This paper attempts to reinterpret the landforms of deglaciation in the Cairngorm mountains. Most deglacial features are fluvioglacial in origin. Ice-directed meltwater channels and associated eskers are often discordant with the older relief and their orientation reflects the dominant direction of ice movement. On the floors of the glens (valley), valley-controlled fluvioglacial forms accord more closely with the underlying bedrock surface and represent meltwater deposition in stagnant ice. These latter depositional features, formerly described as 'hummocky moraine' and related to a separate valley-glacier readvance, are interpreted merely as marking the last stages of decay of a progressively downwasting ice sheet. True ice-dropped moraines are rare in the Cairngorms, but those that do exist are thought to represent no more than minor fluctuations of the ice margin during the last deglaciation. The problem of the Cairngorm equivalent of the Zone III readvance (Younger Dryas) is discussed. Tentatively, it is suggested that it might be represented by a relatively extensive east Grampian ice cap directly inherited from the last ice sheet. However, without further work over a wider area, the possibility that it might be represented by a minor readvance in a restricted number of cirques cannot be excluded. AN IRREGULAR mantle of glacial drift clothes the lower slopes of the Cairngorm mountains, a granite massif comprising the culminating summits of the eastern Grampians (c. 13 10 m). In the flatter parts of the mountain glens (valleys), the drift commonly attains a thickness of 10-30 m, while near the foot of the northern flank of the mountains thicknesses of about 1oo m occur in places. The drift thins with increasing altitude until it becomes indistinguishable from the regolith which covers most of the gently rounded mountain summits. In addition to the drift, meltwater channels incised into both solid rock and drift are common; especially distinctive are those that cut the skylines of many higher spurs, particularly around the northern periphery of the massif. There have been many attempts to explain the pattern of the drift and channels. T. F. Jamieson (I908) and officers of the Geological Survey (L. W. Hinxman et al., I896, I915; G. Barrow, I912) described the landforms and suggested that, following the last glacial maximum when an ice sheet covered the mountains, the drift was deposited by a series of valley glaciers whose snouts withdrew up the glens to the corries (cirques). Essentially the same idea was envisaged by A. Bremner (I929) when he described the characteristic landforms of Glen Derry, and by J. K. Charlesworth (I955) who described the retreat of each valley glacier glen by glen. More recently, J. B. Sissons (1967) has suggested that morainic forms in the glens down to an altitude of 425-480 m are related to the Loch Lomond readvance when, as in other massifs in the eastern Grampians, valley glaciers a few kilometres long flowed out radially from the higher mountains. He argues on a wider scale that the hummocky drift associated with this readvance is quite distinct from the drift associated with the wasting of the last major Scottish ice sheet. This paper, which is based on mapping at a scale of I : 10 560 with the aid of vertical