Fig. 307: Cyprides from the Lower Purbeck. Fig. 308: Mantellia nidiformis. _Dirt-bed or ancient Surface-soil._—The most remarkable of all the varied succession of beds enumerated in the above list is that called by the quarrymen “the dirt,” or “black dirt,” which was evidently an ancient vegetable soil. It is from 12 to 18 inches thick, is of a dark brown or black colour, and contains a large proportion of earthy lignite. Through it are dispersed rounded and sub-angular fragments of stone, from 3 to 9 inches in diameter, in such numbers that it almost deserves the name of gravel. I also saw in 1866, in Portland, a smaller dirt-bed six feet below the principal one, six inches thick, consisting of brown earth with upright _Cycads_ of the same species, _Mantellia nidiformis,_ as those found in the upper bed, but no _Coniferæ._ The weight of the incumbent strata squeezing down the compressible dirt-bed has caused the _Cycads_ to assume that form which has led the quarrymen to call them “petrified birds’ nests,” which suggested to Brongniart the specific name of _nidiformis._ I am indebted to Mr. Carruthers for Figure 308 of one of these Purbeck specimens, in which the original cylindrical figure has been less distorted than usual by pressure. Many silicified trunks of coniferous trees, and the remains of plants allied to _Zamia_ and _Cycas,_ are buried in this dirt-bed, and must have become fossil on the spots where they grew. The stumps of the trees stand erect for a height of from one to three feet, and even in one instance to six feet, with their roots attached to the soil at about the same distances from one another as the trees in a modern forest. The carbonaceous matter is most abundant immediately around the stumps, and round the remains of fossil _Cycadeæ._ Besides the upright stumps above mentioned, the dirt-bed contains the stems of silicified trees laid prostrate. These are partly sunk into the black earth, and partly enveloped by a calcareous slate which covers the dirt-bed. The fragments of the prostrate trees are rarely more than three or four feet in length; but by joining many of them together, trunks have been restored, having a length from the root to the branches of from 20 to 23 feet, the stems being undivided for 17 or 20 feet, and then forked. The diameter of these near the root is about one foot; but I measured one myself, in 1866, which was 3½ feet in diameter, said by the quarrymen to be unusually large. Root-shaped cavities were observed by Professor Henslow to descend from the bottom of the dirt-bed into the subjacent fresh-water stone, which, though now solid, must have been in a soft and penetrable state when the trees grew. The thin layers of calcareous slate (Fig. 309) were evidently deposited tranquilly, and would have been horizontal but for the protrusion of the stumps of the trees, around the top of each of which they form hemispherical concretions. Fig. 309: Section in Isle of Portland, Dorset.