the same sound one should be written as a substitute for the other,
as we have done in this book.

2. Train the Eye. No doubt the most effective way of learning spelling
is to train the eye carefully to observe the forms of the words we read
in newspapers and in books. If this habit is formed, and the habit of
general reading accompanies it, it is sufficient to make a nearly
perfect speller. The great question is, how to acquire it.

Of course in order to read we are obliged to observe the forms of words
in a general way, and if this were all that is needed, we should all
be good spellers if we were able to read fluently. But it is not all.
The observation of the general form of a word is not the observation
that teaches spelling. We must have the habit of observing every
letter in every word, and this we are not likely to have unless
we give special attention to acquiring it.

The "visualization" method of teaching spelling now in use in the
schools is along the line of training the eye to observe every letter
in a word. It is good so far as it goes; but it does not go very far.
The reason is that there is a limit to the powers of the memory,
especially in the observation of arbitrary combinations of letters.
What habits of visualization would enable the ordinary person to
glance at such a combination as the following and write it ten minutes
afterward with no aid but the single glance: _hwgufhtbizwskoplmne?_
It would require some minutes' study to memorize such a combination,
because there is nothing to aid us but the sheer succession of forms.
The memory works by association. We build up a vast structure of
knowledge, and each new fact or form must be as securely attached
to this as the new wing of a building; and the more points at which
attachment can be formed the more easily is the addition made.

The Mastery of Irregular Words.

Here, then, we have the real reason for a long study of principles,
analogies, and classifications. They help us to remember.
If I come to the word _colonnade_ in reading, I observe at once that
the double _n_ is an irregularity. It catches my eye immediately.
"Ah!" I reflect almost in the fraction of a second as I read in
continuous flow, "here is another of those exceptions." Building on
what I already know perfectly well, I master this word with the very
slightest effort. If we can build up a system which will serve the
memory by way of association, so that the sligh