When using artificial satellites to discover more about the outside world, optics have a very special, fundamental role to play. Whether observing the Earth in detail, including its vegetation and atmosphere or detecting pollution, analysing the chemical composition of stars or galaxies, measuring angular deviations or the velocity of celestial objects, the natural option is to use visible, ultraviolet or infrared instrumentation in order to extend our knowledge. Since its creation in the sixties, the French Space Agency CNES, just like other space agencies, has been heavily involved in observation projects and those aiming to physically characterize natural, terrestrial, planetary, stellar or galactic phenomena. At first fairly straightforward, experiments quickly benefitted from the huge advances made in technology, making it possible to construct and send into space veritable observatories stable in both attitude and temperature. Such observatories are also equipped with computing facilities and a transmission downlink capability. After the first scientific experiments, more ambitious space astronomy projects such as D2B were given the go-ahead by CNES. This satellite flew five experiments investigating solar and stellar radiation in the far ultraviolet: spectral analysis and photometry of stars located around the ecliptic, study of solar activity and determination of atmospheric absorption. Photometers and spectrometers working in the visible or far ultraviolet were orbited for this purpose (with a wavelength of up to 1 70 A when observing the iron line for example, and commonly in Lyman's range). The satellite flying these experiments weighed 1 05 kg and was developed within 5 years (it was launched in 1975). The total mass allocated for the instruments on board was around a third of the total mass. Obviously, the equipment flown was compact in the extreme. The seventies saw the start of large-scale space application projects. The harnessing of space for the study of natural resources and meteorological purposes began with the Meteosat and Spot projects. The latter used an optical instrument of a new design. Indeed, because the satellites on which the instruments were flown were heavier (a satellite launched by Ariane or Thor Delta being at least ten times heavier than one launched by a Diamant BP4 launch vehicle) the instruments themselves could aim for more ambitious missions and become more complex. The Meteosat radiometer thus provides a complete visible and near infrared image of the Earth every half an hour. The image has a spatial resolution of both 3 km and 5 km and the instrument uses two infrared channels: an atmospheric transmission window between 0.5 and 12.5 jim, and a water vapour absorption band between approximately 5 and 6.5 jim. The Meteosat satellites, designed in the early seventies by CNES then developed by the European Space Agency, have provided an unbroken service. They form part of a worldwide meteorological watch network and will be replaced later this decade by more efficient imaging instruments and sounders as part ofthe European 2nd Generation Meteosat project. As far as Earth observation with optical instruments is concerned, CNES has played a major role by setting up and developing the Spot programme, which consists in detecting, recording, processing and distributing high-resolution multispectral images ofEarth. The years leading up to the decision to proceed with this programme were spent developing the basic techniques used, such as the push-broom principle behind linear array CCDs. A multispectral linear array CCD camera was thus developed and tested during airborne experiments in 1975 . The Spot mission is highly reliant upon optical techniques: telescope nearly diffraction-limited and extremely stable in orbit, alignment of 6000 photodetectors over 78 mm with a precision of around one micron, high degree of integration for the optoelectronic components (dichroic filters, electronics) all built into a detection unit with extremely precise positioning constraints, calibration device using gradian-index fibres etc. Spot 1, the first satellite in the Spot family, was launched in 1986 and still produces highquality images. Spot 2 and 3 have now been in orbit since 1993 . The HRV (High Resolution Visible) instrument includes a mirror which can change the viewing direction by 13.5° along a plane perpendicular to the ground track. Three of the instrument's spectral bands have a resolution of 20 m and another, wider band, has a resolution ofjust 10 m. There are in fact two HRV instruments on the Spot platform, both identical yet independent from each other. One of Spot's original features is its stereoscopic vision over two different orbits, retransmitting data on relief invaluable for cartography purposes. Other quite original optical instruments dedicated to specific missions have been used to explore the solar system. We could quote, for example, the mission to observe and analyse Halley's comet in 1986. Several different instruments developed by European science laboratories, including visible and ultraviolet spectrometers and infrared imagers, made a significant contribution to extending our knowledge of the structure and chemical composition of the comet's core, and provided outstanding images of the comet itself (both core and tail).