fourth-page advertisement in four different cities in four
local papers, each of which has 100,000 readers. But if he uses the
same paper in one town, he would produce a much greater effect by
printing a fourth of a page four times than by using a full-page
advertisement once only.

As a matter of course this would hold true only as far as size and
repetition are concerned. Many other factors have to be considered
besides. Some of these could even be studied with our material. We
could study from our results what memory-value is attached to the
various forms of type or suggestive words, what influence to
illustrations, how far they reinforce the impressiveness and how far
they draw away the attention from the name and the object, how these
various factors influence men and women differently, and so on. Other
questions, however, demand entirely different forms of experiment. We
may examine the effects of special contrast phenomena, of unusual
background, of irregular borders and original headings. The particular
position of the advertisement also deserves our psychological
interest. The magazines receive higher prices for the cover pages and
the newspapers for advertisements which are surrounded by reading
matter. In both cases obvious practical motives are decisive. The
cover page comes into the field of vision more frequently. What is
surrounded by reading matter is less easily overlooked.

But the newspaper world hardly realizes how much other variations of
position influence the psychological effect. Starch[51] made
experiments in which he did not use real advertisements, but
meaningless syllables so as to exclude the influence of familiarity
with any announcement. He arranged little booklets, each of 12 pages,
on which a syllable such as _lod_, _zan_, _mep_, _dut_, _yib_, and so
on was printed in the middle of each page. Each of his 50 subjects
glanced over the book and then wrote down what syllables remained in
memory. He found that the syllables which stood on the first and last
page were remembered by 34 persons, those on the second and eleventh
by about 26, and those on the eight other pages by an average of 17
persons. In the next experiment he printed one syllable in the middle
of the upper and one in the middle of the lower half of each page. The
results now showed that of those syllables which were remembered 54
per cent stood on the upper half and 46 per cent on the lower half of
the page. Finally, he div