The prospect of using steam power in printing prompted research into means by which the different operations of the printing process could be joined together in a single cycle.

In 1803, in Germany, Friedrich Koenig envisaged a press in which the raising and lowering of the platen, the to-and-fro movement of the bed, and the inking of the form by a series of rollers were controlled by a system of gear wheels. Early trials in London in 1811 were unsuccessful.

Presses with a mechanized platen produced satisfactory results after the perfection, in the United States, of the “Liberty” (1857), in which the action of a pedal caused the platen to be held against the bed by the arms of a clamp.

Though Nicholson very early took out patents for a printing process using a cylinder to which the composed type pieces were attached, he was never able to develop the necessary technology involved.

The cylinder was in fact the most logical geometric form to use in a cyclical process. It was also the one capable of providing the greatest output. Given an equal amount of energy, the pressure exerted by a platen had to be spread over the whole of the surface to be printed, whereas the pressure exerted by a cylinder could be concentrated on the strip of surface actually in contact with the cylinder at any one instant.

A limited demonstration of the efficiency of the cylinder had been made as early as 1784 on a French press for books for the blind.

In 1811 Koenig and an associate, Andreas Bauer, in another approach to the rotary principle, designed a cylinder as a platen bearing the sheet of paper and pressing it against the typeform placed on a flatbed that moved to and fro. The rotation of the cylinder was linked to the forward movement of the bed but was disengaged when the bed moved back to go under the inking rollers.

In 1814 the first stop-cylinder press of this kind to be driven by a steam engine was put into service at the Times of London. It had two cylinders, which revolved one after the other according to the to-and-fro motion of the bed so as to double the number of copies printed; a speed of 1,100 sheets per hour was achieved.

In 1818 Koenig and Bauer designed a double press in which a sheet of paper printed on one side under one of the cylinders passed to the other cylinder, to be printed on the other side. This was called a perfecting machine. In 1824 William Church added grippers to the cylinder to pick up, hold, and then automatically release the sheet of paper.

The to-and-fro movement of the bed that was retained in these early cylinder presses constituted an element of discontinuity; to make the cycle completely continuous, not only would the platen have to be cylindrical but the typeform also. In 1844 Richard Hoe in the United States patented his type revolving press, the first rotary to be based on this principle. It consisted of a cylinder of large diameter, bearing columns of type bracketed together on its outer surface; pressure was provided by several small cylinders, each of which was fed sheets of paper by hand. This system gave speeds of more than 8,000 copies per hour; its only drawback was its fragility; faulty locking up of the forms caused the type to fall out of the cylinder.

This defect was remedied by applying stereotypy to the production; that is, forming curved plates by making an impression of the typeform on strong pasteboard, the flong, or mat, which was fixed against the inside surface of a rounded mold, which was injected with lead alloy. In France, from 1849 onward, experiments were conducted with this process; it was regularly used in London by the Times from 1856 onward and after 1858 was in general use.

But feeding the press with paper still remained outside the mechanized cycle. Mechanization of this step was accomplished by the use of a continuous roll of paper supplied on reels instead of sheets. Techniques for producing paper in a continuous roll had been known since the beginning of the century. The first roll-fed rotary press was made by William Bullock of the United States in 1865. It included a device for cutting the paper after printing and produced 12,000 complete newspapers per hour. Automatic folding devices, the first of which were designed by Bullock and Hoe, were incorporated into rotaries after 1870.

Later, numerous other types of curved stereotype plates were used on rotary presses. These included electrotype plates that are curved before being backed; rubber or plastic plates made by molding or by a photomechanical process; and metal wraparound plates made by photoengraving or electronic engraving.

Attempts to mechanize composition (mid-19th century)

Unlike the mechanization of the printing process, mechanization of the composition process was difficult to achieve in the 19th century. The invention of a compression mold in 1806 opened prospects for the mechanization of the production of type. In 1822 William Church of Boston patented a typesetting machine consisting of a keyboard on which each key released a piece of type of the corresponding letter stored in channels in a magazine. The pieces of type thus obtained had to be assembled by hand and the line justified. Church had avoided the problem of distribution and shown an intuition as to its solution by annexing to the magazine a device for constantly casting new pieces of type.

Numerous machines based on the same principle and with the further addition of a mechanism that placed the type pieces selected the right way round appeared in the course of the next 50 years. On one of these the more than 10,000 pages of the ninth edition of Encyclopædia Britannica were composed. These machines produced type at the rate of 5,000 to 12,000 pieces per hour, as opposed to about 1,500 by hand composition. But in all of them the type was simply delivered in a continuous row, which had to be divided into lines and justified.

These machines were completed by the introduction of a mechanical distributor, which was a sort of reverse compositor: pieces of type from lines that had been used passed before the operator, who pressed the corresponding key on his keyboard for the appropriate channel in the magazine to be opened up. The speed of mechanized distribution did not exceed 5,000 pieces of type per hour and was, thus, no faster than hand distribution.

Mechanization of letterpress composition faced two difficulties: first, justification, which required intelligent estimation of the size of spaces to be provided between words; and, second, the time taken during which the pieces of type were used for printing, which delay kept composition and distribution from being integrated into one cycle.