Moraines marking an Early Pleistocene glaciation on Mt. Kenya are known from several valleys along the eastern and southeastern flanks of the mountain. The most prominent group of end moraines delimiting the lowermost extent of the Early Pleistocene Gorges Glaciation (2900 m above sea level) overlie either older till or weathered bedrock, the latter composed of thin pedostratigraphic remnants covered with locally derived aeolian sediments truncated by the ingress of ice. The loess, tills and interbedded palaeosols of the Gorges Glaciation crop out adjacent to the Nithi River, which drains the Gorges Valley, and abut the upper timberline of montane forest. The lowermost palaeosols at these sites were formed either in till predating the Gorges Glaciation or in weathered bedrock of trachytic textured lavas, similar to the lithology forming the base of the Mt. Kenya volcanic series of Miocene or Pliocene age. Palaeomagnetism and weathering characteristics are here used to refine the age of sediments assigned to the Gorges Glaciation. These normally magnetized deposits carry a persistent reversed overprint, suggesting that they were deposited during one of the normal subchrons within the Matuyama Reversed Chron. They are underlain either by a reversely magnetized till, or by weathered bedrock and lower palaeosol, the latter exhibiting normal magnetization (Gauss?) with reversed overprint (Matuyama?). The sediments of the Gorges Glaciation are overlain at all three sites by normally magnetized loesses and palaeosols, presumably of Brunhes age but carrying a high percentage of well-weathered recycled grains. The normal magnetization and reversed overprint recorded in sediments of the Gorges Glaciation most probably span a considerable portion of the Olduvai subchron (1.78–1.950 Ma), which persisted for sufficient time to accommodate an extensive montane glaciation followed by a prolonged period of weathering and soil formation. The importance of these sediments in global palaeoclimate reconstruction and insolation changes is discussed. The stratigraphic evidence contradicts the cosmogenic exposure ages presented by previous workers attempting to obtain true ages of deposits using exhumed coarse clastic sediment without analysis of complex pedostratigraphic complexes.
Tables 1S–5S are available at www.geolsoc.org.uk/SUP18592.