6 Conclusions

The rather modest scale of the miniature hyper-telescope sufficed to verify the theory of its operation and to confirm the imaging performances to be expected with future interferometers at the kilometres scale. We have verified that arrays in which the entrance and exit pupil are non homothetic can form direct images, contrary to what was commonly believed. We have also measured an intensity gain of $24\pm 3\times$ with respect to an equivalent Fizeau array and this is considered as supporting the $100 \times$ value obtainable in principle with perfectly phased components.

Hyper-telescope architectures are candidates for the next generation of extremely large ground-based telescopes. Aperture sizes even larger than the 30 to 100 metres, currently considered for large steerable mosaic mirrors (Gilmozzi et al. 1998), can be implemented in the form of sparse mosaics. Rather than large pointing mounts, these can extend the principle of the Arecibo radio-telescope, with a fixed diluted primary mirror and moving focal optics, possibly carried by stationary balloons. This configuration appears of interest even to achieve kilometre size apertures.

In space, the situation seems more favourable and many sub-apertures can be combined in formation flight to achieve kilometre size apertures (Boccaletti et al. 2000) for detecting Earth-like exo-planets (Labeyrie 1999).

Acknowledgements

This work is based on observations with the equatorial table at the Observatoire de Haute Provence and used hardware and software of the Observatoire de la Côte d'Azur for data reduction. We thank the referee, Jacques Beckers, for his useful criticism. One of us (Ettore Pedretti) would like to thank Farrokh Vakili for his encouragement and many useful discussions.