8 



PROCEEDINGS OF THE AMERICAN ACADEMY 



RESULTS. 



Hence seven hundred-thousandths of a milligramme is ahout 

 the limit of visibility under these conditions. It is remarkable 

 that this result should be essentially identical with Bunsen's re- 

 sult obtained by a different method.* 



It has long been known f that the best method for detecting 

 small quantities of strontium and calcium in the presence of barium 

 is to evaporate the solution of the cblorides to small bulk, precipi- 

 tate most of the baric chloride by means of alcohol, and test the 

 filtrate. From this filtrate, by means of two or three repetitions of 

 the fractionation with alcohol, it is possible to eliminate nearly all 

 the barium. It is evident, on the other hand, that if the precipitate 

 is dissolved and reprecipitated several times, all the calcium and 

 strontium must go into the mother liquors. This is one of the 

 most rapid methods of obtaining pure baric chloride; it served for 

 the preparation of the material used in the succeeding experiments. 



In the first experiment half a milligramme of calcium was added 

 to the solution of three grammes of very pure baric chloride. 

 Upon the usual fractional treatment a most brilliant calcium spec- 

 trum was obtained from the mother liquor. One fifth of a milli- 

 gramme of calcium in another experiment gave similar results. For 

 a third experiment one fiftieth of a milligramme each of calcium 



* Compare Vogel's " Spectralanalyse irdischer Stoffe," 1877, pp. 92, 94. 

 t Ibid., p. 99. 



