88 ATOMIC WEIGHT DETERMINATIONS. 



foimcl as the mean of many trials, that 50.77 calcic carbo- 

 nate are equivalent to 56.1 calcic chloride. Wollaston, 

 taking the equivalent of calcic carbonate at 630, and that 

 of calcium at 255, calculates the equivalent of chlorine at 

 441 for O = 100. Wollaston cites Berzelius as having ob- 

 tained the same number by the conversion of plumbic car- 

 bonate into chloride. {Fhil Trans., 97, 1807, 301; lOJ^, 

 1814, 20.) 



J. J. Berzelius : 35413 (O = 16) ; 221.327 (0 = 

 100). 



The molecular weight of potassium chloride was ascer- 

 tained from four experiments on the decomposition of 

 potassium chlorate, which on being heated lost 39.15 per 

 cent, oxygen. This gives for the chloride 932.567, (0 = 

 100.) 100 parts of potassium chloride were further found 

 equivalent to 192.4 parts argentic chloride, and 100 parts of 

 silver to 132.75 argentic chloride. The value follows. 

 Berzelius in his Lehrbuch accepts Marignac's determination 

 and ascribes the error of the value he had obtained to the 

 imperfect decomposition of that portion of the chlorate 

 which was carried off as dust during the experiment. {Pog- 

 gend. AnnaL, 8, 1826, 17; also Lehrbuch der Chemie,3, 1189, 

 1191.) 



E. Turner : 35.42 (0 = 16). 



Turner made two experiments on the decomposition of 

 plumbic chloride with argentic nitrate. Assuming the 

 atomic weight of lead at 103.6, and that 100 silver = 132.8 

 chloride, these analyses gave CI = 35.43 and 35.48. Turner 

 also decomposed corrosive sublimate with calcic oxide neu- 

 tralized with nitric acid and precipitated with argentic 

 nitrate. If mercury = 201, these analyses give a maximum 

 of 35.28, and a minimum of 35.21, of which Turner selects 

 the largest. From calomel treated in the same way, he 

 arrived at the value 85.35. From his experiments on the 

 composition of argentic chloride (and apparently compari- 

 son with potassic chloride and chlorate) Turner got 

 35.45. The mean of the other experiments was 35.35, but 

 Turner considers 35.42 as being the most likely value. The 

 plumbum chloride was prepared from the carbonate, and 

 was purified by recr^^stallization, as was also the corrosive 

 sublimate. The calomel was " prepared by Mr. Howard," 

 and retained traces of moisture at 300°, which would make 

 the atomic weight derived from its analysis too small. The 

 values are for vacuum. {Phil. Trans., 123, 1833, 529.) 



