398 Mr. W. H. Pendlebmy and Miss M. Seward. 



potassium chlorate was more often employed, being easily obtained 

 and kept in a state of purity. 



Bunsen gives the following hypothetical equations for a reaction 

 between hydrogen chloride and any chlorate (the equations are given 

 with hydrogen chlorate, for simplicity). But in these the number of 

 molecules of chlorate reacting is arbitrarily limited to two. Without 

 this limitation, it is obvious that the list of possible reactions may be 

 indefinitely extended. 



Relative proportion. 



1:1 2HC10 3 + 2HC1 = C1 2 + C1 2 3 + H 0. 



1:2 2HC10 3 + 4HC1 = 3C1 2 + 3H 2 O. 



1:3 2HC10 3 + 6HC1 = 2C1 2 + 2C1 2 + 4H 2 0. 



1:4 2HC10 3 + 8HC1 = C1 2 + 4C1 2 + 5H 2 0. 



1:5 2HC1O 3 +10HC1 = 6C\ -f6H 2 0. 



In our experiments the quantity of reacting substances was always 

 such that, except for change in sodium thiosulphate, the composition 

 of the mixture was sensibly the same at the end of the experiment 

 as at the beginning. Each experiment was not carried to any definite 

 limit, but was concluded as soon as the constant velocity of change in 

 the mixture had been ascertained by the observation of several 

 intervals corresponding to successive additions of thiosulphate. 



The following considerations show the constancy of the composition 

 of the mixture throughout an experiment. Each drop of thio- 

 sulphate corresponded, on an average, to the decomposition of three- 

 millionths of a gram of potassium chlorate in each cubic centimetre 

 of the mixture. Now, the smallest amount of potassium chlorate 

 ever used was O01263 gram in each cubic centimetre, and of this 

 only 0*000003 gram would have disappeared when as many as 

 10 drops of sodium thiosulphate had been added. This is an 

 alteration of about 0'02 per cent. Or, to state it otherwise, in the 

 case of one of the greatest velocities observed, when each interval 

 was hardly greater than a minute, there was 0'03788 gram potassium 

 chlorate in each cubic centimetre, and this was disappearing at the 

 rate of 1'826-millionths of a gram per minute. Speaking roughly, 

 it would take about 24 hours, proceeding at this rate, to cause a 

 difference of 1 per cent, in the amount of salt present. 



Messrs. Harcourt and Esson represented the variation of the 

 intervals they observed with the mass present, ?/, as a logarithmic 

 curve with asymptote meeting it when y = co. The constant 

 intervals obtained in the present investigation would be represented 

 in a portion of the curve produced to a great distance in the direction 

 of the asymptote, this portion being sensibly a straight line parallel 

 to the asymptote, so that the time observed for each interval is 

 constant. 



