ents, each of definite amount. Either a whole drop emerges or
nothing emerges from the spout. If all change went thus drop-wise,
so to speak, if real time sprouted or grew by units of duration of
determinate amount, just as our perceptions of it grow by pulses,
there would be no zenonian paradoxes or kantian antinomies to trouble
us. All our sensible experiences, as we get them immediately, do thus
change by discrete pulses of perception, each of which keeps us saying
'more, more, more,' or 'less, less, less,' as the definite increments
or diminutions make themselves felt. The discreteness is still more
obvious when, instead of old things changing, they cease, or when
altogether new things come. Fechner's term of the 'threshold,' which
has played such a part in the psychology of perception, is only one
way of naming the quantitative discreteness in the change of all our
sensible experiences. They come to us in drops. Time itself comes in
drops.

Our ideal decomposition of the drops which are all that we feel into
still finer fractions is but an incident in that great transformation
of the perceptual order into a conceptual order of which I spoke in
my last lecture. It is made in the interest of our rationalizing
intellect solely. The times directly _felt_ in the experiences of
living subjects have originally no common measure. Let a lump of sugar
melt in a glass, to use one of M. Bergson's instances. We feel the
time to be long while waiting for the process to end, but who knows
how long or how short it feels to the sugar? All _felt_ times coexist
and overlap or compenetrate each other thus vaguely, but the artifice
of plotting them on a common scale helps us to reduce their aboriginal
confusion, and it helps us still more to plot, against the same scale,
the successive possible steps into which nature's various changes may
be resolved, either sensibly or conceivably. We thus straighten out
the aboriginal privacy and vagueness, and can date things publicly, as
it were, and by each other. The notion of one objective and 'evenly
flowing' time, cut into numbered instants, applies itself as a common
measure to all the steps and phases, no matter how many, into which we
cut the processes of nature. They are now definitely contemporary,
or later or earlier one than another, and we can handle them
mathematically, as we say, and far better, practically as well as
theoretically, for having thus correlated them one to one with ea