The hypertelescope construction initiated in the Southern Alps (Labeyrie et al., this conference) has
provided some preliminary operating experience indicating that larger versions, up to perhaps
1200m, are probably feasible at suitable sites. The Arecibo-like architecture of such instruments
does not require the large mount and dome which dominate the cost of a 40m ELT. For the same
cost, an "Extremely Large Hyper Telescope” ( ELHyT) may therefore have a larger collecting area.
It may thus in principle reach higher limiting magnitudes, both for seeing-limited and, if equipped
with a Laser Guide Star and adaptive phasing, for high-resolution imaging with gain as the size ratio,
i.e. about 30 with respect to a 40m ELT. Like the radio arrays of antennas, such instruments can be
grown progressively. Also, they can be up-graded with several focal gondolas, independently
tracking different sources. Candidate sites have been identified in the Himalaya and the Andes. We
describe several design options and compare the science achievable for both instruments, ELTs and
ELHyTs. The broad science addressed by an ELHyT covers stellar chromospheres, transiting exoplanets
and those requiring a high dynamic range, achieved by array apodization or coronagraphy.
With a Laser Guide Star, it extends to faint compact sources beyond the limits of telescopes having a
smaller collecting area, supernovae, active galactic nuclei, gamma ray bursts. The sparse content of
remote galaxies seen in the Hubble Deep Field appears compatible with the crowding limitations of
an ELHyT having 1000 apertures.