72 PROCEEDINGS OF THE AMERICAN ACADEMY. 



amount of nitric acid, besides the measured amount of standard 

 hydrochloric acid. Two experiments, after filtration, ignition, and 

 titration as before, indicated that 1.01 and 1.02 milligrams of silver 

 were precipitated by one cubic centimeter of hydrochloric acid, instead 

 of just one milligram. Again, with all the other conditions just the 

 same, 0.50 c.c. of hydrochloric acid precipitated in two cases 0.62 and 

 0.59 milligram of silver. In order to determine whether the pres- 

 ence of large amounts of baric nitrate might make any essential differ- 

 ence, these experiments were repeated with the addition of five grams 

 of the purest preparation of this salt,* quite free from chlorine. In 

 this way 0.50 c.c. of hydrochloric acid yielded 0.57 milligram of silver 

 and 0.20 c.c. precipitated 0.27 milligram of silver. The last six 

 experiments showed that the presence of a considerable excess of silver 

 in solution causes the precipitation of a slight excess of silver with the 

 chloride. 



Two experiments with neutral solutions treated in the same way 

 showed much more marked gains ; one exposed to the air during 

 evaporation for eight hours gained 0.13 milligram of silver ; and the 

 other, evaporated for twenty-four hours, the water being replaced from 

 time to time, gained 0.32 milligram. These gains might have been 

 due to the action of organic matter in the absence of nitric acid, or the 

 action of alkali dissolved from the glass ; in any case, the necessity of 

 the presence of free acid was shown. 



In the work which follows it was several times necessary to evapo- 

 rate strongly acid solutions of argentic chloride containing only a very 

 slight excess of silver, hence parallel determinations under these con- 

 ditions were very essential. In this way three successive experiments 

 yielded respectively 0.38, 0.38, and 0.37 milligram of silver, instead 

 of 0.50 milligram, which should have been obtained ; one experiment 

 gave 0.52 instead of 0.60, and two more gave respectively 0.90 and 

 0.93, instead of 1.00. Hence, when much nitric acid and very little 

 silver are present, the amount of metal found is too small instead of 

 too great. 



These determinations all together show that the evaporation of solu- 

 tions of argentic chloride upon the water bath is a process carrying 

 with it small errors, some plus and some minus, which are reasonably 

 constant under constant conditions. They show also that the presence 

 of baric nitrate does not in the least alter the accuracy of the deter- 



* These Proceedings, XXVIII. 16, 19. 



