e been expected to occur if the
neighbouring volcanoes had been in activity during its existence. Hence
it may be supposed that, as Scrope suggested, the waters of the lake
were drained off owing to the disturbance in the levels of the country
caused by the first explosions of the Auvergne volcanoes.[4] If this be
so, then we possess a key by which to determine the period of the first
formation of volcanoes in Central France; for, as the animal remains
enclosed in the lacustrine deposits of the Vale of Clermont belong to
the early Miocene stage, and the earliest traces of contemporaneous
volcanic _ejecta_ are found only in the uppermost deposits, we may
conclude that the first outburst of volcanic action occurred towards the
close of the Miocene period--a period remarkable for similar exhibitions
of internal igneous action in other parts of the world.

(_c._) _Successive Stages of Volcanic Action in Auvergne._--The volcanic
region here described, which has an area of about one hundred square
miles, does not appear to have been at one and the same period of time
the theatre of volcanic action over its whole extent. On the contrary,
this action appears to have commenced at the southern border of the
region in the Cantal, and travelling northwards, to have broken out in
the Mont Dore region; finally terminating its outward manifestations
among the craters and domes of the Puy de Dome. In a similar manner the
volcanic eruptions of the Haute Loire and Ardeche, lying to the
eastward, and separated from those of the Cantal by the granitoid ridge
of the Montagnes de Margeride, belong to two successive periods
referable very closely to those of the Mont Dore and the Puy de Dome
groups.[5] The evidence in support of this view is very clear and
conclusive; for, while the volcanic craters formed of ash, lapilli, and
scoriae, together with the rounded domes of trachytic rock of which the
Puy de Dome group is composed, preserve the form and surface indications
of recently extinguished volcanoes, those which we may assume to have
been piled up in the region of Mont Dore and Cantal have been entirely
swept away by prolonged rain and river action, and the sites of the
ancient craters and cones of eruption are only to be determined by
tracing the great sheets of lava up the sides of the valleys to their
sources, generally situated at the culminating points of their
respective groups. Other points of evidence of the great antiquity of
the