In 1914, six years into production of the Washington-Franklin series, the perforations were altered to gauge 10 on all sides. The change was made in response to complaints from postmasters and users that the perf 12 stamps separated too easily, a problem encountered on earlier coil stamps. The weakened and separated sheets made it difficult for local postmasters to properly account for stock, and businesses lost time putting stamps on parcels when multiples fell apart.
The Bureau had previously experimented with perf 81/2 stamps for coils to resolve the same problem. These were found to be too hard to separate, so the Bureau settled on perf 10 for the new gauge. Johl notes that the first machine was set to gauge 10 on September 4, 1914, and that the last machine was set on November 4 of the same year. This month-long change gave philatelists some of the rarest Washington-Franklin varieties, the compound-perf 10 x 12 or 12 x 10 stamps.
In order for stamps to be perforated on all sides, they were put through the perforating machine twice—once for horizontal and once for vertical perforations. Some sheets were perforated at the different gauges during the transitional period. Three different denominations exist with compound perforations: 1c, 2c and 5c. The Zoellner collection contains one of only two covers known with the compound-perf (lot 616, a postal card), a superb 2c with plate number (lot 618), and an example of the very rare 5c.
The Scott Catalogue lists the compound-perf stamps as small-letter (a, b, c, d,) varieties of the basic Perf 10 stamps. This treatment is inconsistent with other perforation varieties, such as the 1923 Rotary Perf 11 stamps (Scott 544, 594, 596 and 613), which were created when a non-standard gauge of perforation was applied to a small quantity of sheets. In our opinion the 1914 compound-perf stamps deserve full status as major Scott numbers, in recognition of their place in the transition from gauge 12 to 10 perforations. If they were given appropriate major listings, the Scott numbers would be 424A(1c—P12 x 10). 424B(1c—P10 x 12), 425A (2c—P10 x 12), 425B (2c—P12 x 10) and 428A(5c—P12 x 10). One can imagine how much more valuable these stamps would become if collectors were not misled into thinking that they are something less significant than the 1923 Rotary Perf 11 rarities.
By 1914 demand for coil stamps had increased to the point where the Post Office Dept. began to examine alternative, less labor-intensive printing methods. Increased plate size eventually gave way to the new Stickney rotary press, named for its inventor who worked at the Bureau of Engraving and Printing. The rotary press was able to print using paper in long rolls, with plates wrapped around a cylinder. The machine also gummed stamps. This was ideal for coil production, as pasting sheets together was no longer necessary.
The new process worked well, except that the first rotary press coils were found to have less sharp impressions than flat plate printings. New dies were created with deeper shading lines to overcome the problem. This partly accounts for the rarity of the 2c Type I vertical coil (Scott 449). The Zoellner collection contains a joint line strip of four of this rarity (lot 626).
Some interesting errors were also printed during this time. The 5c Rose and Carmine color errors—Scott Nos. 467, 485 and 505—had nothing to do with ink colors. As plate number 7942 was produced for the 2c value and proofs were taken, it became necessary to re-enter three positions. The error was created when the 5c transfer roll was mistakenly used and three 5c subjects were entered on a plate of 2c stamps. The error was not noticed until the plate had been in use for some time. The 5c Rose and Carmine errors exist as Perf 10, Perf 11 and the extremely rare Imperforate stamps. The Zoellner collection contains outstanding examples of each, including an Imperforate Mint N.H. block containing the double 5c error positions (lot 659).