.504 



THE BELL SYSTEM TEfHXirAL JOURXAL, MARCH 1957 



7 

 6 



5 



>- 

 U 



z 



LU 4 



D 

 O 



LU 



DC 3 



U- 



2 

 1 



1.5 



2.0 2.5 3.0 



WORDS PER SECOND 



3.5 



4.0 



17 



23 



28 34 



BITS PER SECOND 



40 



45 



Fig. 5 — Distribution of reading rates for preferred vocabulary. 



indicates that the optmial list would be somewhat different for reader A 

 than for reader B. ]\Iore extensive data would, however, be required to 

 confirm this hj^pothesis. 



Experiments not described here in detail showed that reading rates 

 for digrams (successive pairs of words related as in English text) are 

 intermediate between those for prose and discrete words. 



Experiment 5: Effect of Multiple Channels 



Licklider^ has found that when the reader attempts simultaneous} j' to 

 perform a tracking operation while he is reading, his reading rate re- 

 mains almost unimpaired, and the tracking information is added to 

 that of reading alone. This two-channel transmission gave him his 

 highest rate of transmission. We obtained the reverse finding. Reading 

 the preferred list gave us our highest transmission rate. Simultaneous 



