not grow with a further
reduction of the size. Two thirds of the subjects were men and one
third women. On the whole, the same relation exists for both groups,
but the climax of psychical efficiency was reached in the case of the
men by the 4 times repeated fourth-page, in the case of the women by
the 8 times repeated eighth-page. The 4 times repeated fourth-page in
the case of the women was 0.45, in the case of the men, 0.51, the 8
times repeated eighth-page, women, 0.53, men, 0.37.

I am inclined to believe that the ascent of the curve of the
memory-value from the full-page to the fourth-page or eighth-page
would have been still more continuous, if the whole-page
advertisements had not naturally been such as are best known to the
American reader. The whole-page announcement, therefore, had a certain
natural advantage. But when we come to another calculation, even the
effect of this advantage is lost. We examined the relations for the
first 10 names and articles, which every one of the 30 persons wrote
down. These first 10 were mostly dashed down quickly without special
thought. They also included only a few half reproductions. When we
study these 300 answers which the 30 persons wrote as their first 10
reproductions, and calculate from them the chances which every one of
the 60 advertisements had for being remembered, we obtain the
following values: The probability of being remembered among the first
10 was for the full-page advertisement, 0.5, for the half-page 2 times
repeated, 1.2, for the fourth-page 4 times repeated, 2.9, for the
eighth-page 8 times repeated, 2.3, and for the twelfth-page 12 times
repeated, 2.4. The superiority of repetition over mere size appears
most impressively in this form, but we see again in this series that
the effect decreases even with increased number of repetitions as soon
as the single advertisement sinks below a certain relative size, so
that the 12 times repeated twelfth-page advertisement does not possess
the memory-value of the 4 times repeated fourth-page advertisement. If
Scott's experiments concerning the size and these experiments of mine
concerning the repetition are right, the memory-value of the
advertisements for economic purposes is dependent upon complicated
conditions. A business man who brings out a full-page advertisement
once in a paper which has 100,000 readers would leave the desired
memory-impression on a larger number of individuals than if he were to
print a