Now in order to provide a practical telegraph system permitting the interconnection of apparatus throughout the world, the first basic requisite is a standard signaling code; this code, moreover, must be of such nature as to use the simplest form of electrical signaling (such as make and break signals, or positive and negative signals), and the code should be of equal length for all characters.

The permutation code, where combinations of five plus and minus pulses will give thirty-two selective positions, was suggested as far back as 1833 by Gauss and Weber. Whitehouse, in 1854, and Barnett, in 1860, experimented further, using the permutation code to operate a recording mechanism. However, no practical means for actually printing letters and figures was found until 1882 when Jean Maurice Emile Baudot designed a multiplex system to permit the transmission of four messages in each direction over a single line circuit. Later, Baudot designed a tape printer in which the selecting and printing mechanisms comprised an ingenious arrangement of cooperating parts, including a rotating typewheel associated with a coded combiner wheel and five stationary elements selectively movable to received code positions, each element having an extending finger to be brought into contact with the periphery of the coded combiner wheel upon the completion of a selective code setting. When the code combination of the set fingers matched a code on the wheel, a print roller was released to press the recording tape against the typewheel and print the selected letter.

Although the Baudot Multiplex was used extensively in Europe, engineers and inventors in the United States had not produced a practical 5-unit permutation-code teleprinter system. They were constantly searching for a unit code system requiring a minimum number of electrical impulses to operate a telegraph printer. Various types of relay systems using distinctive signal pulses were proposed. A three-unit-code system that showed promise, and did not require synchronism, used four different electrical pulse conditions: a high voltage positive or negative pulse, and a low voltage positive or negative pulse. Records show that many inventors played with this code and that John Burry, C. L. Krum, G. A. Cardwell, and J. C. Barclay, among others, built operable equipment using such a code arrangement. The following excerpt from Cardwell’s patent No. 905,497 of December 1, 1908, may be of interest. It describes the code arrangement of four different line conditions in three signal groups to produce 36 different code combinations.

In order to energize the controller magnets in proper sequence to position the type wheel for printing a desired letter, a predetermined code or system of sending in the impulses is essential. In an instrument constructed by me in accordance with the present invention I have used the following:

1-2-1	letter space
1-2-3	carriage return
1-2-4	line space
1-3-1	type wheel shift
1-3-2	type wheel release
1-3-4	A
1-4-1	I
1-4-2	O
1-4-3	D
2-1-2	E
2-1-3	H
2-1-4	N
2-3-1	W
2-3-2	R
2-3-4	S
2-4-1	T
2-4-2	V
2-4-3	U
3-1-2	-
3-1-3	,
3-1-4	?
3-2-1	Y
3-2-3	C
3-2-4	F
3-4-1	G
3-4-2	Q
3-4-3	L
4-1-2	M
4-1-3	J
4-1-4	B
4-2-1	X
4-2-3	Z
4-2-4	.
4-3-1	K
4-3-2	P
4-3-4

The numerals in the above code or system indicate the sequence of the impulses through the relay contacts 1, 2, 3 and 4.

Cardwell’s backers formed the American Telegraph Typewriter Company and sold shares, claiming a great profitable future for their apparatus. A few printers were built, but, on extended service tests, the high-low-voltage feature proved to be impractical and the company folded.

J. C. Barclay of the Western Union Telegraph Company designed printing apparatus and perforated tape transmitting equipment for this type of system, but, after limited use of the high-low-voltage principle, he changed this feature to long and short pulses with discriminating relays. With later improvements, this system was put into service on a number of Western Union message circuits.

Note: Today, in the 1960s, the techniques of frequency division and electronics could be used for transmitting three different line conditions from a group of four different frequencies. In such a system, synchronism between send and receive terminals, or the start-stop method to control correct timing, is not needed. The transmission of any selected group of three, when received in succession at a teleprinter, will cause printing of a letter when the third receiving relay completes a circuit to the selected printing magnet. Transistor circuits with associated control elements could replace the relays.

Telegraph companies in the United States were mostly using the Morse and increased its efficiency through development of quadruplex operation and high-speed transmission systems. However, the quest for a more efficient printing telegraph system persisted and various types were proposed and tried. All made use of codes that were impractical for intercommunicating systems and, while some were used in message service by the telegraph companies, they did not appear to have any great advantage over the Morse telegraph.

In 1911, Western Union began investigating systems used in Europe. The first to be tested was the Creed, a system for transmitting and printing Continental or Morse code signals at high speed. After observing several other systems, the Murray Multiplex, an improved and modernized version of the Baudot Multiplex, was found to show better operating features and, due to the use of the five-unit permutation code, more efficient use of telegraph lines.

By agreement with Donald Murray, and with his cooperation, the so-called Western Union Multiplex was developed, using Murray’s phonic wheel drive and synchronous correction. Western Union engineers designed tape transmitters and an autocontrol device for transmitting service signals. A keyboard-operated, five-unit-code, tape punch was supplied by the Morkrum Company. A page printer, which was later converted to print on gummed tape, was supplied by the Kleinschmidt Electric Company. The Western Union Multiplex was effective in increasing operating efficiency and came into extensive use throughout the system.

From that time on, development of the Baudot permutation-code printing telegraph apparatus was furthered by various companies, such as Bell Laboratories, Western Electric, the Morkrum Company, the Kleinschmidt Electric Company, as well as the companies of J. E. Wright, L. M. Potts, and others.

It appears that all of these inventors experimented with the idea of operating all transmitting and receiving apparatus at identical speeds to transmit a five-unit-code combination by first transmitting a pulse to start both transmitter and the distant receiver at the same time. Synchronous operation was obtained by electric motors equipped with governors to maintain correct speed; some used tuning-fork-controlled impulse motors. A system of five relays with progressive contact arrangements to cause cascade operation was also used to provide correct timing. Later, when correctly-timed, 60-cycle alternating current became available (see page 27), synchronous motors did the job.

Teleprinter apparatus using this synchronous method worked very well for point-to-point transmission as was used by the telegraph companies and news-distributing organizations to carry local traffic. It did not, however, solve the sought-for plan for a teleprinter intercommunicating system. When attempts were made to connect printers at different distant points in a telegraph intercommunicating circuit while using the synchronous principle, false reception and printing errors would occur, due to variation in line circuit conditions and to a variation of the mechanical start operation at outlying teleprinters. This required frequent overline adjustment to keep the connected apparatus in phase.

The following is an excerpt from a patent application of Dr. Louis M. Potts (later to become research engineer for the Morkrum-Kleinschmidt Corporation), filed June 17, 1909, giving his idea for achieving a teleprinter intercommunicating system on the synchronous principle. Dr. Potts was a very capable telegraph engineer. His early association was with the telegraph system of the Rowland Telegraphic Company which had limited use in the early 1900s.

In those synchronous telegraph systems wherein the sending and the receiving commutator brushes constantly rotate, it is necessary to adjust these brushes so that they will approximately simultaneously engage corresponding segments. To effect this adjustment there is usually provided a special device embodying an additional segment on each commutator, and the adjustment consists in causing the brushes to arrive approximately simultaneously on this segment at the two ends of the line. Such adjustment is known as “finding the letter” and has to be made every time synchronism has been interrupted. According to the present invention, the necessity of providing additional means for performing the so-called operation of “finding the letter” is rendered unnecessary, since the transmitting and receiving commutator brushes at the two ends of the line start up afresh as it were for each signal.

In a later patent application, filed October 11, 1913, Dr. Potts stated:

This machine has also the advantage of being able to operate with a very short code. In order to adapt it to Morse circuits, it should be capable of operating with a code at least as short as the Morse code. In the present machine, I preferably employ a five unit code which, together with the starting impulse and the lag between signals, is actually shorter than the Morse code. Counting the five code impulse periods, the starting impulse period and the period of lag between impulses, each signal in my case, may be considered in comparison with the Morse code as being seven units long, whereas the Morse code is 8.5 units long. The average signal being shorter than the Morse, the delicacy of adjustment will be less and the distance of operation greater than a Morse telegraph for the same rate of transmission in words per minute.

Dr. Potts, however, still depended on like synchronous operation at both terminals, and dual start, so he did not have the answer after all!