In 1932, an engineer at Bell Laboratories in the US named Karl Jansky (1905:1950) was searching for sources of radio interference in the sky. He created an array of antennas mounted on a circular track to see if sources could be found in specific directions, and detected radio emissions coming from what he concluded in 1935 to be the center of our Galaxy. He had made one of the first useful observations of the cosmos outside the narrow visible-light window.

To the extent that anyone had considered natural radio sources in the sky, the assumption was that the Sun would be the predominant source, and Jansky's finding was surprising. Actually, it might be better said that it would have been surprising -- if anyone had paid much attention to it. Bell Labs publicized Jansky's findings, but few astronomers noticed. Jansky wanted to go farther, constructing a 30 meter (100 foot) dish antenna to get a real map of the radio sky, but Bell Labs wasn't in the astronomy business and, reasonably, didn't see that the expenditure was justified. Jansky went on to other fields of research.

Jansky's findings were published in the popular radio press and did attract the attention of an American amateur radio enthusiast named Grote Reber (1911:2002). Reber was a very clever and energetic tinkerer, and put up a dish antenna with a diameter of 9 meters in the backyard of his mother's house in Wheaton, Illinois. It was an "elevation only" instrument, capable of being shifted up and down, but not rotated around. Reber used the first real radio telescope to make the first real radio map of the sky, and mailed his results to Subrahmanyan Chandrasekhar (1910:1995) of the University of Chicago, a prominent astrophysicist who was then editor of the ASTROPHYSICAL JOURNAL.

Chandrasekhar wasn't too sure of what to make of Reber and so he suggested that a few astronomers drive to Wheaton and look things over. They weren't too sure of what to make of things either, but after they arrived they realized that Reber knew exactly what he was doing. Chandrasekhar approved Reber's article for publication. Jesse Greenstein (1909:2002) of the University of Chicago had been one of the few astronomers to pay any attention to Karl Jansky's work, and Greenstein really wanted to encourage Reber, who he described as the "ideal American inventor." Greenstein tried to arrange a position for Reber at the University of Chicago, but Reber was "an independent cuss", as Greenstein put it, and didn't want to put up with any of the bureaucracy involved.

The American way of invention sometimes seems to be to come up with a brilliant new idea, then essentially not bother to follow up on it. Other nations take the next big steps and then the Americans get in a hurry to catch back up. This more or less happened with submarines and airplanes; it certainly happened with radio astronomy. Although a British researcher, J.S. Hey, had monitored radio waves from the Sun in 1942 while trying to track down sources of interference for the military, nobody had time or resources to pursue radio astronomy during World War II. However, after the war a number of researchers who had been working on radar and other military electronics during the conflict went back to academia with heads full of bright ideas. They also had access to a large amount of new radio gear that had been developed during the war, to be sold off as surplus as it rapidly became obsolescent or was thrown out in defense cuts.

A number of radio astronomy research teams were formed, mostly in Britain and Australia, with three teams leading the way: a group under Bernard Lovell (born 1913) at the University of Manchester in the UK and operating at the Jodrell Bank site; a group under Martin Ryle (1918:1984) at the University of Cambridge in the UK; and a team under J.L. Pawsey (1908:1962) and John Bolton in Australia. They worked mostly with arrays of simple antennas, wired together through electronic delay lines to give the position of cosmic radio emitters through interferometry. These arrays were cheap and could give locations, though they couldn't provide a real image of the radio source.

By 1949 the Australians had achieved a resolution of 10 arc-minutes, and had pinned down three cosmic radio source, the most prominent of them being the Crab Nebula. By 1951 Ryle's team had achieved a resolution of 1 arc-minute, leading to the discovery of a radio-emitting galaxy, designated "Cygnus A". Astronomers had also learned to use the radio emissions of various atoms and molecules as "tracers" of cosmic structures. One of the first, and still the most important, was the 21 centimeter radio emission of monatomic hydrogen. This relatively longwave radiation could penetrate through cosmic dust clouds, and since neutral hydrogen was an important constituent of the gas clouds that mark the spiral arms of our Galaxy, the signal could be used to trace the spiral arms. The first map of the structure of our Galaxy was published in 1951 by an American-Australian-Dutch group; it was a feat that would have astounded astronomers of previous eras.

As the British and Australians blazed paths in radio astronomy, the Americans continued to sit on their hands. Jesse Greenstein found the situation frustrating and decided to take action. At the time, in the early 1950s, Greenstein was at Caltech, working on the organization of an observing program for the new Hale telescope. Greenstein wanted Caltech to fund a radio interferometer array, but the authorities were cautious about the matter. To push the issue, Greenstein set up an international conference on radio astronomy in Washington DC in early 1954. The conference brought emerging American interest in radio astronomy to a boil. With backing from the US National Science Foundation (NSF), a group of researchers set up a government-backed "National Radio Astronomy Observatory (NRAO)" in Greenbank, West Virginia. Of course, Greenstein got his radio interferometer array as well, which was built by Caltech in Owens Valley, California. * While the Americans were getting up to speed, radio astronomers elsewhere were not idle. In 1957, the Jodrell Bank site put Bernard Lovell's "Mark I" radio telescope into operation. It was the biggest fully steerable radio telescope, 76.2 meters in diameter. It was used to track early space shots and occasionally showed up as a background prop in sci-fi shows like DR WHO. It was upgraded in 1970:1971 and 2001:2003; it was renamed the "Lovell Telescope" after the second upgrade. A number of other, smaller radio telescopes have been set up at or near the Jodrell Bank site, which is open to the public as an arboretum.

The Australians, who had been having a running competition with the British, set up the largest steerable radio telescope in the Southern Hemisphere at Parkes in New South Wales in 1961. It has a diameter of 64 meters.

One of the other things about the odd American tendency to start something and then drop is that, once they pick it up again, they overcompensate and try to build something bigger and better than anything else. In the case of radio astronomy, the US funded construction of the world's biggest radio telescope, at Arecibo, Puerto Rico, with a diameter of 305 meters. The Arecibo instrument was dug into a depression between a set of hills. The antenna bowl originally consisted of a wire mesh grid, set up with a spherical curvature, with three towers around the edges that used cables to support a central receiving platform. Spherical curvature was used because the telescope was, obviously, a transit instrument; it could be steered to an extent, however, by adjusting the position of the receiving platform, allowing the telescope to scan the sky 20 degrees around the central axis of the dish. The spherical curvature permitted a reasonable focus no matter what the position of the receiving platform was. The receiver system was designed to compensate for the spherical aberration; as with most radio telescopes, it included several different receivers to permit it to "tune in" on different wavelength bands.

The Arecibo facility went online in November 1963. Interestingly, it was officially designated as a facility for ionospheric studies, apparently as a way to get US government funding. The Arecibo telescope has been updated since its establishment. In the early 1970s, the dish was resurfaced with aluminum panels to permit reception of shorter wavelengths. The panels were perforated to let sun and light through; vegetation is allowed to grow under the dish to prevent soil erosion. It was upgraded again in the 1990s, tripling its sensitivity.

The Arecibo dish was a grand project, but other groups in the US managed to perform radio astronomy experiments on a budget. One of the most interesting was the Ohio State University (OSU) "Big Ear" radio telescope, which went online in 1963. The Big Ear was the brainchild of Dr. John D. Krause, who was a professor of both astronomy and electrical engineering, making him ideally suited to developing such a project.

The Big Ear consisted of a two billboard arrays, with a mesh surface, arranged on the ends of a metal-covered flat court that acted as a ground plane. The transmitter part of the array was a fixed parabolic array, 110 meters wide and 21 meters tall and with twin moveable feed horns. It stared flat across the court to a tiltable flat reflector array 104 meters wide and 30 meters tall when fully erect. The court was made of concrete covered by aluminum panels, with a width of 110 meters and a spacing between the two arrays of 152 meters.

The Big Ear could be aimed in altitude but not in azimuth, but it was intended as a survey instrument and that wasn't a major limitation. It was originally used to map wideband radio sources through the sky, but due to funding cutbacks that mission had to be abandoned. In 1973, it was converted to conduct a pioneering search for extra- terrestial signals, with the effort continuing into 1985. The Big Ear was finally dismantled in 1998. * Although the Arecibo instrument was and remains impressive, at the time the British still retained the largest steerable dish, the Jodrell Bank instrument. The Germans topped it in 1971 by setting up a 100 meter (328 foot) fully steerable dish at Effelsburg, near Bonn. The biggest competing American instrument was a partly steerable (elevation only) NRAO instrument with a diameter of 92 meters at Green Bank, West Virginia. It collapsed in a storm in 1988, and was replaced about a decade later by a fully steerable telescope with a diameter of 100 meters at the same site. The new instrument is known as the "Green Bank Telescope (GBT)". It has an "offset" feed / receiver assembly that minimizes obstruction of the sky. It also features a surface consisting of 2,204 moveable panels that are adjusted by computer to ensure an accurate surface, allowing it to operate effectively at wavelengths shorter than a centimeter.

In any case, the establishment of more powerful and capable radio telescopes in the 1960s led to a series of major discoveries. One was the discovery of "quasi-stellar radio sources" or "quasars" in 1963, radio-bright objects that appeared to be stars but which were later determined to be distant galaxies with highly energetic core regions. The third catalog of radio objects put together by the Cambridge group had noticed a very bright radio source, with the catalog entry of "3C 273", but couldn't pin down its exact location. The Parkes group realized that the Moon would pass through the region of the sky where 3C 273; the Moon would block out the source at a precise instant, allowing its location to be pinned down accurately enough to allow a big optical telescope to take a picture of it. 3C 273 was so low on the horizon that the Parkes staff had to saw off the stops on their telescope, but the trouble was worth it. The location of 3C 273 was pinned down, the object was photographed, and its spectrum indicated that it heavily redshifted, implying that it was an extremely bright source in the very distant Universe.

Another was the 1967 discovery of radio pulsing stars or "pulsars", which were quickly determined to be superdense "neutron stars", objects a few kilometers across with the mass of a full-size star, composed mostly of densely-packed neutrons. Pulsars were discovered by a British team under Anthony Hewish (born 1924), with his grad student Jocelyn Bell (born 1943) being the first to notice the radio pulsations. They were using a special dedicated array of 2,048 antennas, arranged in a grid over 1.8 hectares (4.5 acres) and designed to perform wide sky surveys for rapid radio transients in the meter wavelengths. It was a thoroughly cheap home-brew system, little more than a field full of light wooden poles strung with wires, but it proved very effective, and won Hewish the Nobel prize in 1974, sharing the prize with Martin Ryle. In 1974, two American astronomers, Russell A. Hulse (born 1950) and Joseph H. Taylor (born 1941), spotted the first binary neutron star system using the Arecibo instrument. They would also get the Nobel prize for their work, in 1993. Theoreticians had determined that neutron stars were possible decades earlier, but they seemed so preposterous that not everybody was convinced they were for real. The discovery of pulsars indicated they were. Given that encouragement, astronomers became more willing to believe in the existence of "black holes", objects denser than neutron stars, so dense that light cannot escape from them.

With the introduction of large, powerful dish radio telescopes, astronomers began to think of perform radar observations of the planets. Radar involves transmitting radio pulses and then measuring the time it takes for the pulses to return from a target. Obviously, this requires a powerful transmitter and a sensitive receiver since even the nearest planets are far away.

Radar pulses had been reflected from the Moon not long after the end of World War II, but this was basically a stunt. In fact, when the big US "Ballistic Missile Early Warning System (BMEWS)" military radars were set up in Greenland and Alaska in the early 1960s, the first time the Moon rose in the line of sight of one of the radars it set off alarms; the designers hadn't factored the Moon into their considerations, and some changes had to be made to ignore the returns from the Moon.

In 1961 astronomers actually performed the first useful radar observations of another world, bouncing radar pulses off of Venus during a conjunction of Earth and Venus. The pulses were in the form of "pseudorandom noise" trains that provided long and distinctive patterns, allowing the astronomers to sort out the returns. The radar studies showed Venus had a slow rotation rate, and that it rotated in the reverse direction of other planets, as if it had suffered some monster impact in the distant past that reversed its spin.

In the 1970s, improved radar observations using the Arecibo instrument actually permitted some mapping of gross surface detail on the planet as well, with the radar penetrating the unending thick cloud cover of the planet. This was just a prelude for much more detailed and comprehensive radar maps of the planet, produced by US and Soviet Venus orbiting spacecraft in the 1980s.

By that time, radar was being used to observe asteroids passing by the Earth, leading to the imaging of the asteroid Castalia in 1989. A number of other asteroids have been imaged by radar since that time, and in fact radar observations of large asteroids have been performed into the asteroid belt itself, beyond the orbit of Mars.

Radio interferometry has been around from almost the beginning of radio astronomy. In radio interferometry, two separated radio antennas pick up radio waves from a single cosmic source, with phase interference occurring if the paths from the source to the antennas are not both an integer multiple of the wavelength being observed. Measurement of the phase difference gives highly precise position information on a point source. The resolution of radio interferometry is dependent on the baseline between the two antennas, though the ability to collect radio waves is only proportional to the sum of the areas of the two antennas, no matter how far they are apart.

Radio interferometry is not too different in principle from optical interferometry, but the long wavelengths observed by radio astronomers make radio interferometry less demanding than optical interferometry, and the radio measurements are somewhat easier to process as well. This is why radio interferometry is well-established but optical interferomentry is only now becoming so.

The early radio interferometry arrays, as mentioned, were only capable of finding the locations of radio emitters, being just simple antennas wired together through delay lines, with the signals from the antennas electrically summed to find nulls and maximums. In the 1950s, the Australians came up with a straightforward extension of this idea, using two such arrays at right angles to each other to pin down sources in two dimensions. A series of such "cross interferometers" were built at Fleurs, not far from Sydney, beginning with the 1954 "Mills cross", followed by the larger "Shain cross" in 1958.

Much more dramatically, in the 1960s, Martin Ryle came up with a scheme known as "aperture synthesis" that instead of merely locating cosmic radio sources provided an image of them as well. He won the Nobel prize for his invention.

The patterns of phase interference observed in radio interferometry define concentric circles that shrink to a point at the precise location of the point source. The positions of the two telescopes changes as the Earth moves, defining a path through the map of concentric circles. The hints provided by the patterns along this path can be used to reveal the exact position of the source.

This is the case for a point source, but for an extended source the object can be regarded as a matrix of point sources, with the number of points in the image related to the resolution of the interferometry array. A process known as "Fourier analysis" allows the individual sources to be sorted out. As one astronomer put it, the data obtained from an interferometer array is analogous to a series of pictures obtained through a slowly rotating colander. The pictures can be mathematically combined to rebuild the actual image. This process is known as "filling the aperture". As more telescopes are added to an interferometer array, the observational data yields more information. Aperture synthesis wasn't really practical until the computers became relatively common, since it requires a good amount of number-crunching.

Early experiments in imaging radio interferometry were conducted in the late 1940s using war surplus radio dishes. By the 1960s, small imaging radio interferometry arrays had been developed, such as the British Cambridge "One Mile" (1.6 kilometer) array, which provided resolutions on the order of a few arc-seconds. The Cambridge array was later increased to a size of 5 kilometers (3.1 miles). There are other significant imaging radio interferometers at Narrabi in Australia; the Westerbork Radio Observatory in the Netherlands; and the United Kingdom. One of the UK telescopes, known as "Merlin", consists of seven antennas spread over 200 kilometers in western and central England.

One of the most important dedicated imaging radio interferometry arrays is the 27- element "Very Large Array (VLA)" at Socorro in New Mexico, completed in the 1970s, which is built up of 27 25-meter dishes that can be moved on rails to provide a maximum baseline of 40 kilometers (25 miles). It was built by the NRAO, using NSF funding. It performed its first observations in 1975 and was fully operational in 1980. Like the Jodrell Bank instrument, the VLA occasionally makes appearances in science- fiction videos.

One of the traditional problems with radio astronomy was that for wavelengths longer than about 30 centimeters (1 gigahertz), ionospheric interference makes it difficult to obtain a decent map of a radio source. In 1991, astronomers figured out algorithms to correct for ionospheric interference, and managed to obtain radio maps from the VLA at a record long wavelength of about four meters (74 megahertz).

One of the latest imaging radio interferometers is at Narayangaon, near Pune in India. The "Giant Meterwave Radio Telescope (GMRT)" came online in 2000; it was designed to focus on long wavelengths generally neglected by other radio telescopes. It has thirty 45 meter dishes with a maximum baseline of 25 kilometers, performs observations in six bands from 38 MHz to 1.420 GHz (7.9 meters and 21 centimeters). The GMRT was built by India's Tata Institute of Fundamental Research. The long wavelengths mean the dishes didn't need to have the surface precision of radio telescopes that operate at shorter wavelengths; they are wire-mesh reflectors mounted on steel-tube frames kept in shape with steel cables, greatly reducing cost and incidentally giving them an attractive, delicate appearance. Its many tasks include searches and observations of "protoclusters" of galaxies in the early Universe, and also studies of pulsars. Its near-equatorial position gives it a good view of the plane of the Milky Way, allowing it to hunt down large numbers of pulsars.

Arrays now seem to be replacing large unitary radio telescopes. Scaling up structures, such as a radio telescope, tends to run into diminishing returns: in very simple terms, doubling the linear dimensions means increasing the structure's cross section by a factor of four while mass and volume increase by a factor of eight. That means doubling the stress on the structure, which has to be made all that much more massive.

Martin Ryle no doubt never predicted at the outset that the computing power that made aperture synthesis possible would evolve so fast; by the 21st century, an average household would have a more powerful computer than almost any available in the 1960s. Given modern computing power, it is much simpler and cheaper to build an array of small antennas and link them together with aperture synthesis; some astronomers believe the big unitary telescope is now a thing of the past. One of the pioneering examples of this new approach, the "Allen Telescope Array (ATA)", is gradually being put in place in the Lassen National Forest in northern California, a somewhat isolated area where radio noise is relatively subdued. The ATA will ultimately consist of an array of 350 six meter (20 foot) dishes; at last notice, it was up to 32 dishes, but new dishes are being installed on an ongoing basis. It was originally called the "One Hectare Telescope (1HT)" since a hectare (2.47 acres) was the objective for total collecting area.

In completion, it will have the same effective collecting area as the 100 meter NRAO Green Bank telescope, but will cost only about a third as much. All the sources of funding are private -- it named after its primary patron, Paul Allen (born 1953), a senior official at Microsoft -- and the bureaucratic overhead is minimal. It is operated by a private group in collaboration with the University of California at Berkeley and is primarily focused on the search for other civilizations in space, but of course it can be used for less speculative purposes as well.

Other new radio telescope arrays are being set up elsewhere. A group of Chinese, Canadian, and American scientists has set up a pilot version of the "Primeval Structure Telescope (PaST)" in the barren regions of western China. The pilot array consists of 25 "pods", with 127 antennas per pod; the antennas are simple "Yagi" types, nothing more than a set of dipoles of increasing span mounted on a bar, very similar to the ordinary broadcast TV antenna. Dish antennas don't work very well for wavelengths much longer than about two meters (150 meters); dipole arrays are not only cheap, but they are actually superior to dishes for low-frequency work. At such long wavelengths, they don't even need precision manufacture. The ultimate PaST array will have 10,000 antennas covering three hectares (7.4 acres); the group hoped to complete the instrument by the end of 2006.

An international group of radio astronomers became interested in building such "low frequency arrays (LOFARs)" after the low-frequency VLA breakthrough in the early 1990s. PaST is only one result of this effort; others include:

• A Dutch-German collaboration is now setting up a 25,000 antenna LOFAR array distributed over an area about 350 kilometers (220 miles) in diameter, straddling the border between the two countries.
• A US group is working on a "Long Wavelength Array (LWA)" consisting of 10,000 antennas strung over 400 kilometers (250 miles) of the US SouthWest.
• US and Australian researchers are now planning the "Mileura" array of 3,000 antennas in Western Australia.

Not only are these LOFARs cheap and powerful, they also have a wide field of view, consisting effectively of the entire sky above them. Computing power can be used to sort out the actual direction of any signal arriving at them. This makes them ideal for picking up "transient" events, such as the radio component of short-lived high-energy events in the distant Universe.

One of the most ambitious schemes for a next-generation instrument is the "Square Kilometer Array (SKA)", an array of antennas that will actually have a total kilometer of area, making it two orders of magnitude more sensitive than any previous radio telescope. The SKA will operate at relatively short wavelengths and will be used to probe the distant, early Universe; map the magnetic fields of galaxies; monitor thousands of pulsars to track gravitational waves; and search for intelligent life in the Universe.

he SKA will be cheap for its capabilities, but it will still cost up to about $1.6 billion USD. It is now being promoted as an international collaboration, with the main selling point at present being mapping of galactic and extragalactic neutral hydrogen clouds through the 21 centimeter emission. Of course, the SKA will have other capabilities. It will be able to pick up the signals from carbon monoxide and other cosmic molecules to map out the structures of distant galaxies; examine the violent cores of active galaxies, where supermassive black holes may lurk, using the coherent microwave emissions of clouds of stimulated water molecules called "mega-masers"; and trace out the magnetic fields of nearby galaxies, giving hints of galactic structure. The SKA is expected to have about 4,000 10-meter dish antennas operating over 1 to 25 GHz, supplemented by a set of fixed planar lowband radio receiver arrays operating over 100 MHz to 1 GHz. Candidate sites have been narrowed down to Mileura in Western Australia and Karoo in South Africa. The nation that hosts the site will be required to control communications emissions to prevent unnecessary interference with the operation of the SKA.

Both nations are building small demonstrator arrays at the proposed sites to evaluate technologies. Detail design of the SKA is expected to begin in 2008. The first SKA elements are expected to be in operation by 2014, with completion of the full array in 2020.

Along with the development of large radio interferometry arrays like the VLA, techniques were developed to link radio telescopes spanning continents or oceans into interferometry arrays, a technique known as "very long baseline interferometry (VLBI)".

Even before the construction of the VLA, radio astronomers were considering VLBI. The VLA is a "connected element" interferometer, with direct signal connections between the dishes, but for longer baselines direct connections are not practical. Observations had to be recorded on magnetic tape, along with the necessary precise timing information, which required accuracies better than a microsecond. This meant that any observatory participating in a VLBI network had to have recording equipment linked to a precise local atomic clock.

The first experimental VLBI measurements were made in 1967 by several Canadian and American research teams. A year later, a Swedish group joined the effort, providing an intercontinental baseline for observations. By the early 1970s, VLBI observations were providing details of the cosmic jets emitted by distant quasars. The VLBI pioneers managed to perform some significant observations of radio objects, but the logistics of conducting VLBI were burdensome and quickly exceeded the capability of informal groups of researchers. In particular, trying to schedule observations on multiple radio telescopes at the same time was troublesome, and argued for a set of radio telescopes dedicated strictly to VLBI.

VLBI advocates joined together in the "Network Users Group", which lobbied for a formal and properly funded VLBI effort in 1980 report on priorities for astronomical research. The result was the American "Very Long Baseline Array (VLBA)", an array of ten identical radio telescopes, each with a diameter of 25 meters, scattered across the North American continent. The VLBA was funded by the NSF at a cost of $84 million US in 1989 dollars, and was formally dedicated in 1993. The VLBA provides the highest resolution of any astronomical observatory ever built, up to a tenth of an arc- millisecond.

The VLBA is controlled by the NRAO Array Operations Center (AOC) in Socorro, New Mexico. The ten radio dishes, or network "nodes", are located at:

• Kitt Peak, Arizona.
• Pie Town, New Mexico.
• Mauna Kea, Hawaii.
• Owens Valley, California.
• Brewster, Washington.
• Los Alamos, New Mexico.
• Fort Davis, Texas.
• North Liberty, Iowa.
• Hancock, New Hampshire.
• aint Croix, Virgin Islands.

The baselines between the ten nodes range from 200 to 8,000 kilometers (125 to 5,000 miles). Each node site has a small local staff to handle observations and routine maintenance. The AOC provides central control over the entire network and a dedicated VLBI processor, or "correlator", that crunches the observations from the ten nodes. The AOC also maintains a spares stockpile, and an engineering staff to provide high level repairs or upgrades for the nodes.

Early experiments with radio interferometry revealed a sky that was very different from the one seen at visible wavelengths. The brightest objects were remote active galaxies and quasars, and the visible counterparts of these bright objects could only be found with powerful optical telescopes. The VLBA has greatly extended these early observations.

While VLBI can provide the resolution equivalent to a radio telescope with a diameter the size of the baselines between the individual telescopes, as mentioned the amount of radio energy a VLBI network can pick up is still limited by the sum of the apertures of the telescopes. This means that VLBI is best suited to relatively small cosmic objects that are energetic emitters at radio wavelengths.

The best targets for VLBI have been the cores of active galaxies and quasars, where violent and highly energetic events take place in a very small region. Astronomers believe that the energy is being released by matter falling into central supermassive black holes. This process releases 10 to 100 times more energy than the fusion processes that takes place in stars, causing huge cosmic jets to flow from the core. The interaction of these jets with the medium they pass through leads to strong radio emission.

Optical and infrared telescopes generally observe radiation released by thermal processes, such as light emitted by stars or heat released by the contraction of cool cosmic dust clouds. Radio telescopes, in contrast, more generally observe radiation caused by nonthermal processes, such as the continuous-spectrum "synchrotron radiation" emitted by electrons curving through strong magnetic fields, or the line- spectrum radiation emitted by gas clouds acting as "cosmic masers".

Clouds of water molecules can act as cosmic masers under some circumstances, and maser emission from such clouds in the cores of distant galaxies has been used to study their motion around central supermassive black holes. Using a few simple physics calculations, the motions have revealed the mass of these black holes to within a few percent.

Another useful target for radio astronomy observations is the 21 centimeter radio wavelength emitted by cold hydrogen. The 21 centimeter radiation allows the creation of maps of hydrogen clouds in our Galaxy and other galaxies. The VLBA has also been used to observe pulsars and events in our galactic core.

The VLBA is a work in progress. In 1997, the Japanese launched the HALCA (High Altitude Laboratory for Communications & Astronomy) satellite, which is a radio telescope platform with a folding antenna 8 meters in diameter. Although an experiments was performed with a communications satellite in the late 1980s to demonstrate that satellites could be used as part of a VLBI network, HALCA was the first spacecraft to be specifically designed for the purpose.

HALCA had a highly elliptical orbit around the Earth that allows it to perform VLBI observations over a baseline as long as 30,000 kilometers. Although HALCA's communication system suffered a failure that has limited the usefulness of its observations, it performed practical validations of a space-based VLBI system. The SKA, described previously, will actually be a VLBI system. Only about half the array, as determined by collecting area, will be at a central location. The other elements will be placed at sites up to thousands of kilometers away to provide VLBI capability.

The large radiotelescopes and interferometers typically operate at the meter and centimeter wavelengths. The signals are broadband, continuous, and often generated by "synchrotron emission", the radiation emitted by electrons spiraling in a magnetic field.

Beginning in the 1980s, radio astronomers began to move into wavelengths of a few millimeters or less. Such "millimeter" and "submillimeter" emissions, nudging up to the infrared spectrum at the short end of the scale, are generally from the different molecules that make up cold interstellar clouds of gas and dust, such as carbon monoxide, water, or a surprisingly wide range of organic molecules. Carbon monoxide is a particularly important target, because it is the second most common molecule in such clouds, next to diatomic hydrogen (H2), and can be used as a "tracer" to reveal the presence of H2. Such observations can be used to map the structures of galaxies or probe in detail into the inner regions of clouds where stars are being formed.

There are two problems with observing the Universe at millimeter and submillimeter wavelengths. First, the emissions are weak, requiring receivers cooled to a few degrees Kelvin to reduce thermal and electronic noise. Second, these emissions are absorbed by atmospheric water vapor. Telescopes designed to observe this region of the spectrum have to be placed on top of high, dry mountains.

One of the best millimeter-wave telescope in the world is the French-German-Spanish Institute de Radio Astronomie Millimetrique's 30 meter (100 foot) diameter dish high in the mountains of Spain's Sierra Nevada, not far from Granada. This instrument has been in operation since the mid-1980s, and makes extensive use of insulation, fans, and other components to control temperature and minimize thermal deformations.

Other notable millimeter-wave instruments include the 45 meter (150 foot) Nobeyama telescope in Japan, and a 12 meter (40 foot) dish operated by the US NRAO on Kitt Peak, Arizona. The Kitt Peak telescope covers almost the entire millimeter spectrum, and has discovered a large number of the molecules currently known to exist in deep space. There are over a dozen smaller instruments scattered around the world, and several projects are underway to build new instruments.

One of the best submillimeter telescopes is the 15 meter (50 meter) James Clerk Maxwell telescope on Mauna Kea, a British-Dutch-Canadian venture. It is accompanied on Mauna Kea by the California Institute of Technology 10.4 meter (34 foot) Submillimeter Observatory. A 10 meter (33 foot) submillimeter telescope is now in operation on Mount Graham, near Tucson. It is run jointly by the Max Planck Institut fur Radioastronomie in Bonn, Germany, and the Steward Observatory of the University of Arizona.

A 50 meter millimeter-wave telescope is now in operation on the top of the mountain of Sierra Negre in central Mexico, at an altitude of 4,640 meters, with a good view of both northern and southern skies. The "Large Millimeter Telescope (LMT)" was a collaborative project of Mexico's Instituto Nacional de Astrofisica, Optica y Electronica (INAOE) and the University of Massachusetts at Amherst, with substantial funding from the US Defense Advanced Research Projects Agency (DARPA). DARPA provided funding because the LMT promised to advance technologies useful for tracking near- space objects and for millimeter-wave radars, though there were protests in Mexico over the involvement of a US defense agency in the project.

Interferometry is being used in millimeter astronomy, with four now in operation. The Owens Valley Radio Observatory in California has six antennas and is run by the California Institute of Technology. A consortium of the Universities of California, Illinois, and Maryland also operate a six-antenna interferometer at Hat Creek in California, and expect to upgrade this instrument to nine antennas soon.

In addition, there are millimeter-wave interferometers at Nobeyama, and on the Plateau de Bure in France, where the Institut de Radio Astronomie Millimetrique operates an instrument comprising four 15-meter antennas, with a fifth to be added soon. A submillimeter array, with six 6-meter dishes, is being built on Mauna Kea by the Smithsonian Astrophysical Observatory, based at Harvard University. This array will consist of six 6-meters dishes arranged in a pattern 500 meters across, and will have a resolution down to 0.1 arc-seconds.

None of these arrays is on the scale of the VLA, but in November 2003 ground was broken on the Atacama plateau in Chile for the "Atacama Large Millimeter Array (ALMA)", which in completion will be the largest millimeter wave astronomical facility in the world. It will consist of 50 (originally 64) precision dish antennas, each with a diameter of 12 meters. The antennas are steerable and moveable, and will be used to form interferometer arrays ranging from 150 meters to 14 kilometers across. The dry climate and the site altitude of 5,000 meters provides excellent "seeing"; the control station is on lower ground, where breathing is not as difficult.

ALMA is being built by a consortium of US, ESO, Japan, Canadian, and Spanish research institutions. The group has set up and is operating a 12 meter submillimeter telescope at Atacama named the "Atacama Pathfinder Experiment (APEX)". As its name implies, it is intended to validate technologies and procedures for ALMA, but it is also the largest submillimeter instrument in the southern hemisphere. The Japanese are to build a secondary array, the "Atacama Compact Array (ACA)", with four 12- meter dishes and twelve 7-meter dishes, which will be optimized to observe wide-area, diffuse objects. The full ALMA array is expected to go into preliminary operation in 2007 and be completed in 2012.