he carried the spool home with him, to try once more the
possibility of educating his water-wheel into a watchman.

That night the pull did indeed come, but, alas before he had even fallen
asleep.

Something seemed to be always going wrong! He concluded already that it
was a difficult thing to make a machine which should do just what the
maker wished. The spool had gone flying round, and had swallowed up the
thread incredibly fast. He made haste to get the end off his wrist, and
saw it fly through the little hole in the window frame, and away after
the rest of it, to be wound on the whirling spool.

Disappointing as this was, however, there was progress in it: he had got
the thing to work, and all that remained was to regulate it. But this
turned out the most difficult part of the affair by far. He saw at once
that if he were only to make the thread longer, which was the first mode
that suggested itself, he would increase the constant danger there was
of its getting fouled, not to mention the awkwardness of using such a
quantity of it. If the kitten were to get into the room, for instance,
after he had laid it down, she would ruin his every hope for the time
being; and in Willie's eyes sixpence was a huge sum to ask from his
father. But if, on the contrary, he could find out any mode of making
the machine wind more slowly, he might then be able to shorten instead
of lengthening the string.

At length, after much pondering, he came to see that if, instead of the
spool, he were to fix on the axis a small cogged wheel--that is, a wheel
with teeth--and then make these cogs fit into the cogs of a much larger
wheel, the small wheel, which would turn once with every turn of the
water-wheel, must turn a great many times before it could turn the big
wheel once. Then he must fix the spool on the axis of this great slow
wheel, when, turning only as often as the wheel turned, the spool would
wind the thread so much the more slowly.

I will not weary my reader with any further detail of Willie's efforts
and failures. It is enough to say that he was at last so entirely
successful in timing his machine, for the run of the water was always
the same, that he could tell exactly how much thread it would wind in a
given time. Having then measured off the thread with a mark of ink for
the first hour, two for the second, and so on, he was able to set his
alarum according to the time at which he wished to be woke by the pull
at his wrist