Mutualistic asexual endophytes in a native grass are usually parasitic

Asexual systemic fungi that live symbiotically within grasses are viewed as strong mutualists on the basis of theory and empirical studies of introduced agronomic grasses. Evolutionary theory predicts that microbial symbionts that lose sexuality and rely on propagules of their hosts for transmission should evolve to benefit their hosts. Fungal endophytes of some cultivated turf and pasture grasses are well known for increasing plant performance and competitive abilities, especially under stress, and increasing resistance to herbivores, pathogens, and root-feeders by virtue of fungal alkaloids. The assumption of mutualism, however, has rarely been tested in native grasses, which often harbor high but variable frequencies of systemic asexual endophytes. We tested the effect of Neotyphodium infections for the native grass Arizona fescue in a 3-yr field experiment. We strictly controlled host genotype and manipulated soil moisture and nutrients. Infection generally decreased host growth in terms of plant volume, number of tillers, and dry mass of shoots and roots. Infected plants also showed decreased reproduction in terms of number and mass of seeds, and the seeds produced by infected plants had lower germination success than plants without their endophytes, suggesting that the negative effects of the symbiont are transferred to the next generation. Plant genotype strongly influenced host's growth and reproduction and interacted with the presence of the endophyte, but the interaction was usually in the direction of negative effects. Our results challenge the notion that systemic asexual endophytes must be plant mutualists for infections to persist in nature. We propose other hypotheses to explain the variable but usually high endophyte frequencies in natural populations of grasses.