11 



to vinitv, hence the unit of time in each case is the time which 

 would l)e required for complete decomposition if the whole reaction 

 progressed at a uniform rate ecjual to the maximum velocity. 

 In other words, the units of time chosen are inversely i)ro})orti()nal 

 to the maximum velocities, which, therefore, in the new scale 

 become numerically equal and fall on a Single point. 



The four curves thus ])lotted in fig. 3 come astonishingly near to 

 coincidence; in fact, they do coincide within the limits of exper- 

 imental error. 



Among all the cases in which reaction velocity has l)een studied 

 the present phenomenon occupies a wholly unique position. 

 Exceedingly few autocatalytic reactions have been quantitatively 

 investigated. In a homogeneous system the study of the reaction 

 of oxalic acid and permanganate by Schilow* may be mentioned; 

 in heterogeneous systems the study of the action of acids on sodium 

 thiosulphate l)y Foussereau'' and the decomposition of hydrogen 

 arsenide, hydrogen selenide, and hydrogen antimonide, respec- 

 tively, by Cohen,^ Bodenstein,* and Stock and Guttmann.^ In all 

 these cases in heterogeneous systems a reaction in a licpiid or gas- 

 eous phase is catalyzed by a solid product. Bodenstein,* recalcu- 

 lating the work of Stock and Guttmann, points out the surprising 

 simplicity and regularity of the decomposition studied by them. 

 But how much more surprising it is that in our present case the 

 decomposition of one solid is catalyzed even more regularly by the 

 the presence of another solid w'ith which, in the very nature of 

 solid substances, it can not l)e in very intimate contact. The 

 decomi)Osition behaves, indeed, almost like a sim]ilo homogeneous 

 reaction. 



In a homogeneous catalytic reaction the simplest case is that in 

 which, for a given volume, the reaction velocity is proportional to 

 the quantity of the reacting substance and to the (juantity of the 

 catalyzer. Let us see how nearly tiie decomposition of silvin- 

 oxide conforms to these conditions. 



^Ber. C'hnn. Gex., 36, 2735 (1903). 

 ^\nn. Chim. Phy.-., (6) 15, 533 (1888). 

 'Z/'iV. P///y.x. Chnn., 25, 483 (1898). 

 *yyil. I'In/s. Chnn., 29, 429 (1899). 

 'B>r. Chnn. CV."., 37, 901 (1904). 

 ^Zclt. I'hyK. Chnn., 49,41 (1904). 



