Ch. xxxii] teetiary volcanic rocks. 549 



When we follow the course of the river Couze, <from its source in the 

 lake of Charabon, to the termination of the lava-current at Nechers, a dis- 

 tance of thirteen miles, we find that the torrent has in most places cut a 

 deep channel through the lava, the lower portion of which is columnar. 

 In some narrow gorges the water has even had power to remove the 

 entire mass of basaltic rock, though the work of erosion must have been 

 very slow, as the basalt is tough and hard, and one column after another 

 must have been undermined and reduced to pebbles, and then to sand. 

 During the time required for this operation, the perishable cone of Tar- 

 taret, composed of sand and ashes, has stood uninjured, proving that no 

 great flood or deluge can have passed over this region in the interval 

 between the eruption of Tartaret and our own times. 



If we now return to the section (fig. 676), we may observe that the 

 lava-current of Tartaret, which has diminished greatly in height and 

 volume near its termination, presents here a steep and perpendicular 

 face 25 feet in height towards the river. Beneath it is the alluvium 

 No. 9, consisting of a red sandy clay, which must have covered the 

 bottom of the valley when the current of melted rock flowed down. 

 The bones found in this alluvium, which I obtained myself, consisted 

 of a species of field-mouse, Arvicola, and the molar tooth of an extinct 

 horse, Equus fossilis. The other species, obtained from the same bed, 

 are referable to the genera Sus, Bos, Ccrvm, Fells, Canis, Maries, Talpa, 

 Sorex, Lepus, Sciurus, Mus, and Logomys, in all no less than forty- 

 three species, all closely allied to recent animals, yet nearly all of them, 

 according to M. Bravard, showing some points of difierence, like those 

 which Mr. Owen discovered in the case of the horse above alluded to. 

 The bones, also, of a frog, snake, and lizard, and of several birds, were 

 associated with the fossils before enumerated, and several recent land 

 shells, such as Cydostoma elegans. Helix liortensis, H. nemoralis, H. Za- 

 picida^ and Clausilia rugosa. If the animals were drowned by floods, 

 which accompanied the eruptions of the Puy de Tartaret, they would give 

 an exceedingly modern geological date to that event, which must, in that 

 case, have belonged to the Newer-Pliocene, or, perhaps, the Post-Plio- 

 cene period. That the current, which has issued from the Puy de Tar- 

 taret, may nevertheless be very ancient in reference to the events of 

 human history, we may conclude, not only from the divergence of the 

 mammiferous fauna from that of our day, but from the fact that a Roman 

 bridge of such form and consb'uction as continued in use down to the 

 fifth century, but which may be older, is now seen at a place about a 

 mile and a half from St. Nectaire. This ancient bridge spans the river 

 Couze with two arches, each about 14 feet wide. These arches spring 

 from the lava of Tartaret, on both banks, showing that a ravine pre- 

 cisely like that now existing, had already been excavated by the river 

 through that lava thirteen or fourteen centuries ago. 



In Central France there are several hundred minor cones, like that 

 of Tartaret, a great number of which, like Monte Nuovo, near Naples, 

 may have been principally due to a single eruption. Most of these cones 



