Hence we are led by an examination of the internal structure of volcanic mountains to conclude that scoriæ- and tuff-cones, and cones formed of very liquid lavas, increase by an _exogenous_ mode of growth, all new materials being added to them from without; in the cones formed of very viscid lavas, on the other hand, the growth is _endogenous_, taking place by successive accretions within it. The composite cones owe their origin to both the _exogenous_ and the _endogenous_ modes of growth, but in a much greater degree to the former than the latter. The layers of scoriæ, tuff, and dust, and the successive lava-streams are added to the mass from without, and the lava forming the dykes from within it. [Sidenote: THEORY OF ELEVATION CRATERS.] There are doubtless cases in which, when a tuff-cone is formed, a mass of very viscid lavas is extruded into its interior, and the mass is distended like a gigantic bubble. But inasmuch as the very viscid lavas do not appear to give rise to scoriæ to anything like the same extent as the more liquid kinds, such 'cupolas,' as they have been called by some German geologists, are probably not very numerous, and may be regarded as constituting the exception rather than the rule. The idea which was formerly entertained by some geologists that all great volcanic mountains were formed of masses originally deposited in a horizontal position, and subsequently blown up into a conical form, has been effectually disposed of by the observations of Lyell and Scrope. The condition of the great fluid masses which underlie volcanic vents is another point on which much light has been thrown by the study of naturally-dissected volcanoes. In some cases, as was shown by Hochstetter during his admirable researches among the New Zealand volcanoes, the rising lavas form a great chamber for themselves in the midst of a volcanic cinder-cone, taking the place of loose materials which are re-ejected from the vent, or have been re-fused and absorbed into the mass of lava itself. From this central reservoir of lava, eruptions are kept up for some time, but when the volcano sinks into a state of quiescence the lava slowly consolidates. In such slowly solidified masses of lava, very beautiful groups of radiating columns are often exhibited Northern Germany abounds with examples of such basaltic masses, which have once formed the centres of great cinder-cones; but in consequence of the removal of the loose materials and the surrounding strata by denudation, these central reservoirs of the volcanoes have been left standing above the surface, and exhibit the peculiar arrangements of the columns formed in them during the process of cooling. [Sidenote: INTRUSIVE LAVA-SHEETS.]