Abstract: The rhythmic, arrangement of vegetation into -dense bands separated by nearly bare ground (Phots. 1, 2 and 3) was first appreciated by Macfadyen (1950a, b) from a study of aerial photographs during the course of a geological and water-supply survey in the Somaliland Protectorate. He showed that such patterns occurred over very large areas of the flat Haud and Sol Haud plateaus and that the bands, which he called 'vegetation arcs', ran approximately parallel to the ground contours and appeared concave downslope in valleys. The pattern had already been noticed by Gillett (1941) and later by Gilliland (1952) but neither appears to have recognized that it might be extensive. Gillett noted the pattern as a 'grove-glade' sequence in the Acacia bussei association of the western part of the Somaliland Protectorate, and stated that all the groves were elongated in the same direction. Gilliland, working in the eastern half of the Protectorate, described a 'chevron pattern' as typical in the Acacia bussei-Dactyloctenium robecchi association, attributing the regularity to the constancy of the wind direction. However, he noted that similar arcs also occurred on alluvial flats and that the chevrons ran across the valleys and had no connection with the wind direction. Since the publication of Macfadyen's papers, Greenwood (1957) has observed that such vegetation arcs probably occur chiefly on clay soils in Somaliland. Arcs have also been seen in other regions, notably those described by Worrall (1959, 1960b) from the Sudan. Macfadyen (195Gb) states that arcs have been seen in the Kalahari desert, and less impressive arcs appear on aerial photographs of parts of Kenya and Tanganyika (Greenwood 1957). In the Wadi Houran area of Iraq small crescentic patches of vegetation have been seen on the ground by R. S. Seale (private communication) and arcs were seen by one of the authors from an airliner while flying over the Syrian desert. Apparently similar formations have been noted in West Africa in the curve of the River Niger (Clos-Arceduc 1956), while in Arabia what may be arcs are described by Vesey-Fitzgerald (1957) from the Jedda region, where he observed the ponding of water in vegetation patches in a way that resembles the accumulation of rainwater in arcs in Somaliland. More recently arcs have been recognized in Acacia aneura woodland in central and western Australia (Slatyer 1961; Litchfield & Mabbutt 1962). Thus arcs are widespread in arid regions but, as far as is known at present, are most characteristically developed in Somaliland. The areas of the Northern Region of the Somali Republic, formerly the Somaliland Protectorate, where they are found are shown in Fig. 1, compiled from aerial photographs. During the years 1955-59 the authors were