t himself
away from the outer world and fall asleep, does his soul break the
dendritic contacts between cell and cell; and when he awakes, does it
make contacts and switch the impulses evoked by sense stimuli on to
one or other tract of the axons, or axis cylinder processes, which
form the association pathways? Such a hypothesis is no explanation; it
simply puts back the whole question a step further, and leaves it
wrapped in mystery. It cannot be fatigue that produces the
hypothetical interruptions of the dendritic synapses and then induces
sleep, for sleep can follow after fatigue of a very limited kind. A
man may sleep equally well after a day spent in scientific research as
after one spent in mountain climbing, or after another passed in
idling by the seashore. He may spend a whole day engaged in
mathematical calculation or in painting a landscape. He fatigues--if
we admit the localization of function to definite parts of the
brain--but one set of association tracts, but one group of cells, and
yet, when he falls asleep, consciousness is not partially, but totally
suspended.

We must admit that the withdrawal of stimuli, or their monotonous
repetition, are factors which do undoubtedly stand out as primary
causes of sleep. We may suppose, if we like, that consciousness
depends upon a certain rate of vibration which takes place in the
brain structure. This vibration is maintained by the stimuli of the
present, which awaken memories of former stimuli, and are themselves
at the same time modified by these. By each impulse streaming into the
brain from the sense organs, we can imagine the structure of the
cerebral cortex to be more or less permanently altered. The impulses
of the present, as they sweep through the association pathways, arouse
memories of the past; but in what way this is brought about is outside
the range of explanation. Perhaps an impulse vibrating at a certain
rate may arouse cells or fibrils tuned by past stimuli to respond to
this particular rate of vibration. Thus may be evoked a chain of
memories, while by an impulse of a different rate quite another set of
memories may be started. Tracts of association are probably formed in
definite lines through the nervous system, as during the life of a
child repeated waves of sense impulses beat against and overcome
resistances, and make smooth pathways here and there through the brain
structure. Thus may be produced growth of axons in certain directions,
a