646 MISCELLANEOUH PAPERS. 



which the determination of the moments of the pendulum required. On 

 the other hand it was desirable to make surer the empirical elimination 

 or ascertaining of these constants by increasing and modifying' the ob- 

 servations. I have therefore made a series of observations, using on 

 the pendulum balls of lead the same size but heavier than the brass 

 balls, which, combined with the series made without any balls, gave an- 

 other result independent of the result obtained when the brass balls were 

 used. 



In all other respects the method of reduction was the same as dui ing 

 the first determination. 



The value found for the mean density is 5.579i 0.012. 



