On a cold, clear night last May the shutters rolled away from the aperture window. The Pierre Auger Observatory's newly completed air-fluorescence telescope was now overlooking a vast expanse of Argentine desert. Experimenters switched on the photomultiplier tubes and moments later they watched as the first cosmic-ray air shower appeared on the event display. The beautiful, nearly noise-free images represented an important milestone for the Auger collaboration in its study of the mysteries of the highest-energy cosmic rays.

Seven months later an even more dramatic event demonstrated the unique strength of the Auger Observatory. By then the first particle detectors of an extensive surface array were in operation. In the early hours of 8 December, the fluorescence detector and the surface array recorded a single shower as it cascaded through the air and splashed into the array below. This shower, captured by two quite different but complementary techniques, was transmitted to collaborators around the world. It demonstrated beyond doubt that the detectors, trigger, timing, data communications and data acquisition systems were working as designed.

However, there could only be a brief moment of celebration. Ahead lay the daunting task of deploying 1600 surface detector stations over 3000 sq. km of desert, and a total of 30 fluorescence telescopes to overlook this array. The observatory needs such a large aperture to gather enough of the very-high-energy cosmic-ray events to probe their origin. At such energies, cosmic-ray particles are extremely rare. Above 1019 eV there is just one cosmic-ray particle per square kilometre, per steradian, per year. Above 1020 eV there is only one per square kilometre, per steradian, per century.

The quest for the source of the highest-energy cosmic rays is one of the most interesting problems in astrophysics. Following the discovery of the cosmic microwave background in 1965, Greisen and, independently, Zatsepin and Kuzmin realized that this background radiation would make space opaque to cosmic rays of very high energy. Nevertheless, over the past 30 years, several tens of events have been recorded with energies of more than the Greisen, Zatsepin, Kuzmin (GZK) cut-off, which starts at about 5 x 1019 eV, including a number above 1020 eV. The cosmic acceleration mechanism for achieving these energies is not known. Because of the GZK cut-off, these particles must come from nearby - less than about 50 megaparsecs. Yet even though particles of these energies are only slightly deflected by galactic and extragalactic magnetic fields, none clearly points to a source sufficiently violent to be a candidate.

The two common detection techniques for cosmic rays both use the Earth's atmosphere, a remarkably effective calorimeter for capturing the great energy of these particles. The traditional means is an array of particle detectors that measure ionizing radiation in a plastic scintillator, or Cerenkov light in a water tank. The other technique was developed in the 1980s by a University of Utah-led group. They used a set of stationary telescopes in their "fly's eye" to record the development of showers by gathering the faint fluorescence produced as the shower passes through the atmosphere. Each of these techniques has its own strengths and limitations. The surface detector array depends on a comparison of the shower density distribution to simulations to obtain the energy scale. On the other hand, the surface array has a well defined aperture and can measure several features of the shower. The shower density, the electromagnetic and muon components, the shape of the shower front and the time structure are all useful in obtaining the composition of the primary particles. The fluorescence detector records the development of the shower and can make a more nearly calorimetric measurement of the energy. It can only be used on dark nights, however, and it depends on very careful calibration and an understanding of the attenuation of light in the atmosphere. The Auger Observatory combines the strengths of both techniques.

Two cosmic-ray air shower detectors are currently active. The AGASA surface array near Akeno, Japan, has been taking data for about 20 years. The other detector is the High Resolution Fly's Eye fluorescence detector in Utah, US. Although the data from the two experiments seem to agree in many respects, recent results show significant differences in the shape of the high end of the energy spectrum. In the next few years the Auger Observatory should be able to resolve these differences using the power of the hybrid detector system to collect a large number of events around the GZK cut-off.

The Auger Observatory has other new and important features. The fluorescence telescope uses Schmidt optics, which, with their aperture stop and corrector lens, allow greater light collection and reduced coma aberration with a spherical mirror. This aperture is sealed with a window that is also an ultraviolet filter for selecting the nitrogen fluorescence lines. As a result, the camera, mirror and all of the electronics are contained in a clean, controlled environment.

The surface detector stations are 10,000-litre water Cerenkov detectors, each equipped with three 220 mm hemispherical photomultipliers. Each is self-contained, with its own data processing unit, radio transceiver and solar power system. Event triggers indicate the possibility that a large air shower has struck the array. These move by radio to the central data acquisition system, which examines them for interesting events.

The central data acquisition system is on the Auger campus, located at the edge of the array in the town of Malargue. The campus also contains the detector assembly building with electronics shops, mechanical shops and a water purification plant. Besides the data acquisition system, the handsome new Auger centre building contains offices for staff and Auger collaborators, and a visitors' centre. For the scientists and engineers from 50 institutions in 18 countries working at the observatory, Malargue has begun to feel like home. At the inauguration of the new office building, the provincial governor, the mayor and a thousand townspeople came to hear speeches and tour the buildings.

In late October, an international review committee chaired by Werner Hoffman of the Max Planck Institute, Heidelberg, Germany, assembled at the Auger Observatory to evaluate progress. Its report was then received by the Auger Project Finance Board in Washington, which voted to proceed to completion. The collaboration hopes to finish the observatory by the end of 2004.