The gardens of suburban homes in a low-density housing estate in Hong Kong, Fairview Park, provided opportunities for the cultivation of private trees. The hypothesis that residential tree populations are different from public ones in terms of composition, structure, spatial pattern, and tree-habitat relationship was explored. A study was initiated to evaluate the nature of this tree population, and its resulting landscape and planning implications. Data on habitat conditions, species composition, tree structure and performance were collected through a field survey of front gardens. Some 1087 houses scattered throughout 18 streets at a sampling intensity of 22% with 2345 trees were studied in detail. The data were analyzed with the help of microcomputer database and statistical programs, and a number of tree and habitat indices, correlation and chi-squared association computations were performed. Some results were compared with those of a previous local street-tree survey. Most species were introduced rather than indigenous. The biotic diversity suprisingly high, totalling 65 species, and was strongly represented by conifers, fruit trees and palms, but weakly by shade, foliage and flowering trees. Ten dominant species were accompanied by many minor ones, mainly of exotic origin, seldom planted elsewhere in urban Hong Kong. At present, most trees are small and comfortably accomodated by the plantable spaces, but a number of potentially large species may in time overtax the small garden niches. The majority were well cared for and showed little signs of arboricultural maladies. Most gardens held less than five trees. There is a tendency to adopt a variety of species within individual households. Gardens with less trees had a relatively higher species diversity, and half of the species occurred as solitary specimens (one tree per garden). Species planted as multiples (over one tree per garden) were of two contrasting groups, namely a few popular and many uncommon. Such a spatial pattern could possibly be explained by the operation of two opposing determinants, namely, conformity versus individualism, within the overriding confines of restricted growing spaces. Tree distribution by streets indicated a clear pattern of spatial congregation. Propinquity of homes (in the same street) could lead to the adoption of similar species. More species in a street resulted in higher inter-house but not intra-house variability. Analysis of tree-habitat association confirmed that the major controlling factor was garden size, and the ill effects of cramping too much biomass within the limited niches. The long-term prospects of the trees and the treescape were assessed in the light of the findings. Implications for landscape planning and management of the estate, and of similar suburban developments locally and in the region, were discussed. Copyright © 1993 Published by Elsevier B.V.