at first than later. More is lost in the first
hour than in the second hour, and more in the first week than in the
second week. Few of the experiments have been continued long enough to
bring the curve actually to the zero line, but it has come very close
to that line in tests conducted after an interval of two to four
months.

PER CENT. OF WORDS RECOGNIZED AT DIFFERENT INTERVALS AFTER BEING SEEN
(From Strong)

Interval between Per cent. recognized with
exposure and test certainty and correctness

15 secs. 84

5 min. 73

15 min. 62

30 min. 58

1 hour 56

2 hours 50

4 hours 47

8 hours 40

12 hours 38

1 day 29

2 days 24

4 days 19

7 days 10

The subject read a list of 20 disconnected words once through,
giving careful attention to each word. Immediately at the close of
the reading he performed an example in mental arithmetic, to prevent
his reviewing the list of words mentally. After an interval, he was
shown these {352} twenty words mixed with twenty others, and had to
pick out those he surely recognized as having been shown before.
Many lists were used, for testing after the different intervals.
Five adult subjects took part in the experiment, and in all 15 lists
were used with each interval; the per cents. given in the table are
the averages for the 15 lists.

THE PER CENT. OF ERROR IN RECALLING DETAILS OF A
PICTURE AFTER DIFFERENT INTERVALS OF TIME
(From Dallenbach)

Time of test Per cent, of error Per cent of error
in spontaneous in answering
recall questions regarding
the picture

Immediately after exposure 10 14

After 5 days 14 18

After 15 days 18 20

After 46 days 22 22

The picture was placed in the subject's hands, and he examined it
for one minute, at the end of which time he wrote down as complete a
description of the picture as possible, and then answered a set of
sixty questions covering all t