LABORATOEY NOTES. 299 



These di.scre});iricies nve not .snrpri.siiio-. We know how varia1)Ie 

 are tlie thermoelectric properties of .stähle alloy.s^^ intended t(_) ha^•e tlie 

 .same composition, and h(3w a very sh<^-ht eliange in com[)0.sition may 

 he accompanied hy a very large change in thermoelectric position. The 

 present experiments mnst therefore he judged of •dt(_)gether on their 

 own merits. A simple comparison shows us that Professor Tuit's c()l)alt 

 will fit in to the regi(3n hetween lead and l^i.smatli ^•ery much as 

 Matthiessen's cohaJt fits in io his own .series. Thus tlie cohalt inves- 

 tigated liere seems to ditter from tliese other specimens in much tlie 

 aame wa3\ The new col)alt indeed hes so high in tlie diagnim that 

 its line is higher than the line of Ta it'. s nickel, for which A = — 21.S. 



This unexpected result was at once tested. A rougli experiment 

 was made witli the couple nickel-cohalt and a neutral point was oh- 

 tained a.t a temperature below 100° C. This cohalt line therefore, at 

 ordinary temperatures of the air, is ahove nickel ; l)ut because of its 

 greater downward inclination 'j:et:^ below it at temperatures above 



100° c. 



As regard.-; the inclination <jf the coljalt line, the present result 

 agrees as well with the earlier result as ci^uld ]'easonal)ly Ije expected 

 with two quite dillerent specimen;' of the metal. I'hus, expressed in 

 tlie same units, the thermoelectric power ;/f Professor Tait's coljalt is 

 given by the formula. 



p=—26:S — i).\ Mi / 

 Avhereas for the present specimen 



2j=-13.f8-U.lo86 t 

 \\'ith tlie exception of the .sharp upward Ijend in ni'^kcL ihi.- gives the 

 greatest inclination yet ohtained for a thermoelectric line. It would 



1) See tlie pnpei- by MacGre^'or aud myself already referred to, also my paper ou T/«' 

 Electrical Proi)crtu'^ of Uydrogcnisicd Falladinia (Traus. \l. S. E., Vol. XXXIII., 18SGJ— abstract 

 in this Journal, Vol. I. 



