ring, soon discovered it could
be manipulated to isochronism, i.e., so arcs of different extent would
be formed in equal time. Of course, the friction-rest escapement
requiring a spring to possess different properties from one which would
be isochronal with a perfectly detached escapement, these two frictional
escapements also differing, the duplex requiring other properties from
what would isochronize a spring for a cylinder escapement. Although
pocket watches with duplex and cylinder escapements having balances
compensated for heat and cold and balance springs adjusted to
isochronism gave very good results, careful makers were satisfied that
an escapement in which the balance was detached and free to act during
the greater proportion of the arc of vibration and uncontrolled by any
cause, would do still better, and this led to the detent escapement.

FAULTS IN THE DETENT ESCAPEMENT.

As stated previously, the detent escapement having pronounced faults in
positions which held it back, it is probable it would never have been
employed in pocket watches to any extent if it had not acquired such a
high reputation in marine chronometers. Let us now analyze the
influences which surround the detent escapement in a marine chronometer
and take account of the causes which are combined to make it an accurate
time measurer, and also take cognizance of other interfering causes
which have a tendency to prevent desired results. First, we will imagine
a balance with its spring such as we find in fine marine chronometers.
It has small pivots running in highly-polished jewels; such pivots are
perfectly cylindrical, and no larger than are absolutely necessary to
endure the task imposed upon them--of carrying the weight of the balance
and endure careful handling.

To afford the necessary vibrations a spring is fitted, usually of a
helical form, so disposed as to cause the balance to vibrate in arcs
back and forth in equal time, _provided these arcs are of equal extent_.
It is now to be taken note of that we have it at our disposal and option
to make these arcs equal in time duration, i.e., to make the long or
short arcs the quickest or to synchronize them. We can readily
comprehend we have now established a very perfect measure of short
intervals of time. We can also see if we provide the means of
maintaining these vibrations and counting them we should possess the
means of counting the flights of time with great accuracy. The
conditi