240 



SCIEJS'CE. 



[Vol. VIII., No. 188 



namely, to read the denomination of each card, 

 yet it takes longer to do so the greater the number 

 of denominations to which it may belong.^ One 

 must take a longer look at a card to tell that it is 

 a four-spot, for instance, where it may be a one, 

 two, four, six, eight, or ten, than when it may be 

 a two, four, or sis. This difference was most 

 marked with me in passing from two to three 

 kinds. The increasing number of possibilities is 

 more puzzling to the little girl than to the others ; 

 for it takes her as much as ^ of a second to tell 

 the cards when five denominations are used, 

 whereas it takes the others only about ^^g- of a 

 second. 



A few words of caution must be added for those 

 who intend to repeat the experiments. Do not 

 expect very constant results at first ; the famil- 

 iarity which one acquii-es after the second or third 

 trial very much reduces the time ; after this there 

 is a more gradual reduction, due to practice. The 

 numbers in the table are regular only because 

 founded on many sets of experiments, and the 

 first few records of each kind of reaction are 

 omitted in a few cases. 



3. Choice time. — This time is obtained by an 

 indirect process. We have already become ac- 

 quainted with the ' throwing time.' This time has 

 no particular psychological interest, as it simply 

 tells how long it takes one to throw out cards. 

 This time will differ very largely in different 

 persons, and is much reduced by practice. It took 

 I. I of a second, II. j^^ of a second, and III. ^ of 

 a second, to throw a card uj)on one of two heaps. 

 It takes longer to distribute the cai'ds, the more 

 numerous the heaps among which they are to be 

 divided ; but the increase in time is slight. It 

 took I. less than | of a second to place a card in 

 one of five heaps, and II. and III. | and i of a 

 second respectively when six heaps were used. Of 

 course, the time refers to the simple operation of 

 placing the cards, without reference to their de- 

 nomination, in one of a certain number of heaps. 

 Each of these counts has a different mode of reac- 

 tion. 



Having gotten the throwing time, the next step 

 is to distribute the cards among the heaps in such 

 a way that each heap will contain but one kind of 

 cards. If we are throwing five-spots and nine- 

 spots, then all the five-spots must be put on one 

 heap, and all the nine-spots on the other. If we 

 are usmg two, four, six, and eight spots, then there 

 will be four heaps, each containing all the cards 

 of one denomination. In addition to the time con- 



1 The only comparable experiment (and the similarity is 

 not very close) I can find is one recently published by Dr. 

 Cattell, in which he finds that it takes only about 1-160 of a 

 second longer to distinguish one out of ten than one out 

 of two colors. 



sumed by the manual operation of taking the card 

 and placing it on the pack, part of the time is con- 

 sumed in recognizing the denomination of the card, 

 and the rest in placing it on its appropriate pack. 

 In other words, if from the time occupied by this 

 operation we subtract the throwing time, we have 

 left the distinction time together with the choice 

 time. But we know the value of the distinction 

 time by our previous experiments. Simple sub- 

 traction yields the choice time. I will again put 

 the results in the form of a table. 



If we compare this table with the former one, 

 we see at once that, as before, the time increases 

 with the complexity of the operation ; but the 

 increase is more rapid in this table than in the 

 former one. This is just what we should expect ; 

 for in the former case it was the same process to 

 be done under different conditions, while here the 

 nature of the reaction is changed with each addi- 

 tional kind of card. "When we deal with but two 

 kinds of cards, the choice time and the distinction 

 time are about equal. This agrees well with Pro- 

 fessor Wundt's results.' The process readily be- 

 comes at least partly automatic. But as we pass 

 to a choice between three kinds of I'eactions, it 

 would seem that a distinct exertion of the will is 

 necessary in each case. The time undergoes a 

 marked increase. From that point on, the increase 

 in time with the complexity of the operation is 

 more gradual. But, as before, the little girl finds 

 great difficulty in distributing the cards appropri- 

 ately when many kinds are used. It takes her 

 over 1^ of a second to determine upon which of five 

 heaps to put a card after she knows its denomina- 

 tion, while it only takes the others i and -^^ of a 

 second to perform the same operation with six 

 heaps. 



A comparison of the first and last columns of 

 the table shows the regularity of the phenomena 

 we are studying. The choice time ought not, of 

 course, to be affected by the nature of the distinc- 

 tion upon which it is founded ; and the choice 

 time for five-spots and nine-spots and that for 

 green and blue ought to be and are (approximately) 



1 It is again difficult to find comparable results. But the 

 distinction plus choice time can be compared with similar 

 results of Dr. Cattell. His figure is .078; mine is 0.81. 



