Several measures that aim to conserve the genetic resources of crop plants and their wild relatives have been implemented worldwide (Wilkes, 1995). In addition to seed genebanks, these include ex situ gardens, field genebanks and in situ methods of conservation. Several programmes to conserve the genetic resources of crop plants in situ have also been initiated wherein the resources are conserved in their native habitats where these resources originated. In addition, some 12,000 to 15,000 threatened plant species are being cultivated in ex situ gardens such as in botanical gardens, genebanks and arboreta (Scarascia-Mugnozza and Perrino, 2000). While these strategies are meeting the needs of crop species, there is concern that they might be less suited for the conservation of forest tree species. This might be particularly the case for forest trees that are either endangered or threatened and for species with recalcitrant seeds. Ex situ gardens cannot accommodate large populations of tree species and consequently their genetic base is narrow. Further, they should be deliberately managed to minimize drift and selection for adaptation to the garden environment (Frankel et al., 1995). In situ conservation can complement such ex situ approaches and allow for adaptive evolution. Ideally in situ programmes require information on where to conserve coupled with long-term support and commitment, and this has limited the adoption of this method (Uma Shaanker and Ganeshaiah, 1997). Obtaining spatially explicit information on the distribution of intraspecific genetic variability and the identification of ‘hotspots’ of genetic variability is very tedious and for tropical species has only recently been attempted (Uma Shaanker and Ganeshaiah, 1997). In some species, only a few sites can be set aside for conservation, as contending pressure on land use restricts the proportion of the total genetic variability in a given species that can be set aside.Genetic diversity of populations may erode over time in in situ sites, especially if the populations are small and fragmented or disjunct. A decrease in genetic diversity parameters from adult to progeny population has been reported in several forest tree species whose populations are sparse or isolated with little or no gene flow among the populations or patches (Doligez and Joly, 1997). We suggest an additional approach, referred to as forest genebanks, for the conservation of forest genetic resources. The forest genebank combines benefits from several conventional conservation protocols and is dynamic and evolutionary. We discuss the various steps involved in the establishment of the forest genebanks with particular reference to the conservation of the genetic resources of Phyllanthus emblica, an important medicinal plant species of India.