Temporal Responses of Propagule Banks during Ecological Restoration in the United Kingdom J. Ghorbani, 1,2,3 M. G. Le Duc, 1 H. A. McAllister, 1 R. J. Pakeman, 4 and R. H. Marrs 1 Abstract Successful ecological restoration is expected to be accom- panied by change in the propagule bank. We tested for temporal change in the soil propagule bank at two Bracken (Pteridium aquilinum)-infested sites in the United Kingdom (acid grassland and heathland), each with replicate experiments. A combination of bracken control (cutting, spraying, and combinations) and vegeta- tion restoration treatments (seeding, fertilizer, harrowing) were applied. Soil propagule banks were sampled in 1998, 4–5 years after the start, and in 2003 after a further 5 years. We used univariate and multivariate statistical methods to investigate the response of the propagule bank to experimental treatment in space and time. Few effects were found in 1998, but after a further 5 years, the propa- gule bank size and composition changed significantly, implying that propagule bank development lags behind vegetation development. The effect of treatment on the propagule bank differed within sites. Thus, ecosystem development is occurring at different speeds and direc- tions even in closely adjacent areas. Coupling between the propagule bank and vegetation changed between sampling times with either some new coupling or decoupling. At both sites, the propagule bank contained propagules of the target community and thus propagule bank develop- ment during restoration provides increased potential for vegetation recovery. However, it can take a considerable time for management effects to be detected in the pro- pagule bank. Moreover, the effect or speed of effect is spatially variable. Continuing application of restoration treatments is recommended at heathland where there is a deep bracken litter layer. Key words: Agrostis capillaris, bracken control, Calluna vulgaris, diaspore bank, ecosystem recovery, invasive species, Pteridium aquilinum, seed bank, spore bank. Introduction In ecological restoration, there is a need to understand the key components of the target ecosystem, and especially how they change during ecosystem development (Hobbs & Norton 1996). One of these key components is the soil propagule bank. Although species in this bank can play an important role in the early phases of vegetation develop- ment (Bakker et al. 1996), the species composition is itself affected by successional change during the restoration process, as the processes affecting the propagule bank are dynamic (Harper 1977). A functioning soil propagule bank must, therefore, be considered part of the outcome of successful restoration in temperate ecosystems (Willems & Bik 1998). However, processes affecting the propagule bank may be slow, and often lag well behind changes in the vegetation during ecological restoration (Bekker et al. 1997; Akinola et al. 1998; Smith et al. 2002; Ghorbani et al. 2003). This is especially true when the propagule bank of the starting conditions is dominated by species with long-lived seeds that are perhaps considered weeds in the target community, and contributions of new species from the vegetation is low (Williams 1985; Milberg 1995; Grandin 2001). Bracken (Pteridium aquilinum (L.) Kuhn) invasion is a serious cause of species decline and habitat degradation in the United Kingdom (Pakeman & Marrs 1992). Dense stands with deep litter layers displace plant communities with a greater conservation value (Marrs et al. 2000). Con- ditions under a bracken canopy may influence species composition within the underlying soil propagule bank and its dynamics through (1) a low input of propagules, either from a reduced production by depauperate vegeta- tion or due to the trapping of propagules in the inhospita- ble litter and (2) a reduced germination from the propagule bank because of an unfavorable microclimate (Pakeman & Hay 1996). The balance between these pro- cesses determines the propagule turnover rate, which is generally low under bracken. The propagule bank under bracken tends to comprises mostly of very long-lived seeds, of species with a low representation in the vegeta- tion; that is, there has been a decoupling between the propagule bank and the vegetation (Pakeman & Hay 1996; Ghorbani et al. 2003). Previous studies have shown that bracken control by herbicide (asulam) is initially very successful, but there is often rapid recovery unless follow-up treatments are applied (Marrs et al. 1998a; Le Duc et al. 2000). Regular cutting tends to be the most successful treatment in 1 Applied Vegetation Dynamics Laboratory, School of Biological Sciences, University of Liverpool, Liverpool L69 7ZB, U.K. 2 Address correspondence to J. Ghorbani, email jamshid@liv.ac.uk 3 Department of Rangeland and Watershed Management, Faculty of Natural Resources, University of Mazandaran, PO Box 737, Sari, Iran. 4 Macaulay Land Use Research Institute, Craigiebuckler, Aberdeen AB15 8Q, U.K. Ó 2007 Society for Ecological Restoration International MARCH 2007 Restoration Ecology Vol. 15, No. 1, pp. 103–117 103