largely from that brought down by the Clarence River in the past and comprises greywacke, argillite, Amuri limestone, and various igneous types. Here windkanter are found at several different levels corresponding with the development of bedded sands. These sands, on weathering, yield worn boulders which form the raw material upon which facets are cut. The levels higher and farther inland usually exhibit specimens with the largest and most perfect facets. On the lowest levels only small fragments show facetting, and only the relatively soft limestone has developed into fully shaped (mature) specimens, the harder greywacke and igneous material apparently taking longer to cut. On the intermediate levels these latter also assume in the case of smaller specimens, the typical windkanter shapes. The higher levels are sometimes characterized by an almost entire absence of limestone specimens, while quite large boulders of greywacke a foot in diameter and measuring six inches or more along the facet-slopes may show fully developed facets. The absence of limestone examples may possibly be due to an original lack of such material; but the presence in localized patches of abundant facetted limestone pebbles indicates rather that, where they are rare or absent, they have been abraded and dissolved away while the grreywacke boulders were being cut, and that their presence in other patches is due to being covered for long periods by sand-dunes. Conditions of wind abrasion are very similar to those at the Ure, the north-east and south-west winds exercising a far greater power of abrasion than any others.* King, L. C., loc. cit., p. 202. The Material. Comparison with the specimens from the Ure District reveals many points of difference, some of which can readily be referred to a general contrast in the shape of the fragments provided. Whereas those north of the Ure River are mainly cut from polygonal plates characterized by sharp angles, typical of material freshly shed from an outcrop, the majority of the specimens at the Clarence locality are developed from originally rounded, river-worn boulders exposed by the weathering of the bedded sands in which they occur. This fundamental difference is in great measure reflected in the form of the subsequent windkanter. Consequently, some of the principles enunciated in the writer's previous paper require some adaptation when applied to the new material. In particular no discussion was there instituted as to the sequence of forms which might be expected to develop from rounded fragments or stream-worn boulders. Rounded material undoubtedly offers greater opportunity for the direction of the facets to be governed by wind-direction than does angular material on which plane surfaces are already developed; and in this connection it may be noted that “roof-shaped” pebbles (Firstkanter) are by far the commonest type developed at the Clarence locality (Table V) (Figs. 2, 4). In the writer's previous