IN my paper on the Lizard this rock is mentioned, but without a specific name, for I had not then seen the figure given in 1898 by Rosenbusch (Elemente der Gesteinslehre, 1898, p.51). He calls it Beerbachite, and describes it on p. 219 as one of his Ganggesteine, among those von malchitischem Habitus, quoting an analysis. The name appears to have been given by Chelius to a rock from Frankenstein in the northern Odenwald, where it forms small veins in a gabbro, and has the following analysis: SiO2 = 47·21, Al2O3 = 20·52, Fe2O3 = 7·48, FeO = 5·32, MgO = 4·16, CaO = 8·63, Na2O = 5·17, K2O = 0·33, H2 = 0·34, P2O5 = 0·46, with 0·19 FeS2 and 0·10 hygroscopic water. He states that the rock is fine-grained to compact, not lustrous, grey to pale grey in colour, consisting of a panidiomorphic mixture of labradorite and diallage, with a variable amount of hypersthene and magnetite, and that olivine-bearing varieties occur. In these commonly a brown hornblende replaces the diallage, in which the above-named constituents are pæcilitic. He states that beerbachite also occurs in gabbro near Harzburg and in the Isle of Mull.

The diagnosis of this species given by Dr. Woodward in 1881 may be summarized as follows: Body with little or no signs of trilobation. Glabella oblong, with three pairs of lateral spines and three median spines, the anterior one apparently double. No evidence of cheek-spines. Eyes nearer genal border than in H. armatus, and rostrum more prominently developed. Thorax composed of thirteen segments with broadly expanded pleuræ; each axial ring with pair of widely separated spines. Pygidium imperfectly known. Total length of type-specimen 8 inches; breadth (measured along curve of thorax at widest part) 3¾ inches. Spines on thorax about 1 inch apart.

THERE are two hills named Castle-an-Dinas in West Cornwall, one near Penzance and the one on which the Wolfram Mine is situated, which lies about 2 miles east of St. Columb Major and about 4 miles to the north of the St. Austell granite mass.

Description.—This specimen is characterized by a subrectangular whorl-section with depressed venter, which is peculiar to Fontannes's group of “Phylloceras” tortisulcatum. The greatest thickness of the whorls is near the umbilical margin, but the sides are almost parallel and the ventral area is evenly convex, not flattened. The ratio of height to thickness is about 7: 6, as compared with 4: 3 in Sowerbyceras tortisulcatum, according to Pompeckj, and with 8: 7 at most in S. protortisulcatum Pompeckj sp. The umbilicus is comparatively large and deep; its walls are rounded and not abrupt. There are about four constrictions to the whorl, beginning as conspicuous grooves with a forward bend at the umbilicus, but becoming less distinct on the lateral area. After bending back again they become once more pronounced, passing across the periphery with a sinus that is concave forwards.

IN giving graphical expression to the data obtained by mechanical analyses of sediments, the use of curves was found to be very satisfactory. After some consideration the method finally adopted was to mark off along a horizontal abscissa a scale of percentage-weights, from 0 to 100, and upwards, along a vertical ordinate, a scale of grade-size limits from 0 to ·2 mm. and beyond. In cases where the range of the diameters of the grains in a sample is very great it would be necessary to employ a method of reducing the representative lengths, and this has sometimes been done by marking off lengths proportional to the logarithms of the diameters; but in the case of the arenaceous sediments dealt with by the writer, the necessity for this did not arise, and in the curves illustrating this paper, the ordinate lengths shown are directly proportional to the diameters represented. The length along an abscissa, intercepted between the first ordinate and the curve, represents the total percentage weight of the sample which, up to the limit of size specified, has been washed off, and the remaining length signifies the percentage-weight yet remaining to be dealt with. A tendency towards verticality on the part of the curve denotes relative absence or paucity of material of the sizes specified, and, correspondingly, a tendency towards horizontality signifies a preponderance of material of the size specified. Hence, a fine-grained sediment will give a curve occupying the lower part of the diagram, and with increasing coarseness of sediment the curve climbs successively higher and higher. (See Figs. 2, 3, and 6.)