Marignac on the Equivalents of Barium, Strontium, and Lead. 21 7 



was unalterable in the air. For the mode of purifying this and 

 the other salts used in the experiment, as well as for the precau- 

 tions and corrections necessary in the experiment, the original 

 memoir must be consulted*. 



The method employed consisted in precipitating a known 

 weight of chloride of barium by nitrate of silver, and then pre- 

 cipitating the barium as sulphate, and from the relation between 

 the chloride of silver and the sulphate of barium thus formed, 

 deducing the equivalent of barium. The result of these experi- 

 ments gave the numbers 6861, 68 59, and 68 55, or as the mean 

 68-58, which is almost the same as that formerly obtained, 68-57. 

 And Marignac considers that the sources of error which cannot 

 be eliminated are sufficiently large to justify the adoption of the 

 number 68-5. The method used for the determination of the 

 atomic weight of strontium was essentially the same. The 

 results of experiments made with perfectly pure salts gave for 

 the equivalent of strontium the numbers 43-77, 43-74, and 43-76. 

 The mean of these, 43-76, is exactly the mean of the numbers 

 43-67 and 43-85 found by Berzelius and by Pelouze. The num- 

 bers hitherto adopted have been 43-5 and 44; but Marignac 

 thinks it impossible that the errors of experiment would justify 

 the adoption of either of these numbers. The equivalent of 

 strontium is almost certainly 43-75 ; and if we wish to assume 

 the existence of simple relations between equivalents, we must 

 lower still further the unit which serves as their common mea- 

 sure ; and having halved it for chlorine, and perhaps for barium, 

 copper, and lead, we must again halve it for strontium. But 

 if this common relation be no longer the actual equivalent of 

 hydrogen, but a fraction of that equivalent, there is no reason 

 why this fraction should be f rather than \, or than any other 

 smaller fraction, yi^ for instance. That is to say, that this 

 question of the existence of a unit, of which the weights of all 

 chemical elements would be simple multiples, will always be in- 

 capable of solution by experiment. 



The equivalent of lead was determined by Berzelius by the 

 reduction of oxide of lead by hydi-ogen. He found the number 

 103-56. jMarignac applied the method previously used for the 

 determination of barium and strontium to this metal. He found 

 the numbers 103-57, 103-49, 103-55, and 103-46; the mean 

 of which, 103-52, would confirm Berzelius's numbers. We may 

 accept 103-5 without exceeding the limits of errors of observa- 

 tion. 



V. Hauert has been engaged in determining the equivalents 

 * Bihl. Univ. de Gentve, March 1858, j). 209. Chemical Gazette, 

 April \f,,M:iy 1, 1858. 



t Journal fur Prukl. Chemie, February 1858. 



