Pollen diagrams from four neighboring sites are examined to try and elucidate the relationship between topography and erosion in southern Pennine blanket peats. Variations in the depths of peat formed in the time intervals between six probably synchronous horizons are shown to parallel closely the variations in the pattern of erosion. The indications are that erosion of some sort may have been operating for many centuries, but that it is unlikely that a single cause can be invoked to "explain" the erosion. Erosion probably started along pre-peat stream curves in zone VIIb and increased in intensity following the climatic deteriorations of 1200 B.C. and 600 B.C. Erosion may have set in widely after A.D. 400 and become accentuated after A.D. 1000. In recent centuries there has been extensive sheet erosion of the peat caused probably by active human interference with the vegetation.