he period of revolution is much slower. Position errors
average out as certainly if not as frequently. In Bonniksen's "Karrusel"
watch of 1893[2] the duration of a cycle is 52.5 minutes[3] while in
the Auburndale Rotary which we are about to discuss the period of each
revolution is 2-1/2 hours.

-----
[1] Paul M. Chamberlain, _It's about time_, New York, 1947,
p. 362.

[2] British patent 21421, granted January 21, 1893.

[3] Chamberlain, _op. cit._ (footnote 1), pp. 229, 230.

The Invention

The patent model of Jason R. Hopkins' revolving watch, now in the U. S.
National Museum,[4] was not the first in which the entire train
revolved but it was a very novel conception intended to reduce greatly
the number of parts usually associated with any watch. This may be seen
from figures 2 and 3, where everything shown inside the ring gear
revolves slowly as the main spring runs down. This spring is prevented
from running down at its own speed by the train pinion seen in mesh with
the ring gear. Through this pinion motion is imparted to the escape wheel
and balance, where the rate of the watch is controlled. The balance,
being planted at the center of revolution, travels around its own axis,
as in the tourbillon, at the speed with which the entire train revolves
around the barrel arbor. This arbor turns only during winding. No dial or
dial gearing is shown in the patent or exists in the patent model. The
patent merely says, casually, "By means of dial wheels the motion of the
barrel may be communicated to hands and the time indicated in the usual
manner." No fine finish or jeweling has been lavished on the model, the
only jewels present being in the balance cock which was utilized as it
came from its original watch with only minor modification to the shape of
its foot. Apparently the balance wheel itself is also a relic of the same
or a similar conventional watch. There is no jeweling in the escapement
or on the other end of the balance staff. In spite of this the model runs
very actively and will overbank if wound up very far. The beat of the
escapement is two per second and the movement revolves once in 20
minutes.

There are two great faults in the model. First is the lack of an adequate
bearing for the barrel to turn on. There is only one very short bearing a
long way removed from the point of engagement between the pinion and
internal gear, and no adequate support is given the barrel, with the
resul