WORK ON ACID POTASSIUM 



503 



obtained after four crystallizations from hot water. In normal 

 times pure sublimed phthalic anhydride and pure potassium 

 carbonate are cheap, and this standard is, therefore, one of the 

 most cheaply and easily obtained. 



We next directed our attention to the hygroscopic character of 

 acid potassium phthalate. It is very desirable that a substance 

 used as a standard should have the minimum of hygroscopicity. 

 To test this character we carried out two sets of experiments. The 

 first consisted in heating the air dried salt in an electric oven 

 at different temperatures and determining the losses; the secDnd 

 in exposing the salt thus dried in air of known moisture content 

 and determining the amount of moisture taken up from the air. 



In carrying out the first series of experiments about twenty-one 

 grams of the salt was heated in the electric oven in a platinum 

 dish. At the end of each period the loss was determined and a 

 portion of the salt was taken out without loss and titrated. The 

 losses and titrations would thus form checks on each other. There 

 were eleven such time periods with their corresponding losses 

 and titrations. Since only four different temperatures were 

 used I have reduced the results to this numlber. 



Weight of Salt Time 



Temp. 



Loss 



Value of HCl 



This is a very remarkable showing. We all know the difificulty 

 of securing accurate weights of substances owing to their hygro 

 scopic character or that of the containing vessels. Here an air 

 dried substance weighing twenty-one grams and in a platinum 

 dish lost no more on heating at 110° C. than could be accounted 

 for on the basis of change in temperature 'a the balance case 

 and the hygToscopic character of the dish itself. As stated there 

 were eleven determinations. The total loss of weight was about 

 1 part in 3.000, which is within the limits of accuracy of volu- 

 metric analysis in any case. The loss was comparatively con- 

 stant and apparently independent of the temperature and time 

 of heating. It was probably due in part to a constant error in 

 the method. This consisted in letting the dish rnd contents stand 

 in the air of the room while taking the sample for titration and 

 weighing it. The dish was then weighed for the next heating. 



