lOMXTltOMOTIVi; I'OIK i;s IN TIIK VOI.TMC CKI-L. 



r/iT 



amalj^am fit a Icinporiitun' just below llio multiiif? point of tin, and obtain 

 (ir |)ii.ssil)Ii') tliu iii'tt ovdliititin of liciit llii'ii. 



Siipposo tlio lieat of C()ml)iiiaiiou of llu) ll'l pfninimoH of tin witli 

 rncrciiry to bo somehow or oilier ch^lei'iiiiued, wo havo next to suppose tiio 

 amalgam miulo otlicrwise, briii^inu; tlio mo'eoules tof^etber iti ji, reiisoiied 

 wav. JiL't tbo sanu) (|imntity of tin bo brou^'bt to witliiii jnoleciihir 

 ilistanc'o of tho inerciiry in successivo pi(;ee.s of veiy thin foil, first mailo 

 to touch at one corner and then laid (h)wn. 



It is (p.utc triK! tliat eaeh ilakc; would bo charjjfed with a Volta E.Af.K. 

 of, say, M) volt, and so would attra(!t tho niercnry and do a certain amount 

 ol' work in layijif^ itself down. iJut it is not. fail' to compare! an operation 

 tlius conducted in air with tho drop]>in!^ of a sohM nlas.^ of tin into 

 mercury; to bo ablo to compare tho two operations one must perform the 

 feil experiment in absolute vacuum. This l)oinf^donc, tho contact J'l.M.K. 

 is no longer "(I volt, but only about •MdOl.') volt accordin;^- to tho cxpori- 

 ments of Matthicssen. Good data for this quantity arc however wantiiiij ; 

 mercury is not ono of tho metals included in J'i'oi'essor Tait's series, ft 

 was observed by Ganfi;ain; and l)y rather hypothetical deduction from lii.'i 

 nnmbers, as <^iven pictorially in Wiedemann's ' Kloktricitiit,' I make tho 

 tin-mercury Peltier force I'T-'i millivolts at 10°.' 



Taking ono of theso nundjers (loi'ilJU or 175,000 in C.G.S. units), or a 

 better ono when determined, we can calculate how near tho given nmss of 

 tin must bo brought to tho mercury in order to gcnei'ato tho actual heat 

 of combination, provided ono knows tbo specitie inductive capacity of 

 ab.solute vacuum.- But I do not know it. Thus tho sup^Jy of data for 

 this case is distinctly unsatisfactory. 



' Siiici' tliis was in ty))(;. a ]iu])vv Ijy C. L. Weber lias a|)p('ar('il in AN'ieiU uiiiin's 

 Ainmli'ii for Novcmbcu' 18SI, on the theriaoelectrio ])r(>p('rtk's oi anial^^'iiiua, in wiiicli 

 mercury itself was cxuiuinoil ; and from tlie data tliero recorded, to.uctlipi' with 

 Tait's vuluo for copper, 1 rcclvou the tlioniioeleutric value of mercury at t' Ciiiii- 

 ,uni(lc as 



■llil + -Tit absolute eloctro-uiagni'tie units. 

 ■WiiL'iice tlio Peltier force at the same temperature is 



1 ISl + ."JMiS^H- -tJO.-j/'- microvolts. 

 Till' i'eltier force between tin and niercurv at 10° is tlu^rcforo iL'S.SOO absolute uniN, 

 or t'l't millivolts, which agrees well I'uouiih with tho rouifh estimate above. 



- 'i'likiu;,'' this as I ami iissumiiiji; t hi' estimate of molecular dimensions hereaflei- 

 I'stablislii'd and workinn' backwards, one can show that tlii' Peltier force of tin and 

 iMiTcuryat 10^ is connected with the heat of combination of our L'l j,'-rammes of tin 

 with the ~>0'2 grammes of mercury by the relation, 



.rn = 3C> ;: 10" vn. 



The two rouf^-h estimates of ,1 n deduced from ^lattliiessen and (laui^'aiu rcsp(>etively 

 (l.",ii()0 and 17.5,000) tlius uive 11 as about | and -j'^: of a unit resjiectivelj'. JCither nf 

 these is too small a (juantity tobe observedinthe process of dissolving tin in mercury ; 

 so neglecting it wo get, i'rnm that exjieriment, the latent heat of molten tin at. 

 10^ C. as il()-4. Another experiment made in .n similar way gave li)'G, 



If the above reasoning bo regarded as legitimatis a (■ond)ination of tliermoeleelrii? 

 ine;isurcments with observed heats of solution in mercury n.aj' furnisli a means u[ 

 c-tiiiiating latent heats of fusion at various low tempenilures in general. 



Working back similarly to the heat of combination of 1 grimnne of copper with 

 I graumie of zinc, we calculate -077 unit as the heat dt^velopeil at ordinary tcmjiera- 

 tures; onlj enough to raise tho mass of brass formed through throe-eighths of a degree 

 Centigrade. At a higher teiupcrature such as -100° C. tho I'cltier force for tliose 

 raetals is greater, being 4,(J00 microvolts, and the calculated heat of combin;ition. 

 IS then i of a unit per gramme of each ; sutHcient to raise the whole mass of metal 

 tlao«gh nearly 2 degrees Centigrade. This, then, is the sort of elevation of tempeni- 

 ture one nay expect in making brass at a temperature of 400°. 



I 



