Miscellaneous Papers. 



285 



The testing of the balance and weights has thus far consisted 

 of the following determinations: 



1. The testing of the weights by the Bureau of Standards. 



2. The sensibility curve. 



3. The values of certain weights from laboratory sets, both for 

 purposes of calibration and to determine how closely values may be 

 repeated. 



4. The ratio of the arms. 



The details of the above determinations are as follows : 

 1. Testing of the Weights. — The weights were also by Bunge, 

 supplied with the balance, and of the usual type. They were sent 

 to the Bureau of Standards for comparison with the United States 

 standards. The following is taken direct from the certificate of the 

 bureau : 



"The weights designated below have been compared with the 

 United States standards and have the following corrections : 



Designation. 



100 grams 



50 " 



20 " 



101 " 



102 " 



5 " 



21 " 



22 " 



1 " 



"A ( + ) correction means that the weight is heavier than the 

 nominal value; a ( — ) correction that it is lighter by the amount 

 indicated. When applied to the nominal value of a weight, the 

 correction will give its mass referred to the international kilogram." 



2. /Sensibility Curve. — Readings for the sensibility curve were 

 made in the usual way, using loads of even numbers of grams 

 from to 200. The microscopic scale was used, sensibility being 

 taken as the number of divisions displacement for one mg. addi- 

 tional load. In the tabulation of results, "L" indicates the load; 

 "S," divisions displacement per mg.; "M," means of five preceding 

 values of "S." The values of "S" are the means of independent 

 determinations by each of the authors of this paper. However, 

 but one set of values was gotten by each observer, and it is proba- 

 ble that the irregularities are due quite as much, or more, to acci- 

 dents of observation as to lack of uniform action of the balance. 

 The curve traced from the values of "S" given shows a rather 

 rapid drop from to about 30, a tolerably uniform value slightly be- 

 low 5 from 30 to about 130, and from this point to the maximum load 



