Notes and records Rodent-mediated dispersal of Acacia seeds in Kalahari savannah rangelands – implications for bush encroachment Helga Ursula Kuechly, Julia Sophie Mueller, Vanessa Luisa Reinfelder, Stephan Wiedemann and Niels Blaum* Department of Plant Ecology and Nature Conservation, Univer sity of Potsdam, Maulbeerallee 2, 14469 Potsdam, Germany Introduction Bush encroachment caused by heavy livestock grazing is a serious form of land degradation in savannahs worldwide (Sankaran et al., 2005), reducing livestock carrying capacity and occasionally also leading to habitat frag- mentation and loss of species diversity (e.g. Blaum, Ross- manith & Jeltsch, 2007a; Blaum et al., 2007b, 2009). In African savannahs, Acacias play a significant role in bush encroachment (Wiegand, Ward & Saltz, 2005). Despite the ecological and economic relevance of bush encroach- ment, the impact of rodent-mediated dispersal and caching of seeds to bush encroachment remains poorly understood. This is surprising as rodents can have a major impact on plant succession (e.g. Brown & Heske, 1990) and seed caching by rodents occurs across ecosystems (e.g. Vander Wall, Kuhn & Beck, 2005). The importance of rodent dis- persal is not limited to northern ecosystems but has also been found broadly in southern Africa (e.g. Miller 1994, 1996, Midgley et al., 2002). This study assessed the role of rodents in Acacia seed dispersal in Kalahari savannah rangelands (including two Acacia species listed as bush encroachers, Department of Agriculture, South Africa, 2010). Methods Study area The study area is in the southern Kalahari, on the farm Loch Lomond (27°15¢S, 20°63¢E), Northern Cape Prov- ince, South Africa. The vegetation is described as the Aoub dune veld an open shrub land, with trees and shrubs (Acacia erioloba, E. Mey, Acacia haematoxylon, Willd., Acacia hebeclada Willd., Rhigozum trichotomum, Burch. and Acacia mellifera, Benth.) scattered in a grassy landscape (Stipa- grostis spp, Eragrostis spp and Schmidtia kalahariensis, Stent) (Mucina & Rutherford, 2006). Seed dispersal experiment Seed dispersal of four Acacias (A. erioloba and A. haema- toxylon with indehiscent seed pods adapted for ungulate seed dispersal, and A. mellifera and A. hebeclada with dehiscent seed pods nonadapted for ungulate seed dis- persal, Miller, 1994) was analysed on a 50 m · 60 m experimental plot (divided into nine 16 m · 20 m sub- plots) during April 2009, a time when both rodent (Blaum, Rossmanith & Jeltsch, 2007a) and seed rain would be expected to be at their highest. Vegetation cover was re- corded on three 4 · 4 m squares per subplot by measuring the maximum perimeter of each plant. Bushes and trees were determined to species level. Total vegetation cover was 28 ± 3% (A. mellifera: 6 ± 3%, R. trichotomum: 7 ± 3% A. erioloba and A. haematoxylon: <1%, perennial grasses: 11.2 ± 2%, annual grasses: 3 ± 1%, and herba- ceous vegetation: <1%). For each subplot, Acacia seeds were offered in four treatments for three consecutive days between half moon and new moon similar to Walters et al. (2005) in dishes (i) mounted on a pole 1.5 m above ground (accessible by birds and flying insects), (ii) on the ground covered with mesh wire, mesh size 1.5 cm to exclude birds and rodents and (iii) mesh size 2.5 cm to exclude birds only, and (iv) in open dishes on the ground serving as a control. Each seed dish was filled with twenty Acacia seeds (five seeds per species). In total, 720 Acacia seeds were offered daily (20 seeds · 4 treatments · 9 subplots). Treatments (i–iii) were arranged in the centre of each subplot and treatment (iv) at the lower right hand corner. Fluorescent dye (JST43, http://www.radiantcolor.be) was applied around the seed dishes of treatments (iii) and (iv) to follow fate of seeds dispersed by rodents. *Correspondence: E-mail: blaum@uni-potsdam.de In memory of Katharina Anne Osterried. Ó 2010 Blackwell Publishing Ltd, Afr. J. Ecol., 49, 119–121 119