The classical external shape for a telescope has been for a long time a cylinder topped by a rotating hemispherical dome, which has a slit made such that it can have at any time only the opening strictly required for the view of the telescope. It is interesting to note that this concept was purely based on the intuition that a curved external shape would create less turbulence and therefore limit local seeing. The phenomenon of local seeing was very mysterious and the absence of hard evidence on its causes left dome designers relying on intuitive beliefs, such as that a smooth laminar air flow on the dome would be most favorable for the observation. The dimension of the dome was generally oversized with respect to the telescope, to protect it from wind loads and it was largely believed that any airflow on the telescope should be avoided. It was also thought that the telescope should be raised high above the ground level in order not to be affected by the near ground turbulence of the surface layer. Such ideas, although they were not quantified experimentally, determined the design of most existing telescope buildings.
Let us consider for instance the ESO 3.6-m telescope dome at La Silla
(fig. 
).
Its designers believed that they were giving their telescope all best
chances by locating it on the highest peak of the ridge; for good
measure they built a massive 30-meter high tower topped with an
oversize spherical dome (to prevent wind buffeting on
the telescope). As a result the whole building is very impressive and
the rotating dome is a particularly fine piece of heavy steel
engineering. But, as in other examples of 4-meter telescope domes
of the same generation, building and dome costed about 60 million DM
(of 20 years ago) and
accounted for about 60% of the entire telescope project cost.
Some evidence that something may have been overdone began to appear
a few years later when the ESO 2.2-m telescope was erected at La
Silla. Its building is much more modest than its large predecessor
(fig. ). It is also located on a lower location on the main
La Silla ridge.
The telescope is placed almost at ground level and the dome fits the
telescope with a minimal clearance.
These choices were at the time dictated by financial budget
constraints, nevertheless they did not seem to affect performance:
the 2.2-m telescope was reported to have
consistently better seeing than the 3.6-m ([Pedersen]).
While a proper comparison
should in particular assess the different thermal designs
of the two buildings, it is today quite obvious that there was
an evident disproportion in the previous dome between
means employed and results achieved.
