range of states of the system that resulted when he passed a bar magnet nearby:
Fig. 9.3. Linear array of magnetic compasses (14)
.
The idea was first to set [the compasses] nearly touching in a row. The individual needles have a
time constant, in pointing somewhere near to the magnetic north pole, of the order of a second. When
they are close to one another, however, they interact to an extent that overshadows the field of the
earth, and the time constant is of the order of, say, a tenth of a second. Thus they will point north to
south, north to south, on down the line [Fig. 9.3a]. The experiment was set up so that north was normal
to the axis of the array, and that gives a very stable sort of array. Bringing up a south pole gives a
repulsion that will tend to displace the end needle. You can see, I am sure, that a quasi-static system
will result, where we get something as shown in [Fig. 9.3b]. The angles of displacement will decrease,
so that after the initial impulse a dynamic situation is established and the signal moves along, not too
fast.
You can bring up the bar magnet slowly and dose, and maintain a static situation where
equilibrium is propagated, so that the needles assume angles equally. The behavior is an
exact analog of a gear train; that is, one turns this way, one that way, and so on [Fig. 9.3c].
It is very much like the bar where you turn one end and observe that the other end turns too.
That is not too interesting.
However, if you look upon this as a dynamic rather than a quasi-static system, you can get
some extraordinary phenomena that I cannot draw. With a little practice, bringing a south
pole up just right, you can make the first compass spin all around and nothing is propagated
down the line. The skill in my hand automatically introduces some random numbers, so the
experiments were not reproducible. I can tell, nevertheless, of several things that can
happen. If you bring the south pole up in a certain fashion, a nice signal goes along, with a
complete flip-flop of every needle in the row, and a truly binary, bistable system exists.
On the other hand, if you do not do it in quite the same way, the signal will go down only
so far, sometimes apparently even amplified through resonance, and somewhere along the
line one of the needles will turn all the way around, and the signal will be reflected and go
back again, never getting past a certain point. In other instances-you can run several
ELECTROMAGNETISM & LIFE - 130