Perclilorates associated with Chlorides, Chlorates, etc. 39 



composition of hydriodic acid at the temperature necessary to 

 effect the decomposition of the perchlorate. 



We have, however, succeeded in developing a simple and 

 delicate method of detecting perchlorates, and one which may 

 be applied without great sacrifice in delicacy to mixtures of 

 the perchlorates with chlorides, chlorates and nitrates. It is 

 evident that for a rapid qualitative test conditions should be 

 so chosen that the effect of atmospheric air shall not interfere 

 with the certainty of the indication. Of the various salts 

 which we have employed we choose by preference fused zinc 

 chloride, chiefly because, while sufficiently energetic in its 

 action upon the perchlorate, it does not, like manganese 

 chloride or the double chloride of aluminum and sodium, 

 evolve chlorine under the influence of ordinary air at the high 

 temperature of the reaction. 



In the experiments recorded in Table I varying portions of 

 a solution of potassium perchlorate were evaporated to, dry- 

 ness in a test tube and fused with anhydrous zinc chloride. A 

 trap made by cutting off an ordinary two-bulbed straight dry- 

 ing tube was hung with the larger end downward in the test 

 tube, after moistening the interior of the bulbs with a solution 

 of potassium iodide. The chlorine evolved during the heat- 

 ing was registered by the iodine set free from the iodide and 

 subsequently washed with a little water from the trap and 

 tested with starch emulsion. 



Table I. 



Indication by 

 KCIO4 taken. the starch, test. 



0-00100 grm. Strong. 



0-00050 " " 



0-00020 " • " 



0-00010 " " 



0-00010 " " 



0-00005 " Distinct, 



0-00005 " " 



0-00003 " Trace. 



0-00003 " . None. 



0-00001 " " 



0-00000 " " 



The test for 0*00005 grm. of potassium perchlorate is sure 

 and distinct ; and it is, of course, evident that the presence of 

 a chloride in the original test can in no way interfere with the 

 certainty of the indication. All substances, however, which 

 yield chlorine by decomposition or by the action of the air 

 must be removed or destroyed before the application of the 

 test. "We find by experiment that 0*1 grm. of potassium 

 chlorate is completely broken up by treatment with 5 cm 3 of 



