It is said that about one thousand of the "Rotaries" were made. The highest serial number to come to the author's attention, 507, may indicate that only a part of the watches started were finished.
Accounts agree[35] that the next product of the factory was a "Timer" containing a novel escapement patented on May 28, 1878,[36] by William A. Wales. Early specimens are marked "Pat. Applied For," but one with the serial number 996[37] bears no reference at all to a patent, presumably because issuance of the patent or patents was imminent. Apparently the timer was in full production before the patent was issued on May 28. Specimens with higher serial numbers are stamped with three patent dates, May 28, 1878,[38] June 24, 1879, and September 30, 1879, as seen in figure 13, which also shows the arrangement of the train. In this escapement the escape wheel (fig. 14) carries in the rim any suitable number of steel pins all on the same radius from, and parallel to, the axis of wheel rotation. In all cases the wheel makes one revolution per second. The verge (figs. 14 and 15) is so proportioned that the distance between the points of repose on the entrance and exit pallets will stop the wheel at intervals equal to half the angular distance between the pins.
In other words, with two pins in the escape wheel the escapement will beat quarters of a second, because starting from a point of repose the wheel will be arrested on the other point of repose after turning through 90°. With four pins in the escape wheel and a suitably proportioned verge the escape wheel advances in steps of 45° and beats eighths of a second. The growing trend in this period to standardize the timing of sporting events in intervals of fifths of a second is reflected in still another model having five pins in the escape wheel and beating tenths of a second. By the nature of the verge in this escapement, it will be seen that the number of beats must be twice the number of pins in the escape wheel, leaving no way to secure an odd number of beats per second, hence the 1/10-second model. This must have been a less desirable form because of the much smaller verge required, plus the problem of accelerating so much mass from a dead stop 600 times per minute.