Strowger and the Automatic Exchange — Telephone Switching History

Figure 1 — A Strowger Automatic dial candlestick telephone of 1905, Telephone Museum, Waltham, Massachusetts. The dial in the column is the finger-wheel calling device patented by Keith and the Erickson broth… — Figure 1 — A Strowger Automatic dial candlestick telephone of 1905, Telephone Museum, Waltham, Massachusetts. The dial in the column is the finger-wheel calling device patented by Keith and the Erickson brothers in 1898. Each turn of it sent a train of pulses that stepped the switches at the exchange directly, with no operator anywhere in the call. — Daderot, via Wikimedia Commons (CC0)

2.1 An Undertaker in Kansas City

Almon Brown Strowger was born at Penfield, New York, in 1839. The sources disagree on the day, giving either 11 February or 19 October. He taught school, served in the Civil War with the 8th New York Cavalry (enlisting as a musician, wounded at Third Winchester in 1864, ending as a second lieutenant), and by the late 1880s was an undertaker in Kansas City, Missouri. His patent describes him as “of Kansas City, Missouri” (US 447,918).

The story goes that a rival undertaker’s wife worked as an operator at the local exchange, and that calls asking for an undertaker were being put through to her husband instead of to Strowger. He resolved to build an exchange that needed no operator at all. It is a good story and it may be true, but it is worth knowing how thin its support is. The earliest printed version that can be traced is a newspaper piece, “Originator Sought Revenge,” in the Sioux City Journal of 19 March 1913. That is more than twenty years after the invention and eleven years after Strowger’s death. R. B. Hill’s careful 1953 history of the Strowger system for the Bell Laboratories Record calls Strowger an undertaker but does not tell the story at all (Hill, “The Early Years of the Strowger System,” 1953). The IEEE history calls it “a commonly told story” (ETHW). No source names the rival or the operator. It is best treated as tradition.

What Hill does record is how the idea first took physical form: “Strowger is said to have constructed” a crude model from a cardboard collar box and a paper of pins, with ten rows of ten pins pushed through the box wall and a lead pencil for the shaft. A contact arm on the pencil could be moved along the shaft to a row and turned to a pin. That is the whole principle of the switch that followed: one motion to choose a row, a second, at right angles, to choose a contact in it.

2.2 The Patent: US 447,918

Strowger filed his application on 12 March 1889, and it issued on 10 March 1891 as US Patent 447,918, “Automatic Telephone-Exchange” (Google Patents). The Bell Laboratories history misprints the issue date as 10 May 1891, and at least one modern document mis-cites the number as 447,919. The patent itself is the authority on both.

It is sometimes said that Strowger’s patent did not describe a two-motion switch. It did, though not the switch that became standard:

Each subscriber’s line had its own switch at the exchange. The contacts were the ends of the line wires, held in curved rows inside a hollow insulating cylinder.
A central rod carrying a contact needle moved lengthwise from row to row and rotated along a row. Pawls and ratchets driven by electromagnets did the stepping. Separate magnets released it, and gravity and a spring returned it.
The subscriber controlled it with push keys, not a dial: one key for hundreds, one for tens, one for units, and one to release. To reach number 315, the caller pressed the hundreds key three times, the tens key once and the units key five times. That took four control wires plus a talking wire from every telephone (“five wires and ground”), and a heavy battery at every subscriber’s premises.
There was no busy test: nothing stopped a caller from connecting to a line already in use.

Hill states plainly that “this particular arrangement was never used commercially.” What survived was the idea that a subscriber’s own signals could move a switch directly, row first and then contact. Kempster Miller’s 1905 textbook credits Strowger’s key contribution in exactly those terms: moving the selecting arm “in one direction to select a given series of contacts … and then in a given direction at right angles to the first” (Miller, American Telephone Practice, 1905, ch. XXXV).

Strowger was not the first to patent an automatic exchange. Connolly and McTighe’s US 222,458 of December 1879 has that distinction, and their eight-station system was shown at the Paris Exposition of 1881. Twenty-six automatic-exchange patents issued between theirs and Strowger’s (Hill, “Early Work on Dial Telephone Systems,” 1953). The Bell System itself had installed E. T. Gilliland’s “Village System” at Leicester, Massachusetts, in September 1885. It switched at the telephones, with no central office. Strowger’s distinction is that his line of development led to something that worked commercially, at scale, for a century.

Figure 2 — Strowger's US Patent 447,918, drawing sheet 1. The subscriber's station with its push keys (left) is wired to an individual switch at the exchange. The patent's switch moved a needle lengthwise and… — Figure 2 — Strowger's US Patent 447,918, drawing sheet 1. The subscriber's station with its push keys (left) is wired to an individual switch at the exchange. The patent's switch moved a needle lengthwise and around inside a cylinder lined with rows of contacts, the two motions of the later switch in a different geometry. — A. B. Strowger, US Patent 447,918 (1891), via Google Patents

Figure 3 — An early electromagnetic selector of Strowger's type, on museum display with a German-language label. Commons does not name the museum, and it is probably a model or reconstruction rather than an o… — Figure 3 — An early electromagnetic selector of Strowger's type, on museum display with a German-language label. Commons does not name the museum, and it is probably a model or reconstruction rather than an original. It shows the essential parts in their most naked form: magnets, ratchet, and a contact arm. — Reinraum, via Wikimedia Commons (CC0)

2.3 The Company and Its People

Strowger was the inventor. The machine was largely made practical by others. The Strowger Automatic Telephone Exchange was incorporated on 30 October 1891, with Strowger, the Chicago financier M. A. Meyer and Joseph Harris among its founders (Hill 1953; MacMeal, The Story of Independent Telephony, 1934, excerpt). Harris is credited with seeing the switch’s potential and persuading Strowger to move to Chicago. The engineers who mattered most arrived soon after:

Alexander E. Keith joined early in 1892 from the Brush Electric Company of Baltimore, and was “for many years thereafter one of its leading technicians and inventors” (Hill).
John Erickson, his brother Charles J. Erickson, and Frank A. Lundquist, all from Lindsborg, Kansas, joined around the turn of 1893–94. Lundquist had worked for the Chicago Telephone Company and had built a 100-line automatic model with the Ericksons in 1892. John Erickson eventually held 115 patents and Charles 35 (Lindquist, Kansas Historical Quarterly, 1957).

Strowger himself left the company in the fall of 1896 “for his health,” moved to Florida, “took no part thereafter,” and died in St. Petersburg in May 1902 (Hill). He is reported to have sold his patents in 1896 for $1,800 and his remaining share for $10,000 in 1898, but those figures come from secondary sources only.

2.4 La Porte, Indiana: 3 November 1892

“On November 3, 1892, the first Strowger exchange was opened for public service at La Porte, Ind., with about seventy-five subscribers. This was the first automatic telephone exchange to be installed anywhere, and a considerable amount of ceremony was attached to the affair, with a special train run from Chicago and a brass band” (Hill 1953). With Gilliland’s station-switched Village System in mind, it is more exactly the first automatic central-office exchange. Smith and Campbell’s 1921 textbook gives the year as 1893, but Hill, working from company records, Lindquist, and a contemporary newspaper report of 5 November 1892 (cited by Wikipedia, not seen for this dive) all agree on 1892.

The La Porte switch was not yet the two-motion switch. Because the office served fewer than 100 lines, Hill records, “a flat rubber disc type of switch was employed, with but one (rotary) movement for the wiper arm and one circular row of contacts.” The subscribers still had five line wires and four push buttons (the hundreds button went unused). The switch battery had moved to the central office, but there was still no protection against calling a busy line.

2.5 From Push Buttons to the Dial

The next six years produced almost every essential of the step-by-step system:

1893. Keith’s first automatic release (US 573,884): a mercury switch on the telephone’s hook restored the exchange switch when the caller hung up. A small Strowger equipment was exhibited at the Chicago World’s Fair.
Fall 1894. “Piano-wire” terminal banks with a rotary and lengthwise switch arm (US 540,168) replaced the original switches at La Porte and went into Michigan City, Indiana, with the first simple busy signal: the calling bell stayed silent if the called line was busy.
1895. “A return was made to the principle of Strowger’s original patent.” Keith and the Ericksons’ switch (US 638,249) used semicircular banks and the “familiar ‘up-and-around’ motion of the selector rod and wipers,” the first switch to resemble the modern step-by-step switch. Hundred-line switches of this type replaced the piano-wire ones at La Porte in June 1895, and 200-line switches went into Michigan City that August (US 638,249).
Summer 1896. The finger-wheel dial. Numbered finger holes on a rotating plate wound a spring as the plate was turned, and an escapement governed its return, sending the pulses on the way back. Keith and the Erickson brothers filed on 20 August 1896, and the patent issued on 11 January 1898 as US 597,062, “Calling Device for Telephone Exchanges” (Google Patents). The dial replaced the push buttons with a single calling device. Its pulses at first grounded one line wire or the other (Hill); only with the straight two-wire system of 1908 did the dial simply open the loop once per pulse, the signal every later switch in this dive obeys.

Figure 4 — An Automatic Electric Strowger wall telephone with an early dial, about 1905. Early dials had finger holes but no number ring; later models put the digits under the finger wheel. Telephone Museum, … — Figure 4 — An Automatic Electric Strowger wall telephone with an early dial, about 1905. Early dials had finger holes but no number ring; later models put the digits under the finger wheel. Telephone Museum, Waltham. — Daderot, via Wikimedia Commons (CC0)

Figure 5 — The Automatic Electric calling device (dial) mechanism, from Smith and Campbell's Automatic Telephony (1921). The spring wound by the finger wheel, the governor that sets the return speed, and the … — Figure 5 — The Automatic Electric calling device (dial) mechanism, from Smith and Campbell's Automatic Telephony (1921). The spring wound by the finger wheel, the governor that sets the return speed, and the cam that opens the pulse springs once per digit step are the parts Volume 5 relies on. — A. B. Smith and W. L. Campbell, Automatic Telephony (1921), via Internet Archive

2.6 Trunking: How the System Grew Past 100 Lines

A single up-and-around switch reaches 100 lines: ten levels of ten contacts. To serve more, the system had to be built in stages. An early switch chooses a group of lines, and a later switch chooses the line within that group. Keith and the Ericksons began a 1,000-line trunking system in 1896 (US 672,942). It went into Augusta, Georgia, in March 1897, about 400 lines, and was operable but not commercially practical, because it had only one trunk per selector group. The missing idea was automatic hunting: a switch, sent to a level, that steps along it by itself to find a free trunk. Work on it began in the summer of 1897. Lundquist was later awarded US 776,524 (1904) for the automatic selection of an idle trunk.

The first exchange with “a practicable form of trunking system” opened at New Bedford, Massachusetts, in November 1900, about 3,600 lines with first and second selectors (Hill 1953). From there the essentials of the mature system fell into place over a decade:

Table 1 — The first exchange with "a practicable form of trunking system" opened at New Bedford, Massachusetts, in November 1900, about 3,600 lines with first and second selectors (Hill 1953). From there the essentials of the mature system fell into place over a decade

Date	Place	What was new	Source
Nov 1900	New Bedford, Mass.	practical trunking; first and second selectors; ~3,600 lines	Hill 1953
1901	Chicago	Automatic Electric Company organised to make and sell Strowger equipment	Hill 1953; BTL Early Years p. 545
Feb 1903	Chicago “Loop”	~1,500 subscribers; first Strowger office with message-rate meters	Hill 1953
Jan 1904	Grand Rapids, Mich.	5,300 subscribers, replacing a manual board	Hill 1953
May 1905	South Bend, Ind.	first common-battery Strowger office (~700 subscribers)	Hill 1953
1905–07	—	Keith’s line switch (filed 29 April 1905), in general use by about 1907; it “eliminated about 90 per cent of the first selector switches”	Hill 1953; US 1,304,324
early 1908	Pontiac, Ill.	first straight two-wire Strowger system, using a slow-release relay to detect the pause between digits	Hill 1953

Hill calls Pontiac “the beginning of the modern phase of the Strowger (or ‘step-by-step’) system.” It is also the beginning of the circuit that Volume 5 dissects. The slow-release relay that waits out the gap between digits is the selector’s C relay, and every later Strowger selector descends from it.

A note on dates: Automatic Electric is sometimes said to have been founded in 1905. Every primary source consulted, including Hill, the Bell Laboratories history, MacMeal and Lindquist, gives 1901. The Strowger company carried on as a patent-holding firm until 1908, when Automatic Electric acquired its stock.

Figure 6 — A section of the automatic switchboard at Grand Rapids, Michigan, from Miller (1905). The office opened in January 1904 with 5,300 subscribers, replacing a manual board, and shows the Strowger syst… — Figure 6 — A section of the automatic switchboard at Grand Rapids, Michigan, from Miller (1905). The office opened in January 1904 with 5,300 subscribers, replacing a manual board, and shows the Strowger system's early ranks of switches, each a vertical shaft over a semicircular bank. — Kempster B. Miller, American Telephone Practice, 4th ed. (1905), via Internet Archive

Figure 7 — A Keith plunger line switch with its bank unwired, from Smith and Campbell (1921). Every subscriber line had one. When the subscriber lifted the receiver, the plunger was thrust into the bank to se… — Figure 7 — A Keith plunger line switch with its bank unwired, from Smith and Campbell (1921). Every subscriber line had one. When the subscriber lifted the receiver, the plunger was thrust into the bank to seize an idle trunk to a first selector. Shared selectors replaced one-per-line selectors, and a 1,000-subscriber office needed only around a hundred of them. — A. B. Smith and W. L. Campbell, Automatic Telephony (1921), via Internet Archive

2.7 “Strowger” and “Step-by-Step”

The two names refer to one system. Automatic Electric and its licensees called it Strowger, after the founder; the Bell System called it step-by-step, abbreviated SxS. Hill’s 1953 history writes “the Strowger (or ‘step-by-step’) system.” Whatever the badge, the defining principle is direct progressive control: each switch in the train is stepped by the caller’s own dial pulses as they arrive, and each one, when it has done its part, hands the line on to the next. No central intelligence plans the route. The dialled number is the route.

2.8 The Bell System Comes Round

For two decades the Bell System watched automatic switching from a distance. It had followed dial patents “from the very beginning” but saw “how little would be gained from dial switching under the existing conditions” (Hill, “Early Work”). At the end of 1904 Hammond V. Hayes reported that the future of automatic exchanges in the Bell System, “particularly in sizes over 100 lines, was uncertain” (BTL Early Years, p. 547). Meanwhile the independent telephone companies, competing with Bell town by town, were Automatic Electric’s best customers. By 1904 about 4 per cent of independent lines were automatic.

The war years changed the arithmetic. Operators’ wages rose with traffic. In May 1916 AT&T and Automatic Electric signed a patent agreement giving Western Electric the right to manufacture Strowger-type equipment, and in 1917 an order was placed with Automatic Electric for a Bell office at Norfolk, Virginia (BTL Early Years). In 1919 the Bell System “decided … to adopt the step-by-step dial system for single and small multioffice exchange areas,” reserving panel for the largest cities (BTL Switching Technology, ch. 3). Norfolk’s first Bell System dial telephones went into service on 8 November 1919, according to AT&T’s own centenary page. A secondary source dates a $2.5 million licence to 1919; the May 1916 agreement and the 1919 decision appear to be separate events, and no primary contract was seen.

Bell had already acquired about 100,000 lines of Automatic Electric equipment by buying independent companies, and under the 1919 arrangement about 2.3 million more lines of AE-made equipment, built to Bell Laboratories specifications, went into Bell service. Western Electric did not begin making step-by-step equipment itself until 1926. Its first job, Springfield, Massachusetts, cut over in January 1927 (BTL Switching Technology, ch. 3). Step-by-step then became the Bell System’s workhorse outside the big cities. It peaked at 24,440,000 Bell System lines in 1973: “the step-by-step system provided automatic service for more Bell System lines than any other switching system.”

Figure 8 — An Illinois Tunnel Company advertisement for its Automatic telephone service in Chicago, promising "unmeasured, unlimited, and secret service". Secrecy (no operator listening) was a selling point f… — Figure 8 — An Illinois Tunnel Company advertisement for its Automatic telephone service in Chicago, promising "unmeasured, unlimited, and secret service". Secrecy (no operator listening) was a selling point from the start. Undated; probably 1905–1915. — Illinois Tunnel Company, via Wikimedia Commons (public domain)

2.9 Strowger Abroad

Automatic Electric licensed the system widely. Harris signed with Siemens & Halske in Germany in 1909 and with Thomson-Houston in France in 1911. In Britain the Automatic Telephone Manufacturing Company of Liverpool was founded in 1911 to make Strowger equipment under licence (Science Museum Group). The Post Office, having taken over the National Telephone Company in January 1912, opened Britain’s first public automatic exchange at Epsom, Surrey, in May 1912, a Strowger office. After trials of rival systems it standardised on Strowger in 1922.

London posed the same problem as New York: too many offices for a dialled number to be the route. The British answer kept step-by-step switches but added a director, a register-translator that stored the dialled office code and sent the switches whatever digits the routing actually required. Holborn, the first director exchange, cut over at midnight on 12 November 1927 (Wikipedia, Director telephone system, citing The Times). Britain’s last Strowger exchanges were withdrawn in 1995: Crawford, Crawfordjohn and Elvanfoot in Scotland on 23 June, and the uniselector exchange on Foula, Shetland, the last electromechanical system of all, on 12 July (BT Archives).

Figure 9 — An engineer checks one of the 176 directors at a London automatic exchange in 1945. The director stored the dialled exchange code and translated it into the routing digits the step-by-step switches… — Figure 9 — An engineer checks one of the 176 directors at a London automatic exchange in 1945. The director stored the dialled exchange code and translated it into the routing digits the step-by-step switches needed, which let Strowger equipment work in a city too large for the dialled number to be the route. — Ministry of Information Photo Division, via Wikimedia Commons (public domain)

2.10 Automatic Electric After Strowger

Automatic Electric grew into the largest independent maker of telephone switching equipment. Its plant at Morgan and Van Buren Streets on Chicago’s West Side employed about 850 people by 1910. Theodore Gary’s interests took control in 1919. In 1955 General Telephone acquired Theodore Gary & Co., and with it Automatic Electric, and in 1957 the company moved to a 35-acre site in Northlake, Illinois. The parent became General Telephone & Electronics in 1959 and adopted the short “GTE” identity in 1971. Automatic Electric’s own technical journal changed its title to GTE Automatic Electric Technical Journal with the January 1971 issue. In 1983 Automatic Electric was merged into GTE Network Systems, and in 1989 it became part of AG Communication Systems, an AT&T–GTE joint venture (Wikipedia, Automatic Electric; catalogue records of the journal at Texas A&M and the ITU).

That chronology dates my own Type 26A Switch Tester, whose red nameplate reads “GTE Automatic Electric, Northlake, Ill., U.S.A.” Northlake places it after 1957, and the “GTE Automatic Electric” style after 1971; the name disappeared when Automatic Electric was merged out of existence in 1983. So the plate points to roughly 1971–1983. That is an inference from the company’s naming, not a documented rule for nameplates. The tester has its own dive.

2.11 In the Reference Library

Sources

R. B. Hill, “The Early Years of the Strowger System,” Bell Laboratories Record, March 1953, pp. 95–103. https://www.historyofphonephreaking.org/docs/blr-1953-03-early-years-of-strowger.pdf
R. B. Hill, “Early Work on Dial Telephone Systems,” Bell Laboratories Record, January 1953, pp. 22–27. https://www.historyofphonephreaking.org/docs/blr-1953-01-early-work-on-dial.pdf
A History of Engineering and Science in the Bell System: The Early Years (1875–1925) (1975) and Switching Technology (1925–1975) (1982). https://archive.org/details/bell-system-science-history
Kempster B. Miller, American Telephone Practice, 4th ed., 1905. https://archive.org/details/americantelephon00mill
A. B. Smith and W. L. Campbell, Automatic Telephony, 2nd ed., McGraw-Hill, 1921. https://archive.org/details/automatictelepho00smit
Emory Lindquist, “The Invention and Development of the Dial Telephone: The Contribution of Three Lindsborg Inventors,” Kansas Historical Quarterly 23:1, 1957. https://www.kansashistory.gov/p/the-invention-and-development-of-the-dial-telephone/13134
H. B. MacMeal, The Story of Independent Telephony, 1934, pp. 271–276 (excerpt). https://www.britishtelephones.com/histstro.htm
BT Archives, “Digital exchanges” information sheet.
Patents (Google Patents): US 222,458 (Connolly & McTighe, 1879); US 447,918 (Strowger, 1891); US 540,168 (1895); US 573,884 (Keith); US 597,062 (Keith & Ericksons, the dial, 1898); US 638,249 (Keith & Ericksons, up-and-around switch, 1899); US 672,942 (trunking, 1901); US 776,524 (Lundquist, idle-trunk selection, 1904); US 1,304,324 (Keith line switch, filed 1905).