The church at Sauges has an early and remarkable belfry. An immense arch, richly moulded, admits to a porch. Above this is a still larger relieving arch to sustain the octagonal tower that is on two stages. Granite and black basalt are employed in bands and in the arches of the windows, two-light in the tower story, single in that above, and the whole is capped by a dwarf spire. Near Sauges is the Tour de la Clauze, erected on a protuberant mass of granite fissured into blocks. The rest of the castle is completely ruined. But that which is most curious at Sauges is a monumental structure composed of a cubical base, on which stand four pillars supporting arches and a vault with groined ribs. This goes by the name of the Tombeau du Général Anglais, and is supposed to have been set up in honour of a Captain MacHarren, who commanded one of the mixed companies of English and Gascons that held the land or harried it for the English Crown nominally, actually for themselves. This MacHarren was probably one of the English garrison that held Sauges till 1360, when they were driven out by the Viscount Polignac. La Voute-Chilhac down the river stands on a peninsula between the Allier and the Avesne that here debouches into it. It possesses a church of the fifteenth century that has taken the place of one erected by S. Odilo of Cluny in 1075. The original door-valves remain, but injured by cutting to make them fit the ogee portal. In the midst it bears the inscription:— "Hic tibi rex regum hoc condidit Odilo templum Agminibus superis quem miscuit arbiter orbis." There were other inscriptions, but they have been mutilated. Chilhac stands on a rock composed in the lower portion of beautiful prismatic columnar basalt, capped with an amorphous flow. It is curious how sharp the line of demarcation is between the two beds. The situation is pretty, the church Romanesque. The course of the Allier above Langeac presents many faces like organ fronts of basalt; in places the pillars form a _pavé de géants_. The prisms are employed along the roads to mark distances, and might easily be supposed to have been specially cut for the purpose. But all lava does not crystallise into prisms; under pressure it does. When not squeezed by superincumbent beds it is cinderous. But there is another form it assumes, that of phonolith or clinkstone, flakes that can be cut like slates and divided into laminæ. As slates they are employed extensively in Velay. But why the ejected lava should form films here and prismatic pillars there, I do not comprehend.