Responses of legumes to herbivores and nutrients during succession on a nitrogen-poor soil

Ecology, Dec, 1995 by Mark E. Ritchie, David Tilman

INTRODUCTION

Because of their ability to fix atmospheric nitrogen, leguminous plants (legumes) can be important in nutrient cycling and soil development processes on nitrogen-poor soils (Groffman et al. 1986, 1987, Van Sanbeek et al. 1986, Virginia 1986, Ledgard and Steele 1992). Thus, legumes' prevalence within plant communities can be an important factor in determining plant succession and ecosystem functioning (Chapin et al. 1986, Mooney et al. 1987, Vitousek et al. 1987, Sheehy 1989, Coates et al. 1993). In grasslands on nitrogen-poor soils, legumes might be expected to dominate because they can persist at lower available soil nitrogen levels than non-legumes (Lodge 1991, Hein and Vinall 1993, Posler et al. 1993). In many natural grasslands, however, legumes are rare ([less than]5% of aboveground phytomass; Gadgil et al. 1986, Bartolome and McClaren 1992, Osman and Cocks 1992).

One or more of three hypotheses might explain this unexpected rarity of legumes. First, native herbivores may consume legumes preferentially because legumes often have a higher tissue nitrogen content than surrounding plant species on low N soils (Boller and Hei-chel 1983, Power and Zacchariassen 1993). Thus, intensity of herbivory may be much higher for legumes than for other plants, such that herbivores may reduce legumes to very low abundances (Jones and Mott 1980, Gutteridge 1985, Gadgil et al. 1986, Hein and Vinall 1993).

A second hypothesis is that legumes might be limited by nutrients other than nitrogen (Grime and Curtis 1976, Tilman 1982, Whitehead 1987, Fenner and Lee 1989). Because traits often trade off in their benefits for species facing conflicting demands (Tilman 1990), legumes might be expected to have higher requirements for nutrients other than nitrogen as a cost of maintaining their nitrogen-fixing capability, and might not compete effectively for these other nutrients (Tilman 1982). Addition of nutrients such as calcium or manganese are often necessary for optimal performance of both crop (Whitehead 1987, Fenner and Lee 1989) and wild legumes (Foote and Jackobs 1966, Tilman 1982). Therefore, a shortage of nutrients other than nitrogen may limit legume abundance.

A third hypothesis suggests that legumes may be rare in environments with high disturbance levels because they cannot quickly recolonize following disturbance. Many legumes produce large seeds, which may disperse poorly (Osman and Cocks 1992), or herbivores may heavily depredate reproductive parts, which may reduce seed supply (Gadgil et al. 1986, Ehrlen 1992). Areas with a recent history of soil disturbance, such as old-field grasslands, often have low legume abundance unless legumes are seeded.

Previous studies of successional grasslands and savannas at Cedar Creek Natural History Area (CCNHA) clearly show that soils are nitrogen poor ([less than]800 mg/kg) and that plant production is limited by nitrogen (Tilman 1984, 1987). However, legumes occur in very low abundance and diversity in these grasslands (Tilman 1988). Herbivores, particularly white-tailed deer (Odocoileus virginianus), occur at high densities and can greatly reduce the abundance of preferred plant species (Allison 1990, Inouye et al. 1994). In successional grasslands, legumes may also not have had sufficient time to recolonize after agricultural soil disturbances.

We specifically addressed the first two of these hypotheses by measuring abundance of legumes (as a group and by species) within existing long-term (13 and 5 yr) mammalian herbivore exclusion and nutrient addition experiments at CCNHA. This allowed us to test the hypotheses that (1) mammalian herbivory reduces the abundance of legumes, and (2) legume abundance is limited by the availability of one or more nutrients other than nitrogen. We also compared legume abundance and legume responses to non-N nutrient addition among three plant communities that differed in their disturbance history (savanna, undisked 60-yr-old field, and a 13-yr-old disked portion of the 60-yr-old field).

METHODS

The study was conducted at Cedar Creek Natural History Area (CCNHA), located 60 km north of Minneapolis in east central Minnesota. Detailed descriptions of the savannas and successional grasslands at CCNHA are available elsewhere (Tilman 1987, 1988, Faber-Langendoen and Tester 1993). We assessed legume abundance in three plant communities: undisked (in 1982) and disked portions of a single 60-yr post-cultivation old-field prairie grassland (Field C, Tilman 1987), and an oak savanna maintained by burning two of every three years (Field D, Tilman 1987). These communities are dominated by prairie grasses Schizachyrium scoparium, Andropogon gerardi, and/or Sorghastrum nutans, and include a diverse array of forbs and some woody species. The savanna contains scattered pin oaks (Quercus ellipsoidalis) and bur oaks (Quercus macrocarpa). The major (99% by biomass) legume species in these fields are Lathyrus venosus, Lespedeza capitata, and Amorpha canescens (hereafter referred to by genus), but these typically represent [less than]5% of aboveground phytomass (Tilman 1987, 1988).