Discrimination of shades of gray for different intervals of time. 19 
for 15 sec. — 0,74, for 30 sec. — 0,74, for 60 sec. — 0,71. It is 
probable therefore that for this reagent, at any rate, the longest 
intervals are judged most quickly and the shortest intervals most 
slowly. The difference would be greater if the highly ambiguous class 
of doubtful judgments were omitted. 
The results of the time measurements are in accord with the 
explanation of these judgment processes. The reaction time for 60 sec. 
is shorter than for any other time interval; it is the interval also in 
which one would rely most on the formation of a scale of bright¬ 
nesses and on free judgments. 
As regards the reactions for 5 sec., it has been observed that 
the demeanor of reagents for the shorter intervals is commonly, not 
invariably, different from what it is in the longer. In the 30 and 
60 sec. intervals, the reagent is apt, soon after the exposure of the 
norm, to relax the trunk muscles, settle himself into an easy attitude, 
to breathe easily and to move the eyes from time to time over the 
background. During the shorter periods as has already been observed, 
the reagents usually try to maintain uniform sensory conditions for 
both norm and comparison: the tension sensations from the trunk, 
respiration and eye muscles are kept constant, in order apparently, 
to make the conditions of comparison as much alike as possible. 
Accordingly we have, for most reagents, a much larger mass or back¬ 
ground of sensation entering into the comparisons of the shorter 
intervals, and in all probability more genuine acts of comparison. 
This position is strengthened by the longer reaction time for 5 sec., 
— so far as it is permissible to draw conclusions from a few ex¬ 
periments. 
As has been remarked the percentage of right cases for N = V 
decreases with the time interval and the law of forgetting for sen¬ 
sory impressions has been drawn from the like cases1). The judg¬ 
ments of ->like« differ from those »unlike« in being in great part 
negative. They are, so to speak, a function of the »unlike« judgments. 
The greater the difference between N and F, and the easier it is 
to mark differences, the easier it is to judge »like« when N — V. 
Accordingly, judgments of »like« result, in great part, from failure 
1) Wolfe, in Philos. Studien, III, p. 552. 
2*