MiscellaneoiLs Papers. 



287 



Each weight given is the means of weighings right and left — 

 not the square root of the product. But one determination of each 

 weight was made by each observer. The adjustment of the bal- 

 ance was for the lower sensitiveness of ^^ mg. It was planned to 

 make determinations for a sensitiveness of y^ also, but lack of 

 time prevented. All values have been carried out at least one 

 place further than shown in the above figures. 



4. Eatio of the Arms. — The ratio of the arms has been computed 

 from the^values from 20 to 170, with the results given : 



Mean value of l/r, 0.999997. 

 Further determinations are desired, especially for the higher 

 sensitiveness, no values of which are given, and but few of which 

 have been secured. 



The weighings were all made without any precautions to secure 

 dryness of the air in the balance case. 



