208 KANSAS UNIVERSITY SCIENCE BULLETIN. 



through an apperture in irregular order. The difficulty with 

 the latter is that it involves a large proportion of motor ac- 

 tivity to the degree of attention required. While the attention 

 wave with the minimal light has in general proved satisfactory 

 in the Michigan laboratory, special conditions in the present 

 series made an alteration of the test seem advisable. The 

 sounder test was first suggested by Professor Shepard. Four 

 telegraph sounders were selected with as nearly the same 

 quality of sound as possible, and adjusted to approximately 

 the same intensity. A short piece of rubber tubing was slipped 

 on the end of the hammer of each sounder to render the back 

 stroke inaudible. In parallel with each sounder was a recorder 

 and a rheostat. The recorders were adjusted to write less than 

 a quarter inch apart (in later experiments, one-eighth inch 

 apart) on the smoked paper of the kymograph. A fifth re- 

 corder wrote in the same vertical line with these. It was con- 

 nected with the subject's key and recorded his reactions. The 

 sounders were operated by four keys arranged in a convenient 

 row. The subject to be tested sat at a table several feet from 

 the sounders, with his hand upon his key. The experimenter 

 operated one of the sounders several times in rapid succession 

 (six times in two seconds) as a signal. He then operated the 

 four sounders one second apart in irregular order for one 

 minute. The subject was to react to the "signal" sounder, 

 whenever he recognized it, by pressing his key. At the end of 

 the minute the operator signaled with another sounder, to 

 which the subject was to react for the next minute. This was 

 continued for sixteen minutes, the four sounders being used as 

 the "signal" sounder once in each four-minute period, but not 

 in the same order. 



The records were preserved upon smoked paper belts about 

 ten feet long. As the markers wrote closely together and the 

 paper traveled fairly fast, and thus spread out the marks 

 horizontally, it was an easy matter to count the number of 

 omissions and errors (wrong reactions). The signal at the 

 beginning of each minute required two seconds. There were, 

 therefore, fifty-eight clicks each minute. The one exception 

 was in the records of Do. The experimenter in this series did 

 not use a stop watch or metronome as a guide, as in the other 

 series, but operated the sounders at a convenient rate, which 

 was somewhat faster than one per second. In all the series the 



