Total Pageviews

Sunday, 14 August 2011

On This Day in Typewriter History (LXXXIV)

100 Years of
Influences on the Selectric
1856 to 1956
Whenever I give a typewriter presentation, I am invariably asked about the IBM golfball. Until now I have usually responded by handing around Blickensderfer typewheels or Hammond typeshuttles, in an effort to make the point that there was nothing really new in the golfball – it was “very old typewriter technology”.
Blickensderfer typewheel
Yet many people today do, oddly enough, identify typewriters in general with the golfball. It’s almost as if it seems, to some, that the story of the typewriter began and ended with the golfball. And in some ways, that’s actually quite true.
But not the way IBM tells it. According to that mighty conglomerate, “for the first 90 years of typewriter technology, typebars proved to be the only efficient way to make an impression on a page”. Try telling that to James Bartlett Hammond and George Canfield Blickensderfer, if you dare!
Hammond typeshuttle
IBM goes on, “In 1961, IBM introduced a revolutionary way to make a typewriter work, and work well. This was the IBM Selectric typewriter, which replaced typebars and moving carriages with a printing element, a sphere no larger than a golfball, which bears all alphabet characters, numbers and punctuation symbols.”
Wow! Revolutionary? “A sphere no larger than a golfball, which bears all alphabet characters, numbers and punctuation symbols” Um, didn’t Robert Thomas do that in 1854? And John Pratt in 1866? And James Hammond in 1881? And Lucien Stephen Crandall later that same year? And Charles Spiro (below) in 1885? And George Blickensderfer in 1893? And Charles Bennett on a pocket typewriter in 1907?
Crandall typesleeve
But wait, there’s more: “The IBM Selectric typewriter, the machine that switched typeface styles by changing its golfball … quite simply revolutionised the way in which the world typed. By using various Selectric typing elements, a user could type away in Hebrew (done many years before on manual portable typewriters) or in modern Greek (done years before), or in Japanese phonetic Katakana (done years before).”
Come on, IBM, tell us what really was so new about the Selectric! An electric typewriter with a typewheel? Blickensderfer did that in 1902, almost six decades before the Selectric.
Well, OK, in terms of the “end” of typewriters, IBM’s introduction of the Selectric a little more than half a century ago, on July 31, 1961 - almost exactly 87 years from the day the Sholes and Glidden first went on sale - probably did spell the beginning of the end of production of the manual typewriters that we all know and love.
In that regard, it's not surprising some people today see the Selectric as the be all and end all of typewriters.
Many seem to have been brainwashed by the bombardment of 1960s IBM propaganda, its heavy duty promotion and advertising for the Selectric, the so-called “Cadillac of typewriters”. Indeed, the US Postal Service featured the Selectric in a “Pioneers of American Industrial Design” stamp series.
But as for the “beginning” of typewriters, regardless of what IBM might say now, the IBM golfball’s true genesis is officially recorded as dating back to John Jonathan Pratt, inventor of the 1866 “Pterotype” (“winged type”).
Pratt (above) patented his machine in the US on yesterday’s date, August 11, in 1868, describing it as a “mechanical typographer”. Some historians regard the Pterotype as the first practical typewriter ever produced, predating the Sholes and Glidden by as many as five years. See
In its history of the creation of the golfballed Selectric, IBM passingly acknowledges the influence of the truly revolutionary Blickensderfer, the classic styling lines of Olivetti (see and its need to buy the design patent for Louis Marx’s 1940s toy typewriter.
The Marx toy typewriter in question, by the way, was designed by Ray Lohr, who was educated Xavier University in Cincinnati, where noted typewriter collector and historian Richard Polt, one of oztypewriter’s best friends, is now professor of philosophy.
IBM even overlooks the efforts of one of its own finest servants, the brilliant and uber prolific Frederick Merchant Carroll. As the Selectric was being developed by much younger men in Poughkeepsie, a little further north-west in New York, at IBM’s laboratory in Binghamton, the then 86-year-old Carroll was working on a “moveable drum typewriter”, using a much more conventional typewheel on an IBM electric typewriter. Carroll’s design (below) was patented on this day in 1958, when he was 89.
Carroll died on October 30, 1960, aged 91. The Selectric was still nine months away from the marketplace. In the excitement of that final period of a 15-year development program to produce the Selectric, Carroll’s input, the last ripening of fruit from 64 years of inventing, was sadly forgotten.
Other than its slight nod to the Blick, Olivetti and Marx, IBM keeps the background information on the Selectric “in-house”, in Poughkeepsie, mentioning the IBM design and mechanical engineers involved but ignoring the original ideas that flowed into the Selectric from such legendary great typewriter inventors as John Pratt and Charles Spiro.
IBM’s official history has the genesis of the golfball dating from 1946. IBM says: “The pedigree for the golfball element goes back to 1946 and an umbrella-shaped development model in Poughkeepsie, New York, called a ‘mushroom printer’.” But that is only partially true.
The people who actually patented the golfball itself, on July 21, 1959, referenced a design Pratt had assigned to James Bartlett Hammond on June 21, 1892 (below).
Pratt had, of course, used a typewheel in his Pterotype and sold the rights to it to Hammond in 1878. Ensuing legal action by Hammond against Lucien Stephen Crandall forced Crandall to adapt the typewheel into an elongated typesleeve, coming up with something typewriter historian Michael Adler links to Robert Thomas’s 1854 Typograph (below). See
The “inventors” of the golfball were John E. Hickerson, Ralph E. Page and James A. Weldenhammer, working for IBM in Poughkeepsie. They had applied for the patent on November 17, 1955 (below). IBM workers initially referred to it as the “flying walnut”, or “bouncing ball”.
Apart from Pratt’s design for the Hammond, this team also referenced a typewheel typewriter designed by George H. Treadgold (1859-1925), of Port Huron, Michigan, in 1898, and a typographic machine designed in 1902 by the famous Hartford, Connecticut, inventor Francis Henry Richards (born 1850) which Richards had assigned to the American Typographic Corporation of New Jersey.
That’s another part of the real heritage of the Selectric, at least as far as this particular style of single type element is concerned. The technology of how to operate it is an entirely different story. IBM largely gives the credit for this to Horace Smart "Bud" Beattie (1909-1993).
According to a 1961 advertisement for the Selectric, Beattie's Ploughkeepsie team “began their search by forgetting the past 50 years of typewriter design”. Judging by its patent references, however, that is a very long way from the truth! Or maybe it forgot the immediately preceding 50 years, then remembered the previous 50 years, going back to the 1850s?
Under Beattie’s direction, Hickerson "revised the typehead toward its ultimate spherical configuration".
Beattie (above), who was to become vice-president of engineering in the office products division of IBM, graduated from the Massachusetts Institute of Technology. He started with IBM in June 1933, working under Carroll, and retired after more than 40 years. Beattie worked subsequently as a consultant. He received the American Society of Mechanical Engineers Medal in 1971 and the Engineering Citation of the Society of Manufacturing Engineers in 1973. He was elected to the National Academy of Engineering in 1976.
Described as “the proverbial engineer's engineer [and] well ahead of his time”, Beattie was IBM’s lab director at its new Electric Typewriter Division in Lexington, Kentucky, in 1957. There he completed what IBM president Thomas John Watson (below) called "the most totally distinct invention we've ever made as a company" — the single-element printer. "This," Watson said, "is the most significant change in typing in 120 years."
The Selectric was said to have been “the basis for countless printers and led to IBM's memory typewriters and correcting typewriters. In a 1986 article on office communications, USA Today said Beattie ‘changed the office landscape forever’.”
Watson rated Beattie one of the three best design engineers and inventors ever to work for IBM. I wonder if Carroll was another?
Beattie recalled later the moment Watson first saw the golfball operate: “We were in the old Kenyon House in Poughkeepsie. [Watson] came in and asked me what I was doing. I said, ‘I’ve got something I’d like to show you.’ We went down into the bowels of that old mansion, to the little room where John Hickerson was tinkering with the mushroom [typehead] … John turned it on, and it started to bob and wobble around. Mr Watson sat there mesmerised. Then he broke out laughing and said, ‘Bud, you must have been drunk when you designed that thing!’”
What Watson saw in 1954 was the operation of a development of the “mushroom printer” (above) to which the IBM history refers - the “umbrella-shaped development model in Poughkeepsie”. IBM says, “The intent had been to use it as a printer on an IBM accounting machine. Earlier versions did see some action with the IBM Stretch computer [before being applied to the Selectric typewriter].”
Here’s where things get interesting. Beattie had applied for a patent for this device, calling it a “high speed printing mechanism”, on February 17, 1948, and the patent was issued on December 8, 1953.
In filing his application, Beattie referred to many earlier inventions, starting with some astonishing typewriter designs from a Charles Hire Perry (born Vermont, June 25, 1844), a doctor in Oneida, New York. The designs (above) date from 1887, and Perry had assigned them to Thomas Redfield Proctor (1844-1920, below), a wealthy Utica, New York, banker and businessman who backed Perry’s enterprises.
Compare the Beattie and Perry configurations:

Beattie (left)                                Perry (right)
Other earlier typewriter designs to which Beattie referred were by the noted Freeport, Illinois, inventor, manufacturer and banker Daniel Carroll Stover (1839-1908, below) and London-based Scottish politician James Dundas White (1866-1951), as well as another quite remarkable machine, this one from Lithuanian-born Saul Aronson of New York.
Stover’s designs date from 1888 (IBM failed to cite Stover's original "mushroom" patent, below), White’s date from 1902 and Aronson’s from 1907.
Stover 1888
Stover's second patent, 1891
Put aside the claims that Beattie got the idea of the mushroom-shaped type element from changing a light bulb at home. The Perry, Stover and Aronson machines all used convex type elements, giving Beattie the basis for his “mushroom” printer. In fact, they are all almost identical to Beattie’s design.
Adler in Antique Typewriters (1997) called the Perry design (above) “interesting” and said the idea was “an index of lenticular section comparable to the Lambert but manipulated by means of vertical joy-stick”. Adler said of White (who he mistakenly called “J.G.Whyte”) “A machine which typed around a vertical platen riding up on a spiral groove”.
Before Hickerson was issued with a patent for a “key-operated machine” [the Selectric, sans an inking component] on April 4, 1961 (above), Hickerson and Page had been part of a team headed by Leon E. Palmer which patented a “single element printing machine” on March 31, 1959 (below).
This shows clear evidence of the way in which the golfball “wobbled” on its stem. Because the Selectric’s platen remained stationary, the type element’s face needed to be curved so that all three lines of type could strike the platen squarely. A straight-edged typewheel could not have achieved this without the platen moving up and down to accommodate each line.
It is in these areas that the Lohr-Marx patent (above) came into play. It was very much still in force at the time the Palmer team’s patent was applied for on November 17, 1955. At that time, Palmer and Co referenced an unusual 1931 typewheel typewriter designed by Gustave Teissedre of Strasbourg-Alsace in France and first patented in Germany (below). The typewheel was rotated by a curved arm.
Palmer and Co also referenced a typewriter (below) designed by Charles A. Fuchs for the Burnell Laboratory Company of Locust Valley, New York, in 1935. This used as typewheel on a pivoted arm.
The Palmer team’s patent has a rounded type element inked by adjoining ribbon spools. This method was used, for example, on the Teletype Model 26 (see below).
Palmer developed the character-selection system for the Selectric. IBM says, "This system provided for fixed units of motion controlled by mathematical ratios. Palmer was made an IBM Fellow — the highest honor IBM bestows on a scientist, engineer or programmer, and an elite group handpicked by the CEO — for his contributions to the Selectric. He holds the majority of the patents that resulted from the development of the original machine."

Meanwhile, “back at the ranch”, as they say in the classics: While Beattie and his team in Poughkeepsie were working on their “mushroom printer”, in Binghamton, one of IBM’s oldest and most gifted hands was coming up with one last, magnificent design.
For Frederick Merchant Carroll (above), this “movable drum typewriter” would be a parting gesture for the company he had served so well, and so profitably, for 41 years.
Carroll’s patent, for which he had made an application on January 5, 1955, was aimed in the same direction as Beattie’s single type element printing, except it was for a typewriter – and he was using a three-line typewheel, identical to those used on Blicks and Bennetts.
Carroll wrote in his application that his typewriter would have “type characters .. arranged on a drum or other printing member which is moved under control of selector elements, such as type keys …”
Bear in the mind that at this time, the Poughkeepsie team was headed toward an accounting machine or the massive Stretch computer, not a typebar typewriter. Also, Carroll was using a typewheel (or typeball) that rotated and pivoted.
For his references in his design, Carroll pointed way back, to Charles Spiro’s 1885 typewheel typewriter (below).
During a career spanning 61 years, Carroll took out 97 patents. These included an IBM typewriter in 1943. He was still being issued with patents a year after his death.
Carroll was born in Union City, Ohio, in 1869, the son of an inventor (David Carroll), and he started taking out his own patents in 1896. His first was for a cyclometer.
At the turn of the century, Carroll moved to New Haven, Connecticut, were he worked on adding machines and calculators, then to Dayton, Ohio, to work for the National Cash Register Company. In 1916 he joined the Tabulating Machine Company, one of the organisations which later emerged into IBM, at the Product Development Laboratory in Endicott, New York. He subsequently worked in Yorkers then Binghamton. During his time at IBM, Carroll developed rotary type presses, an automatic high-speed rotary card manufacturing machine, the automatic ledger posting “compensating” carriage, a "unit counter", a small electromechanical device that acted as a single adding wheel, and other accounting machines. He introduced a photoelectric sensing apparatus. It is said that he “helped to revolutionise the entire business machines industry”.
Among the people he had worked closely with at IBM was Eugene Amzi Ford, inventor of the Ford typewriter (above). See
Fred M. Carroll, left, Eugene A. Ford, right
Finally, we come to the Marx toy typewriter (below), designed in 1939 by Raymond John Lohr of Erie, Pennsylvania (ironically, George Canfield Blickensderfer’s home town).
Robert William Bemer (above), a computer scientist at IBM, recalled, “One day during initial Selectric production at Lexington, Kentucky, the honoured touring guest was Mr [Louis] Marx, of Marx Toys [see the cover of Time, below], grinning ear to ear. The reason? IBM had purchased the rights to his toy typewriter for children as a large basis for their [Selectric] design.”
Lohr’s design is for an enlarged version of the well-known Junior toy typewriter, the patent for which had been held in Germany. But bear this in mind: Marx did not own the patent for the more familiar “stemmed” Marx toy typewriter, which it produced under licence from Samuel I. Berger of Newark (apparently Marx and Berger “swapped” a lot of designs in the 1930s). The patent on the Berger typewriter had expired in 1949, anyway, while the Lohr patent was still very much in force in 1957, at a time when development of the Selectric was in full flow.
Lohr, who also designed the Big Wheel and Rockem-Sockem Robots for Marx, was born in Cincinnati in 1912. At 23, he was hired as a mechanical designer by Louis Marx and Company Toys in Erie, retiring after a 37-year career in 1972 as chief toy designer in 1972. He died in Erie on October 3, 2005, aged 93.
The Selectric’s body was the work of influential American designer Eliot Fette Noyes (1910-1977), of New Canaan, Connecticut (below).
Boston-born Noyes created a number of design projects for IBM. He studied architecture at Harvard (1932-1938) and worked as a curator at the Museum of Modern Art and as director of industrial design. After World War II, Noyes was design director with Norman Bel Geddes, for whom Royal typewriter designer Henry Dreyfuss had served his apprenticeship in the early 1920s. When Bel Geddes' office closed in 1947, one of the unfinished projects was a Model A electric typewriter for IBM. Noyes opened his own office and finished the job for IBM. He was retained in 1948 by Thomas Watson for IBM product design and in 1956 developed a unique IBM corporate style and image similar to Olivetti. He did so with help from Paul Rand, Marcel Breuer and Charles Eames.
In his design for the Selectric, Noyes referenced earlier IBM machines and the work of Frederic S. Grover, Theodore G. Clement and Carl W. Sundberg. Notably, however, he also referred to Laird Fortune Covey’s FP design for Royal. See
Selectrics and their descendants eventually captured 75 per cent of the United States market for electric typewriters used in business. It is estimated 13 million were sold worldwide up to 2002 (though production officially ended in 1987). The Selectric was succeeded by the Wheelwriter in 1984 and IBM transferred its typewriter business to the newly formed Lexmark in 1991.
At the end of the day, the IBM Selectric was first and foremost a team effort, involving not just Beattie but Hickerson, Palmer and many other IBM design, mechanical and electronic engineers. But there was very little, if anything, in its make-up that was entirely unique. Its genes go back to Robert Thomas, John Pratt and James Hammond – in other words, more than 100 years before the Selectric was actually made.


Richard P said...

One of your most impressive research articles so far! It far exceeds anything else I've seen about the Selectric.

It's quite interesting to see the early patents for "wobbling" type elements that tilt to use different rows of characters. The only such machine that I was familiar with is the Lambert.

I think from now on I'll call the IBM element a "flying walnut."

Robert Messenger said...

Thank you, Richard, this is very kind (and of course encouraging). As you may imagine, a lot of elements, so a lot of time and effort, putting me way behind schedule again. But your comments help make it worthwhile.
I have added in an 1888 Stover 'mushroom' design which the IBM people failed to cite, for some reason. By sheer luck I stumbled across it while researching August 13.
I was a bit surprised they didn't reference the Lambert. But it is fascinating to see such unfamiliar designs as those by Perry, Aronson etc, and uncover characters like Ray Lohr of Cincinnati.
Like you, when I saw it I thought "flying walnut" was a far better name than golfball.

notagain said...

Fabulous research. I get a lot out of these posts. You should bind them together in book (or calendar) form after a year.
I find it interesting that they continued to produce typebar machines competing with the selectric, like the one I found yesterday.

Robert Messenger said...

Thank you, notagain, you are very kind and I'm naturally delighted and encouraged to know that you get something from these posts. I have been thinking along the lines of a biographical dictionary on typewriter inventors for 2014, to mark 300 years since Henry Mill. Might have to start saving up something apart from words for that, though.
Yes, it is interesting about the manual IBM. Apparently Bemer wanted to expand the golfball, with ASCI in mind, but IBM told him the Selectric would only ever be a "correspondence" machine! IBM was actually taken by surprise by its immediate sales success, though it doesn't like to admit that now!

Eric Rucker said...

There's a few reasons why IBM continued selling typebar machines after the debut of the Selectric.

1. Price - a Model C with a 13" carriage cost $460 in 1962, a Selectric 723 (13" platen, 11" writing line) cost $475, and a 725 (15" platen, 13" writing line) cost $495. Fast forward to the Model D (with 12.5" carriage)'s debut in 1967, at $490, and the 723 was $500, 725 was $520. And, in 1969, it was $510 for the D, $520 for the 723, and $540 for the 725.

2. Speed - While the Selectric is faster for ordinary typists that aren't trained in how to run a typebar typewriter properly (myself included), a fast typist can get a burst of speed that out-types a Selectric. 15.5 characters per second is the most a Selectric can do - that means that if it takes you less than 64.5 ms for a character (or, the other way, if you burst at over 186 wpm), you're out-typing the machine. (If you do a 2-character burst and then slow after that, it's no problem, the machine can absorb that in the interposer system, but a *3*-character burst - that is, selecting the third character less than 129 ms after selecting the first character - and now you're slowed down by the machine.) Experienced typebar typists could keep the cadence such that multiple typebars were in flight but not colliding, and world records were set on IBM typebar machines (but never the Selectric AFAIK). IBM didn't want to lose that market.

3. Executive - while the Selectric was IBM's "flagship", it was also a mid-range machine (although the cheapest machine they sold, from the day the Selectric debuted, was the Selectric 721/711). The bottom of the line was shared by the 11" platen Selectrics and the typebar machines, and the top of the line was exclusively the domain of the proportional spacing Executive typebar machines. The typebar machines didn't go away until IBM came up with a proportional spacing Selectric that was suitable for office use (the Mag Card Executive).

Eric Rucker said...

And, of course, I post that before checking...

The CURRENT world record typing speed (over 1 minute in this case) is 216 wpm.

Set by Stella Pajunas, in 1946... on an IBM electric. That thing was being driven harder than a Selectric's theoretical max.

quadibloc said...

I have used a VariTyper, and so I know that typing on it is slow compared to typing on a conventional manual typewriter, and the keys require considerably more force to press. So, although I have no direct experience with the Blickensderfer, I would dare to tell Mr. Hammond, at least, that the Selectric did bring something new to the table, even if it certainly was not the very idea of a single element typewriter.