788 



MR J. D. HAMILTON DICKSON ON 



the constants so small as '01, '005, and '006 per cent, are great enough to produce errors 

 of as much as 1 per cent, in the results aimed at ; or, in other words, to produce errors 

 from 100 to 200 times their own relative magnitudes instead of being merely propor- 

 tional to them. 



In conclusion, the thermo-electric diagram, investigated now from the experimental 

 researches of Professors Dewar and Fleming, is to be taken as dealing only with the 

 range of temperature from — 200° C. to + 100° C. The curves calculated have been shown 

 extending from —400° C. to + 300° C, but not for extrapolation ; the purpose is to show 

 how necessary it is to determine the real nature of a relation between two correlated 

 quantities, and to guard against being misled by what appears to be the case instead of 

 the reality — as, for example, in the case of copper. 



I have appended, for the metals examined, a Table of the values of the observed 

 E.M.F.s referred to centigrade temperature. 



Note. — All the diagrams, except those for Al and Sb, were originally drawn on millimetre paper. 

 This can be made out from the figs, by recalling that a degree of temperature was always represented by 

 1 mm. Hence in the corresponding figs, the small squares are the photographic reductions of square 

 centimetres in the original diagrams. Those for Al and Sb were drawn on inch paper, hence the squares 

 in figs. 18 and 19 represent inch-squares in the original diagrams. 



INDEX OF METALS. 



Pt (Callendar), 1898 







pa 



ge 743 



fig. 3 



1908 









, 746 



,, 4 



Germai 



i silver . 









, 750 



„ 6 



Cu 











, 752 



„ 7 



Zn 



. 









, 756 



„ 9 



Cd 











, 758 



„ 10 



Ni 











, 764 



„ 12 



Mg 











, 765 



„ 13 



Au . 











, 766 



„ 14 



Pt(l). 











, 769 



„ 15 



Pd 











, 772 



,, 17 



Al 











, 774 



„ 18 



Sb 











, 776 



„ 19 



Ag 











, 780 



„ 20 



Thermo-electric diagram 







, 782 



„ 21 



Tait's observations 







, 783 



„ 22 



Holbon 

 Barus' < 



i and Wien's ob 

 )bservations 



servatio 



ns ) 





. 785 



„ 23 



