Plant Ecology 172: 133–141,2004. 133 © 2004 Kluwer Academic Publishers. Printed in the Netherlands. Mycorrhizae transfer carbon from a native grass to an invasive weed: evidence from stable isotopes and physiology Eileen V. Carey1,2,*, Marilyn J. Marler1 and Ragan M. Callaway1 1Division of Biological Sciences, University of Montana, Missoula, MT 59812, USA; 2Current address: Department of Forest Resources, University of Minnesota, 115 Green Hall, 1530 Cleveland Avenue N., St. Paul, MN 55108, USA; *Author for correspondence (e-mail: [email protected]; fax: 612-625-5212) Received 3 April 2002; accepted in revised form 12 February 2003 Key words: Arbuscular mycorrhizae, Bouteloua gracilis, Carbon transfer, Centaurea maculosa, Festuca idahoe- nsis, Invasive weeds Abstract Invasive exotic weeds pose one of the earth’s most pressing environmental problems. Although many invaders completely eliminate native plant species from some communities, ecologists know little about the mechanisms by which these exotics competitively exclude other species. Mycorrhizal fungi radically alter competitive inter- actions between plants within natural communities, and a recent study has shown that arbuscular mycorrhizal (AM) fungi provide a substantial competitive advantage to spotted knapweed, Centaurea maculosa, a noxious perennial plant that has spread throughout much of the native prairie in the northwestern U.S. Here we present evidence that this advantage is potentially due to mycorrhizally mediated transfer of carbon from a native bunch- grass, Festuca idahoensis,toCentaurea. Centaurea maculosa, Festuca idahoensis (Idaho fescue, C3), and Boute- loua gracilis (blue gramma, C4) were grown in the greenhouse either alone or with Centaurea in an incomplete factorial design with and without AM fungi. Centaurea biomass was 87–168% greater in all treatments when mycorrhizae were present in the soil (P < 0.0001). However, Centaurea biomass was significantly higher in the treatment with both mycorrhizae and Festuca present together than in any other treatment combination (P < 0.0001). This high biomass was attained even though Centaurea photosynthetic rates were 14% lower when grown with Festuca and mycorrhizae together than when grown with Festuca without mycorrhizae. Neither bio- mass nor photosynthetic rates of Centaurea were affected by competition with the C4 grass Bouteloua either with or without mycorrhizae. The stable isotope signature of Centaurea leaves grown with Festuca and mycorrhizae was more similar to that of Festuca, than when Centaurea was grown alone with mycorrhizae (P = 0.06), or with Festuca but without mycorrhizae (P = 0.09). This suggests that carbon was transferred from Festuca to the in- vasive weed. We estimated that carbon transferred from Festuca by mycorrhizae contributed up to 15% of the aboveground carbon in Centaurea plants. Our results indicate that carbon parasitism via AM soil fungi may be an important mechanism by which invasive plants out compete their neighbors, but that this interaction is highly species-specific. Introduction mycorrhizal mutualisms, thereby improving their competitive ability against non-mycorrhizal species Mycorrhizal can modify or even reverse competitive or those that have weak or facultative mutualisms interactions among plant species (Grime et al. 1987; (Caldwell et al. 1985; Allen and Allen 1990; Hartnett Hetrick et al. 1989; Hartnett et al. 1993; Moora and et al. 1993). Mycorrhizae may also alter competition Zobel 1996). The effects of mycorrhizae on the bal- by transferring nutrients and photosynthetically fixed ance of competition is often attributed to direct and carbon among plants (Chiariello et al. 1982; Grime et disproportional enhancement of species that rely on al. 1987; Hetrick et al. 1990; Francis and Read 1994; 134 Walter et al. 1996), but this process remains contro- tissue (Waters and Borowicz 1995; Watkins et al. versial (Robinson and Fitter 1999). 1996; Graves et al. 1997; Fitter et al. 1998). Regard- Little is known about the role of mycorrhizae in less of the direction of transfer, AM fungi have the the intense competition that occurs among exotic, in- potential to significantly alter competitive interactions vasive weeds and native plants. Goodwin (1992) within plant communities by providing a conduit for speculated that invasive species would be at a com- the transfer of carbon and nutrients from one plant to petitive disadvantage against natives if they required another. Unidirectional transfer of carbon via mycor- specific fungal mutualists to maximize their fitness. rhizal connections would constitute parasitism and However, Marler et al. (1999a) demonstrated that ar- could promote species richness by permitting under- buscular mycorrhizal (AM) fungi strongly enhanced story components or individuals in resource-poor mi- the competitive effect of the noxious exotic, Centau- crosites to persist (Grime et al. 1987; Simard et al. rea maculosa, (spotted knapweed) on Festuca ida- 1997). Alternatively these interactions could decrease hoensis (Idaho fescue), a bunchgrass common species richness if they enhance the competitive abil- throughout the northwestern United States. They ity of dominant species. demonstrated that the direct positive effects of AM A limitation of many previous experiments is the fungi on both Centaurea and Festuca were much use of experimental pulses of radioactive or stable weaker than the indirect negative effect of Centaurea isotope labels to detect carbon transfer, a process per- on Festuca in the presence of AM fungi, and sug- formed over very short time periods (hours to a few gested that mycorrhizal transfer of carbon or re- days). Radioactive isotopes decay and the application sources would explain these results. Net transfer of of labels must be performed in enclosed chambers carbon has been shown for ectomycorrhizae using that alter ambient CO2 concentrations and leaf-to-air stable and radioactive isotopes (Simard et al. 1997), vapor gradients potentially confounding results. but net transfer of carbon among plants by AM fungi Therefore, the short application-to-harvest period of remains a topic of debate (Robinson and Fitter 1999). labels provides only a snapshot of the long-term in- Grime et al. (1987) established microcosms in which teractions between the plants and does not provide AM fungi were either present or eliminated and found enough time for large amounts of carbon to move that the growth of many competitive subordinates was through fungi or for carbon to move from fungal tis- many times higher in the presence of AM fungi. Si- sues into plants. Moreover, if labeled CO2 is leaked multaneously, the growth of Festuca ovina, the com- from chambers or if chambers are left in place for petitive dominant, was reduced. When a radioactive longer periods of time, leaked label or labeled CO2 14C label was applied to F. ovina, 14C was later de- that is fixed and then respired from the soil may be tected in the shoots of the same subordinate species refixed by unenclosed vegetation leading to ambigu- for which large increases in growth were detected, but ous results (Robinson and Fitter 1999). To avoid these only in microcosms with AM fungi. potential problems others have tracked carbon trans- Despite the demonstration of reciprocal increases fer between plants by growing species that naturally and decreases in growth of receiver and donor spe- differ in 12C:13C isotope ratios as a function of dif- cies and the corresponding movement of 14Ctothe ferent photosynthetic pathways (Press et al. 1987; shoots of the subordinate species, Bergelson and Watkins et al. 1996), hypothesizing that if mycor- Crawly (1988) and Watkins et al. (1996), Robinson rhizae transfer carbon among plants, fungal connec- and Fitter (1999) have argued that the detection of the tions among plants that differ in stable isotopic ratios isotope in the receiver plant does not necessarily in- would lead to an alteration of the isotopic ratios of dicate net transfer since there may have been balanc- the species receiving carbon. Plants using the C3 and 14 ing movement to the donor, and that the C experi- C4 photosynthetic pathways differ greatly in their ments do not indicate the direction of movement. 12C:13C ratios reflecting differences in discrimination Furthermore, Grime et al. (1987) did not manipulate against the heavier isotope of carbon by their respec- the presence and absence of the donor species. The tive carboxylating enzymes ribulose-1,5-bisphosphate debate has been fueled by reports that AM fungal carboxylase oxygenase (Rubisco) and phosphoe- connections only produce increases in either radioac- nolpyruvate carboxylase (PEPCase). As a result of 13 tive or stable ( C) carbon labels in roots not in greater discrimination by Rubisco C3 plants have a shoots, suggesting that the label may remain in the signature that is more depleted in 13C, as compared fungal tissue rather than being transferred to the plant to the atmosphere, than the signature of C4 plants; the 135 isotopic ratios of C3 and C4 species typically range inoculum, a 2-cm layer of field soil mixed with sand between −20 and −35‰ and −7 and −15‰, respec- in the ratio 1:1 by volume was added two cm beneath tively (Ehleringer and Osmond 1991). Furthermore, the surface (Hartnett et al. 1993; Klironomos 2002). within C3 plants individual species can be distin- After germination, plants were thinned to one grass guished by their isotopic signatures, which reflect dif- seedling per pot. The native grasses were given