry, it would 

 be wholly unseemly to do so. But I must confess that were 

 I unacquainted with the fact that such views are held by 



at 



whence it follows that 



™=§ <M 



And all the relations are established. 



They may be exhibited in one line, along with the approximate formula 

 of experiment, thus : — 



dt 



= j = $fM~f t {t) = &a-hXto-t)> • • • • (V) 



