The Great Bridge

by David McCullough

  The initial view of the caisson interior was generally something of a shock, once the eyes had adjusted to the light. The six big chambers looked something like vast cellars from which a flood had only recently receded. Every post and partition, every outside wall, and the entire ceiling were covered with a slimy skim of mud. Every man in the place wore rubber boots and got about on planks laid from one section to another and between the planks the muck and water were sometimes a foot deep or more. Most days the work force would be concentrated in a few locations, leaving some of the huge chambers as dark and silent as subterranean caves.

  Where there was light it came from calcium lamps, limelights as they were also called, which threw steaming, blue-white, luminous jets into the corners where the men worked, or from squat sperm candles that blazed like torches at the end of iron rods planted alongside the plank walkways. “The subject of illuminating a caisson in a satisfactory manner is rather a difficult problem to solve,” Roebling remarked in his report to the directors of the Bridge Company. At first the candles had burned with such vigor in the compressed air and sent up such clouds of smoke that the air had become intolerable. This had been overcome somewhat by reducing the size of the wick and of the candle and by mixing alum with the tallow and drenching the wicks in vinegar. Even so Roebling worried about the quantities of floating carbon the men were breathing into their lungs.

  Kerosene lamps had to be ruled out from the start. They smoked even worse than candles, and with fire a constant hazard in such a charged atmosphere, Roebling did not want the risk of spilled oil. So he had hit upon the idea of limelights, of the kind ordinarily used for stage lighting or nighttime political rallies. He had the gas—a combination of compressed oxygen and coal gas—piped into the caisson, put burners in every chamber, and found two lamps per chamber would do the job. One small explosion had singed the beard off an attendant, but other than that the system had worked most satisfactorily. Ordinary street gas would have been about five times less expensive, but when that had been tried, the heat inside the caisson had built up to the point where no one was able to take it.

  The air as it was, besides being heavy and dank, was uncomfortably warm. On the way from the compressors it passed through a cooling spray of water. Even so, winter or summer, regardless of the time of day or the weather outside, the temperature inside stayed at 80 degrees or more and the air was so saturated with water that under the best conditions the chambers seemed continuously shrouded in mist. Visitors who did not have to exert themselves in any way soon found they were wringing-wet with perspiration.

  Most of the people who visited the caisson—newspapermen, local politicians, an artist from Harper’s Weekly, editors from some of the professional journals—came out with their clothes thoroughly mud spattered and quite relieved to have the experience behind them. Many of them also expressed open amazement that men could actually work in such a place day after day.

  The first load of rock and mud was hauled out of the caisson by clamshell dredge buckets on July 5. Most of the effort inside was spent removing the sharp-edged boulders that threatened to damage the frames and shoe as the caisson began to come down on them with crushing force. Boulders under the water shafts were the most serious initial problem, for if the caisson were to settle suddenly, the shafts might be blocked shut or badly damaged. And there was no way to get the boulders out of there except to chip away laboriously hour by hour, by hand, with long steel bars and sledge hammers.

  In the middle chambers the ground was nearly all traprock, packed like gravel and joined by what Roebling described as a natural cement made of decomposed fragments of green serpentine rock. Every boulder was coated with this unyielding substance, upon which a steel-pointed pick had virtually no effect. Only by driving in steel-pointed crowbars with heavy sledges were the men able to make the slightest headway.

  In chambers No. 1 and 2, those nearest the ferry slip, there was clay and gravel between the rocks, which made the going easier, while in Nos. 5 and 6, those at the upstream end of the caisson, there was a gummy blue clay that extended down forty feet, just as indicated by earlier soundings. This made the digging there relatively easy, of course, but it also meant that the caisson would have to go down at least forty feet—or beyond the clay. As Roebling said, no better foundation could have been wished for than what they were finding in chambers No. 3 and 4, but only if it had extended all over. And with the nature of the material so vastly irregular from one chamber to the other, lowering the caisson uniformly seemed practically an impossibility.

  Roebling kept careful track of the rock uncovered. Nine-tenths of it, he found, was of Hudson River Palisades origin, transported, like all of Long Island, millions of years before by the glaciers. This traprock, as it was commonly called, was basalt, an igneous rock, like granite, and nearly as hard. As the men dug into the caisson floor, the traprock emerged in chunks the size of paving blocks or in monstrous boulders, but when a shovel or pick first struck one of them, with a sharp metallic clink, there was no telling which size it would turn out to be.

  Boulders of quartz and gneiss occurred here and there, but rarely. Two big boulders of red sandstone were also found. But a collection made by Roebling of all the different varieties of smaller rocks uncovered, most of which had been worn down to pebble size, presented a complete series of the rocks to be found for a hundred miles to the north and northeast of Brooklyn.

  The idea of driving the cutting edge of the caisson through such material by building weight overhead had to be abandoned at the start. The pressure needed to do that would crush the cast-iron shoe and smash the bearing frames. So the cutting edge would simply not cut. Instead, every boulder, every rock of any size, had to be removed before the shoe or frames began bearing down on them. And all such work had to be done by probing underwater since there were trenches along the inside edge of the shoe, clear around, and these were nearly always brimful of water that seeped in from the outside. (This water flowed in turn into cross trenches at the foot of the frames, which supplied the big pools under the water shafts.)

  Just finding the boulders under the shoe, let alone removing them, was an unbelievably tough and disagreeable task. The full perimeter of the cutting edge was 540 feet. This added to the five frames, each 102 feet long, brought the caisson’s full bearing surface to 1,050 lineal feet, or a distance greater than the length of three football fields, every inch of which had to be probed beneath with a steel sounding bar twice daily with each shift. Whenever a new shift came down, the work accomplished in the preceding eight hours had to be carefully explained; and since most of the trouble spots discovered would be underwater, there was no way simply to point them out—the information had to be written down or memorized. “Moreover,” as Roebling wrote, “a settling of the caisson of six inches or a foot would bring to light an entirely fresh crop of boulders in new positions, and very often half without and half within the caisson.”

  To keep weight off the shoe, and so off any such boulders protruding under the shoe, it was necessary that the frames, or chamber partitions, take up that part of the load not balanced off by the compressed air. And with the frames thus the prime structural supports of the whole enormous burden, there had to be a way to lower them as the caisson descended.

  The system used at first seemed the simplest solution, but it did not work well at all. Small pillars of earth were left under the frames, each one about three feet square and from six to eight feet apart. These pillars were then to be dug away systematically and the caisson lowered in that fashion. But the earth pillars often concealed a boulder that had to be removed, or they would be eroded away by water, or still more often, the workers in adjacent chambers, not working in unison, would undermine them and destroy their usefulness.

  The system next adopted worked extremely well and was used until the end. Beneath each partition, every eight feet or so, two wooden blocks, a foot square and two feet long, were placed, one on top of the other, with oak wedges
jammed between them and the bottom edge of the partition. Whenever the shoe had been cleared of all obstructions to a depth of several inches the entire way around the caisson, the wedges were knocked loose with sledge hammers, one by one, frame by frame, until the whole caisson settled. New blocks were then put in beside the old ones, which, if the descent of the caisson had been sudden, were split in two or crushed to a pulp. “The noise made by splitting blocks and posts was rather ominous,” Roebling commented dryly, “and inclined to make the reflecting mind nervous in view of the impending mass of thirty thousand tons overhead.”

  Collingwood and Paine were in charge of clearing boulders from beneath the shoe and seeing that the caisson settled properly. “Levels were taken every morning on the masonry above,” Collingwood wrote later, “and a copy furnished the general foreman…. If the caisson were level, the usual method followed in lowering was to begin at the central frame, and loosen the wedges regularly from the center towards the ends. The two frames next to these were then treated in like manner, and finally the outer two. When no obstructions occurred, the blocks would all be gone over several times in the course of a day, and the caisson would settle easily, at the rate of three or four inches in 24 hours.”

  At first, however, things had not gone that way at all. Through July and on into early August, the rate of descent had been less than six inches a week, and the boulders, instead of diminishing in number, as had been expected, became more plentiful. It was a hopeless rate of progress Roebling reported to his directors. At this rate it would take nearly two years to sink just the one caisson.

  Boulders within the work chambers were the lesser evil. Before they could be hauled up the water shaft, they had to be split into manageable pieces, never an easy job, but at least they could be dealt with under comparatively reasonable conditions. Boulders under the shoe, however, or those found beneath the frames, were each a major undertaking. The removal of a boulder from under the shoe, for example, went as follows.

  The ground around the inner side of the boulder had to be dug away with pick and shovel, with the excavation filling with water as fast as the men worked. Then the boulder had to be drilled by hand, underwater, and a lewis inserted, a dovetailed iron eyebolt to which a hoisting rig of some kind could be attached. In the early days of the work, double sets of block and tackle were tried, with a gang of thirty or forty men hauling at the ropes, while others worked furiously with winches and crowbars. But very often the boulder refused to budge. So Roebling had hydraulic jacks lowered through the supply shafts. These were of a kind designed for pulling instead of lifting and had a capacity of two to three tons. The water chamber on such a jack was above, not below, the piston, and the piston rod had a big hook at the end instead of a lifting shoulder. This hook was attached to the iron eye in the boulder, while the opposite end of the jack was chained fast to the nearest substantial timber or, better still, to the ceiling. The jack pump was then set in motion and, as Roebling said, it would prove itself a “very effective instrument.” There would be an immense momentary strain, then the boulder would give way and come sliding into the caisson, where it would be broken up..

  When a boulder appeared to extend several feet outside the caisson, no attempt was made to haul it in. Rather the part inside was slowly split up until enough had been removed for the caisson edge to clear.

  But whichever way they were handled, a few good-sized boulders beneath the shoe could hold up everything for three or four days. Such delays were maddening, and there were more and more of them as time passed and equipment began to break down or the water shafts failed to function as they were supposed to. The big clamshell buckets, armed with seven-inch teeth, were formidable-looking affairs and under normal conditions one of them could dredge up more than a thousand yards a day. This, in theory at least, meant that the equipment in use should have been able to haul out the whole volume of material that had to be removed for the tower foundation in about a month’s time. But the buckets kept breaking down or getting caught under the bottom edges of the water shafts. As it was, the job would take five months, and these, as Roebling wrote, were “five months of incessant toil and worry, everlasting breaking down and repairing, and constant study to make improvements wherever possible.” Bucket teeth that worked well for scooping would not last a day at grappling with stones. For every two buckets in working order, three were being repaired. “There was, indeed, one period,” Roebling said, “when we were almost tempted to throw the buckets overboard…”

  One of the greatest early disappointments was to find that the buckets were unable even to dig their own hole under the water shafts, as they were supposed to. Much of the time the buckets failed even to bite into the material dumped into the hole unless a couple of men were kept constantly stirring the pool, “to keep the stuff alive,” as Roebling said. But even then the bottom of the pool kept filling in and had to be dug out by hand repeatedly.

  Stone and clay would pack solid and actually fill the hole in a few hours, such was the incredible nature of the material being excavated. So it became necessary to feed all the stones into the pool at one time, separately, then the clay by itself. The kind of bucket in use could lift any stone it could catch hold of, but such a stone, or a chunk of split boulder, had to be placed just so in the hole for the bucket to get a proper hold, and the stones could only be taken out one at a time. Whenever a badly placed stone got wedged under a shaft, which happened fairly regularly, somebody had to dive under to see what could be done. “When the lungs are filled with compressed air,” Roebling wrote, “a person can remain under water from three to four minutes.” He knew this, it seems, from personal experience.

  Any material fed into the pools from other parts of the caisson could be properly prepared, as it were, for the dredges to handle, but when the trouble was inside the water-shaft pools, as often happened, or when the pool had filled in, one to two days would be lost while the shaft was sealed off on top, with an iron cap, the water forced out by compressed air, the pool pumped dry, and the pit dug out by hand to a depth of six to eight feet. And the whole time this was going on (about two days on the average), the other shaft had to handle all the work. There were, in fact, so many occasions when the pools had to be cleared in just this fashion, so many repairs needed on the buckets, that most of the time only one water shaft was in operation.

  Furthermore, whenever the work was held up a day or two, and the caisson stopped settling, its movement immediately afterward could be quite erratic, coming in sudden, unpredictable, uncontrollable starts. This, Collingwood explained, was due to the earth compacting around the caisson, as it does around a pile when driven slowly. “At such times it would seem impossible to get it started, and when once movement began, it was almost sure to split a set of blocks before it was arrested.”

  Once, after the caisson had been at rest for several days because of breakdowns in equipment, all the usual steps were taken to get it started again, but to no avail. The blocking was eased, the shoe was cleared of obstructions, and still the caisson just hung there, motionless, with nothing holding it. The men did not know quite what to make of this. The only real significance of the episode, however, was that it gave the engineers a chance to compute roughly how much side friction the caisson had to overcome during its descent. As Collingwood figured it and reported later to a meeting of the American Society of Civil Engineers, the average pressure in the chamber at the time was seventeen pounds per square inch, giving a lifting force from the compressed air only of 20,400 tons. The bearing surface (posts and frames) was carrying about 625 tons and estimates were that the whole outer edge was probably carrying about that much again, which gave a total upholding force of somewhere near 21,650 tons. But the total weight of the caisson then, including the stonework on top, was judged to be 27,500 tons. Therefore, when it failed to move, the weight being carried by side friction alone was 5,850 tons. So this meant that along with everything else that had to be overcome to get the caisson
down even a single inch, there was about 900 pounds of friction working against every square foot of the exterior surface.

  To add to the over-all physical discomfort of everyone involved, blowouts continued and with greater frequency than Roebling had figured on. After each initial rush of air out one side or other, a returning wave would follow, inflowing river water that would stand knee-deep over the work surface until the air pressure eventually forced it out. Blowouts were usually caused by changes in the tide, which in the early stages affected the balance of pressures inside and out and which apparently Roebling had anticipated. But even the wake of a passing steamer could cause enough of a change in the water level to bring on a blowout, and this came as quite a surprise.

  To build up additional weight on the caisson, some of the excavated material was dumped on top, in the spaces not taken up by the masonry. The rest of the material was dumped into carts that ran on inclined tracks down to big scows tied up on the riverside. (Once the caisson was in position this side had been closed in with a cofferdam, as the others were, and docks and tracks and turnarounds for the stone carts had been built.)

  Eventually, when the caisson got down about ten feet below the river bottom, water ceased to come in at all, so tightly was the ground packed about the outer sides. Now, much to their amazement, the boulder crews encountered a new phenomenon. As Collingwood wrote, “It was not an uncommon occurrence in removing a large boulder, that an opening would be made entirely outside the caisson, for three or four feet.” Sometimes when this happened a man might crawl inside, beyond the limits of the caisson, that is, to dramatize the uncanny nature of such a space, not to mention his own nerve.