Why is Abundance of Herbaceous Legumes Low in African Savanna? A Test with Two Model Species Patrick G. Cech 1 , Peter J. Edwards, and Harry Olde Venterink Institute of Integrative Biology, ETH Zurich, Universit ¨ atsstrasse 16, CH-8092, Zurich, Switzerland ABSTRACT Although fire is frequent in African savanna ecosystems and may cause considerable loss of nitrogen (N), N 2 -fixing herbaceous legumes—which could be expected to benefit from low N conditions—are usually not abundant. To investigate possible reasons for this scarcity, we conducted a pot experiment using two common plants of humid African savannas as model species, the legume Cassia mimosoides and the C 4 grass Hyperthelia dissoluta. These species were grown at different levels of water, N and phosphorus (P), both in monoculture and in competition with each other. In the monocultures, yields were significantly increased by the combined addition of N and P in pots receiving high water supply. In pots with interspecific competition, the legume grew poorly unless P was added. Foliar d 15 N values of legume plants grown in mixtures were considerably lower than those in monocultures, suggesting that rates of symbiotic N-fixation were higher in the presence of the grass. Grass d 15 N values, however, were also lower in mixtures, while N concentrations were higher, indicating a rapid transfer of N from the legume to the grass. We conclude that the main reason for the low abundance of C. mimosoides is not low P availability as such, but a greater ability of H. dissoluta to compete for soil N and P, and a much higher N-use efficiency. If other C 4 grasses have a similar competitive advantage, it could explain why herbaceous legumes are generally sparse in African savannas. We encourage others to test these findings using species from other types of savanna vegetation. Key words:C 4 grass; Cassia mimosoides; competition; fire; Hyperthelia dissoluta; nitrogen fixation; phosphorus; Tanzania. ONE-FIFTH OF THE WORLD’S LAND SURFACE is covered by savanna vegetation (Scholes & Archer 1997). These ecosystems are strongly influenced by seasonality in water availability, fire and herbivory (Bond et al. 2005, Keeley & Rundel 2005, Sankaran et al. 2005). In addition, many tropical savannas have infertile soils, often because nutrients have been lost from the ecosystem through frequent fire (Medina 1987). Because the proportion of nitrogen (N) lost by burning is particularly high (Villecourt et al. 1980, Pivello & Coutinho 1992, Cook 1994, Cech 2008), it is commonly sup- posed that legumes and other plants with access to atmospheric N through symbiotic N 2 -fixation have a competitive advantage where fire is frequent (Medina & Bilbao 1991, Vitousek & Howarth 1991, Vitousek & Field 1999). A high abundance of herbaceous legumes has, indeed, been reported in fire-prone forest ecosystems in Australia, Europe and North America (Bell & Koch 1980, Kazanis & Arianoutsou 1994, Hendricks & Boring 1999), and Vitousek and Field (1999) have proposed a simple model showing how legumes may coexist with nonfixing species under conditions of frequent fires. In African savannas, however, which are among the most frequently burnt ecosystems in the world (Bond et al. 2005), herbaceous legumes are usually scarce, often making up o 1 percent of the biomass (Ezedinma et al. 1979, Menaut & Cesar 1979, Huntley 1982, Isichei 1995, Laclau et al. 2002). A similar scarcity of herbaceous legumes has also been reported for prairies and semiarid grasslands in the United States (Woodmansee 1978, Collins et al. 1995, Ritchie & Tilman 1995), and for savan- nas in northern part of South America (San Jose et al. 1985, Me- dina 1987). Another South American savanna region, the Cerrado, forms an exception since it can contain a high abundance of herba- ceous legumes (Sprent et al. 1996, Bustamante et al. 2004). While the woody legumes such as Acacia trees are very common in African savannas, and are also relatively well studied, it is largely unknown which factors determine the abundance of herbaceous species. We hypothesize four possible reasons for the low abundance of herbaceous legumes in African savannas. First, N 2 -fixation and growth of leguminous plants may be limited by the availability of another nutrient, perhaps phosphorus (P), as has been demon- strated for other tropical ecosystems such as rain forest (D¨ obereiner 1978, Medina 1982, Perreijn 2002, Vitousek et al. 2002). Second, growth of herbaceous legumes may be limited by lack of water (Medina 1987, McKey 1994, Hendricks & Boring 1999, Khadka & Tatsumi 2006). Third, herbaceous legumes may be outcompeted by grasses (Norman 1966; Van Auken 1994; Viera-Vargas et al. 1995; Hu & Jones 2001, 2004; del-Val & Crawley 2005), which could occur if grasses are stronger competitors for nutrients (Nor- man 1966, Van Auken 1994, Viera-Vargas et al. 1995, Hu & Jones 2001), or if they have a higher N-use efficiency (NUE) (Ludlow 1985, Del Pozo et al. 2000, Tjoelker et al. 2005). Fourth, legumes may be preferred by herbivores due to the high nutritional quality of their tissues (Ritchie & Tilman 1995, Hulme 1996). These effects, and particularly interactions among them, have been poorly studied, especially for noncrop species in tropical savannas. Know- ledge about the constraints to symbiotic N 2 -fixation by herbaceous legumes is, however, important for understanding vegetation and ecosystem level properties such as species composition, nutrient turnover and resilience. To investigate possible causes for the low abundance of herba- ceous legumes, we conducted a pot experiment using two model species: the legume Cassia mimosoides (which, though widespread, is rarely abundant), and the tall C 4 grass Hyperthelia dissoluta. Both species are widely distributed in humid African savannas (from Received 10 May 2009; revision accepted 18 November 2009. 1 Corresponding author; e-mail: patrick_georges.cech@alumni.ethz.ch BIOTROPICA 42(5): 580–589 2010 10.1111/j.1744-7429.2009.00622.x 580 r 2010 The Author(s) Journal compilation r 2010 by The Association for Tropical Biology and Conservation