Symbiotic seed germination is a critical stage in orchid life histories. Natural selection may act to favor plants that efficiently use mycorrhizal fungi. However, the necessary conditions for natural selection -variation, heritability, and differences in fitness -have not been demonstrated for either orchid or fungus. With the epiphytic orchid Tolumnia variegata as a model system, we ask the following questions: (1) Do seeds from different individuals in a population differ in germination and seedling development in the presence of the same fungi? (2) Do different mycorrhizal fungi (Ceratobasidium spp.) differ in ability to stimulate seed germination and growth in T. variegata? And (3) are the Ceratobasidium isolates that best induce seed germination and seedling development more closely related to each other than to isolates that are less effective? We performed symbiotic seed germination experiments in vitro. The experiments were done using mycorrhizal fungi isolated from T. variegata; relationships among the fungi were inferred from nuclear ribosomal ITS sequences. We found significant variation for both symbiotic germination and seedling growth among biparental seed crops obtained from a population of T. variegata plants. Differences among Ceratobasidium fungi in seed germination were significant. The fungi that induced highest seed germination and seedling development belonged to two of four clades of Ceratobasidium. The two experiments show that there is potential for natural selection to act on orchid-fungus relationships. Given that orchids vary in performance, and that mycorrhizal fungi are not geographically distributed homogeneously, mycorrhizae may affect population size, distribution and evolution of orchids.