greg goebel / public domain
* VECTORS is an original newsletter of fact and commentary on aerospace, technology, science, and historical topics.
* In the early days of aviation, all flight was by the seat of the pants. Pilots navigated by following railroad tracks on the ground. Nobody but a fool flew in really bad weather or at night. It was dangerous on the face of it; worse, when visibility went to zero, pilots simply couldn't fly straight or level and had a nasty tendency to fly into the ground.
A visionary named Elmer A. Sperry wanted to change things. Sperry, a farm boy from New York state with a knack for making useful inventions in a range of fields, had made his name in putting the spinning gyroscope into harness. The "gyro" had been invented in the mid-19th century, and given its name by the well-known French physicist Leon Foucault -- "gyroscope" means "turn seer", by the way. In 1894, an Austrian inventor named Ludwig Obry used the gyroscope to build a system that allowed a torpedo to stay on course after launch. Sperry thought that gyros would be useful to help make ships compensate for rolling seas -- a particularly important requirement for naval gunnery.
He filed a patent on the concept in 1908 and got support from the US Navy for further development, and set up the Sperry Gyroscope Company in Brooklyn in 1910 to conduct work on gyrostabilization systems and gyrocompasses. His gyrostabilization system featured a pendulum as a sensor to sense the beginning of a roll, activating an electric motor to swing the gyro preemptively. This "active" response proved much more effective than the "passive" response of competing gyrostabilization systems that lacked the pendulum.
* By this time, aircraft were becoming commonplace, though only daredevils flew them. That label fit Elmer Sperry's son Lawrence perfectly. Elmer was an inventor; Lawrence was a man of action, becoming an enthusiastic aviator, even marrying a glamorous movie star, Winifred Allen -- who commented that life with Lawrence was "more exciting than anything in the movies." It might have seemed that this drastic difference in temperament would have set father and son against each other, but in practice it produced a highly effective team.
In 1912, influenced by his son, Elmer Sperry began work on a gyrostabilization system for aircraft, the particular application being the new flying boats that the US Navy was obtaining from the Glenn Curtiss company. While the Sperry shipboard gyrostabilization system only compensated for roll, his aircraft stabilization system would also have to compensate for "pitch", the nose-up / nose-down movement of an aircraft. That meant a dual or "two axis" stabilization system. The pendulum active response system was retained in the aircraft gyrostabilization system, though its relationship with the gyros was more complicated, since an airplane often banks in flight. In any case, the gyrostabilization system was linked by servomotors to the aircraft's flight control system to keep the aircraft flying right.
Lawrence Sperry performed a "hands off" demonstration to the public at the controls of a Curtiss flying boat at an exhibition in Bezons, France in 1914. His mechanic, Emile Cachin, actually got out of the back seat and walked out on the wing to show how the aircraft could maintain the straight and level. The demonstration won a 50,000 franc prize.
* That was about as far as the idea went at the moment. One reason was because the Sperry aircraft autostabilization system was a classic demonstrator, in no way ready for practical use. Another reason was that few aviators of the era were really keen on the idea of letting a machine pilot for them. Sperry and his sponsors focused on the idea of using the gyrostabilization system provide indicator outputs to the pilot to tell him how to fly right.
That basic concept had already been embodied in the "turn indicator", which was introduced in 1917. It consisted of an air-driven gyro attached to a spring. On a turn, the gyro would stay level with where it had been before the turn, only to be pulled back to the new heading by the spring. A dashboard indicator needle shifted during the turns and then came back to normal. The turn indicator did not catch on quickly, mostly because it was contrary to the seat-of-the-pants aviation culture of the era. However, everybody was aware that seat-of-the-pants flying was dangerous, and postwar experiments showed precisely why. The experiments involved two-seat aircraft with dual controls. One cockpit was covered with a hood to see how a pilot flew blind, while the other cockpit carried a backup pilot to keep the blind pilot out of trouble.
Why blind flying was so dangerous would be obvious to anyone who's ever been on a "virtual thrill ride", the kind of theme-park theatre in which the seats tilt and shift around while a screen displays the "movements" of the thrill ride. From a strict physics point of view, there's no distinction between the force of gravity and the forces set up by acceleration: the virtual thrill ride can give the impression of a hard turn simply by tilting the seat to the side, or of acceleration forward by tilting the seat back. Aircraft don't generally tend to fly straight and level on their own, and in the blind flying experiments the hooded pilots simply could not sort out the influence of gravity and acceleration. Using a turn indicator meant that a pilot had to "trust the instruments" and not "trust his instincts". That was a very hard conversion to make. Pioneering aviator Charles Lindbergh managed to save his skin twice in bad weather on mail flights in 1926 by trusting the turn indicator, and he couldn't have made his solo trans-Atlantic flight in 1927 without it.
* The turn indicator was really not enough, though -- it just didn't provide enough information to make blind flying really safe. By that time work was moving forward on something better. In 1926 Harry Guggenheim, a US Navy pilot in World War I and heir to the copper mining fortune established by his father Daniel, set up the Daniel Guggenheim Fund for the Promotion of Aeronautics. Harry was president. The fund provided grants to universities for aeronautical research and sent Lindbergh on a US tour to promote the field. He helped fund work on the emerging airline business, and in 1928 set up a flight laboratory at Mitchel Field on Long Island, not too far from the Sperry plant in Brooklyn. The lab's main emphasis was on blind flight. The test pilot was James H. Doolittle.
Jimmy Doolittle was both a wild barnstormer and one of the pioneering aeronautical engineers, both a man of action and a scholar. His goal in the project was to conduct a completely blind flight from takeoff to landing. He figured out how to use radio-direction finding to get a bearing on an airfield, though he had to use a long field and have reinforced landing gear to handle landing by guesswork. He also obtained a vastly improved barometric altimeter from the German-American instrument maker Paul Kollsman. A new "turn and bank" indicator had been invented, adding a ball in a curved tube that would move out of center position when a banked turn wasn't done right. Doolittle didn't think it was adequate, and was also not happy with magnetic compasses. He wanted a gyroscopic system that could show bank and pitch, and provide an accurate heading reference in a maneuvering airplane.
Elmer Sperry had his son Elmer JR develop two new instruments. (Lawrence was dead by this time, finally having had a crash he didn't walk away from; his father would never stop mourning him.) The first new instrument was the "artificial horizon", which had a nose-first profile of an airplane backed up by a bar that tilted to show the angle of the aircraft to the horizon. The beauty of the artificial horizon was not just its ability to perform the feat of showing where the true level was, but in presenting it to the pilot in an intuitive fashion -- the instrument providing a seat-of-the-pants sort of input to the pilot. It was much easier to trust than the old wavering needle of the turn indicator. Doolittle was impressed, saying that it was "like cutting a porthole through the fog to look at the real horizon." The second innovation was a gyrocompass, linked to a magnetic compass. The direction reference could be set on the gyrocompass using the magnetic compass while the aircraft was flying straight and level under benign conditions; when the aircraft was maneuvering and the magnetic compass was unstable, the gyrocompass could be used to give a heading reference.
The full system was ready for flight test on 24 September 1929. Doolittle took off hooded, flew a racetrack course, and then landed 15 minutes after takeoff. His safety pilot kept his hands up to make sure observers knew Doolittle was flying the plane. It was a persuasive enough demonstration to ensure that the new blind-flying instruments caught on.
* Aircraft now had the ability to fly blind, but much more work still needed to be done. Radio technology had evolved by this time to the point where it could be put into any airplane, and radio communications as well as radio-direction finding also helped pilots fly under difficult conditions. In addition, landing aids were developed in which two radio beams were transmitted in an overlapping fashion. One radio beam transmitted dots, the other dashes, and when the pilot heard the two turn into one continuous tone, he knew he was on the proper "beam" for landing.
The increasing amount of instrumentation threatened to overload pilots, however. Elmer Sperry JR responded to the problem by developing the first practical autopilot system. The new Sperry autopilot compensated for the tendency of gyros to "drift" from their reference over time due to friction through use of a pendulum system that provided corrections when the aircraft drifted off its heading. In 1933, aviator Wiley Post flew around the world in 8 days in a Lockheed Vega 5C, the Winnie Mae. It was fitted with a prototype of the Sperry A2 gyroscope, allowing post to navigate or even take a nap while the plane flew itself. Autopilots began to appear in airliners the next year. They couldn't keep an aircraft on course for long periods of time, but they could allow the pilot to take his hands off the controls when necessary.
* From World War II, aircraft blind-flying systems increased greatly in sophistication and flexibility. Jimmy Doolittle's blind landing scheme amounted to groping his way down until hit ground, and something better was needed for general use. The military developed a "ground controlled approach (GCA)" system in which the aircraft was tracked by radar and an analog computer system, with the ground controller talking the aircraft down to the runway. The civilian "instrument landing system (ILS)" was also developed and is still in common use today, in which the end of a runway features two radio arrays: one to tell the pilot which side of the runway centerline the aircraft is on, the other to tell if the aircraft is on a proper descent path to the runway.
More sophisticated gyrocompass and gyrostabilization systems were also developed after the war, mostly to support the long-range missiles developed during the Cold War: there was no sense in blasting a missile from continent to continent if it couldn't even hit a city when it came down. Improved autopilot systems using electronic controls instead of electromechanical controls had been developed during the war; Charles Stark Draper then developed an improved gyro system, using precision bearings and supported in a fluid, to greatly reduce gyro drift. In 1955, a Boeing B-29 Superfortress bomber flew all the way across America using a Draper navigation system, and was only about 3.2 kilometers (2 miles) off course when it got to Los Angeles.
Draper would later develop even more accurate gyros using gas bearings, and in the late 1960s gyrocompass systems using laser gyros -- basically lasers operating in a loop -- went a step farther beyond that. Another useful postwar innovation was the "VHF Omnirange / Distance Measuring Equipment (VOR/DME)" radio navigation aid, which involved a radiocompass system whose signal could actually give the compass bearing of the aircraft relative to the transmitter, along with a system to give the range of the aircraft. Yet another innovation of the 1950s was "positive control", with the US Federal Aviation Administration (FAA) marking out three transcontinental corridors in 1958 in which visual flight was not permitted. Aircraft were tracked on radar and kept on course by ground controllers.
In the 21st century, aircraft navigation is being revolutionized by the use of the "Global Positioning System (GPS)" navigation satellite constellation, which can give accurate locations of airliners for long distance flight. A "Wide Area Aircraft Surveillance System (WAAS)" is now being implemented for general aircraft navigation, with an enhanced "Local Area Aircraft System (LAAS)" being developed to permit GPS-aided landings. Aircraft positions are to be tracked with a GPS receiver and a satellite relay system, known as "ADS-B", instead of radar. The technology is available now to permit a jetliner to conduct a complete flight from start to finish without a pilot on board -- though it may be some time before passengers will be comfortable with that idea.
* In the 1990s, chemists learned the trick of rolling up the hexagonal grid of carbon atoms known as "graphite" into tiny balls and tubes. It was an impressive accomplishment, though at the outset the reaction was sometimes along the line of: "Cute, but what good is it?" Those working in the field understood perfectly well that it takes time to exploit a major technical breakthrough, and believed that such a leap in technology was going to have a wide range of applications.
We are now beginning to see applications emerge. Carbon "nanotubes" seem to have a particularly promising future in electronics. Chemists can tweak them to act as conductors, just like copper threads, or semiconductors, like silicon. Fabrication of carbon nanotube devices also appears to be straightforward: a solution of the tubes can be sprayed on various substrates, including plastic to form a nanotube layer, and additional layers can be applied over the base layer to provide electronic functions, such as emitting light. While nobody expects to make a quad-core computer processor chip with carbon nanotube technology any time soon, it does seem well-suited to applications such as "electronic paper" for roll-up displays; chemical sensors; wearable electronic devices fabricated into clothing; solar power cells printed onto rooftop tiles; or radio-frequency identification (RFID) devices. These would be relatively uncomplicated, built using cheap, flexible, lightweight, and (when needed) transparent electronic circuitry.
* Carbon nanotubes were actually discovered several decades ago, but nobody paid any attention. In 1991 Sumio Ijima, a chemist at Nippon Electric Corporation (NEC) in Japan, rediscovered them and realized their potential. Carbon nanotubes can have conductivities comparable to copper and current densities (maximum current capability per unit of cross section) a hundred times better. The nanotubes are also physically robust, capable of being bent without damage and relatively inert chemically.
Carbon nanotubes are fabricated by vaporizing coal with an electric arc or a laser, and then using a catalytic process to convert this "carbon plume" into various sorts of carbon molecules that amount to soot. This soot includes various forms of spherical graphite assemblages, known as "buckyballs" or "fullerenes" after inventor R. Buckminster Fuller, and carbon nanotubes -- not just those with single tube walls but with multiple concentric tube walls. Sorting out this mix is troublesome and so carbon nanotubes are currently very expensive, but given a mass production application there will be an incentive to scale up carbon nanotube manufacture to an industrial process.
Prototype carbon nanotube transistors are actually as fast as or faster than silicon transistors, but they can be very difficult to make, and don't have uniform properties because of variations in nanotube configuration. One way to get around the variations is to lay down a thin film layer of carbon nanotubes. The conductivity of this "nanonet" is much more predictable, since it's an average of the nanotubes that make it up, providing conductive paths from one side of the nanonet to the other. If some of the nanotubes are broken, the others easily take up the load. The nanonet is so thin and sparse in its coverage that its transparency can easily exceed 99%. It is also flexible and resists damage from physical shocks.
* Laying down a carbon nanonet is not a trivial proposition, and early efforts to do so left much to be desired. The idea of simply dissolving the nanotubes in alcohol or some other solvent and then spraying them on didn't work very well in practice, since the nanotubes tended to clump together. Adding surfactants -- soaps -- helped keep the tubes apart by coating the tubes with a non-stick layer. However, once sprayed on the surfactants tended to interfere with the operation of the nanolayer. Several years of tinkering with combinations of solvents, surfactants, and processing technologies has finally led to the ability to produce workable nanonets on substrates at near room temperatures. The transparency of the nanonet can be controlled by adjusting the thinness of the layer.
The first nanontube-based field-effect transistors were built in 2003, though they required high temperatures that ruled out plastic substrates. Transistors on flexible plastic substrates followed quickly, followed by transparent devices useful for displays. These devices required relatively high voltages for operation, making them unsuitable for battery-operated systems, but low-voltage devices have now been fabricated.
Considerable research is being performed on solar power cells incorporating nanotube electrode layers sandwiching a photovoltaic material layer made of polymers or other flexible materials. Such devices have low efficiencies, only about 6% at most, but they are potentially very cheap and could be fabricated on a wide variety of substrates, such as roofing tiles. Other devices being investigated include touchscreen overlays for displays, consisting of two transparent conductive nanotube layers separated by insulating spacers; and batteries or other power technologies. Devices will be made by inkjet or offset printing technologies. Right now, carbon nanotube technology is where silicon transistors were in the 1950s. Progress in the field is proving rapid and carbon nanotubes appear poised to take off.
* CHICAGO PILE ONE: By late June 1942, the OSRD and the US Army Corps of Engineers were beginning to consider the groundwork for a full-scale atomic bomb project. Civilian contractors were tapped to build production facilities. On 27 June, Compton conducted a meeting of Met Lab research personnel in Chicago -- including Fermi, Seaborg, Szilard, Teller, Wigner, and others in attendance -- to inform them of developments. Not everybody in the audience was entirely happy with the situation as reported, since the scientists were now being confronted with the idea of having to knuckle under to an industrial, or worse a military, organization. Szilard in particular raised objections.
The work went ahead regardless. Seaborg's group continued to work on isotope separation, while Fermi and Szilard worked on improved piles. Szilard's persistent efforts in obtaining purer graphite, uranium oxide, and uranium metal began to pay off, with a test run in July obtaining a value of k = 1.004 and another in August giving a value of k = 1.014.
In the meantime, Robert Oppenheimer was assembling a study group at the University of California, Berkeley, to lay the theoretical groundwork for building an atomic bomb. Hans Bethe and Edward Teller were in attendance; by this time, Teller had reconsidered his rejection of the fusion bomb, and ideas were also floated for how to build the "Super Bomb". By the end of summer, a report was produced that went upstairs to Vannevar Bush. The report said that an atomic bomb could be available as early as March 1944, and that it could have a yield of up to 100,000 tons of TNT. The paper also suggested that a "Super" bomb could be a hundred times more powerful. The days of equivocation about the bomb program were now completely over. On 17 September, the Army assigned Colonel Leslie Groves to take charge of the atomic bomb project.
Groves was a big, beefy, man who suggested a bulldozer, with a streak of arrogance a mile wide. Most who worked for or with him hated him, but they also respected him, because he was boundlessly energetic, highly decisive, and extremely competent. He had been working as chief of operations for General Brehon Somervell on construction of the Pentagon building. Now that the Pentagon project was winding down, was looking for a combat command and the prestige that came with it. Being told he had to remain stateside and work on engineering projects was a big let-down.
When Somervell told Groves: "If you do the job right, it will win the war." -- Groves replied: "Oh." Groves was promoted to brigadier general to give him the necessary authority, which took some of the sting out of an unwanted duty assignment. The effort had already been given the cover name of the "Manhattan Engineering District", which to any snoops would probably suggest a water or harbor project for New York City. It would become more informally known as the "Manhattan Project".
Groves appeared before a group of senior officials in Secretary of War Stimson's office on 23 September, the day his promotion became official. Bush, Conant, Somervell, and Army Chief of Staff George Marshall were among the attendees, and Groves' take-charge no-nonsense attitude made an excellent impression. When Stimson suggested a nine-man oversight committee, Groves talked him down to a committee of three, and when the meeting finally slowed down, Groves asked to be excused to take a train to Tennessee, where he was to inspect a site being lined up for the project. Somervell told him later: "You made me look like a million dollars."
* Between 15 September and 15 November 1942, the Met Lab built at least 16 exponential piles and performed hundreds of tests. At the end of this cycle, the group was obtaining values of k = 1.04 with uranium oxide and k = 1.07 with uranium metal. The work demonstrated that a practical pile capable of supporting a controlled fission reaction could be built with the materials at hand, and it could be operated safely. Work began on the full-scale pile, unimaginatively named "Chicago Pile One", on 16 November 1942.
Pile One was assembled in a squash court under the west grandstand of the University of Chicago's football stadium. The pile was to be about 7.6 meters (25 feet) wide and 6.1 meters (20 feet) high. It would contain 360 tonnes (400 tons) of graphite, 36 tonnes (40 tons) of uranium oxide, and 5.4 tonnes (6 tons) of uranium metal, provided by a group at the University of Iowa that was learning to produce the metal in bulk quantities. Nobody knew how far the Germans were along in their efforts to develop a Nazi bomb, so work went on around the clock in two 12-hour shifts. The day shift was run by Walter Zinn and the night shift by Herbert Anderson, colleagues of Fermi's from his Columbia days.
The graphite came in long blocks and had to be sawed, ground down, and drilled with holes, making the work extremely dirty. The uranium oxide was pressed with a hand tool into 22,000 egg-shaped slugs. The graphite bricks were laid down meticulously inside a wood scaffold, with the uranium oxide slugs inserted in one layer of bricks, followed by a layer of pure graphite, then another layer of bricks with slugs. Channels were cut into the pile to allow the insertion of cadmium control rods, which would strongly absorb neutrons and damp a chain reaction until removed. Fermi worked as an administrator, tracking the construction of the pile on a daily basis. With his accustomed meticulousness, he had careful measurements of the neutron flux from the pile performed at the end of every shift.
The 52nd layer was laid down on 30 November. Fermi's calculations based on the measurements taken indicated the pile would reach k = 1 when the 57th layer was laid down. On the frigid morning of 2 December 1942, Pile One was ready to conduct the first controlled fission experiment in history. Nobody thought the pile was capable of exploding, but there were worries that if the chain reaction got out of control, the pile might break down and scatter radioactive debris over the campus. Compton had not informed the president of the University of Chicago of what exactly was going on, knowing that the university administration would have refused to permit it.
Fermi believed the risk was small and approached the operation with his typical care, directing his team through careful procedures every step of the way. The cadmium control rods were pulled out of the pile, until there was only one left, under the control of George Weil. There was an automatic safety rod that would be plunged into the pile if the chain reaction started to get out of control, plus another that was held in place by a rope that would be cut by a man with an axe. Three men stood over the pile with buckets of a cadmium-salt solution to douse the pile as a last-ditch defense.
The final control rod was pulled out by stages, with measurements at each stage. The neutron flux increased until an automatic safety rod slammed into place. Fermi shut down the pile and broke the session for lunch. After lunch, they resumed the procedures. At 3:42 PM, Weil pulled the final control rod out, and the neutron flux rose in a slow and controlled fashion. Fermi checked his figures with a slide rule, smiled, and said: "The reaction is self sustaining." They ran the pile for 11 minutes and then shut it down. The group broke out a bottle of Chianti that Wigner had brought in, and drank it with paper cups. Pile One was working as expected, providing the physicists with an easily controlled fission source. Its power output was only half a watt, but it was a superb research tool for bigger and better things.
Compton went back to his office and called Conant in Washington: "Jim, you'll be interested to know that the Italian navigator has just landed in the New World. The Earth was not as large as he had estimated, and he arrived at the New World sooner than he had expected."
Conant was excited: "Izzat so? Were the natives friendly?"
"Everyone landed safe and happy." Not everyone was happy, however. Leo Szilard went up to Fermi afterwards and shook his hand, saying: "This day will go down as a black day in the history of mankind."
* CREATING LOS ALAMOS: Les Groves wasn't unhappy with the news about Pile One at all, and in fact he was probing around for a site where bomb development could be put into high gear. He was also looking around for a scientist to become technical head of the project. Partly due to the lack of anyone else, he settled on Robert Oppenheimer.
There were certainly valid reasons for selecting Oppenheimer. A son of a wealthy family of secular Jews, "Oppie", as his friends knew him, was noted for his extreme brilliance and ability to grasp difficult concepts quickly. He tended to be impatient and sloppy with details, but had learned to delegate the "grunt work" to his grad students, many of whom were dedicated to him. He was energetic and could be extremely charismatic.
That, however, was the "good" Robert Oppenheimer. There was also the "bad" Robert Oppenheimer -- a pretentious intellectual with his own streak of arrogance and little patience with inferior intellects, giving him a nasty tendency to make lifelong enemies without good reason. He was smart enough to realize he could be "beastly" and tried to restrain it, though with spotty success. He also a bad diet, poor teeth, liked to drink martinis a bit too much and the details of his sex life didn't bear too much investigation. Worse, from the Army's point of view, he had a background of flirtations with Communism. He'd never been a party member and interviews with Army intelligence folks demonstrated he wasn't a security risk -- though it would turn out later he wasn't completely truthful with the Army men.
There was something tragically confused about Robert Oppenheimer. Physicist Isaac Isidore ("I.I.") Rabi -- a manager at the MIT Rad Lab who would work with the atomic bomb project as a consultant on occasions -- once said of Oppie that he couldn't make up his mind as to whether he wanted to lead the Knights of Columbus parade or head the local chapter of the B'nai B'rith. Oppenheimer flitted around with too many ideas, dabbling in this and that, and those who knew him were frustrated to see a man who they saw as potentially one of the world's great physicists, an American Einstein, failing to really live up to his potential as a scientist.
Oppenheimer's confusions would catch up with him eventually, but war and responsibility brought the good Oppie to the front. Even those scientists who had misgivings about him initially, many thinking that the project technical leader necessarily had to be a Nobelist, would eventually say that Groves had made a brilliant choice. Oppenheimer would demonstrate outstanding leadership capability, proving decisive and balanced while able to hold all the necessary technical details in his head. If he wouldn't go down in history as one of the greatest physicists, he would achieve fame in his own way.
* Groves' staff had put together a list of potential atomic bomb development sites in the US West. When Oppenheimer visited some of the sites in the New Mexico desert regions in mid-November, he fell in love with the scenery at a mesa site, inhabited only by the Los Alamos boy's school. It wasn't technically the best site, but Oppenheimer liked the deserts -- and if it made Oppenheimer happy, Groves felt that was good enough. The land was purchased before the month was out. Groves arranged for the University of California to operate the installation. Construction of barracks and other facilities began immediately; the general was wasting absolutely no time. Oppenheimer began to recruit researchers to the cause, directing people who had signed up to get their hands on the needed equipment.
The recruiting effort quickly ran into a big snag: military security. Groves wanted to draft all the physicists into the Army as officers. They were not as a rule keen on this idea, or in fact on the idea of a security regime in general. Most understood the necessity of it, but felt that it would be impossible to build the atomic bomb if everything was compartmentalized and security snoops were around in swarms. Oppenheimer managed to persuade Groves to relent on drafting the physicists, and arranged for the lab to operate in an open fashion -- internally. It wasn't open to the outside world, being contained by a double barbed-wire fence and security checkpoints. Some of the staff who had fled Europe found it unpleasantly reminiscent of a concentration camp, but it was what had to be done.
Oppenheimer arrived in Santa Fe, New Mexico, in mid-March 1943 with a group of his aides. By the first of April, about thirty researchers of the hundred initial hires were present; the Los Alamos facility was still being thrown together, so Groves had commandeered isolated local guest ranches to house the them. Oppenheimer set up "orientation" sessions, with his student Robert Serber telling the research crew about such progress as had been made on building the atomic bomb to that time. Serber provided sketches of a number of different bomb configurations, the two most significant in hindsight being the "gun" type configuration, in which a slug of fissionable material was shot into a "core" with a matching hole to reach an explosive critical mass; and an "implosion" type configuration, in which wedges of fissionable material were blasted together to form a sphere with a critical mass.
By the end of April 1943, the operation was fully staffed and Los Alamos was a going concern. Many of the researchers didn't like the barbed wire, but were willing to put up with it -- there was a war on, after all. There was also grumbling about delays and Army "hurry up and wait", but not because of the inconvenience. Every day added to the delay in building the atomic bomb and putting an end to the war meant that more people would die.
TO BE CONTINUED
* HALDANE & THE PEPPERED MOTH: Although early genetics research seemed to be at odds with Darwinism, from the mid-1920s the two concepts began to draw together, largely through the work of the Scots biochemist John Burdon Sanderson Haldane (1892:1964), the English geneticist Ronald A. Fisher (1890:1962), and the American geneticist Sewall Wright (1889:1988). Together they formed the quarrelsome triumvirate of a reborn Darwinism.
J.B.S. Haldane was by all accounts overbearing and contrarian, inclined to reject religion and embrace leftist causes. He was also brilliant, and in a series of ten research papers published from 1924 into 1934 showed that Mendelism -- genetics -- and Darwinism were not merely compatible, they fit together like hand in glove. The idea that the two worked together was not new, having been around for about two decades previously, but Haldane was able to demonstrate its plausibility mathematically. He started with the example of the peppered moth.
The pale-colored peppered moth was long known to British insect collectors, but in 1848 one was found with a dark coloration. In 1896, the British naturalist J.W. Tutt (1858:1911) discovered that while the pale peppered moth was still predominant in the English countryside, the dark peppered moth was overwhelmingly predominant in sooty industrial areas. The conclusion was that dark peppered moth was better camouflaged for life in industrial areas, allowing it to avoid predators -- birds in particular -- and giving it a selection advantage over the pale peppered moth.
Haldane applied statistical methods to the peppered moth, pointing out that if dark camouflage gave it a fairly modest survival advantage in an industrial environment, the dark form would outbreed the pale form and predominate within a fairly small space of generations. He also showed that simple mutationism without natural selection couldn't do the job, since it would demand that every fifth moth born to pale parents be a dark mutant, which was absurd.
* Haldane's mathematics, not so incidentally, did not constitute a proof of evolution by natural selection -- a point worth emphasising because people are sometimes impressed by elaborate mathematics without realizing exactly what it can and cannot do. Mathematical modeling, or its modern extension, computer simulation, is merely a description of how the real world is believed to work, not a proof of how it works. Mathematical models start from fundamental assumptions and show what the results will be on the basis of those assumptions. Whether those results are accurate can only be determined by observation of the real world. In the physical sciences, theory is always subordinate to practice. If a model gives results that don't match observations, then something's wrong. Sometimes double-checking the observations shows that earlier observations were bogus and that the model is correct; but if double-checking confirms the earlier observations, then the model's broken.
This is not to imply that mathematical modeling is useless, only that it is constrained from beneath by its assumptions, and from above by its need to conform to observations. Isaac Newton's law of universal gravitation -- that gravity is proportional to the product of two masses and inversely proportional to the square of the distance between them -- is a simple mathematical model derived from observations of the motions of heavenly bodies through Newton's intuition, with its generality being demonstrated by the fact that once Newton had the law, he was able to correctly predict the next arrival of Halley's Comet. It is still used today to plan the flight of interplanetary spacecraft. Newton's law of gravity is extremely useful, but the equation doesn't amount to a proof of anything: it remains no more valid than observations show it to be, and in fact it isn't valid under extreme circumstances that Newton had no knowledge of, such as the "warped space" around a black hole.
It should be noted that when the results of models don't match reality, the model may be tweaked until does. There is nothing inherently wrong with this, the notion not being much different from that of adjusting a gunsight to ensure hitting the bull's eye, or altering a sculpture to make it a better image of its subject. For example, the actual values provided by Newton's law of gravity are subject to a constant factor "G" that effectively gives the strength of the gravitational force, and has to be derived from observations. However, having obtained that value of G, the law of gravity is then generally applicable.
A mathematical model is a description of how the real world is thought to work, and can be useful as a tool, to the extent that it is accurate and usable. If the model is inaccurate, clumsy, excessively complicated, based on unjustifiable assumptions, or has to tweaked in a different way to deal with every different situation, then the model needs to be fixed or, if that's not possible, thrown away.
* In any case, Haldane's math simply showed that natural selection's assumptions could account very reasonably for the rise of the dark peppered moth, and that mutationism couldn't. The assumptions of his model have been examined and are not controversial. Its dismissal of mutationism was decisive, not because of any fancy mathematics but because, in laying out the underlying assumptions of mutationism, it made their absurdity clear, even ignoring the specific calculations.
To account for the conversion of moth populations from a light to a dark form on the basis of mutations alone obviously implied a high level of mutations. Since the mutation rate of organisms is roughly constant, why did populations of moths in non-industrial regions, or in pre-industrial days, remain light? Mutationism seem to imply that a population went from light to dark on a whim, but for some strange reason that whim seemed to be firmly linked to levels of industrial soot.
Might it have something to do with the fact that in sooty areas a dark moth was less vulnerable to predation than a light moth, while in non-industrial areas the reverse was true? It remains odd that the otherwise brilliant and strongly empirical Morgan dismissed natural selection as a fuzzy concept, since on close examination he would have had to make a willful effort to ignore it. His only alternative was to think the change was due to some completely unknown effect -- maybe high levels of soot forced the mutations? But if the soot was a mutagen, why did it only change the color of the moths? A broad-spectrum mutagen would have made changes all up and down the line. Maybe the moths were under the influence of some entirely unknown external source of guidance? Or there a Lamarckian will among the moths for change? The rigorous Morgan would have found such alternatives appalling.
* Incidentally, although the peppered moth is a classic example of Darwinian natural selection, it is also one of the most heavily criticised. One of the later investigators, Bernard Kettlewell (1907:1979), performed experiments in the 1950s that suffered from a number of potential flaws, for example placing moths on tree trunks despite the fact that they don't always perch there. This error was aggravated by texts that showed obviously staged photographs of dark and light peppered moths on tree trunks.
The matter was played up at extreme length by the critics as the scientific scandal of the century, raising the question of whether peppered moths actually lit on tree trunks to a level of importance that made the Piltdown Man hoax seem petty in comparison. In fact, the rise and fall of the peppered moth relative to industrial pollution was already well established, and the differential predation of light versus dark moths remains the obvious driving force. Indeed, one of the criticisms of Kettlewell's work -- that he didn't realize that birds see into the ultraviolet, and so a moth camouflaged to a human might not be camouflaged to a bird -- evaporated when it was determined that the camouflage of the moths worked just as well in the ultraviolet as it did in the visible light range. Accusations that Kettlewell fudged his data haven't been validated by careful reexaminations of his work.
In any case, Kettlewell's experiments are old news. Clean air acts were passed in the UK at about the same time that Kettlewell was performing his experiments. Since that time the British landscape, both rural and urban, has become vastly cleaner. The result is that light-colored peppered moths once again predominate, with the dark form all but becoming extinct.
* FISHER'S STATISTICS / WRIGHT'S FITNESS LANDSCAPE: Haldane's pioneering mathematical studies were paralleled by those of R.A. Fisher. Fisher was a single-minded man, a die-hard eugenicist whose studies were almost entirely focused to that end. The culmination of his work was his GENETICAL THEORY OF NATURAL SELECTION, published in 1930, in which he mathematically demonstrated how a small change would propagate through a population, the rate of propagation being proportional to the advantage of a change, and also showed how shifts in the environment would lead to shifts in the makeup of the population.
Confronted with the work of Haldane and Fisher, geneticists increasingly accepted that the gross changes observed by Mendel with his pea plants were much more the exception than the rule, and that traits influenced by multiple genes could produce what seemed to be a gradual, continuous range of differences. For example, a Swedish biologist named Hermann Nilsson-Ehle showed that if a trait were controlled by ten different genes, that trait would have about 60,000 variations, a range of variation that in practice might well appear to be seamlessly continuous. By the 1930s, genetics had made its peace with Darwinism.
However, field biologists were unconvinced by the musing of theoreticians -- Haldane, incidentally, was an entirely inept practical biologist -- and had a strong tendency to cling to Lamarckism. To them, natural selection just didn't seem able to account for the remarkable variations in organisms that was observed. Sewall Wright, in contrast to Haldane and Fisher, was actually a practical biologist, with a background in animal breeding, who also had a mathematical bent. As a result, Wright had a close focus on the real world: instead of relying on broad generalities, he was able to zero in on real-world scenarios, and as a result field biologists found him much more persuasive.
He was also more persuasive since, instead of hosing down his audience with calculations, he gave them an easily visualizable model: the "adaptive landscape" or "fitness landscape". In a 1932 paper, he suggested that evolution might be compared to a landscape with peaks and valleys, corresponding to adaptations providing good and poor fitness respectively, with mutations in species causing organisms to migrate around the landscape in a blind fashion, one step per generation, guided by the "terrain" to drift up peaks and away from valleys as fitness improved.
In a modern computer simulation, the organism could be seen as a sort of mindless "Darwinbot" that moved over the landscape through genetic changes at random, generation by generation, with the only sort of guidance being a bias towards moving uphill and against moving downhill. Wright saw the landscape as smoothly curved, since the genetic variations from generation to generation were as a rule small and gradual.
It should be emphasized that the fitness landscape is a sheer abstraction. Trying to actually implement a detailed general fitness landscape would be difficult since fitness has many "dimensions" -- it can be a function of the ability of a life-form to deal with climate, to avoid predators, to find prey, to digest foods, to find a mate, to deal with pathogens, and so on in a long list -- making it hard to represent except for specialized cases. Since there are "tradeoffs" in the adaptations of an organism, with an adaptation that gives an advantage sometimes inescapably also providing a disadvantage, what amounts to a "peak" in one dimension of the fitness landscape may be a "valley" in another. For example, the size of a full-grown bull elephant makes it almost invulnerable to predation by lions or the like -- but makes it more vulnerable to starvation in times when forage is scarce. The overall "fitness" of an organism is complicated balance of a great number of factors, reflecting a "fitness peak" that is a composite of many environmental pressures. In addition, the factors are not necessarily fixed, with changes in climate, introduction of competitors and predators, and so on shifting the fitness peaks and valleys around. In fact, if an organism reaches a fitness peak, it will become more common, making it an attractive target for predators and parasites -- inevitably shifting the peak underneath it. Despite its abstract nature, however, the fitness landscape is an elegant visualization tool for various specific scenarios.
There is an implication in this model that if an organism reaches a fitness peak, it is likely to stay there indefinitely, enduring down through the generations with little visible change. The fact that the landscape can, in fact is almost certain to, shift -- the peaks moving or disappearing -- ensures ongoing change. However, Wright pointed out another mechanism to show why organisms didn't remain static indefinitely: "genetic drift". If mutations are occurring in organisms all the time, then they don't really occupy a static position in the fitness landscape anyway, with the locus shifting around at random due to hereditary "noise". In other words, the Darwinbot's motions on the fitness landscape are a bit drunken and unsteady -- becoming steadier when selection pressures are strong, unsteadier when they are weak.
For a small population, the effects of this "noise" are much more visible. In many cases, this "noise" may not have any impact on fitness one way or another, at least not at the outset. Even if it did have a negative effect, if the fitness peaks are not too tall and the (un)fitness valleys not too deep, the noise might well drive the organism towards another fitness peak. If the new fitness peak were "taller" than the old, the new form of the organism might then drive the old form to extinction.
What Wright was suggesting was that small, isolated populations might be the precursors of new species, a notion that appealed to field biologists because it was what they had observed. Wright had practical experience to back up his ideas as well: he had conducted research on inbreeding of guinea pigs and shorthorn cattle. In both cases, he had observed how a particular mutation could be emphasized and then "fixed" in a small population through inbreeding, and then introduced to the wider population. His experiments of course involved artificial selection, but believed that the same process applied to natural selection -- in much the same way that Darwin had leveraged off artificial selection to provide an argument for natural selection.
Fisher, a strict selectionist, didn't like Wright's concept of genetic drift, leading to a more or less polite scientific dispute between the two men -- Wright wasn't as inclined to quarrels as were Fisher or Haldane, though it with such colleagues it was hard for him to avoid squabbling. The feud proved very valuable to population genetics, since it produced volumes of work that did much to finally dismiss all lingering traces of Lamarckism and mutationism in evolutionary thought. By the 1930s, the challenges that Mendelian genetics had seemed to pose to Darwinism had evaporated, resulting in a marriage of the two concepts that became known as "neo-Darwinism". Put simply: Darwin plus Mendel equals evolution.
TO BE CONTINUED
* Website additions for the month include:
Updated documents include:
New reviews include:
This last month's online blog entries include items on: railroad and road infrastructure, California road trip, home hybrid power-heating systems, genetically modified mosquitos, prenatal genetic testing controversy, emissions trading controversy, updating the Panama canal, phone card ripoffs, ICBMs for conventional strike, personal supercomputing, coin hoarding in India, Alpine & Gibraltar tunnels, user interface for XO computer, forensic shoeprint analysis, and the risks of being an archaeologist.
Online update links at: http://www.vectorsite.net/update.html