The Perfect Machine

by Ronald Florence

  There was also a minor scandal at the Mount Wilson Observatory that Hale wasn’t eager to publicize. A bookkeeper, James Herbert, had embezzled $13,000 by forging Adams’s signature. When Adams reported the theft to Hale, Hale reported that Herbert was an “untrustworthy thief” who had previously stolen securities entrusted to him, and that he had only silenced the earlier episode out of concern for Herbert’s wife and children. Herbert wasn’t prosecuted because Merriam feared that “radical elements” in California would pillory the Carnegie Institution for paying a bookkeeper so little as to drive him to theft.

  When Hale heard that Merriam had been lobbying Max Mason, he took another train ride to New York to meet with Mason. Hale explained once again the circumstances that had led to the grant, Merriam’s effort to derail it, the actions of Elihu Root, and the overruling vote of the executive committee that had finally forced the matter. Merriam, he explained, was not opposed to the telescope. He just wanted to “proceed along different lines.” Far from urging stronger cooperation between Mount Wilson and Caltech, he had actually objected that Adams, the director of Mount Wilson, was “too close” to the Caltech people.

  When he got back to Pasadena, in April 1934, Hale answered Merriam’s charges in a sharply worded statement, pointing to the minutes of decision making by the Observatory Council; the de facto membership on the council of Adams, the director of the Mount Wilson Observatory, and his record of attending every meeting; the fine working relationship between the Mount Wilson staffers on the project and the personnel at Caltech; and the use of lenses and other equipment developed as part of the two-hundred-inch project on existing telescopes at Mount Wilson, which had greatly contributed to the research of Hubble and others. “In the future,” Hale wrote, “unless effective cooperation between C.I.W. [Carnegie Institution of Washington] and C.I.T. [California Institute of Technology] is prevented by the action of C.I.W., the advantages to C.I.W. will greatly increase.” Elihu Root congratulated Hale privately on the statement and the progress of the project, commiserating that “an atmosphere of disgust, suspicion, and personal dislike and resentment” will stop the best work anywhere.

  With the depression giving no signs of abating, with 13 million unemployed, with the pitched hatred of the weekly radio broadcasts of Father Charles C. Coughlin and Gerald L. K. Smith drawing even bigger audiences than Amos ‘n’ Andy, and with members of the American Student Union on college campuses openly demonstrating for Communist causes, the Observatory Council’s concerns about publicity for the project were exactly the opposite of Merriam’s craving for public recognition and credit: “In periods of social unrest subversive tendencies may at any moment make their appearance; and in a crisis there might not be time to convince an ignorant public that large sums of money flowing from the great foundations have always been used to useful public purpose.”

  John Merriam had been too consumed with institutional pride to see what was happening outside his Washington office.

  19

  Revelation

  At the end of May 1934 the two-hundred-inch disk was cool enough to view. George McCauley wasn’t willing to predict what he would find when he lifted the cover of the oven.

  After the crowds finally left Corning the night of the pouring, it had taken McCauley and the Corning millwrights and electricians, with advice from the manufacturer, three days to change the lubricants on the screw hoists and retune the linkages so they were able to lift the disk into the annealer. By then the temperature of the disk had fallen to 300°C. McCauley raised the temperature to 530°C, at which annealing normally began, held it for thirty days, then set a rapid cooling schedule, dropping the temperature of the glass by eight degrees Celsius per day.

  With the reporters gone and the attention of the fickle public shifted back to Hitler’s newest threats against Austria and the trial of the accused Lindbergh kidnapper Bruno Hauptmann, McCauley enjoyed the breathing room. He knew the cooling schedule he set, approximately ten times faster than the annealing schedule they had calculated for the disk, wouldn’t produce a satisfactory disk, but with orders for other large disks backed up, he was anxious to free up the annealer. The rapid cooling was an experiment.

  When he opened the annealing oven, the surface of the two-hundred-inch mirror disk was an ugly mask of scars and chunks of refractory brick, the remnants of the floating cores. McCauley wasn’t concerned about the surface. What mattered was the quality of the glass, how well it had annealed. With a polarimeter and a light source, McCauley tested the disk, shining the light source through the glass and measuring the residual strains in different sections of the disk. The strains were approximately ten times what he would have expected for a properly annealed disk. McCauley eagerly reported the good news to Pasadena. If the brief, rapid annealing had reduced the strains to that point, it meant that the full annealing schedule would produce a disk free of strains.

  Enjoying the freedom from outside attention, McCauley began a salvage effort. His tool was a long chisel that had been fabricated in the Corning blacksmith shop. Another man worked alongside him with a portable sandblasting machine. They wore breathing masks and goggles, but after hours of digging and blasting, their ears and hair were filled with fine refractory dust. It took most of a day to dig and blast the remnants of the brick cores out of the surface of the disk. They then vacuumed the surface clean, and the disk was put back into the casting oven and slowly heated for five days, until it reached a temperature of 1100°C, hot enough for the surface to remelt and gradually smooth itself. When the last traces of the broken cores were gone, the disk was again consigned to the annealing oven for another annealing cycle.

  It had taken everyone a long time to unwind from the mad whirl of the casting day. A few reporters still called, and Hostetter, flushed with the attention of that day, had begun intercepting calls, even queries from Pasadena, to answer them himself. Although he had little to do with the technical work on the disk, Hostetter was quick to use the first person plural, or even the singular, in claiming credit. He told the producers of the “March of Time” that he had supervised the pouring, and made sure their broadcast scripts were filled with references to “Dr. Hostetter.” McCauley silently fumed. He wasn’t a man to complain, so friends and even family assumed his sullenness and immersion in work was his way of accepting the responsibility for the accident with the cores.

  For months McCauley worked late at the dining table, reviewing the design of the mold anchors and studying how they could have been built. He went over tables of heat-resisting alloys. When he found a promising alloy, he had Ralph Newman test a sample by heating it in a laboratory furnace. There had been no decision to cast another disk, but McCauley was determined to have the designs and procedures right. He hadn’t been cautious enough before. That wouldn’t happen again.

  A young draftsman named Walter Smith got the assignment of drawing up new rod designs and mold sections. Smith had come to Corning with an EE degree from Ohio State to take a job as an electrician’s helper. His mother and grandmother liked Pyrex pie plates; they thought Corning sounded like a great place to work. Smith was glad to have any job. He had worked on the mirror project earlier, laying out the routing of the monorails that carried the heavy ladles from the glass tank to the mold, then had gone back to the usual design jobs, working on machinery that would be used to mold baby bottles, pie plates, or laboratory flasks. With the crowds gone, the mirror project was another job. Smith had been working on some design problems for a glass dance floor when he was asked to work on the mold anchors. The only difference was that the drawings for the mirror project were done to the unusual accuracy of 1/128 inch to impress the need for accuracy on the mold makers.

  The high-temperature alloy McCauley had selected for new anchor bolts, Calite B-28 from the Calorizing Company, was too hard to shape by milling or forging. Since the anchor rods could not be threaded in a machine shop, Walter Smith drew up plans for rods that would be cast with hooks on the end to a
ttach to the frame that supported the mold. McCauley wanted springs on the anchor shafts to maintain tension as the anchor bolts expanded and shrank in the heat of the mold. The final design needed two springs on each shaft. Walter Smith looked at the specifications and concluded that auto valve springs would do the job, saving the time and money of a special order. He found the smaller ones off the shelf at Rose Buick, the local dealer. No nearby dealer had larger springs in stock, so Smith rummaged for what he needed at Maxx Russ Auto Supply, an infamous junkyard where new, overstock, and used parts—some of dubious origin—shared the boxes and bins. The final design was exactly what McCauley wanted—cautious, overbuilt, using reliable stock parts wherever possible.

  In late September 1934, two months after the disk had been returned to the annealing kiln for a second cooling, it was again cool enough for inspection. The surface scars were gone now, but there were fine checks in some of the lower edges of the ribs. McCauley attributed those to a too-rapid reheating when the surface was remelted; the heat had not been maintained long enough fully to melt and heal the lower extremities of the disk. The clear face of the pale yellow Pyrex showed the missing core holes. Grinding those pockets into the disk would not be an easy job, and the checks in the adjacent ribs were a reminder that a slip of the grinder could fatally crack the disk.

  In the relative quiet of the spring and summer, McCauley and his colleagues weighed the question of whether to pour another disk or to salvage this one. With the new engineering work they had done, McCauley estimated that the chances of successfully casting a new mirror were about the same as the chances of repairing the damaged disk. Neither was certain. In a process as complex as the pouring of a mirror disk, something could always go wrong; the larger the casting, which meant the longer the process, the more likely an accident. The same caution applied to grinding out the pockets in the back of the first disk; a workman with a grinder might discover the fatal check in a glass rib in the last pocket he ground.

  Whichever disk they ultimately used, McCauley had to face the problem of removing the disk from the mold, cleaning away the mold cores from the back, and crating it for shipment to California. The lack of experience at maneuvering a twenty-ton disk of glass suggested that having an expendable dummy to practice the various operations would substantially increase the chances of delivering a usable disk in one piece.

  In the end, the decision was too important for anyone except Amory Houghton to make. Houghton concluded that the cost of making a new disk wasn’t that much greater than the cost of an attempted salvage, and that it was much in the interest of Corning’s reputation as a company to supply a disk to the exact specifications of the Observatory Council rather than a salvaged disk. The Observatory Council agreed. Hale wanted this telescope to be as perfect as they could make it. No formal announcement was made in Corning or Pasadena, but McCauley quietly began plans to cast a second mirror. In addition to the new mold anchors, there would be one other change in the procedure: They would cast this mirror without an audience.

  On the riverbank behind A Factory, the Corning millwrights erected a new structure of steel columns and beams, smaller but similar in construction to the steel building GE had built for its work on fused quartz. Four ten-ton hand-operated chain hoists were suspended from the beams. A careful crew could use the lifts to maneuver forty tons of glass, steel, and brick. At the end of September workmen used the hoists to lift the first two-hundred-inch disk from its mold onto a support of heavy timbers, left over from the dismantling of a section of railroad trestle. There, the two-hundred-inch disk, the beautiful geometry of its back ruined by the filled-in pockets where the cores had broken free, was locked away from prying eyes.

  The 3A melting tank of 715-CF Pyrex had been used through the summer to cast other large mirrors, including an eighty-two-inch mirror for the McDonald Observatory of the University of Texas, auxiliary mirrors for a large telescope Heber Curtis was planning at the University of Michigan, and two sixty-inch mirrors for Harlow Shapley at Harvard. Curtis and Shapley, whose debate had inspired the two-hundred-inch telescope project, both visited Corning, inquiring about the progress of the mirrors, and prying into secrets about the two-hundred-inch mirror.

  By October the Texas and Michigan disks had either been consigned to the smaller annealing oven or sealed away in Sil-o-Cel awaiting a free annealing oven. The big casting oven was cleaned out, the steel flooring of the big mold was rebuilt to accommodate McCauley’s new mold anchors, and the masons went to work building another jigsaw puzzle of bricks to form the pockets and ribs. There had been no publicity about the decision to cast a new mirror, but rumors had been rife. The new official policy was that the company would talk openly about what had been done but make no arrangements for exhibitions or news statements about future plans. Leon Quigley, in the publicity office, did what he could to deflect queries, but with Harlow Shapley carefully leaking what he knew to the press, it soon became impossible to deny the rumors. Houghton and Quigley also thought there should be some documentation of a successful effort, so that the only record of the Corning participation in the project wouldn’t be photos and stories of the failed pouring.

  In late October, Quigley got permission to bring a Pathé News team to the factory with their cameras and power plants. Pathé agreed in advance that they would release no film or printed statements until after a new disk had been successfully cast. The gaffers and crewmen strung heavy cables throughout the blowing room at the west end of A Factory and positioned their cameras to focus on the tank openings. McCauley, trying to be accommodating, watched and explained the various steps in the pouring process. When the equipment was finally ready, the director gave the shout, “Action!”

  “What action?” McCauley asked, explaining that the fires hadn’t been lit in the tank.

  One of the cameramen, eager for some “motion” for his motion pictures, asked to have the mold under the casting oven raised and lowered a few times.

  McCauley shrugged and pressed the start button for the screw hoist. The cameraman shouted, “Faster!” McCauley explained that the screw hoist was operating at its full speed of two inches per minute. The cameramen looked at one another and at McCauley. From the newspaper reports of the earlier pour, they expected something more exciting than a heavy platform moving so slowly the motion would be imperceptible on film. McCauley suggested that the horizontal motion of the mold platform from the casting oven to the annealing oven might be a little more interesting, but the film crew was still disappointed. The director discovered the special building outside where the first two-hundred-inch disk rested on its support of trestle timbers. He suggested that the four hoists could be used to move the disk with enough speed to show up on the film.

  McCauley refused. He wasn’t willing to risk unnecessary movement of the disk for publicity. The compromise was that the Pathé cameramen would run their cameras while the workmen pretended to move the disk. Then with the hoists disconnected from the disk, the cameras would take footage of the chains moving up and down. After a day of filming, the crew left with little to show, and McCauley could settle down to the real business of pouring another disk.

  At the end of October the 3A tank was refired, and the men and equipment were reassembled. The only change to personnel in the pouring crew was the addition of two experienced ladlers from the Blue Ridge Glass Corporation of Kingsport, Tennessee. McCauley had wanted to use an extra crew of ladlers for the first two-hundred-inch disk. Wilson and Ruocco, who had ladled the glass for the earlier disks and were proud of their work, insisted that they could handle the longer workday for the much larger disk, and McCauley had yielded to their arguments. But fatigue had stretched the pouring time for the illfated disk and contributed to a minor accident. This time McCauley wanted another crew to share the ladling. The stakes were high enough to risk hurt pride among the workers.

  The second disk was poured on Sunday, December 2, 1934. There was no special church service that morning. A tiny
group of guests was present, including Heber Curtis, A. G. Ingalls of Scientific American, E. P. Burrell and Henderson from Warner & Swasey, Dr. & Mrs. Francis Pease, and twenty-five representatives of the press, including Howard Blakeslee, science editor at the Associated Press, and William L. Lawrence of the New York Times. Max Mason, who had missed the first pour, came up to watch, along with Warren Weaver from the Rockefeller Foundation. George Hale had been invited as well but was in another of his periodic bouts with the demons. He was well enough, however, to fire off memos urging that the audience be limited lest its presence again cause problems.

  The operation went like clockwork. Hickman and Harris, the new ladlers, filled the ladles with molten glass; Wilson and Ruocco emptied them into the mold. The rest of the crew, now practiced at the operation, opened furnace and oven doors, manipulated the switches on the overhead monorail track, sprayed the ladle rims, filled spray tanks, emptied ladle skins into the special wheelbarrows, and returned the skins to the tank. A total of fifty-six men—carpenters, tinsmiths, laborers, construction workers, millwrights, pattern makers, pipemen, masons, electricians, and foremen—worked on the operation. Their hourly wage ranged from $0.40 for laborers to $1.25 for the experienced mason who rebuilt the openings to the glass tank during the pour. Joe Ruocco, the skilled ladler who poured the glass into the mold, was paid $0.48 per hour for eight hours of work, without a lunch break, that day. His share of the $6 million project was $3.84.

  By two o’clock in the afternoon the work was done. The cores had stayed in place. The two visiting ladlers spent the afternoon seeing the sites in downtown Corning before they went back to Tennessee.

  That evening the disk was transferred to the annealing oven and sealed in. Although the pouring had not been publicized and wasn’t open to the public, everyone in Corning knew what was happening. As a signal that all was well, the Glass Works whistle went off at 11:35 P.M.