the person could learn the new numbering and
gain proficiency in distributing the cards in the new way more quickly
than was the case at first. Similarly, if one learns to run a
typewriter with a certain form of keyboard, one can learn to operate a
different keyboard much more quickly than was the case in learning the
first keyboard.

It is probable that the explanation of this apparent transfer is that
there are common elements in the two cases. Certain bonds established in
the first habit are available in the second. In the case of distributing
the cards, many such common elements can be made out. One gains facility
in reading the numbering of the cards. The actual movement of the hand
in getting to a particular box is the same whatever the number of the
box. One acquires schemes of associating and locating the boxes, schemes
that will work in both cases. But suppose that one spends fifteen days
in distributing cards according to one scheme of numbering, and then
changes the numbering and practices for fifteen days with the new
numbering, at the end of the second fifteen days one has more skill than
at the close of the first fifteen days. In fact, in five days one has as
much skill in the new method as was acquired in fifteen days in the
first method. However, and this is an important point, the speed in the
new way is not so great as the speed acquired in thirty days using one
method or one scheme all the time. Direct practice on the specific habit
involved is always most efficient.

One should probably never learn one thing _just because_ it will help
him in learning something else, for that something else could be more
economically learned by direct practice. Learning one language probably
helps in learning another. A year spent in learning German will probably
help in learning French. But two years spent in learning French will
give more efficiency in French than will be acquired by spending one
year on German and then one year on French. If the only reason for a
study is that it helps in learning something else, then this study
should be left out of the curriculum. If the only reason for studying
Latin, for example, is that it helps in studying English, or French, or
helps in grammar, or gives one a larger vocabulary in English on account
of a knowledge of the Latin roots, then the study of the language cannot
be justified; for all of these results could be much more economically
and better attained by a direct