956 THE BELL SYSTEM TECHNICAL JOURNAL, JULY 1953 



1^ SERIES A -H h- SERIES B -*\ 



asA - 



R5 LIM 



^1 



RANGES 



40 



^« ' I I ' I ' I ' I I I ' ' ' ' I ' I ' I I ' ' I I I I I I ' I I I ' I I I 

 4 8 12 16 20 24 28 32 36 



SAMPLE NUMBER 



Fig. 4 — Control charts for averages and ranges, samples of 5 from a normal 

 universe. 



B, however, a failure to meet Criterion II would have been encountered 

 on the seventh plotted point, which would have caused loss of eligibility 

 to continue using the control chart method. This would have required a 

 reversion to 100 per cent inspection and acceptance by the three-cell 

 method until such time as Criterion I was again met. 



Individual units in a lot accepted by the control chart method are to 

 be selected at random from the lot in groups of 5 each, and the five units 

 of each group are to be placed in a single package or otherwise associated 

 so as to remain physically together until delivered to the user. 



3.2 BATCH METHOD 



For some component elements, units are produced intermittently in 

 relatively large groups or batches and subjected collectively to the 

 same manufacturing processes. Under these conditions individual units 



