Assessing determinants of community biomass composition in two-species plant competition studies

Abstract

A method is proposed for assessing the relative importance of species identity, neighbour species influence and environment as determinants of change in community biomass composition in two-species short-term competition experiments. The method is based on modelling the differences in relative growth rates (RGR) of species (hence called the RGRD method). Using a multiple regression approach it quantifies the effects of initial species' abundance, species identity and environment on RGRD and hence on change in community biomass composition. The RGRD approach is relatively simple to use and deals readily with statistical difficulties associated with correlated responses between species from the same stand. It can be easily adapted to analyse sequential harvest data. An example based on data from two-species mixtures of the annual species Stellaria media and Poa annua is used to illustrate the method. The main determinant of change in community biomass composition was species identity, reflected in the difference in growth rates between the species. Change in community composition was not, in general, significantly affected by the influence of neighbours or fertiliser level. The unimportance of the influence of neighbours in affecting the composition of these communities contrasts with the strong role of intra- and interspecific competition in determining the size of individuals of both species (Connolly et al. in Oecologia 82:513-526, 1990).

Convergence patterns and multiple species interactions in a designed plant mixture of five species

“…Change in species relative abundance in a community depends on the relative performances of all species (Suding et al 2003;Connolly and Wayne 2005;Ramseier et al 2005); yet few studies have explicitly investigated the link between the competitive ability of individual species and the changing proportions of species in a mixture. The partitioning facilitated by the RGRD method shows how the relative sizes of the RGRs for species and the various strengths of intraand interspecific competition on species' RGRs can be interpreted as determinants of proportional changes in Table 3 Regression coefficients for the influences of increased water and nutrient levels on the relative growth rate difference (RGRD) between the species in the column minus the species in the row [for example, RGRD Ce-Le for the cell in the second column and fourth row ( Table S3 to allow a clear interpretation Ce Carex elata, Le Lycopas europaeus, Lv Lysimachia vulgaris, Cf C. flava, Ma Mentha aquatica *P £ 0.05 **P £ 0.01 ***P £ 0.001 Table 4 Regression coefficients for relevant type B influences, that is, the influences of increased initial species biomass on the relative growth rate difference (RGRD) between two other species: the species in the column minus the species in the row [= for example RGRD Lv-Le for the cell in the second column and fourth row ( community.…”

“…Fitted proportions (x) were calculated using a generalisation of the method in Connolly and Wayne (2005). Approximate standard errors of difference between fitted proportions were calculated using a Taylor series approximation method.…”

“…Species proportions in a mixture change when their RGRs differ. Therefore, the RGR difference (RGRD) method was applied (Connolly and Wayne 2005), which models the relationships between the difference in RGRs (each species compared with one of the species chosen as reference) and the initial biomass of species and the abiotic variables, as …”

“…Environmental conditions, such as nutrient availability (Inouye and Tilman 1995), light availability (Lichter 1998) or soil pH (Christensen and Peet 1984), have been suggested as determinants for convergence patterns. Connolly and Wayne (2005) have focused attention on the potential importance of species identity in changing plant communities, apart from intra-and interspecific competitive effects. This is confirmed by our results.…”

“…It has been used by Ramseier et al (2005) as the basis of an experimental design for five annual species in a short-term experiment. In the present study, we applied the simplex design and the modelling approach developed by Connolly and Wayne (2005) to address questions of convergence/divergence in multispecies communities of perennials. Mixtures composed of five wetland species were grown for 2 consecutive years under four environmental regimes produced by combining two levels of water availability and two nutrient levels.…”