over some points caused me to shoot a
piece of bread-and-butter on to the floor. I stooped to pick it up.

"Stop a moment, please!" cried my companion. He jumped to his feet and
examined it. "Ah," said he, "buttered side downward!"

"It's always the same," I said, as I jerked the thing viciously out of
the window. "It's _always_ buttered side downward."

"No, there you fall into a common error," protested the other. "You may
take it that fifty-seven per cent. fall buttered side upward, and only
forty-three per cent. buttered side downward."

"H'm," I said dubiously.

"You must pardon me for my officiousness," he went on, "especially as I
have now no reason to be interested in such things. But habits are
strong."

I looked at him curiously. "Habits?" I said.

"Yes, habits. For years I kept an accurate record of every slice of
bread-and-butter I saw fall to the ground. I had better explain myself.
Nearly all my life, you must understand, I have maintained the view that
the generally accepted theory of the 'cussedness of things' is all
wrong. You know that to most people 'cussedness' is the governing factor
of life."

"Rather!" I agreed.

"Well, I disbelieved it, and I set to work to collect materials for a
book which was to prove my case. For years I incessantly gathered
statistics on the subject. Do I bore you?"

"Not at all," I assured him.

"The results were extraordinary. Take, for example, catching trains. It
is highly important that you should catch a train at short notice. In
nine cases out of ten, you will say, your taxicab breaks down, or your
tram is held up by a block in the traffic, or the current fails on the
Underground."

"Certainly it does."

"On the contrary--I am speaking from memory, but I think my figures are
accurate--the taxicab only breaks down in 1.5 per cent. of cases; with
the tram the percentage rises to 1.8; with the Underground it falls to
.2."

I gasped.

"Or take the case of studs," he went on. "You drop a stud, and it
promptly and inevitably rolls away into some quite impossible
hiding-place. So most of us believe. As a matter of fact it only does so
approximately three times out of a hundred. Or bootlaces. If you are
exceptionally late in the morning; your bootlace always snaps, you say.
Not at all. It breaks in such circumstances only four times out of a
possible hundred. And with bicycles, to take another example. If ever
you get a puncture, you fancy that it always