| || || |
| || |
| || || |
From the book, Technology and Social Change in America.
At a time when the superb drama of exploiting a new continent filled the minds of most Americans, nineteenth-century technical arguments about the shape and number of threads on a screw often have a remote, somewhat comic quality—a bit like Jonathan Swift’s mock epic struggle between the Little Enders and the Big Enders.
But the issue was not a matter of satirical trivia. Industrial development on a national scale demanded that nuts and bolts of the same diameter be interchangeable. Interchangeability, in turn, required that manufacturers conform to a standard system which fixed the contour of screw threads and established for each diameter the number of threads per inch.
For America, in 1860, no such standard existed. Since that era also marked the emergence of this country into the arena of international industrial competition, the search for a standard cast reflections which illuminate such related considerations as a national style of engineering, American industrial practice, and the role of government in technological change.
The most prevalent system—where a system was used at all—was that which had first been proposed in 1841 by England’s Sir Joseph Whitworth. Whitworth’s standard was a synthesis of the best English practice and answered the general case well enough so that it was widely employed.
American usage was disuniform, however, varying from manufacturer to manufacturer and from locality to locality. Some firms developed systems to fit the particular requirements of their own processes.
Others purposely used special threads to prevent outside repairs on their own machinery, in the same vein as the Erie Railroad’s ill-fated use of wide-gauge track.
By the 1860s, the appalling lack of national uniformity clearly called for reform. “If there is any one thing in the transactions of the machine shop more incomprehensible than another,” the editor of ’Scientific American’ claimed in 1863, “it is the want of some settled size or number for screw threads.”
To eliminate the anarchy, the magazine called for some agreement among the country’s principal manufacturers or, failing that, governmental action. Whatever the standard, Whitworth or any other, the adoption of some uniform national system of screw threads was of vital necessity.
But the development of a standard in the first instance posed certain difficulties. As Whitworth had pointed out, a system depended on compromise, not on theory or experimentation. There was simply no way in which all the factors involved in screw-thread design could be stated as a precise rule, applicable to all cases.
Where principle provided no final answer, practice determined the issue. That opened the door to non-technical factors.
A further difficulty in America was that no single agency seemed capable of dealing with the problem.
Editorial injunctions notwithstanding, the federal government had no ready means either to develop a system or to regulate its usage, not even to speak of an inclination to do so. Nor did any professional engineering society exist to propose a standard and advance its adoption.
It was in this apparent vacuum that William Sellers presented a paper before a meeting of the Franklin Institute at Philadelphia in April 1864 outlining a uniform system for American screw threads.
In his paper, Sellers addressed himself immediately to the central question: Why should there be yet another system of threads? Why should not Americans adopt the Whitworth standard and by its consistent usage rationalize current practice?
The English system was the result of several years of study by an outstanding mechanician, who had carefully analyzed the three main factors of screw-thread design—pitch, or the number of threads per inch, thread depth, and thread form.
| || || |
| || |
| || || |
To arrive at his system of pitches, Whitworth had collected bolt samples from all of the principal manufacturers in England, averaged their characteristics, and developed a standard table for pitch and diameter.
The form of his thread, which comprised flat sides at an angle of fifty-five degrees, with rounded tops and bottoms, was also arrived at by the same averaging technique. It was a compromise, but on the basis of the best English screw-thread practice.
Any alternative scheme, Sellers suggested, should clearly “demonstrate its practicability and its superiority.”
The central difference between Whitworth’s system and the one proposed by Sellers was in the form of thread. Sellers leveled three objections to the English form.
First, the fifty-five degree angle was difficult to gauge with consistent accuracy.
Second, in ordinary practice the rounded tops and bottoms of English threads did not fit the corresponding bottoms and tops of nuts, and the thread’s wearing surface was therefore reduced only to its sides. The point was not that a fit was impossible, but that in normal usage it was difficult to achieve contact with that form of thread.
Finally, Sellers objected to Whitworth’s thread form on the grounds that it was more complicated, and therefore more costly to manufacture. According to Sellers, obtaining the rounded top required three kinds of cutters and two lathes to perform what with our practice requires but one cutter and one lathe.”
The core of Sellers’ opposition to the Whitworth standard was that it was complicated, expensive, and difficult to produce with consistent accuracy.
In place of the Whitworth form, Sellers proposed a contour then already in use in several Philadelphia machine shops, including his own. The thread took the shape of an equilateral triangle, sides inclined at an angle of sixty degrees, with the top and bottom flattened one-eighth of the thread depth (see Fig. 1).
| || || |
| || |
| || || |
|Fig. 1. Figure II above shows normal screw-thread practice when no particular standard was employed. Figures III and IV outline the characteristics of the Whitworth and Sellers threads. From Robert Briggs, ’A Uniform System of Screw Threads,’ Journal of the Franklin Institute, LXXIX (February 1865), 124.|
It had, Sellers claimed, precisely those virtues which were defects in the Whitworth form; it could be made with greater ease and less machinery, and could be verified with greater accuracy.
The table of pitches which Sellers proposed varied only slightly from the Whitworth standard, but was notable primarily because those variations allowed for the use of a relatively simple formula to calculate pitch for any diameter.
To his thread form and table of pitches, Sellers also proposed a uniform range of sizes for bolt heads and nuts and a standard thread gauge, thus offering for consideration a complete and standardized system of screw threads, nuts, and bolts.
Since screw threads were not a matter of objective precision, as Whitworth had pointed out, the determination of a system obviously lent itself to bias. Even particular elements of a screw thread reveal the possibilities. For example, assuming a generally V-shaped contour, pitch is the most significant factor in uniformity. It is important that, for any given diameter, the number of threads per inch is the same.
The close similarity between the pitches of Sellers’ system and Whitworth’s suggests that on this issue there was little to separate American from English practice. But the form of Sellers’ screw thread highlights the differences in American engineering and industrial attitudes.
Nowhere was that bias more clearly revealed than in the subsequent competition between the two thread systems.
One of the keynotes in the battle was sounded almost immediately. William Bement, head of the Philadelphia machinery firm of Bement and Dougherty and chairman of the Franklin Institute committee to investigate and report on Sellers’ system, set the tone when he advised the editor of ’Scientific American’ that the committee would be pleased to have the opinions of “all good practical mechanics.”
The emphasis of the word was Bement’s. The magazine reiterated the idea, congratulating American mechanics that their interests had not fallen “into the toils of schemers and theorists who would have confused instead of making the subject plain and practical.”
What was meant by “practical” was made clear in the Institute’s report. The Sellers thread form was based on an angle “more readily obtained than any other.” It was a thread contour which any ordinary workman could make with accuracy.
| || || |
| || |
| || || |
The committee’s conclusion was that a uniform American thread system should embody a shape which would enable “any intelligent mechanic to construct it without any special tools.” There was nothing theoretical or abstruse about it. It sounded democratic and it saved money.
All subsequent consideration of the Whitworth and Sellers thread systems reiterated the basic difference in thread form.
The U.S. Navy’s Bureau of Steam Engineering conducted a fairly extensive analysis of screw-thread systems in 1868, with a view toward the adoption of a standard for the Navy.
“We find but two systems in general use,” the Board noted, “one known as Whitworth’s, deduced many years ago from the general practice of English mechanics, and the other known as that of Mr. William Sellers, deduced more recently from the practice of American mechanics.”
In regard to pitch ranges, strength, and durability, the Board of Naval Engineers found no appreciable difference between the two systems. According to their report, the form of the Sellers thread was one of the major factors which led the Board to recommend his system.
The Whitworth thread required “such skill on the part of the workmen” that uniformity was out of the question. Conversely, the Sellers thread enjoyed the “very important advantage of ease of production.”
In England, uniform screw threads depended on a high level of craft skill. And to Englishmen, such skill seemed a perfectly appropriate foundation. As one editorialist succinctly put it: “With good tools and good men to use them, there is no difficulty whatever in producing the Whitworth cross-section of thread.”
Where labor costs were lower and craft traditions stronger, the system worked. But it was for precisely those reasons that Whitworth’s system was rejected here. In America, interchangeability depended on reducing a job to the point where an ordinary workman could produce the desired uniformity.
It was not Whitworth’s intention to conserve labor. The fifty-five degree angle of his thread was an arithmetic mean; it was mathematically indifferent to production costs. American engineering practice had labor costs as one of its basic considerations, and Sellers’ system reflected the difference.
The genius of his standard, however, was that it was not only easier and cheaper to construct, but easier as well to maintain with accuracy.
Because his thread perpetuated his name, Sellers is probably more widely known for his system of threads than for any other facet of his career. But its acceptance depended in part on his position in American industrial and engineering circles.
William Sellers was Philadelphia’s leading machine-tool builder when that city was the country’s preeminent tool-building center. “Probably no one,” according to Joseph Roe, “has had a greater influence on machine tools in America than William Sellers.”
At a time when it seemed to some that the screw-thread problem lacked a solution simply for want of any leadership, Sellers provided it. Although he had just celebrated his fortieth birthday when he presented his paper on a uniform thread system, Sellers had already established a reputation as one of Philadelphia’s leading machinists.
| || || |
| || |
| || || |
He came to his trade early. It was one of the characteristics of the city’s industrial elite to go directly to business, not college. After a basic education at a private school maintained by his family, Sellers was apprenticed to his uncle’s machine shop in Wilmington, Delaware.
Following his apprenticeship, he superintended the family-connected machine shop of Fairbanks, Bancroft and Company in Providence, Rhode Island, for three years and then, in 1848, returned to Philadelphia where he established a partnership with Bancroft in the manufacture of machine tools and mill gearing. So successful was the firm, a contemporary observer remarked, that within less than a decade it was the city’s leading machine tool company.
Sellers was frequently styled the "American Whitworth,” and there is no reason to believe he shunned the comparison. In fact, there is a remarkable parallelism in the accomplishments for which the two men were remembered.
Both conducted large and successful industrial enterprises. Both advocated high standards of precision in machine-shop practice, and to that end both produced machines whose form depended on function rather than on artificial architectural embellishment.
But there were also important differences between the two. Whitworth’s passion was accuracy. For him, “a true plane and power of measurement” were the outstanding elements of mechanical engineering, and the major thrust of his career was to upgrade English practice along those lines.
Sellers was by no means unconcerned with precision, but it would be fair to say that his work showed greater originality. In the sense that Whitworth’s interest was with the maintenance and extension of precision workmanship, he was a conservator; by comparison, Sellers was an innovator.
From the beginning, Sellers’ machine tools were dominated by the same thinking which characterized his screw-thread system. The catalogues of his firm explicitly stated that position:
"It has been said that good workmen can do good work with poor tools. Skill and ingenuity may indeed accomplish great results, but the problem of the day is not only how to secure more good workmen, but how to enable such workmen as are at our command to do good work, and how to enable the many really skillful mechanics to accomplish more and better work than heretofore; in other words, the attention of engineers is constantly directed to so perfect machine tools as to utilize unskilled labor."
Sellers’ machines embodied that view. His 1857 bolt machine marked the advent of commercially interchangeable nuts and bolts.
The gear-cutting machine, which he showed at the Paris Exhibition of 1867, was one of the earliest to be automatically operated.
And he produced an increasing number of special-purpose lathes on the principle that they would allow “less skillful workmen” to produce more work of better quality.
Sellers was also a grand-scale industrialist. He organized the Edge Moor Iron Company in 1868, which supplied all the structural iron work for Philadelphia’s Centennial Exhibition and all structural materials except the cables for the Brooklyn Bridge. Significantly, at Edge Moor, bridge-making was reduced to a standardized manufacturing process.
| || || |
| || |
| || || |
In 1873 Sellers also became president of Midvale Steel Company, where he supported Frederick W. Taylor’s long series of experiments in metal cutting. By middle age Sellers was already the dean of Philadelphia’s machine industry; at his death, he was recognized as "the greatest tool builder of his day and generation.”
But it was Whitworth himself who bestowed the highest accolade when he described Sellers as "the greatest mechanical engineer of the world.”
By talent and reputation, Sellers was probably better qualified than any other American machinist to effect a substantial change in the country’s screw-thread practice. Proposing a system, however, was only half of the battle.
The utility of a standard lay equally in its widespread acceptance. Whitworth had recognized the dual nature of the problem. He had presented his own system at meetings of the Institution of Civil Engineers, and he was aware of the importance of engineering opinion for the system’s adoption.
Not only was it difficult to achieve agreement on a system, inconvenience in the change-over and the inertia of traditional usage would also inhibit the spread of a standard. These were “obvious” reasons, Whitworth claimed, why engineers and their societies should push for the adoption of a uniform system.
Sellers was faced with similar impediments, but he was closely connected with major elements of private industry and was president of the only technical society in the country which enjoyed a national reputation—the Franklin Institute of the State of Pennsylvania. Therein lay the means for adoption.
A general society in the pre-professional era of American science and technology, the Franklin Institute was founded in 1824 as one of the earliest American mechanics’ institutes, with the aim of educating young apprentices in science and its applications.
In the decades which followed, a series of technical studies projected the organization into national prominence.
- Its experiments into the motive force of water provided an intensive analysis of hydraulic power.
- The Institute’s investigation into the causes of steam boiler explosions was the most exhaustive and detailed study conducted in this country and remained so for another half-century.
- The organization acted as technical advisor to the state of Pennsylvania on a standard of weights and measures and to the city of Philadelphia on the best method of paving streets.
- It participated in the investigation into the U.S.S. Princeton disaster and, along with the National Academy of Sciences, co-operated in a Navy Department study of steam expansion in 1863.
These activities gave the Franklin Institute a reputation for informed judgment, and that prominence was a significant factor in the adoption of Sellers’ system.
Equally important, there was a solid link between the Institute and Philadelphia’s leading machinery firms. The committee which was appointed to consider the subject of screw threads was mainly composed of representatives from those firms.
Chairman of the committee was William Bement, of Bement and Dougherty, a firm second only to William Sellers and Company in the city’s machine-tool hierarchy.
- Baldwin’s Locomotive Works had two men on the committee.
- The Southwark Foundry and Merrick and Sons, sprawling industrial firms established by Samuel V. Merrick, one of the founders of the Franklin Institute, were also represented, as were the firms of
- Morris, Towne and Company, the Pencoyd Iron Works, and of course, William Sellers and Company.
The Institute’s Board of Managers bore the same stamp. When William Sellers was elected president of the organization in 1864, most of the same firms were represented on the Board. Most of them were also using Sellers’ system in their own practices.
| || || |
| || |
| || || |
In their combined interests, these men and their firms wielded enormous influence in the American mechanical community. Promoters of rival thread systems may have felt cause to complain, as one did, that the committee simply confirmed Sellers’ conclusions rather than conduct a thorough investigation of the whole problem.
From the very beginning, the Institute zealously pushed for widespread adoption of the Sellers, or, as it was soon also called, the Franklin Institute system.
A special committee was appointed to consider the subject raised by Sellers’ paper, to the end that the Institute would recommend a system “for the general adoption of American engineers.” At the same time, copies of Sellers’ paper were sent to other mechanics’ institutes, “with a view to promote the introduction into general use of the system advocated.”
When the special committee made its report, it was unanimously in favor of Sellers’ plan.
The Institute then moved to secure general adoption by American engineers. Formal resolutions were sent to the quartermaster general of the Army, chief of the Navy’s Bureau of Steam Engineering, the chiefs of ordnance of both the Army and Navy, the chiefs of the engineer and military railroad corps, and to the superintendents and master mechanics of railroad companies.
The resolutions called for the adoption of the system "by requiring all builders under any new contracts to conform to the proportions recommended.” Similar resolutions were sent to all mechanics’ institutes and to the major machine shops of the country.
Governmental approval was the most obvious and direct means of obtaining conformity, and the Institute’s resolutions were heavily weighted in that direction.
The Navy was the first to respond. In March 1868, the secretary of the Navy authorized the Bureau of Steam Engineering to recommend a standard screw-thread system for the service. Benjamin Isherwood, chief of the Bureau, appointed Theodore Zeller, chief engineer at the Philadelphia Navy Yard, to head a board of naval engineers for that purpose.
After visiting the principal machinery establishments in the country and after formal meetings in Philadelphia, the Board in its report of May 1868 “unhesitatingly” recommended adoption of the Sellers system. By direction of the secretary of the Navy, it was immediately authorized as the Navy standard.
Since the same Philadelphia machinery interests which backed Sellers’ system were then involved in a nasty dispute with Isherwood and Zeller, the Bureau’s quick accord is perhaps surprising.
The quarrel involved used machinery which Zeller had purchased from a New York dealer for use at the Philadelphia Navy Yard. Local machinery manufacturers resented the loss of business to an outsider and raised serious questions about the transaction. The issue led to a congressional investigation which reinforced an already bubbling intra-service squabble, and Isherwood ultimately lost his job.
But the Bureau’s action suggests the importance of the thread system’s identification with the Franklin Institute. The ’Journal of the Franklin Institute’ had long been a favorite vehicle of professional expression for ambitious young Navy engineers. And when both the Navy and the National Academy of Sciences declined to publish Isherwood’s ’Experimental Researches in Steam-Engineering,’ the Franklin Institute did so.
It is also true that the Navy advocated Sellers’ system in the belief that it would probably become the established practice in private industry. As the Board stated in its report, “We were naturally desirous to select a system which, while meeting all the essential requirements of a system, would be most likely to be generally acquiesced in and adopted.”
In other words, the Navy’s concern was with Navy practice; it had no concept of leading a movement for the Sellers standard. Private interests were important only to the extent that they had already reached a consensus in favor of the system.
| || || |
| || |
| || || |
The Board’s conclusions effectively threw the burden of standards enforcement back onto private shoulders, and in the last analysis private interests were the most significant element in the adoption of the Sellers system.
Two factors shaped the response of industry. First, the standard reinforced technical advances in machine tools; it meant maximum utilization of self-acting and automatic machinery.
The laborious process of thread cutting by hand had long since given way to the use of self-acting lathes, which employed change gears to obtain different pitches.
And after 1860, machine shops had automatic bolt cutters at their disposal, further to speed and standardize the thread-cutting process.
Improvements in machinery made it all the easier to manufacture a standard screw. The adoption of a system rationalized those technical advances.
A standard system also had particular benefits for certain segments of industry, and it was from those areas that the most immediate and effective support came.
Interchangeability was critical on railroads, for example. It was obviously important that locomotives and rolling stock be manufactured according to standards, since repairs were frequently required at points distant from a central shop. Steam engines and the equipment of heavy industry fell into the same category, and acceptance came first from these sources.
By 1868, Sellers’ system was used in most Philadelphia firms and had also been widely adopted by New England machine shops and by several manufacturers of machinery in the mid-Atlantic states. The Pennsylvania Railroad took up the system in 1869.
It was adopted by the Master Car-Builders’ Association and the Master Mechanics’ Association (both railroad organizations) shortly afterward and, within the next decade, by most of the country’s railroads.
Standard gauging tools for the system, developed by Pratt and Whitney and by Browne and Sharp, only hastened the process of adoption.
But discussion and controversy did not end. General American acceptance of the system raised the debate to an international level, where it became entangled in the metric system controversy and in engineering competition between nations.
For example, when German engineers proposed the adoption of the Sellers thread form for use in a metric system to replace the Whitworth standard, English technical response was sharp. The effort, according to one editor, was an eccentricity on the part of “a handful of obscure scientific men.” It ran counter to the enlightened practice of thirty years and would bring ruin to Germany’s export trade.
Furthermore, it was claimed that after a brief period of experimentation in America, the Sellers thread contour proved to be a conclusive failure, and like errant children come home, machine makers in the United States were returning to the Whitworth pattern.
Faced with increasing engineering competition from America and from a Continent potentially united by the metric system, English technical opinion was confused and divided.
The Verein Deutscher Ingenieure (Society of German Engineers), also perplexed by the notion that there might not be an American standard thread system, addressed a communication of inquiry to the Franklin Institute, with particular reference to thread form.
The Institute’s secretary, Dr. William H. Wahl, adopted a novel form of response. He directed a circular letter to the officers of the major American railroads and to a number of manufacturing firms, asking whether they employed the Sellers standard and, if so, how successfully the thread form was maintained.
The replies, together with a counter-blast from ’Railway Gazette,’ were all forwarded to the German society.
It is difficult to say which offended the Americans most—English condescension, aspersions on their engineering practice, or technical inaccuracy. ’Railway Gazette’s’ editor retorted hotly: “This is not the first time that facts and figures as to American practice have been ‘evolved from the inner consciousness’ of writers across the water, to suit the occasion, but it is not often that such a complete perversion and reversal of the facts is given currency in a journal of standing.”
| || || |
| || |
| || || |
According to George Bond, of Pratt and Whitney, 90 percent of the orders they received for taps and dies were for the Sellers system. Furthermore, Bond claimed that experimentation and two decades of experience had proved the demonstrable superiority of its thread form.
There was an American standard, Dr. Wahl informed German engineers. It was called the Sellers, or Franklin Institute, standard and was used “throughout the United States, to the exclusion of any other.”
The exchange reflected, once again, the importance of nontechnical elements in the problem of screw-thread standardization. International industrial competition and engineering nationalism were the issues.
The establishment of professional engineering societies in America did not change the equation. Debate continued along much the same lines, though the forum was shifted to specialized groups.
At the first annual meeting of the American Society of Mechanical Engineers, George R. Stetson, of the Morse Twist Drill and Machine Company in New Bedford, pointed out that one of the functions of the new society should be to push the acceptance of the Franklin Institute standard so that “American machinery, by is uniformity in approved design and construction, may be entitled to the most favorable consideration in foreign as well as home markets, and the field of our usefulness and profits correspondingly increased.”
In fact, the A.S.M.E. failed to reach agreement on any standard, primarily on the principle that such questions were best left for solution in the marketplace. As Monte Calvert has noted, it was not that the society was opposed to standards, but that it felt they should be determined by businessmen and business methods.
Trade associations, business firms, and engineers working individually—not government—established national screw-thread practice in nineteenth century America. The Navy’s adoption of Sellers’ system was important insofar as it set a precedent and created a general climate of accord.
Writers on the subject always established the government’s acceptance of the system to suggest that it enjoyed a posture of authority. But there was no reality of governmental enforcement. Implementation was always a job for private hands.
In England, the subject was a matter of debate in the House of Commons. In France, it became settled governmental policy. But in America, private interests, more than any other, determined the outcome.
It was by the efforts of entrepreneurially minded technical groups like the Franklin Institute and the “untiring devotion” of organizations such as the Master Car-Builders Association that uniformity was advanced in America.
To many, it was right that such questions as screw-thread systems should be resolved by practical men rather than by politicians. Coleman Sellers spoke to that point of view when he remarked: “The government of France has always been in the habit of interfering with the private affairs of people.” Conversely, he noted, the American concept of governmental function was “that it should do and enforce justice, and that Liberty in all things innocent, is the birthright of the citizen.”
In time, that attitude changed; simple answers no longer fit the problem. As specialized societies examined the technical requirements of their own groups more closely, the weaknesses of a general system of screw threads became more apparent (see Fig. 2). That called, in turn, for a review of the whole issue, suggesting some governmental action.
| || || |
| || |
| || || |
|Fig. 2. THREADS: V. Thread, English Standdard, U. S. Standard, Bastard, Ratchet, Square, Wood Screw. This illustration shows the most commonly used screw threads of the nineteenth century.|
Whitworth’s thread form is shown as the English standard; Sellers’ thread is labeled as the U.S. standard. The V thread and bastard thread are variants from Whitworth and Sellers models. The remainder are special-purpose threads.
With the establishment of the Bureau or Standards in 1901, the government also had a means of responding to the problem. Subsequent efforts involved combinations of government and technical societies.
Nothing better illustrates the nineteenth-century attitude than George Bond’s frequent comment that, among screw-thread systems, it was “survival of the fittest.” William Sellers’ uniform system was fittest in the sense that it answered contemporary engineering requirements.
Sellers understood the economic as well as the technical aspects of the problem and appreciated their interrelationship as a prerequisite for adoption.
The Franklin Institute’s indorsement and proselytizing activities created a favorable environment for survival. At a time when no similar association existed to perform the function, the organization gave Sellers an institutional framework for his system, providing a platform, a mechanism for its advancement, and an aura of objectivity.
The man and the institution were well mated, and it was the combination which produced the Sellers, or Franklin Institute, standard of uniform screw threads.