(1) A Hobby-Eberly-style 25-m ELT
Amongst current proposals for a leap beyond the 8-m class of ground-based telescopes, the least expensive-looking are those originating from the team associated with the ~9.3m Hobby-Eberly Telescope (HET) for a 25-m using the same design principles (Bash et al. 1996 and references therein). The HET is a minimum-cost VLT currently being commissioned in Texas and likely soon to be cloned in South Africa. It and its proposed ELT offspring use a design resembling that of the Arecibo radio telescope, with the segmented, spherical primary mirror tilted at a fixed zenith angle and therefore subject to a constant gravity vector. Rotation is in azimuth only and is employed only for acquisition: a mobile focal plane feed assembly, providing spherical aberration correction, tracks targets for 45 minutes or more as the entrance pupil moves across the stationary primary mirror. The Hobby-Eberly Telescope was built (though not yet commissioned) for the remarkably low cost of ~$14M, and the team estimate the total cost of their scaled-up 25-m ELT version at around $250M (Sebring et al.1998) .
(2) The 30-m California Extremely Large Telescope (CELT) proposal.
The builders of the highly successful Keck telescopes are currently studying proposals for a 30-m successor. Initial concepts were presented by Nelson & Mast (1999) and a more recent outline of the design concepts was given by Nelson (2000). This differs from the HET-clone and OWL (see below) proposals in several ways, but in particular in adopting a non-spherical primary for a Ritchey-Chretien design, the proposers believing that the complications of fabricating large numbers of aspheric elements can be overcome inside an attainable budget (c.f. Mast, Nelson & Sommargren, 2000). This proposal explicitly includes the capability to deliver a large (~20 arcmin) FOV in order to permit seeing-limited (~0."5 resolution) multi-object spectroscopy as well as near-diffratcion-limited MCAO operation.
(3) 50-m Telescopes: Swedish proposals for XLT and the Maximum Aperture Telescope (MAXAT)
The arguments outlined by Mountain (1996) have tended to drive a number of studies in the direction of a facility significantly larger then the 25-m to 30-m concepts outlined above. Both the Swedish group (c.f. Ardeberg et al., 1996: 25 m; Ardeberg, Andersen & Owner-Petersen, 1998: 25-50 m; Gontcharov & Owner-Petersen, 2000: 50 m) and the US-based studies of MAXAT (Maximum-Aperture Telescope) have been following this trend (though the recent emphatic specification of the Giant Segmented Mirror Telescope (GSMT) as a 30-m facility by the Decadal Survey may have reversed it). This is perhaps a pity, as the science case for ELTs grows ever more attractive with aperture; indeed the 50-m proposals should perhaps be seen as 100-m concepts (see below) constrained by fear of the technical challenges. Even so, 50-m and 100-m propjects have many concerns in common and much synergy between the groups continues.
(4) European ambitions: the OverWhelmingly Large (100-m) OWL
As noted above, the ambitious proposal for a 100-m facility (OWL) was first put forward by Gilmozzi et al. in 1998. In recent incarnations OWL has evolved through a number of possible optical configurations, most assuming a 100-m spherical primary mirror and a 15- to 25-m spherical (Gilmozzi et al.1998) or flat segmented secondary (e.g. Brunetto, Koch & Quattri, 1999). Recent developments include a comprehensive examination of optical design and practicality issues for such a facility and its instrumentation at the UK Astronomy Technology Centre (Atad-Ettedgui et al., 2000). This work has highlighted some of the difficulties in providing a realistic ``seeing-limited" capability for so large a telescope, amongst other instrumental issues, and also the potentially important problem of the degrading effects of atmospheric dispersion, even in the NIR, for image sizes of a few milli-arcsec. An active study group under the aegis of the European Southern Observatory (ESO) (c.f. Dierickx & Gilmozzi, 2000) is continuing the earlier design work.
The telescope is currently envisaged as being partly recessed into the ground, with the primary mirror a little above that level, so that much of the supporting structure is below ground level when observing near the zenith. The telescope would have the ability to tilt its optical axis down to the horizontal to allow engineering on the topend and to permit it to be protected in a roll-away structure when not in use. Informal cost estimates for this facility suggest that it can be built for of order $10e9; but as noted, scepticism about such estimates has in part been responsible for the concentration of other groups on 50-m alternatives.
That said, the vigorous studies under way in Europe are firmly based on demonstrated technololgies and are explicitly oriented towards industrial (mass) production, not only of the optics (see below, section 1.3 (2)) but also of the mechanical structure. Much of the final forming of the components of the latter is envisaged as occurring on site and even in situ ("self machining").