In Part I of this paper the actual frequency-distribution of igneous rocks as revealed by the latest edition of Washington's collection of analyses was examined, and compared with that of earlier records. An attempt wilt now be made to determine the law of this distribution–a necessary preliminary to any petrogenic application of the results. The silica frequency will be considered since this has been found to give a curve more characteristic of igneous rocks than does any other oxide Now Dr. Harker pointed out that the silica distribution of records published in 1903 did not obey any simple law such as that of the probability curve, and the same conclusion applies to the collection of 1917. However, when the empirical method of Karl Pearson was applied to analyse the curve into component curves of error, fairly simple conclusions resulted.

The occurrence of brookite aud anatase within the limits of the Dartmoor granite exposure was first indicated by a sample of stream-sand taken two years ago below the bridge at Dartmeet–a locality which is noteworthy for sands rich ill these and other accessory minerals of the Dartmoor granite.

Anatase, either alone or associated with brookite in small amounts, was found in a majority of the sands taken the following year at selected points along (1) the Plym River, between Cadover Bridge and Hatter Tor, (2) the East Dart and the West Dart, upstream from Dartmeet, (3) the West Webburn, upstream from Ponsworthy, (4) the East Webburn, upstream from Lizwell Meet, and (5) minor streams in the Hamel Down-Bonehill Down area. Rutile proved to be a rare mineral, liable to be overlooked in a residue containing brown cassiterite, from which it cannot readily be distinguished except by microchemical means.

Thin sections afford a very inadequate basis for the determimation of thc accessory minerals present in the Dartmoor granite, and they are of little or no value for assessing the relative proportions of these minerals. A recognizable cryst'd of zircon, for example, is rarely seen iu thin sections of this granite, and uutil zircons were isolated in quantity from the crushed rock, stream-sands, &c., one type of zircon was not detected.

The occurrence of melanterite, pisanite, and chalcanthite as secondary minerals resulting from the oxidation of pyrites masses containing copper is well known. It is not so well known that other metallic sulphates are from time to time found, and a brief description of the mode of occurrence of a few of these, in addition to some details of the composition of pisanite and chalcanflSte, may be of interest.

Pisanite.—This mineral is of common occurrence at all the pyrites mines of Huelva and elsewhere, ibrming stalactites and stalagmites in those parts of the mines that are above drainage level though subject to more or less percolation of water. It is rarely found as well-formed crystals, save occasionally when a pool ef strong sulphate water has been temporarily dammed hack in some way, and then supersaturated through evaporation.