754 BELL SYSTEM TECHNICAL JOURNAL 



by computing, for each of the laboratories, the average tensile strength 

 and the standard deviation of the specimens milled in the four labora- 

 tories and tested in each of the respective laboratories. In each 

 instance the specimens tested or milled by an individual laboratory 

 are grouped together. The dotted lines on the diagrams represent 

 the control limits within which the data should fall without leaving 

 anything to chance, which means that points falling without these 

 control limits indicate variations due to assignable causes, such as 

 errors in measurements, defects in the material, etc. These diagrams 

 show, therefore, that the tensile strength results between laboratories 

 do not reveal significant or assignable difficulties other than those 

 which could be attributed to chance. In other words, the analysis 

 gave no indication of the presence of assignable variations between 

 the testing laboratories. 



These specimens were prepared by cross milling the gage length 

 with a milling cutter shaped to conform to the final shape of the 

 specimen desired. This method, which has been described elsewhere,*^ 

 results in the saving of time and produces specimens of a uniform 

 character. 



Scleroscope Hardness Tests: 



Previous to this investigation Bell System specifications on non- 

 ferrous materials were written in terms of scleroscope hardness, but 

 considerable trouble was encountered between the suppliers and users 

 of metal due to difficulty in checking each other's readings. Because 

 results could not be duplicated on two instruments, it was necessary 

 to allow rather wide limits in each temper of material. This resulted 

 in considerable overlapping in scleroscope limits of the tempers of 

 materials accepted under these specifications. Experience had shown 

 it to be impossible to make correction curves for any two instruments 

 which would hold for any reasonable length of time, and if any re- 

 placements such as new hammers were necessary the calibration 

 was changed. 



In order to obtain more definite information as to what could be 

 expected from the scleroscope, comparisons were made between four 

 type "C" sclerscopes located in four laboratories using the same 

 samples of materials. Great care was taken in preparing the samples 

 so that they were flat and free from any dirt and roughness. The 

 instruments were in good commercial adjustment and were employed 

 in the usual manner, the magnifier hammer being used. A single 



^ R. L. Templin, "Methods for Determining the Tensile Properties of Thin 

 Sheet Metals," Proc. A. S. T. M., Vol. 27, 1927. 



