From cosmic ray studies, there is compelling evidence that the world of low-energy phenomena investigated in great details over several decades is complemented by a high energy world, being a terra incognita with respect to its origin. The origin of the high energetic cosmic radiation can be explored only by neutral, stable probes, i.e by gamma rays and neutrinos pointing back to their sources. Unlike gamma-rays neutrinos can escape regions of particle acceleration strongly shadowed by matter. In addition, neutrinos can reach us from the most distant parts of the observable Universe, whereas high energy gamma-rays are absorbed by interstellar light. On the other hand, the weakness of neutrino interactions makes detection to be extremely difficult and requires huge detectors which should be reliable protected against background.

The history of underground detectors for natural neutrino fluxes started with the pioneering detectors in India (1965), South Africa (1965) and in the Northern Caucasus (1978) and led to armed giants with effective areas of 1000 square meters (MACRO, Italy), and sensitive volume of more than 20000 tons (SUPERKAMIOKANDE, Japan). These detectors allow to investigate a wide spectrum of problems in the fields of neutrino astrophysics, elementary particle physics and cosmic ray physics. Nevertheless existing underground detectors were not able to detect high energy neutrinos from cosmic accelerators. So, a key question of modern astrophysics - what is the nature of cosmic high energy world ? - has still to be considered to be unsolved.

One is forced to considerable increase the sensitivity of neutrino telescopes. This, in turn, dictates the creation of principally larger scales and goes beyond the constructional possibilities of current underground techniques. One of the solutions of this dilemma is to develop underwater detectors which use the giant water masses of natural basins as shield against downward going background, target and detecting medium.

The idea to construct big water Cherenkov detectors for neutrino astrophysics deep underwater was firstly formulated in 1960 by M.A.Markov. These telescopes consist of a lattice of photomultipliers (PMTs) spread over a large open volume in the ocean or in a lake. High energy neutrinos can be detected by upward traveling muons produced in neutrino interactions in water close to the detector. With the angle between parent high energy neutrino and muon being very small astronomy with degree resolution is possible. The direction of the particle is inferred from the measured arrival times and amplitudes of the Cherenkov photons. Technologies for underwater telescopes have been pioneered by the Baikal collaboration and by the DUMAND project.

In the mean time, two underwater/ice detectors are taking data: the Baikal telescope and the AMANDA telescope at the South Pole. Also, the projects NESTOR and ANTARES in the Mediterranean have joined the efforts towards an underwater neutrino telescope.

The scientific goals of underwater telescopes are manifold. The basic motivation is to do high energy neutrino astronomy. Beyond the field of neutrino astronomy, the diverse scientific missions of underwater telescopes include the search for neutrinos emerging from the annihilation of Weakly Interacting Massive Particles (WIMPs) in the center of the Earth or the sun, the search for neutrino oscillations, or for slowly moving, bright particles like GUT magnetic monopoles. They can contribute to such different fields like atmospheric muon physics on the one hand and - for a lake-based telescope - limnology on the other hand.

The possibility to build a neutrino telescope in Lake Baikal was investigated since 1980, with the basic idea to use - instead of a ship - the winter ice cover as a platform for assembly and deployment of instruments. After first small size tests, in 1984-90 single-string arrays GIRLYANDA equipped with 12 - 36 PMTs were deployed and operated via a shore cable. The total life time for these first generation detectors made up 270 days. On the methodical side, underwater and ice technologies were developed, optical properties of the Baikal water as well as the long-term variations of the water luminescence were investigated in great details.

Since 1987, a second generation detector with the capability to identify muons from neutrino interactions was envisaged. According to the approximate number of PMTs this detector was named NT-200 - Neutrino Telescope with 200 PMTs. With estimated effective area of about 2300 square meters and 8500 square meters for 1-TeV and 100-Tev muons, respectively, it is a first stage of a future full-scale telescope, which will be built stepwise, via intermediate detectors of rising size and complexity.

The Baikal Neutrino Telescope NT-200 is being deployed in Lake Baikal, 3.6 km from shore at a depth of 1.1 km. It consist of 192 optical modules (OMs). The umbrella-like frame of 43 m diameter carries the 8 strings with the detector components. The detetector is connected with the shore center by 3 bottom cables.

In April 1993, the first part of NT-200, the detector NT-36 with 36 OMs at 3 short strings, was put into operation and took data up to March 1995. A 72-OMs array, NT-72, run in 1995-96. In 1996 it was replaced by the four-string array NT-96. Summed over 700 days effective life time, 320,000,000 muon events have been collected with NT-36, -72, -96. The first neutrino events have been selected. Since April 6, 1997, NT-144, a six-string array with 144 OMs, taked data in Lake Baikal. NT-200 array is completed in April, 1998.