Memory. 69 



point on the record designated by B where the word was entirely visi- 

 ble to the sub j ect . The first treatment was the one employed by Dodge 

 and Benedict in their series. The time distance between #! and R 2 , 

 provided the second reaction occurred at a point earlier than the first, 

 was taken as the " saving " due to the influence of memory. Complete 

 memorization would mean that the word was spoken the second time 

 before any part of it came into view. The score would be at least 250, 

 and in most cases longer, depending upon the position of the first 

 reaction. 



Using the same series of words each day and on successive days 

 enables, of course, the more or less permanent memorization of the word 

 series; hence we have minus preliminary reaction results, i. e., reactions 

 which show that the subject spoke the word during the first revolution at 

 some time before it had been entirely exposed. This of course might 

 happen in rare instances even with new material. The individual let- 

 ters of a word vary considerably in width, the letter m, for example, in 

 the type used for our lists, being 10 mm. wide, while & is 5 mm. wide. 

 In the exposure of the word "mill" the subject begins to see the final 

 letter, which is the first letter of the word, sooner than he would if the 

 word exposed was "bill." With the latter word he can not be entirely 

 sure of its identity until the very last 2 mm. of the movement exposing 

 the word have been completed, for he is not sure that the word is to 

 be "bill" or "pill," and must wait for the first clue to its identity. 



This phenomenon of speaking the word the first time before the expo- 

 sure has been entirely completed increases from day to day in our data, 

 this occurring the following number of times: 1, 7, 12, 15, 9, and 31 for 

 the six days in order, respectively. The. increase is fairly regular, except 

 on the fifth day, July 3, when there were only 9 of these minus prelimi- 

 nary reactions. 1 The factor of residual memory is therefore evident, even 

 though we employed only two revolutions per test. Obviously, if the 

 subject produces minus reactions and then at the time of the second 

 revolution of the drum waits for a part of the word to appear before the 

 process of recall is complete and he speaks the word, 2 his "saving" as 

 figured from reaction to reaction will be considerably reduced. The 

 #! might be so early that if the subj ect waited for the word to appear in part 

 before speaking it the second time, R 2 would show a very minor "sav- 

 ing" or even minus saving value. Cases of this sort (minus "saving"), 

 where the Bi occurred at a position to the right of # 2 ,are to be found in 

 the series. 



Since the normal day, June 29, was the first experimental day and 

 the measurement on that day would therefore not be affected by the 



1 The mental attitude of a subject not scientifically trained for such experiments will unavoid- 

 ably shift from time to time in spite of instructions. On this day there was evidently less effort 

 to speak the word as quickly as possible in the preliminary reactions. 



2 This was his tendency, for in the experiments here reported only 5 reactions show that he pro- 

 nounced the word before any part of it had been seen. 



