204 KANSAS UNIVERSITY SCIENCE BULLETIN. 



METHODS AND RESULTS. 

 The experiments were undertaken in the psychological lab- 

 oratory of the University of Kansas in 1910, and continued 

 until 1915. During the summer the work was carried on at the 

 University of Michigan. The author is indebted to Professor 

 Pillsbury for suggesting the problem and for many suggestions 

 relating to the procedure. Especially deserving of gratitude 

 are those who have given so faithfully of their time and energy 

 as subjects for these experiments, which were of necessity 

 decidedly irksome. The subjects in the University of Kansas 

 laboratory were Professor D. C. Rogers, E. D. Campbell, in- 

 structor in logic, and George Babb, Raymond Edwards, 

 William Hoyt, Verne Miner, Emily Berger, George Bunn and 

 Carl Brown, all advanced students in psychology, except Mr. 

 Bunn, a sophomore engineer, who was employed in the labora- 

 tory as mechanic. In the Michigan laboratory Professor John 

 F. Shepard, and W. H. Batson, R. I. Cook, J. E. De Camp, and 

 C. S. Wang, graduate students in psychology, were subjects. 



ADDITION TEST. 



The first test to be used was the addition of figures. The 

 test differed from the Kraepelin tests in that the numbers to 

 be added were in single columns of twenty numbers each, and 

 the subject added the whole column and wrote the sum at the 

 bottom. In the Kreapelin tests the subject added only two 

 numbers at a time. He might, or might not, write the sum. 

 There are two advantages in this method: (1) it is slightly 

 more difficult to add and retain the sum of twenty numbers than 

 it is of two numbers as in the Kraepelin method; and (2) it 

 provides an easy means of determining the accuracy as well 

 as the speed of the operations without any serious interference 

 from writing the sums. In fact it was found that the subjects 

 soon acquired the habit of beginning to add the second column 

 before the answer to the first was entirely written. All sub- 

 jects were given a few days' training, though not enough to 

 render further improvement negligible, for it was believed 

 that, while the subjects should be familiar with the method, it 

 was not necessary to eliminate improvement entirely. The 

 tests, of fifteen minutes' duration, preceded and followed an 

 interval of fifteen or twenty minutes. The validity of a test 

 of this length will be discussed later. 



