To find the cause of the acceleration in this case was the next 

 problem. Two possible explanations were left. The phenomenon 

 might be due to a gradual physical cliange in the silver oxide, such 

 as to bring it into a more and more reactive condition, or it might 

 be due to the presence of the silver produced in the reaction. The 

 latter supposition could fortunately l)e tested directly by exjjer- 

 iment, for if the silver formed by the reaction were a catalyzer, 

 then uncjuestionably silver prejjared in the same way and mixed 

 with the oxide before the experiment would produce a similar 

 effect. Therefore, the preceding experiment was repeated, using 

 instead of the pure silver oxide a mixture of 5.5 grams of oxide 

 and 0.1 gram of the powdered metallic silver obtained from the 

 last experiment. These were rubbed together until very intimately 

 mixed. 



Noticeable decomposition began as soon as the mixture was 

 put into the thermostat. The rate of evolution of oxygen after 



2 hours was 1 cubic centimeter in 5 minutes; after 5 hours, in 



3 minutes; after 7 hours, in 1| minutes. After about 8 hours the 

 maximum of 1 cubic centimeter per minute was reached, and in 

 12 hours the reaction was over. In the previous experiment the 

 reaction at the end of 12 hours was just beginning to have a 

 measurable velocity. This difference between the two experiments 

 seemed to point conclusively to catalytic action l)y the silver, yet 

 it was not inconceivable that it might be due to some difference 

 in the treatment of the two samples of silver oxide. In order to 

 remove this doubt the following ex])eriment was tried later under 

 new conditions with a different preparation of silver oxide and 

 at a different temperature: 



Three portions of silver oxide were taken, each weighing o.o 

 grams. To the first (A) 1.1 grams of silver powder was added, 

 to the second ( 15 ) 0.55 of a gram of silver, and the third ( C ) was 

 left i)ure. All three were subjected to jtrecisely the same treat- 

 ment of rul)bing and stirring, and introduced in separate tubes 

 simultaneously into the thermostat. At the end of three-quarters 

 of an hour 1 cubic centimeter of gas was evolved from A in 72 

 seconds, from B in 85 seconds, and from C in 215 seconds; at the 

 end of an hour from A in 42 seconds, from U in 56 seconds, and 

 from C in 210 seconds. The velocity reached its inaxinunn for A 

 in 88 minutes, for B in 110 miinitcs, and for C in lOO niimitcs. 



