460 H. M. JOHNSON 



the work was done. Reference to the primary purpose of the 

 experiment will make it clear why the range of stimulus- variables 

 was narrow, and why certain refinements of control were not 

 employed. 



PRELIMINARY CONSIDERATIONS 



The problem of ascertaining suitable methods of investigating 

 such questions was referred to the writer in the winter of 1915-16 

 by the director of the Nela Research Laboratory. Except for 

 the excellent work of Cobb 2 on the thresholds for detail and 

 for differences in brightness, the field had been scarcely 

 touched. Cobb's results, while interesting in themselves, are 

 perhaps more important in that they demonstrate to the illumi- 

 nating engineer (1) that sensory thresholds are exceedingly un- 

 stable and uncertain things, even when the external situation 

 and the physiological factors are optimal; so that large numbers 

 of observations are necessary to satisfactory evaluation of the 

 results; and (2) that an objective check on the reports of the 

 observer is absolutely essential to intelligent interpretation. 

 These facts have long been known to the psychologists, and they 

 can be readily appreciated by any other scientists who may 

 be brought to consider them. 



EXPLORATORY EXPERIMENTATION 



For ad hominem reasons it was inexpedient to adopt as highly 

 "artificial" a method as that of determining discrimination- 

 times without first testing the adequacy of some " simpler, easier 

 and more natural" method. Accordingly the first exploratory 

 work was in the attempt to locate such a method. 



The most obvious of the possible types of test was first selected 

 for study. It was proposed to assign the subject some definite 

 task, to the performance of which vision is indispensable; and, 



2 Cobb, P. W. and Geissler, L. R. : The effect on foveal vision of bright sur- 

 roundings. Psychological Review, xx, 1913, pp. 425-447; also Cobb, P. W. : same 

 title, ibid., xxi, 1914, pp. 23-32; and Journ. Exper. Psychol., i, 1916, pp. 419-425, 

 and pp. 540-566. 



