which I started by pressing a spring the moment
the word caught my eye, and which stopped of itself the instant I
released the spring; and this I did so soon as about a couple of
ideas in direct association with the word had arisen in my mind. I
found that I could not manage to recollect more than two ideas with
the needed precision, at least not in a general way; but sometimes
several ideas occurred so nearly together that I was able to record
three or even four of them, while sometimes I only managed one. The
second ideas were, as I have already said, never derived from the
first, but always direct from the word itself, for I kept my
attention firmly fixed on the word, and the associated ideas were
seen only by a half glance. When the two ideas had occurred,

I stopped the chronograph and wrote them down, and the time they
occupied. I soon got into the way of doing all this in a very
methodical and automatic manner, keeping the mind perfectly calm and
neutral, but intent and, as it were, at full cock and on hair trigger,
before displaying the word. There was no disturbance occasioned by
thinking of the forthcoming revulsion of the mind the moment before
the chronograph was stopped. My feeling before stopping it was
simply that I had delayed long enough, and this in no way interfered
with the free action of the mind. I found no trouble in ensuring the
complete fairness of the experiment, by using a number of little
precautions, hardly necessary to describe, that practice quickly
suggested, but it was a most repugnant and laborious work, and it
was only by strong self-control that I went through my schedule
according to programme. The list of words that I finally secured was
75 in number, though I began with more. I went through them on four
separate occasions, under very different circumstances, in England
and abroad, and at intervals of about a month. In no case were the
associations governed to any degree worth recording, by remembering
what had occurred to me on previous occasions, for I found that the
process itself had great influence in discharging the memory of what
it had just been engaged in, and I, of course, took care between the
experiments never to let my thoughts revert to the words. The
results seem to me to be as trustworthy as any other statistical
series that has been collected with equal care.

On throwing these results into a common statistical hotch-pot, I
first examined into the rate at which th