392 



THE BELL SYSTEM TECHNICAL JOURNAL, MARCH 1954 



Table XIII — Analysis of Variance — Summary — 

 Without Card Support Bars 



* See Appendix and Reference 6. 



t Significant at 5% level. 



j Significant at 1% level. 



(+) N.S. = Not significant at 5% level. 



creased regardless of the distribution of cards, confirming a conclusion 

 from the first experiment. 



V. CONCLUSIONS OF OVER-ALL STUDY 



When considered in the normal cycle of translator operation with 

 the card support bars functioning, the range of card drop times was ob- 

 served to be from 32 to 40 ms with a mean of 35.1 ms. Based on the results 

 obtained in this study it is predicted the dropping time of a card chosen 

 at random in a new translator will be between 27 and 43 ms. This includes 

 all the known variables except the Ufe of the cards. The test on this 

 variable is continuing and all conclusions reached in this report may be 

 modified somewhat by the results of this life study which will be re- 

 ported at a later date. The relative effect of the several variables is tabu- 

 lated in Table XIV. The only two variables found to have any sig- 



