An Investigation of a Case of gradual Chemical Change. 415 



We have obtained a series in which the amount of potassium chlo- 

 rate employed was as high as six units. Here, as with the case of 

 the corresponding quantity of chloric acid, no approach to an inter- 

 pretation could be attained. 



All these things show that the effect of varying hydrochloric acid 

 with chloric acid or with potassium chlorate is the same, though of 

 what exact nature that effect is, we have not yet fully determined. 

 No doubt the rate obtained for the decomposition of chloric acid 

 alone is too slow to account satisfactorily for the numbers not follow- 

 ing a law similar to that for variation in chloric acid first established. 



Series of experiments were made in which the amount of potassium 

 chlorate used was alone varied, the hydrochloric acid being constant 

 as regards the amount added each time. It will be seen, however, 

 from the potassium chloride results that we were not varying the 

 potassium chlorate only in this case, but really were varying both 

 this salt and the acid. For after saline decomposition 



HC10 3 per cent. = u x 51'5. 



HCl " = <yx 65-11 -wx 51-5. 



The variation of rates in these series, therefore, must follow a very 

 complicated law. We have, however, drawn a series of curves, re- 

 presenting the variation of rate in this part of the investigation (p. 416). 



The curves are thus drawn : 



A series of equidistant base lines (marked by broken lines) are 

 taken, one base line corresponding to each quantity of hydrogen 

 chloride used, and therefore marked at the extremity with a number 

 representing that quantity. 



Along these base lines are marked off lengths corresponding to the 

 quantities of potassium chlorate taken, and then lengths representing 

 the rates are measured perpendicular to these. The distance between 

 two blue lines represents 



000018 (6HC1 + 6KI + KC10 3 = 7KC1 + 3H 2 + 3I 3 ) in millionth - 



grams. 



The lower curves cross the base lines above them, but this does not 

 interfere with their comparison with the others. 



It is worth observing, however, that in nearly all the cases tried, 

 Bt/tt decreased instead of increasing with u, after u = 4 or after u = 5 ; 

 owing of course to the fact that the amount of hydrochloric acid is 

 decreasing as u increases. In fact it may be inferred that dli/du has 

 a root between 4 and 6. 



We now turn our attention to the behaviour of another constituent 

 of our mixture, viz., potassium iodide. We spoke of the part it 

 played in our experiments in the introduction to our paper, but now 



2 E 2 



