Brobeck set the top energy of his Bevatron at 10 BeV because, he wrote in 1947, it seemed "the largest machine that could be made in the near future without departing from the techniques used on machines at present in operation." A more pertinent statement would have replaced '"techniques" with "level of funding": at an estimated price of one to two million dollars for each billion electron volts, his accelerator would cost ten times more than the 184-inch cyclotron. The design presented many technical uncertainties. To help resolve them, the Laboratory built a model to quarter scale, which itself belonged in the same class as Birmingham's synchrotron (1.3 BeV, inaugurated 1953) and Brookhaven's Cosmotron (3 BeV, 1952).

Artist's conception of the Bevatron. The beam injector is at 4 o'clock, the experimental area and emergent beam at 8 o'clock.

Among the problematic features of the Bevatron's design was the size of the gap or aperture in which the magnets constituting the machine's backbone would confine the beam. Forty-four of these vertebrae made up a half doughnut with a mean radius of 18.2 meters; and two such halves, together with the straight sections that joined them, defined the raceway for the particles. (The Bevatron actually built has four curved and four straight sections.) One straight section served to admit the proton beam from a small linear accelerator, which took its feed from a Cockcroft-Walton machine; in the final design, the particles gained 500 keV in the first stage of acceleration and another 9.5 MeV in the second. Brobeck explained that initiation in the Cockcroft-Walton machine would cut loss of the embryonic beam from internal scattering, and that acceleration by the linac would reduce the range of frequency and field strength through which the synchrotron's oscillator and magnets would have to function. The other straight section, which, like the first, would have no focusing magnets, was to contain targets and extractors for the accelerated beam.

Brobeck thought that the mechanisms for injection and extraction and the straight sections without magnetic guidance might cause the beam to oscillate widely around the median orbit through the doughnut halves. Accordingly, he provided for a large aperture between the magnet poles, some 4 feet high and 14 feet wide (in the radial direction); should the beam behave better than expected, the gap could be reduced by changing the pole tips. The plan sacrificed energy for intensity: the bigger the gap, the more particles would survive the many turns necessary to accelerate them; the smaller the gap, the greater the maximum field available to restrain them and the higher the attainable energy.

Despite its advertisement, Brobeck's machine with a 4 ft x 14 ft aperture would have given protons of only 1.5 BeV. Already in the fall of 1947 a British visitor to the Laboratory reported that the plan provided far too large a hole; and in March 1948, at the fixing of the maximum energies of the Bevatron and Cosmotron, Lawrence set his sights on 6 BeV, the threshold for antiproton production. To reach it, however, the aperture that Brobeck planned would have to be reduced by a factor of 14, to 1 ft x 4 ft. Lawrence preferred not to gamble against such odds, and, in ordering the steel for the magnets, bet conservatively on a gap 4 ft x 10 ft. Experience with the quarter-scale model in 1949 inspired another reduction, to 2 ft x 6 ft, for an energy of 3.67 BeV. That was the machine that the Bevatron construction group expected to build in 1950. Subsequent modification might have brought it to 6 BeV, a little above the expected threshold for the creation of the antiproton.

In December 1951 the plan changed again: the gap narrowed to 1 ft x 4 ft; the Laboratory would reach directly for 6 BeV. The decision did not represent a return to recklessness. Experience with the quarter-scale model and with big beams and cyclotron design in the MTA project suggested that the Bevatron would steer protons more accurately than anticipated. Calculations and operating results from Brookhaven's Cosmotron confirmed the inference. The upshot of the two-year interruption of work on the Bevatron by cold-war service was a handsome reward. When completed in 1954, the 10,000 ton synchrotron could accelerate well-behaved protons through 4,000,000 turns in 1.85 seconds without their deviating from the median orbit by more than a few inches. Their journey to 6.2 BeV lasted 300,000 miles.

Huge motor generators with 65-ton flywheels for storing power supplied 100,000 kilowatts to the Bevatron for each accelerating cycle of 1.85 sec.

That stammerer, history, again repeated itself. Just as McMillan's principle doubled the energy of the 184-inch machine and made it competent to make mesons, so the Bevatron's redesign brought it the strength to create antimatter. In both cases the unforeseen in war and science realized Lawrence's goals and guesses. Around the beginning of 1955 several groups, particularly Lofgren's and Segrè's, began to develop instruments to exploit the machine that in retrospect was "purposely planned for forming nucleon-antinucleon pairs." The hunt culminated in October 1955 with the detection of the first antiprotons in the ingenious speed trap designed by Owen Chamberlain, Segrè, Clyde Wiegand, and Thomas Ypsilantis. Of the many messages of congratulations that came, this from Bernard Peters to Lawrence best caught the place of the achievement in the Laboratory's history: "The discovery of the anti-proton is a truly great success of your long-range policy in high-energy work." (See Episode beginning page 80.)

Still another analogy can be drawn between the experiments that detected the pion and the antiproton: both succeeded through application of detector technology then new. Where Gardner adapted C. F. Powell's emulsion technique, Segrè's and Lofgren's groups employed scintillation and Cerenkov counters that had been introduced around 1950, and fast electronic circuits that represented the state of the art. The counters in turn depended on a form of the photomultiplier tube invented during the war for the MED and later developed into a versatile particle detector.

The scintillation counter descends from the visual reading of flashes made by alpha particles striking fluorescent screens, a method long preferred by Rutherford. The Cerenkov counter exploits the fact, discovered in 1934, that charged particles emit light when traveling through a medium faster than the speed of light there. The counter can provide not only a threshold measurement but also a velocity determination, because Cerenkov radiation occurs only in the forward direction and at an angle uniquely determined by the particle's speed and the index of refraction of the medium. The Rad Lab had not been a leader in detection methods before the war and its war work did not, as Robert Thornton put it in 1947, "lead to the establishment of a strong instrument section." The deficiency was corrected and the Laboratory contributed significantly to the improvement of the new detectors and their ancillary electronics. For some purposes, however, the via nova could not replace the slow old way, the cloud chamber and the emulsion. In particle physics as elsewhere a picture may be worth a thousand words or clicks. Students of the tracks of ionized particles had discovered the positron and the mesotron, and cleared up the mess about the mesotron and the Yukon. In 1949 one of Powell's emulsions disclosed still another particle, about half as heavy as the proton, the K meson or kaon. Its strange behavior invited study, its relatives, if any, detection. One of the first achievements of Lofgren and his operating crews was to cause a mixed beam of kaons and pions to issue from the Bevatron. But to realize the full promise of kaon beams and other projectiles from the Bevatron a track detector faster than the cloud chamber and more discriminating than the emulsion was needed.

One day in 1952 Donald Glaser and some colleagues at the University of Michigan were doing physics in a saloon. Someone observed that a stream of beer bubbles made a nice track. Glaser took the suggestion seriously and sought a process in a liquid that could register the path of a charged particle and then quickly expunge the marks. He thought that bubbles might be formed in a superheated liquid much as condensation droplets arise in a cloud chamber. He was right, and in April 1953 he showed a meeting of the American Physical Society pictures of tracks made by the cosmic-ray muons crossing a small vessel filled with hot ether.

The Lab as it appeared about 1955. The Bevatron occupies the central round building, the 184-inch sits under the dome above that.

Alvarez learned about Glaser's device at the meeting and set about developing it in much the same way that Lawrence had transformed Wideröe's accelerator. To make a suitable particle detector, Alvarez had to transcend the intrinsic limitations of Glaser's technique, create several new technologies, walk confidently through unknown terrain, and raise a lot of money.

The first step, the substitution of liquid hydrogen for superheated ether, reduced the detecting medium to one both homogenous and simple, a sea of protons. It brought with it the technical problem of handling large amounts of liquefied gas under high pressure at 20 above absolute zero. By the end of 1953, John Wood of Alvarez's group had made a chamber an inch and a half in diameter and found tracks in liquid hydrogen. He also found that accidental boiling did not impair formation and photographing of the tracks. The point was of great importance: Glaser and others had supposed that useful records can occur only in vessels with smooth glass walls, which give no purchase for formation of unwanted bubbles.

Freed from the size constraints imposed by all-glass vessels, Alvarez's group built a second chamber, 2.5 inches in diameter, with a metal body and glass windows, which worked so quickly that bubbles made at the walls did not reach the active volume before the tracks were photographed. The third model, 4 inches in diameter, began operation at the Bevatron on November 19, 1954. It played a part in chamber development similar to that of the 37-inch machine in cyclotron design. The chamber analogue to the 60-inch cyclotron measured 10 inches, and, like it, was planned by engineers -- in fact by eleven members of the Laboratory's engineering department -- as well as by physicists. The treatment of the large amounts of liquid hydrogen, about six times as much as for the 4-inch chamber, became a special study. It was pursued with the help of the National Bureau of Standards' Cryogenics Engineering Laboratory at Boulder, Colorado, which had been established to help prepare liquid deuterium and tritium for the Eniwetok test of the proto-hydrogen bomb in 1952.

First tracks observed in liquid hydrogen by John Wood, 1954.

While the 10-inch instrument neared completion, Alvarez planned its successor. On January 10, 1955 he proposed a 30-inch rectangular chamber to Lawrence and his senior associates. They agreed, provided development took place "with a minimum of interference" with accelerator work. On reconsideration, Alvarez decided to extend the chamber to 50 inches, long enough to observe the behavior of decay products of new and problematic hyperons. Further consideration altered the dimensions to 72 in. x 20 in. x 15 in.

With this design Alvarez's group jumped from the equivalent of the 60-inch cyclotron to the Bevatron without the benefit of an analogue to the 184-inch synchrocyclotron. The window of the proposed chamber, 8280 cubic inches of optical glass, would have to withstand 100 tons of pressure. The plan required large and dangerous volumes of liquid hydrogen, a huge refrigeration system, metals capable of bearing enormous stresses at temperatures below -253°C and, because only one window was provided, an optical system to illuminate and photograph the interior of the chamber from one side. The plan astonished Lawrence. Even he, accustomed to building on or just over the edge of technology, doubted the bubble chamber could be scaled up from 10 inches to 72 in a single step. "I don't believe in your big machine," Alvarez recalls his saying, "but I do believe in you, and I'll help you to obtain the money."

Alvarez accompanied Lawrence on his next visit to Washington, where they lobbied AEC commissioners Lewis Strauss, Willard Libby, and John von Neumann. In addition to the chamber, the supplicants asked for money to develop a machine capable of reading photographs of tracks and feeding the information into a computer. If the big chamber worked, it would generate data far faster than the unmechanized physicist could digest it. The AEC granted the $750,000 originally estimated. By the end of the year, increases in costs of special equipment, the analyzing magnet, and safety measures had driven the price to $1,250,000, not including $200,000 for an IBM computer to help with data reduction.

A team of engineers and physicists led by James Gow and Paul Hernandez took four years to create the chamber. The optical window, the largest piece of clear optical glass then in existence, was polished by a manufacturer of telescopes. The Rad Lab built the magnet, 115 tons of steel and 20 tons of copper giving 18 kilogauss. The chamber itself, 3.25 tons of austenitic stainless steel, held 12.5 cubic feet of liquid hydrogen. After many trials, Alvarez and Duane Norgren devised the critically important method of one-sided photography, the "coat hangers" retrodirective illumination system. A 15-inch prototype was built to test the optical and refrigerating systems. "It is obvious from whom I learned the 'damn the torpedoes-full speed ahead' attitude that my colleagues and I took in the development of large bubble chambers," Alvarez recalled, tracing his lineage to Lawrence. "It resulted in [our] having an operating 72-inch bubble chamber before Brookhaven -- our most serious competitor -- even had their 20-inch bubble chamber."

The 72-inch chamber removed from its instrumentation.

The 72-inch detector was finished in March 1959. Weighing 240 tons without its refrigeration system, it walked from its place of assembly to its home near the Bevatron on elephant-like hydraulic feet. Its new building had 7500 square feet of space, shop facilities, a crane, two compressors, and safety facilities including a big sphere to catch deuterium released from the chamber in an emergency. A three-megawatt motor-generator set supplied the magnet. The final cost of the project: $2,100,000.

The data reduction program kept pace with the construction of the big chamber. The first successful device, the "Franckenstein" created by a team directed by Jack Franck, worked with stereo pictures from the smaller chambers. The monster projected the tracks, measured them, and punched the results on IBM cards. Its operator first aligned an optical index with the projected track. A scanner, consisting of a photomultiplier and an electronic time discriminator, locked onto the track and directed its own motion along it; the operator controlled the speed of motion and periodically registered track coordinates by pressing a button. The machine could measure five to ten events an hour; the standard method for reducing cloud-chamber data could manage one a day. Franckenstein made many friends when introduced at the Atoms for Peace Exhibition in Geneva in 1958.

An IBM 650 computer, acquired by the Laboratory in 1957, and an IBM 704 on the Berkeley campus completed the data analysis and the first bubble-chamber system. A team led by Frank Solmitz and Arthur Rosenfeld created the programs that reconstructed the tracks supplied by Franckenstein and compared them with those of hypothetical interactions. It took seven physicists two years and more to train IBM's electronic brains to interpret impulses sent through Franckenstein's optic nerve. It was none too soon. In an experiment lasting several months, pions passing through the chamber produced some 80,000 hyperons and 4,000,000 other interactions of possible interest. Still, two machines and 30 persons could analyze only 200 events a day, a very small fraction of the Bevatron's bounty.

Operator maps particle tracks with Alvarez Scanning and Measuring Projector.

The "McCormick Reaper," devised by Bruce McCormick soon after Franckenstein came to life, had the potential to enlarge the harvest. The photomultiplier of its traveling sensor was to send signals directly to a computer. The scheme was so much in advance of the computer art of the day that the project had to be abandoned. A second attempt to realize the potential of the spiral-scan method also foundered on technical difficulties. But in 1963, on the third attempt, with Jack Lloyd as chief engineer, the Spiral Reader began its work. The number of measured events jumped from 80,000 in 1962 to 300,000 in 1965. By 1968 the Alvarez team could measure and analyze 1.5 million events a year. To keep pace, the Laboratory updated and expanded its complement of computers: an IBM 709 in 1960, a 7090 in 1961, a CDC 6600 in 1966, another in 1967, and so on. The 1,500,000 events measured in 1968 were about a thousand times as many as the Laboratory could have handled twenty years earlier.

The bubble-chamber films contain many items of the first importance. Probably the most exciting event in theoretical physics in 1956 was the discovery by T. D. Lee and C. N. Yang that parity does not hold in certain cases. Confirmation came first from examination of beta decay by C. S. Wu and others at the National Bureau of Standards. At the Laboratory, F. S. Crawford, M. L. Stevenson, and others in Alvarez's group passed pions from the Bevatron into their 10-inch chamber and observed that the decay of the resultant A hyperons (created by negative pions striking protons) also violated parity. They found in the bargain that L decay does not respect charge conjugation, which requires a reaction involving a set of particles also to hold if each member of the set is replaced by its antiparticle. The experiment had the inconvenience that both the L and the kaon produced with it have no charge, and so can be detected only by their decay products. The relative ease and frequency with which these ghost-like occurrences could be found on bubble chamber records were a striking demonstration of the value of the new detector. A similar performance occurred in one of the first experiments run with the 15-inch chamber. Alvarez and his associates admitted negative kaons into the vessel and uncovered a new particle, the X0, although neither the X0 nor its decay products (L + ¼0) nor the particle created with it (K0) leaves a track in the chamber.

A negative kaon entering from below produces an uncharged kaon and an uncharged X0 that, in turn, decays into two uncharged particles (L+p0). The dotted lines in the inset follow the trackless participants.

Perhaps the farthest reaching of the discoveries made with the Bevatron were the so-called "resonances" or energies at which fleeting combinations of particles occur. The first case found at Berkeley (Fermi had noticed one earlier) concerned the L hyperon and two pions. Bogdan Maglich's plot of the numbers of the two pions against their kinetic energies showed a strong peak, where, it was supposed, the total energy of the L and one of the mesons allowed them to stay together for the time it takes light to travel a few nuclear diameters. They called this brief encounter (or the compound constituted by it) the Y*(1385), the number signifying its resonant energy. It aroused great interest when reported at the Rochester Conference on High Energy Physics in 1960, for it implied the possibility of creating a spectroscopy for the heavier elementary particles.

In the ensuing rush, the Berkeley group, working with the 15-inch chamber, found the first kaon resonance, K*(890), and another hyperon one, Y*(1405), and still others; and some were detected elsewhere using film from the 72-inch chamber, for example, the X*(1530), discovered by Harold Ticho of UCLA. It has a special interest, since it, like the K* and Y*, perfectly fit the predictions of the "eight-fold way," a scheme of particle spectroscopy developed independently by Murray Gell-Mann and Yuval Ne'eman. From the masses of the Y*(1385) and the X*(1530) and the rules of the way, the mass or energy of another particle, the W-(1676), could be inferred. Since this particle, named in the belief that it would be the last of its kind, would confirm a central point in the eight-fold way, it was eagerly sought. Unfortunately its creation lay beyond the capabilities of the Bevatron; it was found at Brookhaven in 1964.

The great success of the liquid hydrogen bubble chamber overshadowed advances in detectors made elsewhere in the Laboratory around 1960. Wilson Powell's group, for example, made a 30-inch propane bubble chamber, the output of which they analyzed with their own computer programs. Clyde Wiegand of the Segrè-Chamberlain group and others continued to shorten the resolution times of counting systems. W.H. Barkas' group automated measurement and analysis of nuclear emulsions. A team under W.A. Wenzel of the Lofgren group introduced spark chambers to the Laboratory. This technique, first used successfully in 1959, exploits the sparks that mark the passage of a charged particle between closely spaced parallel electrodes. An automatic scanner for the spark chamber was devised by Denis Keefe and Leroy Kerth.

The rise of detecting equipment to something like the status of the accelerators they served may be traced in the organization of the Laboratory. New groups with special missions appeared: one for the bubble chamber operation and development under Gow, another for data analysis development under Hugh Bradner, a third, under H.H. Heckman, to advise on photographic problems. An electronic computer group developed under David Judd. The scanners too were organized into teams, their outputs compared and their operations analyzed. Particle detection in the age of the bubble chamber came to resemble factory production.

Physicist Angelina Galtieri consults log with operator in control room of the Bevatron.

Other institutions followed the Rad Lab's lead. In April 1958, at an international meeting at Imperial College, London, the British declared plans for a 60-inch chamber and the Russians and the CERN physicists contemplated going to 80 inches. They had not yet accomplished much by Berkeley standards. CERN had a working chamber of 4 inches, Italy one of 6 inches, and the Soviet Union one of 15 inches that did not function. The Alvarez group helped by distributing Bubble Chamber Engineering Notes all over the world. As Gow found out at the meeting at Imperial College, however, the Europeans tended to design very conservatively. None gambled in the Berkeley style.