Light partitioning in experimental grass communities Eva Vojtech, Michel Loreau, Shigeo Yachi, Eva M. Spehn and Andy Hector E. Vojtech (vojtech@uwinst.uzh.ch) and A. Hector, Inst. of Environmental Sciences, Univ. of Zurich, Winterthurerstrasse 190, CH8057 Zurich, Switzerland.  M. Loreau, Dept of Biology, McGill Univ., 1205 Ave Docteur Penfield, Montreal, Que´bec, H3A 1B1, Canada.  S. Yachi, Research Inst. for Humanity and Nature, Kita-ku, JP603-8047 Kyoto, Japan.  E. M. Spehn, Inst. of Botany, Univ. of Basel, Scho¨nbeinstrasse 6, CH4056 Basel, Switzerland. Through complementary use of canopy space in mixtures, aboveground niche separation has the potential to promote species coexistence and increase productivity of mixtures as compared to monocultures. We set up an experiment with five perennial grass species which differed in height and their ability to compete for light to test whether plants partition light under conditions where it is a limiting resource, and if this resource partitioning leads to increased biomass production in mixtures (using relative yield-based methods). Further, we present the first application of a new model of light competition in plant communities. We show that under conditions where biomass production was high and light a limiting resource, only a minority of mixtures outperformed monocultures and overyielding was slight. The observed overyielding could not be explained by species differences in canopy structure and height in monoculture and was also not related to changes in the canopy traits of species when grown in mixture rather than monoculture. However, where overyielding occurred, it was associated with higher biomass density and light interception. In the new model of competition for light, greater light use complementarity was related to increased total energy absorption. Future work should address whether greater canopy space-filling is a cause or consequence of overyielding. Competition for light in plant communities is thought to be asymmetric, leading to strong competition and rapid competitive exclusion (Huston and DeAngelis 1994, Schwinning and Fox 1995, Leps ˇ 1999, Vojtech et al. 2007). Does this mean that there is little or no possibility for partitioning of this aboveground resource and that competition for light will always lead to competitive exclusion? A recent theoretical model of light competition among tree species (Adams et al. 2007) predicts that competition for light can lead to competitive exclusion, founder control or coexistence. In several studies (Hirose and Werger 1995, Anten and Hirose 1999, 2003, Werger et al. 2002), it has been shown that subordinate species were able to capture similar amounts of light per unit biomass as dominant species and this equal efficiency in capturing light by species of different stature could also help explain why they were able to coexist (Anten 2005). Further, Anten and Hirose (1999) argue that species may coexist by exhibiting different aboveground growth patterns that enables them to use different positions in time and space and to absorb light efficiently. If species occupy different positions in the canopy space, they might use the space in a more complementary way with some species specializing on upper canopy layers and others on the understory, which could result in canopy structures which can capture more light than monoculture stands. Generally, niche separation reduces competition between co-occurring species via complementary use of resources (Anten and Hirose 1999, Kahmen et al. 2006). Aboveground niche separation and increased light inter- ception could therefore not only lead to species coexistence, but also increase productivity of mixtures as compared to monocultures (Naeem et al. 1994, Spehn et al. 2000, 2005). Fridley (2002, 2003) showed that plant mixtures of grassland species on highly fertile soils overyielded (were more productive than expected based on monoculture yields). Because overyielding disappeared when high fertility plots were shaded, Fridley (2002, 2003) concluded that soil fertility promotes light partitioning by emphasising differ- ences in species’ heights and growth forms. However, direct experimental evidence for local niche complementarity in plants is still scarce in the scientific literature (Silvertown 2004, Kahmen et al. 2006), especially related to light. In a recent paper, Yachi and Loreau (2007) present a simple dynamical model for a light-limited terrestrial ecosystem that can be used to assess the effect of species diversity on light competition and biomass production and to test if it is complementary light use that leads to increased biomass production in mixtures as compared to mono- cultures. Their model shows that reduction in light competi- tion intensity due to differences in canopy architecture among species increases total biomass of mixtures, but that competitive imbalance can reduce mixture biomass. Here we describe an experiment with five perennial grass species found in European fertile meadows which were selected to differ in height as indicated by a local standard Oikos 117: 13511361, 2008 doi: 10.1111/j.2008.0030-1299.16700.x, # 2008 The Authors. Journal compilation # 2008 Oikos Subject Editor: Katharine Suding. Accepted 8 April 2008 1351