We assessed the effects of a range of experimental restoration treatments on the diaspore bank (seed and spores) on an upland moor infested by dense Pteridium aquilinum, 9 years after implementation. Experimental treatment included both Pteridium control (cutting, spraying and combinations) and restoration treatments (grazing, seeding); both univariate and multivariate analyses of variance were used to assess significant effects. The seed bank was made up largely of Calluna vulgaris, Juncus effusus and Agrostis capillaris with the majority germinating from the litter (38 per cent) and upper soil layer (53 per cent). Soil depth influenced species composition and density, with greatest numbers near the surface, in keeping with a negative exponential response. In the lowest soil depth only ferns were abundant. A. capillaris seed density was greater in the most disturbed treatment (cut twice yearly with grazing), whereas C. vulgaris was greater in the low-disturbance treatment (herbicide application only). The diaspore bank composition response identified that certain treatment combinations had a grass-heath composition in the upper soil layer, which may assist future restoration. There was a significant effect of C. vulgaris seed addition detected in the Pteridium litter layer suggesting that this litter acted as a barrier to seed transfer to the soil. There was little evidence of coupling developing between the diaspore bank and vegetation. The most successful treatment combination was cutting twice per yr with low grazing plus C. vulgaris addition as brash. Where seed is added the Pteridium litter should be disturbed. Copyright © 2007 John Wiley & Sons, Ltd.