264 



Messrs V. Harcourt and Esson on the 



[Nov. 22, 



The small measures of hyposulphite consisted of single drops collected 

 under circumstances favourable to their perfect uniformitj^ The peroxide 

 employed was either an acidified solution of sodic peroxide, or dilute 

 hydric peroxide obtained by the distillation of such a solution. In each 

 set of experiments the value of the measures of hyposulphite and of the 

 pipetteful of peroxide could be compared by determining what fraction of a 

 measure of hyposulphite remained in the solution when the set of experi- 

 ments had been brought to an end. These values could also be compared 

 by determining each with a standard solution of potassic permanganate. 

 In presence of an excess of potassic iodide, sodic hyposulphite may be esti- 

 mated by this reagent exactly as it may by standard iodine solution. 



During each set of experiments, then, all the conditions of the' reaction 

 are constant except two, which are progressively modified. One of these is 

 immaterial, the accumulation of a small quantity of sodic tetrathionate ; the 

 other is material, the gradual disappearance of the small quantity of per- 

 oxide upon whose presence the reaction depends. 



The first point, therefore, requiring to be investigated was the law of con- 

 nexion between the amount of change and the amount of peroxide. 



Since the amount of peroxide originally taken is known, and also the 

 amount which corresponds to a measure of hyposulphite, the amount re- 

 maining in the solution at the moment of each observation is also known. 

 Thus the data supplied by each experiment are (1) the amount of peroxide 

 present in the solution at a particular mom.ent of time, (2) the time at 

 which this amount is present, (3) the amount present at a subsequent 

 moment of time, (4) the time at which this amount is present. Kepresent- 

 ing these two amounts of peroxide by y and y' respectively, and the cor- 

 responding moments of tim.e by t and t\ the result of each experiment is 

 that an amount of chemical change y—y^ has been accomplished in an in- 

 terval — t. According to the hypothesis proposed in our former paper, 

 namely that the amount of chemical change varies directly with that of 

 each of the substances partaking in it, these quantities should exhibit 

 throughout a set of experiments the constant relation 



y' 



The numerical results obtained in various sets of experiments performed 

 under different circumstances are compared with those calculated from 

 equations of this form, and the two are shown to agree within narrow 

 limits of experimental error. It is inferred that in this case the amount of 

 chemical change taking place at any moment is proportional to the 

 amount of peroxide present at that moment in the solution. 



The constant a in the preceding equation represents the effect upon the 

 amount of change of those conditions which do not vary in a set of experi- 

 ments; and it is possible by varying one of these conditions in different 

 sets of experiments, and determining the value of a in each, to inquire into 



