The occurrence of iron carbonate as a part of the cement in the Pennant Sandstone is not extraordinary, when the large amount of entombed organic matter is taken into consideration. The presence of organic material and the products of decay, the so-called humus acids, acting as reducing agents, would aid in the formation of iron carbonate and prevent its oxidation. Much of this carbonate probably owes its presence to a metasomatic replacement of calcium carbonate. The granules of siderite may have been derived from rocks similar to those of the Lower Coal Series of South "Wales. The alkaline carbonates which hold the humic substances in solution would readily dissolve silica, thus becoming effective agents in the solution and redeposition of silica in the rock. A solvent of this kind would be formed by the impregnation of alkaline salts by carbonic acid.1 The occurrence of so much secondary silica in the rock is probably the result of a redeposition of silica from such a solution. No oxides of iron have been found in any of the numerous separations, either by means of heavy liquids or the electro-magnet; their absence is easily explained when we consider the strongly reducing environment of the deposition and consolidation of the Pennant Sandstone. The perfectly rounded globules of sphserosiderite may have been precipitated whilst the sandstones were being deposited, but more probably derived from the denudation of fireclays similar to those of the Lower Coal Series of South Wales.2 The dominant mineral chlorite, existing as perfect cleavage flakes, is entirely distinct from the secondary aggregates of chlorite in the rock, and was probably deposited in its present form. I have examined a powdered specimen of the Delabole Slates of Cornwall (Upper Devonian), and have found in it a large proportion of chlorite flakes, exactly like the detrital chlorite of the Pennant Sandstone. The very few flakes of biotite found show little trace of alteration. The paucity of heavy minerals indicates that the Pennant sediments had not a granitic source; they were mainly derived from pre-existing sediments which were rich in chert, had abundant chloritic material present, and yet were capable of furnishing fairly fresh biotite. Metamorphosed shales would account for the presence of cordierite and corundum.