© 2015 The Author 2015. Published by Oxford University Press on behalf of the Annals of Botany Company. All rights reserved. Background Seedling recruitment is essential to the sustainability of any plant population. Due to the minute nature of seeds and early-stage seedlings, orchid germination in situ was for a long time practically impossible to observe, creating an obstacle towards understanding seedling site requirements and fluctuations in orchid populations. The introduction of seed packet techniques for sowing and retrieval in natural sites has brought with it important insights, but many aspects of orchid seed and germination biology remain largely unexplored. Key Considerations The germination niche for orchids is extremely complex, because it is defined by requirements not only for seed lodging and germination, but also for presence of a fungal host and its substrate. A mycobiont that the seedling can parasitize is considered an essential element, and a great diversity of Basidiomycota and Ascomycota have now been identified for their role in orchid seed germination, with fungi identifiable as imperfect Rhizoctonia species predominating. Specificity patterns vary from orchid species employing a single fungal lineage to species associating individually with a limited selection of distantly related fungi. A suitable organic carbon source for the mycobiont constitutes another key requirement. Orchid germination also relies on factors that generally influence the success of plant seeds, both abiotic, such as light/shade, moisture, substrate chemistry and texture, and biotic, such as competitors and antagonists. Complexity is furthermore increased when these factors influence seeds/seedling, fungi and fungal substrate differentially. Conclusions A better understanding of germination and seedling establishment is needed for conservation of orchid populations. Due to the obligate association with a mycobiont, the germination niches in orchid species are extremely complex and varied. Microsites suitable for germination can be small and transient, and direct observation is difficult. An experimental approach using several levels of environmental manipulation/control is recommended.