Interest in quantitative descriptions of forest spatial patterns has increased with the development of distance dependent stand models. The indices based on "quadrat" sampling methods are affected to a greater or lesser extend by "quadrat" size, and it is often impossible to detect non-ramdomness when clumps of individuals are very small. These problems can be overcome by using indices which are based on nearest neighbour measurements.
Distance sampling methods consists of recording the distances from a randomly selected plant to another plant or from a random point to an adjacent plant.
The distance methods are considered to be particularly suited for:
determining the degree of clustering or spatial aggregation of a population point-to-plant index, plant-to-plant index, Hopkins' coefficient of aggregation, Holgate's ratio and correlation tests;
they have been applied as well to the estimation of population densities Morisita (1954), Persson (1973), Patil, Burnham and Kovner (1979), Zeide (1985);
their use in estimation of forest stand volumes is well established;
the impact of these methods on the study of forest regeneration has been more or less limited (MacLeod and Chaudhry, 1979).
Several of the more common nonrandomness measures are discussed. They were applied to data of a series of permanent plots located in Serra da Lousã, Central Portugal. The plots were installed and measured in 19889, with a 2nd measurement in 19934.
In this work special importance is given to the natural mixed regeneration.
Key words: distance methods, mixed natural regeneration.
Correspondence: Maria Emília Pinto da Costa, Departamento de Matemática, Instituto Superior de Agronomia, Universidade Técnica de Lisboa, Tapada da Ajuda, 1399 Lisboa Codex, Portugal
Telefax: 00351-1-3635031