ON BALANCES. 67 



then taken, after adding a small balance weight to either pan, 

 in order to ascertain the value of a division. 



In cases where great accuracy is required, any number of 

 successive comparisons may be made in like manner with the 

 object of taking the mean result of them all. This course 

 not only lessens the probable error of the result, but is also 

 a check against any accidental mistake, either in noting the 

 readings or in the computations. It is important to make 

 the weighing in as short a time as possible, so as to avoid 

 the risk of discordances in the results arising from variations 

 of temperature, of moisture in the air, or other causes ; and 

 it is better to take only two or three comparisons at a time, 

 and to repeat them on subsequent days, taking the mean 

 result of all the comparisons. 



29. In weighing by Gauss's method, it is very desirable to be 

 able to interchange the weights from one side of the balance 

 to the other without opening the balance case, and thus to 

 avoid the risk of changes of temperature of the air of the 

 balance case, and consequent production of currents of air. For 

 this purpose the plan' is adopted of interchanging the pans as 

 well as the weights, and it not only gives the advantage of 

 avoiding the risk of injury to the weights by taking them 

 up in the ordinary way with a pair of tongs, but is especially 

 useful when either of the bodies weighed consists of several 

 separate weights. The pans are readily interchanged by 

 lifting each of them from its place at the end of the beam 

 by means of a brass rod with a curved end introduced through 

 a hole in the side of the balance case, and transferring it to a 

 hook suspended from a brass slider made to move over and 

 parallel to the beam. Thus each pan with its load is slid 

 over and transferred to the other side of the balance. 



30. You will better understand the mode of interchanging 

 the weights and pans used in Gauss's method of weighing by 

 a practical illustration. 



At the Standards Office we use printed forms for noting 

 and recording the weighings, one of which I have brought 

 you, filled up with the actual results of one of the compari- 

 sons of the platinum-iridium lb., designated as P-i., which I 

 have already exhibited to you, with the Imperial Standard lb. 

 of platinum, designated as PS. The following shows the 

 mean results of five observations : 



F 2 



