THE PROCESS OF MEMORIZING. 



129 



Believing that, in English at least, real words represent the least 

 individual differences in apprehension, we abandoned the nonsense 

 syllables for real English words. In some respects at least it is a dis- 

 tinct advantage not to complicate the word series in memory tests by 

 a superposed letter series. In the case of words, we may assume that 

 the association between their spelling and the pronunciation is familiar 

 and fixed. Drawing from an equally well associated and equally 

 revivable mass of possible material, the process of memorizing words 

 is not complicated by the necessity for memorizing the constitution 

 of the members of the series. It is concerned solely with their serial 

 connection. 



APPARATUS AND TECHNIQUE. 



Normal series of 12 four-letter words were printed on strips of white 

 paper, 52 cm. in length, so that each word occupied the same proportion 

 of a 4 cm. space. Such a slip encircled the 50 cm. Blix-Sandstrom 

 kymograph drum, leaving 2 cm. spare space to indicate the beginning 

 of the series, as well as to accommodate the paper clip that held the 

 strip of words to the drum. A circular screen, with a slit large enough 

 to expose two letters at a time, covered the drum and consequently the 

 series of words, except as each word was exposed letter by letter when 

 the drum revolved. The screen and electrical connections are shown 

 diagrammatically in figure 26. 



FIG. 26. Diagram of the connections for memory experiment. 



As each word was perceived the subject spoke it into the voice key 

 (for description of this key, see Chapter III, p. 97), and so broke the 

 circuit of an electric marker that wrote the record of the reaction on 

 the same cylinder which carried the words. Since both stimulus and 

 reaction records were on the same evenly rotating drum, the correlation 

 of the two was permanent and mechanically accurate. Any advance 

 of the reaction record along its base-line on the second revolution of 

 the drum showed the effect of memory and was taken as its measure. 

 Since the cylinder moved at a rate of 10 mm. per second, the change 



