TMS Home Page Presenting a Web-Enhanced
Feature Article from JOM
View Current Issue
LATEST ISSUE

TMS QUICK LINKS:
  • TMS ONLINE
  • MEMBERS ONLY
  • MEMBERSHIP INFO
  • MEETINGS CALENDAR
  • PUBLICATIONS


• TECHNICAL QUESTIONS
• NEWS ROOM
• ABOUT TMS
• SITE MAP
• CONTACT US

JOM QUICK LINKS:
  • JOM HOME PAGE
  • MORE HTML ARTICLES
  • COMPLETE ISSUES
  • BOOKS REVIEWS
  • SUBSCRIBE
• COVER GALLERY
• CLASSIFIED ADS
• SUBJECT INDEXES
• AUTHORS KIT
• ADVERTISE
Overview: Archaeotechnology Vol. 58, No.5, pp. 20-29

How Roebling Did It: Building the World’s First
Wire-Rope Suspension Aqueduct in 1840s Pittsburgh

Donald L. Gibbon
MAY 2006 ISSUE
About this Issue

 

 

 

 

 

ALSO APPEARING IN PRINT

View the Print Version
The print and/or PDF versions of the article can be acquired via the TMS Document Center.

 

 

 

 

FIGURE 1.
Figure 1
Figure 1. Roebling’s three suspension structures in downtown Pittsburgh in 1859, visible in the upper left corner of this image. (Drawn from Nature; lithographed and published by Wm Schuchman Pittsburgh).

 

 

 

 

FIGURE 2.
Figure 2
Figure 2. The population of Pittsburgh, 1800 through 1850.

 

 

 

 

FIGURE 3.
Figure 3
Figure 3. The original Roebling wire ropes on display at the APRR site in Cresson, Pennsylvania. Note that the top rope, from the APRR, is twisted, as per the patent description. This was an early “hemp replacement,” possibly laid up on the Roebling farm. Note also that the bottom rope, purportedly from the Brooklyn Bridge, is much smaller than the “cables” of the bridge and is also twisted. This suggests that it may have been a “hanger” rope, not part of the major suspension cables, which are still in place. Those ropes were either spun on site or made in the factory in Trenton, New Jersey.

 

 

 

FIGURE 4.
Figure 4
Figure 4. The Three Sisters—the 6th, 7th, and 9th Street Bridges, Pittsburgh, 1987. (Photo by D.L. Gibbon.)

 

 

 

 

FIGURE 5.
Figure 5
Figure 5. The Roebling wire-rope suspension bridge across the Monongahela River at Smithfield Street in Pittsburgh, built 1845–46. (From photo collection, Carnegie Library of Pittsburgh, by permission, file #P-3109.)

 

 

 

 

RELATED WEB SITES

 

 

 

 

RELATED ON-LINE READING FROM JOM
UNRESTRICTED ACCESS (FREE TO ALL USERS):

RESTRICTED ACCESS (FREE TO TMS MEMBERS):

 

 

Editor’s Note: In the interest of consistency with historical documents, all units of measure are presented in U.S. rather than metric units.
Questions? Contact jom@tms.org.
© 2006 The Minerals, Metals & Materials Society


The noted bridge designer John Roebling introduced his wire-rope suspension concept in Pittsburgh on a wooden aqueduct. His design was later implemented in bridges in Pittsburgh and elsewhere, including New York’s Brooklyn Bridge. This article describes Roebling’s work based on reviews of his notes and other historical documents.

INTRODUCTION

John Roebling was arguably America’s foremost early bridge-building genius. His first structure using wire-rope suspension was a wooden aqueduct, built in 1844–1845, which carried the Pennsylvania Mainline Canal across the Allegheny River into downtown Pittsburgh, the western terminus of the canal. Philadelphia was the eastern terminus. Three years earlier Roebling had developed wire rope of a substantially different design for the Allegheny Portage Railroad (APRR), another part of the canal. He had hand-laid up (i.e., twisted) those first ropes near Pittsburgh on his farm in Saxonburg. By contrast, the ropes for the aqueduct were laid up in place by traversing 3,800 individual parallel wires across the Allegheny River and bundling the wires into cables. The ropes for the aqueduct, according to Roebling’s records, were 1,100 feet long and were made from 200,000 pounds of #10 “charcoal iron wire” produced in Pittsburgh. The heavy beams were almost entirely white pine. It is important to remember that the structure was built basically by horse and by hand without mechanical equipment other than block and tackle. A significant part of the cost of the structure included keeping smithing fires burning for 200 days.

ROEBLING’S ARRIVAL AND EARLY YEARS IN THE UNITED STATES

John Roebling grew up in the context of early nineteenth-century Prussia, in a society in which answers to engineering questions grew out of a disciplined sense of order. There was definitely a right way and a wrong way to approach problems, and a sense of who was allowed to speak up with solutions. Innovation was not encouraged. To innovate, one had to break out of the mold.

Stifled by that mold, Roebling decided to give up public works engineering and emigrate to the United States. There he and some of his family and friends would establish a new community and a new life as farmers.

In 1831, under John’s leadership, a small group of families made their way to Pittsburgh, Pennsylvania, west of the Appalachian Mountains. John and his brother Carl purchased almost 1,600 acres of uncleared land nearly 25 miles northeast of Pittsburgh and settled what grew into the town of Saxonburg. In 1832, other families joined them and the community grew.

Roebling described the area in letters back to family and friends in Germany as a “rolling plateau with many fine distant views, considerable savannahs, fine meadows alternating with young woodlands and timber forests.” He was trying hard to convince the less venturesome to join them. “In earlier days of Indians and early settlers, great forest fires had destroyed the large dense forest,” he wrote. He said there was “good wood for fences” and encouraged them “above all things to bring an experienced young shepherd, with a pair of shepherd dogs, of which there are none here.”1

A most important aspect of this country, however, was the emotional climate. Roebling said in a letter home that “Every American, even when he is poor and must serve others, feels his innate rights as a man. What a contrast to the oppressed German population.”1

In 1832, Pittsburgh was already known as one of the major manufacturing centers in the United States, with important iron, wool, and cotton works already established. The area was known as the best wood market in America. For the new immigrants, however, it was first things first: clearing the land, building homes, and getting enterprise going, including brick-making. In 1836, John married and in 1837 became a citizen. But he turned out to be no farmer. His heart was not in it. Soon he turned back to engineering and found a job on the Sandy and Beaver Canal, a subunit of the major project in the region, the Pennsylvania Mainline Canal, connecting Philadelphia and Pittsburgh. The Mainline Canal was begun in Philadelphia in 1828 and finished in 1835. Construction actually proceeded from both ends toward the middle.


A BRIEF HISTORY OF SUSPENSION STRUCTURE DESIGN PRIOR TO ROEBLING

The history of suspension structures is complex and confusing and strongly influenced by personal and nationalistic biases. John Roebling wrote up his own version of that history in a “Report to the President and Board of Directors of the Covington and Cincinnati Bridge Company,” dated 1867.4 That particular very successful bridge had been started in 1856, but the Panic of 1857 and the Civil War stopped construction for eight years. Construction resumed in 1865, six months after the war was over, and was completed in 1866. The bridge is still in full operation although its deck has been renovated for automobile traffic. As an illustration of its quality, after the flood of 1937 it was the only bridge remaining open between Steubenville, Ohio, and Cairo, Illinois, a distance of about 1,000 river miles.

Roebling freely acknowledged the prior construction of both chain and wire cable bridges in Europe. All of these were for foot or vehicular traffic, none for transporting freight by water. According to Roebling, most of these were relatively light and were damaged by storms. The bridge at Freibourg in Switzerland was 870 feet long and lasted for some years, but according to Roebling was deficient in both strength and stiffness. Between 1830 and 1840, dozens of wire-cable suspension bridges were built in France, and others in Germany, Poland, England, and other European countries. Other suspension bridges were built during this time using chains.5

Roebling’s main claim was that the Allegheny Aqueduct introduced quite different principles (as discussed in this paper) in both the cable making and in the fundamental bridge design. The point was to make the structure capable of sustaining not only vertical but lateral loads. Stiffness was a vital issue. The French system suspended the bridge floor by a number of light cables. Roebling used only two relatively massive ones, assisted by a network of stays and suspenders. Under the complex stresses of a storm, the larger cables acted as full-strength units, whereas experience had shown that the smaller cables in the earlier designs had failed sequentially, one by one. This is what happened to a suspension bridge in Wheeling, West Virginia, which blew down in 1854; it had been supported by six smaller cables on each side.4

The first cable suspension bridge in the United States was built by Roebling’s rival, Charles Ellet, Jr., in Fairmount, Pennsylvania, over the Schuylkill River in 1842. The French system was used whereby the cables were constructed on land, then dragged to the site and elevated into position. But Roebling contended that in this process the wires in the cable got dislocated and bulged out, assuring that the stresses on the cable would not be equally distributed under load.4 This bridge, however, remained in service until 1875. In contrast, all of Roebling’s bridges used cables “spun in place” or spun in the spatial distribution in which they were to be used and kept under tension until final installation was complete, assuring that the individual wires would share the load equally. Roebling was very sure of the correctness of his principles and the performance of his structures bears him out well.


ENGINEERING PITTSBURGH IN THE 1840s

So what was the state of engineering in Pittsburgh when Roebling signed the contract for the aqueduct?

The town had less than 5,000 residents in 1815, the first year the citizenry’s occupations were documented. Of them, a remarkable percentage was involved in engineering-related businesses.8–10 For example, Robert Fulton’s second steam boat had been built in the city in 1811. In the years following the building of Fulton’s New Orleans, the Pittsburgh Steam Engine Company had become a significant feature of the local mechanical landscape. Associated with this firm one could order anchors and anvils, brass bells and machinery. There was a pattern-maker’s shop and a boring and turning shop. Screws made locally were available.

So community preparation for his debut suspension structure had unknowingly begun years before Roebling needed it. In 1794, the Irwin Rope Walk was operating in Pittsburgh; it moved to Western Avenue in 1813 in Allegheny Town, on the north side of the Allegheny River. The company made everything from wrapping twine to the largest ships cables, with 14 men working there. There was another rope walk in nearby East Liberty. By 1829 there were two rolling mills, the Pennsylvania and the Juniata. There were engine builders and wagon makers; one could buy iron rods and nails easily. R. Townsend was making iron wire nearby in New Brighton; he ultimately supplied Roebling with half of the wire for his ropes. According to the records, William Eichbaum, Sen. (sic) ran a second wire-making operation; he had started this business in 1810.8–10

WHY AN AQUEDUCT IN PITTSBURGH?

While the canal route came down the north shore of the Allegheny River toward Pittsburgh, the big money interests in Pittsburgh wanted the western canal terminus to be downtown, on their side of the river. Thus, the canal had to cross the Allegheny by aqueduct. A wooden structure on seven stone piers had served this function from 1835 to 1844. During the winter of 1844 an ice jam wiped out the structure. Replacing the aqueduct quickly was a high priority for the city because the canal was bringing thousands of tons of freight into the city. At its peak use and during the ice-free season, a canal boat passed a given spot on the canal every 20 minutes.

The Canal Commission in Harrisburg left the execution of the job to the city. Early in 1844, the state legislature passed a bill authorizing the mayor, aldermen, and people of the City of “Pittsburg” (sic) to rebuild or repair the aqueduct at their own expense, and then to exact whatever tolls they needed to pay for it. When the construction expenses had been fully recouped, the toll authority would revert to the Canal Commission.2 Dithering between Pittsburgh’s Aqueduct Committee, the Canal Commission, and Roebling consumed several valuable months of warm-weather construction time. Finally, a $62,000 contract was approved for Roebling, acting as a sole proprietor, to remove the shattered remains of the old wooden aqueduct, repair the seven masonry piers/towers, and build a new aqueduct using wire-rope suspension, a radically new concept. Roebling accepted the responsibility to provide debris removal, erection equipment, accounting, ordering, design, project management, and personal responsibility for the entire project. There had been at least one competitive bid for less money and the loser conducted a fairly vigorous campaign in the editorial pages of the Pittsburgh Gazette to have his design reconsidered. But his efforts did not succeed.3

On August 12, 1844, Roebling filed an acceptable 27-page handwritten set of specifications for the aqueduct with the Canal Commission, detailing the design, the materials, and the calculations on which the design was based. The calculations are remarkably simple, but what he created then are still the basic principles for today’s suspension bridges. By April of the next year, the job was done. Even before that job was complete, Roebling landed a contract to construct his first highway suspension bridge over the Monongahela River less than a mile away, on the site of the present Smithfield Street Bridge. This one was also completed in less than one year. By 1859, Roebling had built a second suspension bridge on the site of the present 6th Street Bridge (one of what are known today as the Three Sisters bridges).6 In that bridge he used I-beams instead of wood for the major structural members.

An 1859 lithograph (Figure 1) is one of the few views of Pittsburgh that shows it with all three Roebling structures spanning its rivers at the same time. Interestingly, the Roebling Smithfield Street Bridge was replaced about 38 years later by the present iron double-lenticular truss structure, designed by Gustav Lindental, who went on to also design the Queensboro Bridge in New York. That bridge is within sight of Roebling’s Brooklyn Bridge. Both are still in use today. Thus, Pittsburgh was the proving ground for two of America’s greatest bridge builders.

WHY ROPES? WHY WIRE ROPES?

Pittsburgh at that time was a major center of boat building, the boats being used for descending the Ohio to the Mississippi and beyond in the westward expansion of the young country. Ropes were a big feature of outfitting the boats and in Pittsburgh there were at least three rope walks—long, open areas in which the labor-intensive mechanical process of laying up the hemp fibers was carried out. The hemp was originally imported from the Philippines or Russia, but increasingly was grown in nearby Kentucky.

During the first years of the operation of the Mainline Canal, hemp ropes were used to haul the boats along the route of the canal. They were the only material available. When the canal reached the Allegheny Mountains, a unique engineering solution had been proposed for traversing this major topographical obstacle. The boats were removed from the canal; each boat was split into two sections and pulled across the mountains on special railroad cars which rode up and down a series of inclined planes. This was the APRR section of the canal, and Roebling had soon found his way to this interesting engineering challenge. (Today at the center of the route of the APRR, along U.S. Route 22 near the town of Cresson, Pennsylvania, there is a U.S. National Park Service site preserving the history of this part of the canal.) The canal boats were drawn up the mountains by stationary steam engines, then moved horizontally along the plateau by mules, then lowered down a second set of inclined planes on the other side. On one unfortunate day, Roebling saw two men die when one of the massive hemp haul ropes broke and the car got away. He reasoned that there were better materials solutions, and that reasoning stimulated Roebling’s plans to make iron wire ropes.

In 1841, Roebling wrote up a patent application for his rope-making process, titled “A New and Improved Mode of Manufacturing Wire Ropes.”7 He had used wire ropes that he had made at his farm in Saxonburg on the Portage Railway and had also begun to sell ropes elsewhere. But there is a great deal more to an aqueduct than just wire ropes.

Two questions present themselves: first, a general one: How does a genius such as Roebling translate his idea into reality? And more locally, what was the cultural milieu in Pittsburgh for performing this sort of major engineering feat? What did Roebling have to work with as he went from concept to finished job? That is, while the availability of iron wire is the first question, what about all the other components Roebling would need in order to do the job when he got the contract to build the structure in this post-frontier 1844 town? This includes not only materials, but skills to help fashion the materials into the finished structure. Specifically, was it possible for Roebling to get wire of the proper quality in a timely fashion?

Consider the flow of history up to that point: Roebling’s work was conceived only about 80 years after the French and Indian War opened in Pittsburgh. Construction of Fort Pitt, a major piece in that war, had begun in 1758. Between then and 1840 the city grew from just a few dozen to more than 40,000 people (Figure 2). Civilization had a firm foothold in Pittsburgh, but it had not been so long since life had been much more tenuous.

THE WIRE-ROPE DESIGN CONTROVERSY

A good deal of confusion has surrounded the specifics of the wire rope Roebling made for use in his aqueduct. Wire rope of strength equal to that of hemp is lighter, smaller gauge, and much more durable, lasting years as opposed to one or two seasons exposed to the elements. The individual wires can be coated with a varnish on manufacturing and then the whole rope oiled once or twice a season to minimize oxidation. An interesting consequence of its smaller cross section is its lesser wind resistance, an important feature in a sailing vessel. It is also much more elastic than hemp, meaning that it withstands a sudden strong load without snapping.

In 1839, when Roebling first began solving the problem of making wire rope from iron wire for the APRR (Figure 3), he was intending to use it for a running line, that is, one that would be in motion, as opposed to a standing line such as the stays on a sailing vessel that hold the masts in place. Thus, his first rope had to be not only strong but flexible, able to pass around a sheave or pulley. It had to be able to bend but not “take a set.” The individual wires and the collected group of wires therefore had to have very particular material properties. Roebling developed and patented11 in 1842 a method for the “spiral laying of the wires around a common axis without twisting the individual wires” while also having them each “under a uniform and forcible tension under all circumstances.” In other words, they would each bear a common share of the load on the rope. They would be “in contact with one another over the entire length of the rope, thus to a great degree excluding air and water, preventing corrosion.”11 Only a few years after the Pittsburgh “aquaduct” (as Roebling spelled it) was built, he wrote up specifications for a structure on the Delaware and Hudson Canal that he said was “in all parts similar to those of the Pittsburgh Aquaduct.” This canal was intended to carry anthracite coal to New York City. The Delaware aqueduct is still standing, although it is now used for automobile traffic. Roebling specified that “none but the best Charcoal iron wire is to be used for the cables. . . . Each strand of the cable (was to) be well varnished before it goes into the cable and the latter as well as the wrapping to be well painted.” The cable would be 8.5 inches in diameter including wrapping.12

Roebling’s first U.S. patent in this area, #2720 mentioned previously, is titled “Method of and Machine for Manufacturing Wire Rope,” dated 1842.11 One would think the patent claims would completely answer any questions about this subject, but they do not. His 1847 patent, #4745 titled “Apparatus for Passing Suspension-Wires for Bridges across Rivers”13 includes the claim that “The above mode of traversing wires, has in its main features been successfully applied in the formation of the cables of the suspension aqueduct in Pittsburgh, constructed by me.” And these two patents seem to be contradictory in the design and construction of the rope itself. The issue is whether the cables for the aqueduct consist of twisted or parallel wires, individual wires, or multi-wire strands. His own writings on this subject use several terms that need definitions:

  • Skein or wire: A skein is a short piece of wire. The skeins are spliced together to make very long wires. The specific manner of splicing at the time of the aqueduct is unclear.
  • Strands: Multi-wire ropes. In the 1842 patent, great stress was placed on the manner of achieving equal tensile stress in each wire within a strand and of laying up the strand without applying torque stress to the individual wires in the process. Thus, in 1842 there is no question that twisted wires were the norm. The drawing in this patent shows a twisted strand being wrapped by the machine Roebling was patenting. The wrapping was done with a single wire wrapped tightly around the twisted strand. The rope is greased while it is being wrapped.
  • Cable: In the 1847 patent describing how to get the wires across the river, the method said to have been used in Pittsburgh, a single endless wire (“composed of a great many skeins, spliced”) is passed across the river as an individual wire, back and forth from reels on either side and anchored at each end by passing around a cast-iron “segment.” In this description there is no mention of twisting or of strands. These individual wires were gathered into a “wire cable,” but that process is not described in detail in the patent. By contrast, in Roebling’s description of the Delaware and Hudson (D&H) Aqueduct in eastern Pennsylvania cables, he details the number of wires in each of seven strands.7

In a long letter written in 1926 by Washington A. Roebling, John’s illustrious son, who executed his father’s plans for the Brooklyn Bridge after his father’s death from lockjaw, Washington Roebling recalls “The Early History of Saxonburg,” the town his father founded. The younger Roebling describes the manufacture of wire rope in a rope walk set up near their family home. He tells briefly of neighbor men hired to work in the rope walk, the joining of individual short wires into longer ones, the laying up of strands of wires, and the final laying up of a completed rope from those strands. However, these ropes were used in projects other than the aqueduct, for which the cable was spun on site. He gives no details of rope design.14

In an unpublished manuscript, including a cover letter to the U.S. Commissioner of Patents, Henry L. Ellsworth, Esq., dated March 27(?)1841, Roebling describes his invention as consisting of “any number of wires laid parallel to each other, so that they form a round cylinder, and occupy the same positions respectively for the whole length; . . . Wire ropes manufactured in the above manner, will likewise be superior to twisted wires ropes for all purposes which do not require short chord bendings over small wheels or pulleys. All the wires being placed parallel to each other, uniformly strained (sic) throughout and not twisted, the greatest strength will be obtained by the least quantity of material.”15

A source of more contemporaneous information on this subject is Roebling’s own unpublished “Notes on Suspension Bridges.” In these he speaks of his design for the D&H aqueduct as being “in all parts similar to those of the Pittsburgh Aquaduct (sic).”16 In the D&H, he used seven strands to form the cable, each strand including from 270 to 325 individual wires. The cables were “spun in place without support,” apparently using the method described in the 1847 patent. Following the completion of the cables, the timber cross frames were hoisted into place from barges in the river below and the remaining suspended structure laid down. Again, Roebling does not mention twisting the wires to form the strands or the strands to form the cable.16

In the November 1845 issue of the Journal of the Franklin Institute, published just a few months after the aqueduct was completed, Roebling reports that for the Allegheny Aqueduct the cables were composed of 1,900 1/8th-inch individual parallel wires, each 1,175 feet long, compacted into 7-inch cables. “Great care has been taken to assure equal tension of the wires,” he wrote. “Oxidation is guarded against by a varnish applied to each wire separately, their preservation, however, is insured for certain by a close, compact and continuous wrapping, made of annealed wire and laid on by machinery in the most perfect manner. . . . for the first time successfully applied.” The oxidation-protection process described here undoubtedly refers to using the cable-wrapping device he patented in 1842, and as mentioned previously, that patent shows twisted cable. Roebling does not mention strands in this article. But the words of the text confirm that the aqueduct was made from parallel wire cables, a “standing” application.17

Finally, to close this issue, Roebling’s article in the American Railroad Journal and Mechanics’ Magazine (November 1843) gives a clue to which type of rope was used in the aqueduct when he distinguishes between ropes to be used for “running” or “standing” purposes, to use the nautical terminology. Wire cables for standing applications (such as in a suspension structure) could be manufactured by laying up parallel wires as they did not need to have any substantial flexibility. On the APRR inclined planes, fundamentally a running application, flexibility had been vital and thus a twisted construction would have been important.18

Roebling states that “wire rope can be spliced in the same manner as hemp rope”18 speaking of the manufacturing process for twisted rope. Splicing in hemp ropes involves opening up the twist on the running ends of the two rope segments to be attached to one another and interweaving the strands of both ropes over a short distance on each rope. This process makes a small bulge in the area of the joint. The strength of the splice comes from the squeezing down of the strands onto one another as the force is applied pulling on the joint, much like the old Chinese finger-trap that becomes tighter and tighter as you pull to get out of it. The problem is that twisted ropes are not used for suspension cables on Roebling bridges. What has to be learned is how to make a single wire “endless,” since they are the basis for making “parallel wire” suspension cables.

The only reference thus far found considering this issue in detail is the unpublished patent application of 1841 for the “new and improved method of manufacturing of wire rope,” handwritten by Roebling. Here he says: “The joining of wire strands, (sic) can be accomplished by annealing from 3 to 6 inches of the ends and twisting them around each other in a spiral manner, while held in a vice, (sic) and then squeezing the joint straight and round. Or the wire ends may be flattened, roughened and united by wrapping fine annealed wire around. Or they may be connected by simply forming loops or tying knots. The first described joint answers the purpose very well.”15

The first method, which sounds like a way to make a crude pressure weld, may provide one of the reasons for the 200 days of smithing fires that had to be maintained during the bridge construction. The smiths were perhaps responsible for creating the endless wires. Some further clues to the way the wire was used for making the cables come in a list of the men involved in the job. This list comes from one of the many small notebooks in the Roebling Collection at Rennselaer Polytechnic Institute:19 16 men splicing, two men filing, two breakmen, four shoemen for Pittsburgh side, three shoemen on Allegheny side, two regulators on center pier, four to lift wires on #1 and #3 piers, one driver, and one foreman. The first three categories of men may have been the “splicing crew.” The function of the “breakmen” is unknown. The shoemen were perhaps involved in attaching the wires to the anchorages.

This makes a total of 35 men for “running out” the wire for the cable. Roebling notes that it took 4–4.5 days to make a strand, or about two months to make the cables for the aqueduct, assuming seven strands per cable as per the reference on the Lackawaksen aqueduct.7

An interesting omission in the 1847 patent on traversing wires across rivers13 is that no mention is made of the process of setting up the endless wire rope from which the whole cable-spinning device is suspended. It appears to be assumed that anyone could do that just using common sense. This patented device actually appears to be a glorified “breeches buoy,” the age-old means of transferring people or goods from ship to shore or to another ship at sea. This is set up by launching a light throw-line from one ship to another, attaching the throw-line to a heavier line capable of carrying whatever the load will be, making that heavier line fast between the two ships, and then sliding back and forth, perhaps on some form of trolley. Presumably in the instance of the Allegheny Aqueduct, someone simply ferried a light line across first, then followed that with heavier lines, finally spinning the cables from individual wires towed back and forth from side to side by horsepower, following Roebling’s patented method.

DESIGN OF THE AQUEDUCT ANCHORAGES

The process of anchoring of the cables was unique and complex. For comparison, consider the Three Sisters bridges in Pittsburgh (Figure 4). These much-admired features of today’s downtown Pittsburgh riverscape were built in 1927–1928 immediately adjacent to the former site of the aqueduct. These bridges are called self-anchored suspension bridges, quite literally freestanding, not anchored into the ground. Their chain-cables are tied into the ends of the arched beams that support the roadway, creating a self-supporting tension/compression unit. Thus there is no heavy anchorage or abutment.20

Roebling’s design, by contrast, was dependent on support from heavy cast-iron and masonry anchorages. The cables themselves did not extend below ground because corrosion was a constant concern. Instead, each cable was attached to a cast-iron anchor chain that followed a curved path below ground to tie into a six-foot square anchor plate which was covered by “700 perches of masonry.” A “perch” in a 1913 version of Webster’s dictionary is given various definitions from 22 to 25 cubic feet of stone, calculated to be twice the total mass needed to resist the greatest possible stress ever to be applied, including the weight of the cables themselves (110 tons), the wooden structure (975 tons), the water (2,100 tons), and the canal boats.21 Of course, the vast majority of that load is carried down into the ground through the cast-iron saddles on the top of the seven stone piers, not transmitted horizontally through the cables into the anchorages. Interestingly, at this point in engineering parlance, the word “strain” was used for “stress.”

THE MATERIALS AND CONSTRUCTION OF THE AQUEDUCT

It is particularly interesting from a materials point of view to look at the components of the Roebling patents cited here.11,13 The “sheaves” or pulleys are made of “wood” or “wood very light.” The groove in the pulley is “made” (presumably this means lined) with “sheet tin.” Although the endless rope could be made of “hemp, a wire rope, which would not stretch, is much preferable.” The suspending arm for carrying the wire across the river, made of “wood or iron,” is attached to the rope by means of “twine.” The attachment of the arm to the guide rolled is made with “inch pine board.” The whole device is driven by a horse turning a vertical shaft. To pull the wire in one direction, the horse goes clockwise, the other direction, counter-clockwise . . . no gears!

Roebling was extremely cognizant of the potential for corrosion of the cast-iron anchor chains. He took pride in designing the materials of the system to avoid this. The anchor chains were painted with red lead, and then embedded in and surrounded by cement. The masonry was sealed with cement and “common lime mortar.” The preservation of bars of the anchor chains was “rendered more certain by the known quality of calcareous cements to prevent oxidation. If moisture should find its way in to the chains, it will be saturated with lime and add another calcareous coating to the iron.” In addition, he stressed that oak beams in the anchorage were not to come in contact with the cast iron, as the tannic acid in the oak would have promoted corrosion. So pine was always the only wood to touch the anchor chains.17

The “trunk” of the structure was attached to the cables through a series of threaded U-shaped iron rods, bolted through the timbers. Wire ropes were not used for this purpose, although rope was so used in many other bridges at other sites. The beams of the structure were of white pine, whereas some of the decking was white oak. The pine beams were up to 27 feet long and 6 inches × 16 inches in cross section.

Because of the newness of suspension construction in the civil engineering field, numerous questions were raised in public to cast doubt on the safety and reliability of the aqueduct. Roebling was nothing if not conservative as an engineer. He calculated and then in some cases actually demonstrated that the structure could stand alone if the wooden part of one or more of the seven spans were consumed by fire; that is, the aqueduct was not structurally dependent on the wooden components of the structure. Roebling was also assuming that the wooden structure might need to be replaced at some point in the future because of deterioration from constant exposure to the water, just as wooden ships’ hulls have to be replaced. Because of the conservative design, the removal of an entire span would not endanger the rest of the structure.

Summary of Materials Used in Aqueduct
The aqueduct, records show, used the following materials: charcoal iron wire; white pine beams; cast iron anchor chains and anchor plates; cast iron “pyramids” and saddles; sandstone masonry and fill rock; mortars and cements; iron bolts and nuts; tar and pitch; twine and hemp rope; oil, red lead, and varnish; water; blacksmiths, smithing tools, and charcoal; picks and shovels, saws; grindstone; forged iron pins; block and tackle; masonry and carpentry tools; and oak woodwork in anchorages.

THE AQUEDUCT CONSTRUCTION PROCESS

In the spring of 1844, the following ad appeared in the Pittsburgh Gazette, the NY Plebian, the Baltimore American, the Philadelphia Pennsylvanian, the Boston Post, the Harrisburg Union, and the Cincinnati Enquirer:

TO ENGINEERS: a premium of $100 will be paid for the best plan and complete specification of an aqueduct with wood or iron Trunk, either suspended or supported, to be constructed on the piers now standing in the Allegheny river opposite this city, provided the same be handed to the Mayor of this City on or before the 20th of June instant. For further particulars, apply in person or by letter to R. Galway, chairman of Aqueduct Committee.22
Just why the phrase “suspended or supported” is included in this widely advertised request for proposals is unknown. Perhaps Roebling helped write the ad. But whatever its origin, Roebling’s suspension design was selected from among many entries, with a contract price of $62,000. Roebling’s wire-rope business was already thriving, quite apart from its successful application on the Pennsylvania Mainline Canal. Shipping interests had immediately recognized its exceptional value in maritime applications. But here was a major departure: Roebling’s expertise would be used to manufacture a structure rather than simply a wire rope. It was quite a gamble for the city council.

On July 30, 1844 the Pennsylvania Canal Commission approved the design adopted by Galway’s Aqueduct Committee of the Pittsburgh City Council. Roebling presented his specifications to the commission for final approval on August 12, 1844. The job was to be completed by April 1, 1845.2 The job included removal of the remains of the demolished prior aqueduct, refurbishing the old stone piers, and constructing the new structure. Never having built a major bridge of any kind before and certainly never having built a suspension structure, Roebling now had just eight months to build his first one with the skeptical eyes of the city and the state’s engineering community on him.

Although Roebling was an obsessive note taker and a scrupulous planner, as far as we know he did not leave any timelines or program evaluation review technique charts for project management on this job. However, he kept his own punctilious accounting ledger.23 Many of the details of project management, personnel, costs, and financing can be inferred from the penny-by-penny double-entry accounts made in this ledger. We know when and what his suppliers brought to the job, so assuming a fairly primitive just-in-time delivery system, we know how the job proceeded.

One of the most interesting things is that there was often a three-month gap between receipt of the merchandise and payment in full. Evidently Roebling’s credit in the community was quite good. A number of employees were not even paid for their labor for two or three months, evidence of a high level of trust . . . or perhaps of a lack of other employment.

Roebling’s Accounting Ledger Highlights
The very first entry is September 2, 1844, acknowledging the receipt of two logs, 50 and 47 feet long from Judge Warner. The second, on September 8, was for delivery of one-half dozen spades, one-quarter dozen shovels, two picks, and one pair blocks. Imagine the opening scene at the construction site here. The canal has been in place up to and past this point for some eight years already. The trunk of the canal comes right up to the edge of the river. The water in the canal must have been cut off, presumably at locks above and below this point, when the aqueduct went out in the previous winter. The seven piers are still standing in the river, topped with the debris from the ice jam. The job now is to reconnect the ends of the canal on both sides of the river, almost a quarter mile apart. Some of the first items brought to the site were the “heavy earth-moving equipment” (the picks, shovels, and spades) and some means of mechanical advantage for picking up those heavy logs, beams, and stones (a pair of blocks).

It may be interesting to note here that the 300-mile ditch between Pittsburgh and Philadelphia (i.e., the Mainline Canal) was said to have been dug largely by hand by 1,000 Irish men. So the digging of the very large aqueduct anchorage holes with picks and shovels was certainly not something particularly notable in this day. Fortunately, the aqueduct site is in the Allegheny’s flood plain, so the anchorages were dug generally in unconsolidated alluvium rather than in the in-place shales and sandstone of the nearby valley walls.

In short order after that starting point in September, Roebling received lumber, cement, posts, hammered links, anchor chains, and 1,040 feet of oak timber, importantly including masts, not for ships but for cranes. Those masts would become the cranes that would use the just-delivered pair of blocks for lifting heavy weights. Evidently, the first order of business was to have the anchorages prepared. Those picks and shovels would dig the holes into which the heavy anchor plates would be placed, the plates to which the anchor chains would be attached and on which the 700 perches of stone would be placed. But needed first would be holes in the ground.

In late September, 140 bushels of cement arrived, delivered by John Linton.

In early October, pins were delivered for attaching the cables to the chains. And soon the stone started arriving. Roebling had specified that the pier repairs would be carried out with the “hardest and best sandstone which can be had on the Allegheny River about Freeport (Pennsylvania).”24 James Stewart, of Allegheny City Quarrying Stone, delivered loads of stone to both the “Allegheny anchorage” and the “Pittsburgh side.” A. Smith of Allegheny delivered 255 perches cut and backing (facing?) stone for piers and abutments—the equivalent of almost 300 cubic yards of stone.

Over the period from October 7 to December 14, Dennis McKelby was credited with weekly deliveries of stone totaling 708 perches of building stone, 12 perches of cut stone, and 44 perches of abutment and pins, all paid for on account.

William Paul was credited with having provided 2,063 cubic yards of excavation on the Allegheny and Pittsburgh sides, apparently at 10 cents per yard, for he was paid $206.30. While this work was done in October and November 1844, the bill was not fully settled until April 10, 1845.

Starting on October 29, 1844, four city bonds for $1,000 each were provided to S.M. Wickersham in beginning payment for about half of the wire for the cables. Ultimately, 104,000 pounds of #10 wire and 6,725.5 pounds of #14 wire were delivered by him at a cost of approximately $11,000.

Another vital page in the ledger is devoted to R. Townsend and Co.—suppliers of the other half of the necessary wire for making the cables. The first entries here are two debits starting on October 30 for a total of $8,000. Townsend delivered 100,246.5 pounds of #10 wire, 3,440 pounds of #14, 150 pounds of #22, and 50 pounds of #13 rivets, for a total of $10,566.44, almost the same amount of wire and the same amount of money as for S.M. Wickersham, listed previously. Perhaps Roebling was being cautious, maintaining two suppliers for this vital ingredient in the project.

One of the most interesting pages in the ledger is headed “Acct with the City.” This page lists a debit of $62,000 on October 26 and bonds for varying amounts credited between then and September 19, 1845, totaling $60,000. Roebling started paying for his wire some three days after these bonds began being issued.

By December, heavy timbers began arriving: 260 beams 27 feet long were brought to the site, paid for by a check on the Merchants and Manufacturers Bank to Sam Frandin, Timber Contractor. These were to be the transverse beams that supported the trunk of the aqueduct. The beams were cut to order for length, but still had to be worked for use in the structure, as described in the following specification submitted to the Canal Commissioners: “The beams will be 27 feet long, 6 × 16 inches, and are to be arranged in pairs, at a distance of 4 feet from centre to centre of pair, each 2 beams to have a space of 4 inches between for the reception of the dovetailed tenons of the posts. The stringers at the corners of the trunk to be 15 inches × 7 inches, rabbited inside for the reception of the first bottom course (of planks) and notched below 4 inches deep, for letting the beams in.”19

Ultimately, about 500,000 “BM” (board measure?= board feet?) of timber were to be used in the aqueduct. About this same time, one grindstone was delivered, a device not present on many construction sites today. It was vital in both making parts and keeping tools in working condition.

On December 21, 740 bushels of lime were delivered, indicating that the mortar and cement were being prepared for sealing the anchorages and grouting the piers. Henry Anshutz was paid in December for patterns and castings for the cast-iron saddles to crown the seven piers and support the cables; they had to be in place before the cables would be started.

On January 3, 1845, 160 gallons of linseed oil were delivered, a sign that the cables were being spun. The cables had to be in place before the wooden structure could be begun, since the cables were designed to support the structure.

On February 7, Isaac Claus delivered 3,200 pounds of thread for caulking the trunk of the aqueduct to make it water tight. About three months remained until the contract deadline. The cables must have been completed or almost so.

As a side note, consider this: in the previous winter a wooden aqueduct built on these same piers had been wiped out by an ice jam. The new aqueduct was not designed to have significantly greater clearance over the water than the old one. Ice was (and still is) a fairly common feature on the Allegheny River.

Now the construction crew was required to be out on the river stringing wire 3,800 times, back and forth, to create two 7-inch cables 1,100 feet long, each with 1,900 wires in them. The timbers were said to be put into place in the structure from underneath, presumably picked up by cranes from above off the boats.

The timbers were fixed in place with wrought-iron spikes. One line at the end of the entries in the “Table of Aquaduct (sic) Exp(enses)” lists “Boats, Cranes, and River Exp.” at $1,000,23 which comes to about 1.6% of the total cost of the job. In the author’s estimate, although only a small percentage of the total monetary cost, this was one of the key challenges in the whole job: doing the river work in the winter time. It would be extremely interesting to know just how this was done—for example, if any of this work was handled by steam boats or if cable ferries were used.

The timing is fitting together here. On February 24, 1,070 screw bolts and nuts were delivered. These were for fastening the suspension rods through the beams. In mid-March, 196 rods were delivered. These apparently were to be fashioned into the suspension rods.

Rope, cords, and twine were brought in February by James Rowley, followed in April by 24 barrels of pitch. In late February, 70 pounds of coal tar were delivered from the City Gas Works, probably a by-product of making “town gas” from coal—the same process used in Britain until the 1970s when natural gas became available from the North Sea. The tar was used for sealing parts of the anchor chains and cable.

One of the last bills is for the caulker who probably used the caulking twine and the pitch to do his job. Also in late March is a substantial bill for the blacksmith and the cost of keeping three smithing fires going for 200 days. The smiths were also responsible for using the grindstone noted earlier to keep woodworking tools sharp. Also near the end of the list is the cost of six wrapping machines to wrap the cables according to the 1843 Roebling patent. The job was almost done.

OPENING OF THE NEW AQUEDUCT

On April 23, 1845, the editor of the Pittsburgh Daily Gazette happened to be walking in the neighborhood of the aqueduct and was surprised to find it nearly complete. The second layer of planking on the floor of the trunk was down on two of the seven spans and a clear idea could be gained about what the structure would look like. The editor commented that in the press of other events, the aqueduct had almost been forgotten by the public.

Most important of those “other events” had been a catastrophic fire in the downtown Pittsburgh area that had destroyed 1,100 buildings, including most of the public offices, the major hotels, many factories, and most of the city’s warehouses. It was a disaster of major proportions. It took out the important covered wooden bridge across the Monongahela River at Smithfield Street. Roebling got the contract to replace it as a suspension structure before the aqueduct was even finished.

On May 5, the paper carried a notice of the near-completion of this “noble structure.” The only thing holding up progress on the project was the difficulty obtaining workmen to complete the “calking” (sic) of the trunk. There were 25 men already working at that job. On Thursday, May 22, the first trial introduction of water into the trunk was attempted to check for leaks. Remaining leaks were closed and the trunk was finally filled. A collective sigh of relief was heard as the 2,100 tons of water poured into the structure and it held under the load. A band was in attendance, playing late into the evening, and a large crowd of people milled about to see the show. The job was complete. Soon canal boats resumed bringing the precious freight into downtown Pittsburgh after over a year of interruption.

Roebling’s ledger shows that the total cost of the aqueduct was $58,297. The paper noted that Roebling was reputed to have made little or nothing for all his effort, but commented that “. . . he has erected a work which will secure him a high reputation, and eventually an ample return in a pecuniary sense. His next contract is for the Monongahela Bridge, which is also on the Wire Suspension plan, and we hope he will have ‘room and verge enough’ to construct a handsome thoroughfare across that stream.”25

That surely happened, as shown in Figure 5. The next 20 years or so were spent piling one engineering achievement upon another, climaxing with his design of the Brooklyn Bridge. An accident led to his death from lockjaw before he could complete that job, but his son took over to fulfill his father’s plans and assure the family a major place in bridge-building history. But it all started right here in Pittsburgh.

REFERENCES

1. J.A. Roebling, English translation of letter to German friends and family (© 1832), published in WPA Hist. Mag., 18 (2) (1935), pp. 73–108.
2. H.M. Cummings, PA Board of Canal Commissioners’ Records w/Allied Records of Canals Chartered by Commonwealth (1959), PA State Archives, Harrisburg, PA, Series 17-516, 17-517, 17-519.
3. Pittsburgh Gazette (July–August 1844), archives of the Historical Society of Western PA, microfilm (1844).
4. J.A. Roebling, “Report to the President and Board of Directors of the Covington and Cincinnati Bridge Company,” privately published (1 April 1867), passim, see esp. pp. 47–53. Available at Rutgers University Roebling Family Archives, Newark, NJ, microfilm reel 8.
5. David Denenberg, Mostly Suspension Bridges (2005), www.bridgemeister.com/list.php?type=full&page=1.
6. D.B. Steinman, Builders of the Bridge (New York: Harcourt Brace, 1945), p. 456.
7. R.M Vogel, Roebling’s Delaware and Hudson Canal Aqueducts (Washington, D.C.: Smithsonian Studies in History and Technology, Smithsonian Institution Press, 1971).
8. J.M. Riddle, “The Pittsburgh directory for 1815: containing the names, professions and residence of the heads of families and persons in business in the borough of Pittsburgh; with an appendix containing a variety of useful information,” printed for James M. Riddle, compiler and publisher (Pittsburgh, PA, 1815).
9. S. Jones, “Pittsburgh in the year eighteen hundred and twenty six: containing sketches topographical, historical and statistical; together with a directory of the city,” (Johnston and Stockton: Pittsburgh, PA, 1826).
10. Isaac Harris, “Harris’ Pittsburgh business directory for the year 1837; including the names of all the merchants, manufacturers, mechanics, professional and men of business of Pittsburgh and its vicinity” (Pittsburgh, PA, 1837).
11. J.A. Roebling, “Method of and Machine for Manufacturing Wire Rope,” U.S. patent 2720 (1842).
12. J.A. Roebling, “Specifications of Delaware and Hudson Canal Company Aqueduct,” Rennselaer Polytechnic Institute (RPI) Roebling Collections, Troy, New York, box 10, folder 6 (date uncertain).
13. J.A. Roebling, “Apparatus for Passing Suspension-Wires for Bridges across Rivers,” U.S. patent 4745 (1847).
14. W.A Roebling, “The Early History of Saxonburg” (1926), Report for Centenary Celebration Committee, Saxonburg, PA, available Saxonburg Borough Library.
15. J.A. Roebling, “Specifications” (27 March, 1841) Patent Application for “New and Improved Mode of Manufacturing Wire Ropes” (patent not granted— perhaps not submitted), Rutgers University, Alexander Library, Roebling Family Archives, Newark, NJ, MS Box, folder 1 microfilm reel 7.
16. J.A. Roebling, “Notes on Suspension Bridges,” unpublished; RPI archives, Roebling Sci-Tech #326,Troy, NY (date uncertain).
17. J.A. Roebling, “The Wire Suspension Aqueduct over the Allegheny River at Pittsburgh,” Journal of the Franklin Institute, 3rd Series, X (Nov. 1845) (5), pp. 306–309.
18. J.A. Roebling, “American Manufacture of Wire Ropes for Inclined Planes, Standing Rigging, Mines, Tillers, Etc.,” American Railroad Journal and Mechanics’ Magazine, Third Series 1 (11) (1843), pp. 321–324.
19. J.A Roebling, RPI Roebling Collections, Troy, NY, book 39, box 14.
20. B.S. Criddlebaugh, website organized and posted by Bruce S. Criddlebaugh providing historical and technical data on “The Bridges and Tunnels of Allegheny County, PA,” http://pghbridges.com/pittsburghW/0584-4477/ninth_st_br.htm.
21. J.A. Roebling, Table of Quantities on Pittsburgh Aqueduct, RPI Roebling Collections, Troy, NY, box 17, folder 120 (1845), pp. 1–4.
22. “Public Announcement of Request for Designs for Aqueduct,” Pittsburgh Gazette (1844) RPI Roebling Collections, Troy, NY, box 10, folder 2.
23. J.A Roebling, Pittsburgh Aquaduct (sic) ledger, Microfilm Reel 9, MS box 7, Rutgers University, Alexander Library, Roebling Family Archives, Newark, NJ (1844–1945).
24. J.A. Roebling, “Specifications,” papers in the Archives of the PA Board of Canal Commissioners records (see Reference 2) (1844).
25. “The New Wire Suspension Aqueduct,” Pittsburgh Daily Gazette (May 24, 1845), p. 2.

For more information, contact Donald L. Gibbon, MATCO Associates, P.O. Box 15580, Pittsburgh, PA 15244; e-mail donald.gibbon@matcoinc.com.