Keller peninsula is an ice-free area situated in the Admiralty Bay, in King George Island, formed by volcanic rocks (basalts, andesites, dacites, tuffs and agglomerates) of ages ranging from Jurassic to Late Tertiary, as part of the Andean orogenic belt. The manifestation of physical weathering in Keller, like elsewhere in the Maritime Antartica is extreme due to a combination of factors: relatively high moisture availability, very low temperatures, fractured volcanic-rock substrates and mountainous topography. The absence of any glacial outlet coming from inland areas of Keller Peninsula hinterland means that most till and related deposits are derived from short-ranged glacial erosion from the upper parts of the peninsula. In this work, we examine the main cryogenic features, soils and landforms of this area. In the volcanic-rock regolith of Keller Peninsula, a number of cryogenic processes leads to physical breakdown, sorting and fineparticles increase. The most important factors are freezing-thawing cycle, ablation, frost and gelifluction. Frost action causes the formation of a thin silty, crusting layer at the suface, especially on andesitic to basaltic materials, but also on acid tuffs and palagonite. These features are particularly frequent in the area just below the Birknmajer Peak Glacial Cirque, where basalts are dominant. During the summer, the permafrost within the soil regolith can be observed at varying depths. In areas along channels of melting water, hollows and depressions, where accumulations of fine materials are greater, the intensity of freezing-thawing cycles is marked. This causes tonguing of sucessive short-distance, lobate moraines as well as active surface geliturbation by melting water down the regolith. In sloping areas, soil creep and regolith gelifluction occurs, being readily distinguished by the redistribution of Usnea- covered gravels and rock fragments at distances of hundred of meters far from the source areas, where Usneaceae cover rock-stable crests. The summer flow of water-saturated regolith is greater in palagonite and andesitic tuffs, due to finer particle sizes, but occurr even in basaltic materials with larger rock fragments. Soil and regolith creep appears to be a slow movement process, as many areas of basalt dykes cutting tuffs ant other andesitc lithologies show only short-range redistribution of rock fragments downslope as thin moraines. Thus, it appears that most rock fragments present at mid to high levels of Keller Peninsula are, in fact, the results of in situ physical breakdown of rock substrates. Gelifluction was found to be an important process for sheet redistribution of soil/regolith materials downslope, particularly where subsurface melting water is abundant. This process of soil movement seems to be greater in acid tuffs and andesitic lithologies, where regolith are usually deeper, and surface rock-fragmentation is of limited importance. On the other hand, Gelifluction is less in basalt and other mafic lithologies, due to a greater stonyness of the surface. Gelifluction accounts for the devolopment of lobate crescentic sheets of stony pavement, at mid and downslope positions. We have also observed that soil creep (and solifluction) is controlled by some structural features of the bedrock, such a fault-lines, presence of unconformites, dykes and, especially near the coast, by uplifted (isosthatic) marine terraces and “structural tabular terraces”. The cryogenic landscape, therefore, is considerably “stable” and little long-range redistribution of debris is actually on going.