Relations of the Alloys of Platinum. 441 



11. Taking the results collectively, they point to a limit be- 

 low which in the case of solid metals and at ordinary tempera- 

 tures, neither electrical conductivity nor temperature-coefficient 

 can be reduced ; whence it appears that a lower limit of both 

 conductivity and temperature-coefficient is among the condi- 

 tions of metallic conduction, not to say of metallic state.* 

 These considerations are suggestive and I shall therefore en- 

 deavor to make what I have in mind clearer. In the case of 

 conduction in metals (solid or liquid) the effect of temperature 

 is a decided decrease of conductivity, continuing apparently, 

 as temperature increases, indefinitely. f In the case of con- 

 duction in non-metallic elements J or in electrolytes (solid or 

 liquid) on the other hand, the effect of temperature is a decided 

 increase of conductivity, which supposing the liquid state to 

 be retained, continues as temperature increases. Hence con- 

 duction in metals is distinguished from conduction in electro- 

 lytes in this respect, that if the temperature coefficient in the 

 one case (electrolytes) be regarded positive, its value in the 

 other case (metals) must be negative. This leads me to in- 

 quire into the possible occurrence or the nature of a class of 

 substances whose temperature-coefficient is zero ; a class of 

 substances in other words in which the metallic and the 

 electrolytic modes of electric conduction may be supposed to 

 converge. § 



The point which I have in view, viz : the possibility of a 

 continuous transition from metallic to electrolytic conductivity 

 gains much in reasonableness by associating with good metallic 

 conductivity the correlative property of optic opacity. Rela- 

 tions between electricity and light have been investigated and 

 many experimental facts are known. Maxwell's electro-mag- 



* Recent researches of v. Ettingshausen and Nernst and of C. L. Weber (Wied. 

 Arm., xxxiv, p. 582, 1888), show that the resistance-temperature coefficient of 

 bismuth is often negative between 0° aud 100°. Edward Westou has made alloys 

 of copper, ferro-manganese and nickel of which this temperature-coefficient is 

 nearly zero or even negative (Science, xii, p. 56, 1888). These exceptions, the 

 underlying cause of which is probably secondary and to be referred to structural 

 or crystalline modification, emphasize the vast amount of evidence in favor of the 

 normal behavior given in the text. I may add, for instance, that the temperature- 

 coefficient of glasshard steel between 0° and 100°, would be nearly zero because 

 of annealing. 



f Following Benoit (C. R., lxxvi. p. 342, 1873) the electrical resistance of all 

 metals increases with temperature at an accelerated rate, except in the case of 

 platinum and palladium, where the rate of increase is retarded. Benoit observes 

 at temperatures limited by the boiling point of zinc. 



\ Matthiessen (Pogg. Ann., ciii, p. 428, 1858), W. Siemens (Wied. Ann., x, p. 

 560, 1880) and others (Bergmann, Kemlein, Muraoka) find this to hold for 

 modifications of carbon. Similar increases of conductivity are usually observed 

 in the case of selenium and tellurium (Hittorf, W. Siemens, Mattheissen. and 

 many others) ; but the relations here are complicated. Quite recenily Duter 

 (C. R., March 19lh, 1888) has shown that sulphur conducts at its boiling point. 

 It is this investigation which I have specially in mind in the text. 



§ Something of the kind may perhaps occur in the case of some natural sul- 

 phides, but it is not open for systematic study and its nature is obscure. 



