It is equally certain that the commencement of their activity was at a late period in the history of the earth. Lava currents are frequently found in France resting upon the early tertiary strata, but no lava current is found below them. The later tertiary strata contain pebbles of volcanic rocks, showing that lavas had been previously ejected, but none are found in the older strata of this formation. We must, therefore, conclude that these volcanic tracts assumed their volcanic character at some intermediate point in the tertiary period. When we find that their activity commenced at so late a period and closed so long ago, we might be led to suppose that it was of very short duration. But a great number of facts, in the present condition of the country, require that we should assign to them a very prolonged activity. A single instance will be sufficient to show the nature of the evidence upon which this conclusion rests. The heavy line (Fig. 6) represents the present form of one of the valleys. A bed of lava forms the highest point of land represented, and a second bed is found in an intermediate part of the slope. The position of the upper bed must have been a valley, when the lava flowed there. We may represent this valley by the line _a b c_. The slow operation of natural denuding causes at length excavated the valley _d e h_, when another lava current flowed through it, covering its bed of pebbles, as before. The same denuding causes have at length produced the present valley, _f g h_. These remnants of lava-currents, as they have formed a very imperishable rock, have protected the subjacent strata from erosion, and furnish evidence of the position of the valley at different periods. When we consider with what extreme slowness denuding causes produce changes on the surface, and what extensive changes they have here nevertheless effected in the interval between the production of the different lava currents, we are compelled to feel that that interval was a very prolonged one. Yet this period, however long it may have been, was evidently less than the period of activity of these volcanoes. [Illustration: Fig. 6.] 3. _Volcanic rocks of an earlier date_ are also found, sometimes as distinct lavas, though generally as volcanic grits. They occur interstratified with the cretaceous rocks, and with every other formation of the fossiliferous series, showing that, from the earliest times, these rocks have been accumulating as they now are. _The trappean rocks_ may, in a general classification, be considered as volcanic. It will be shown, hereafter, that they are the lavas of submarine volcanoes. They do not, however, occur in the form of lava currents, but in great tabular masses, generally between stratified rocks, or in the form of dikes. They are also entirely unconnected with cones or craters.