A Technique for Analysis of Population Density Data

Abstract

The problem of the distribution of individuals of a plant species within a community has not received the same amount of attention as other ecological problems, such as those of recognition and classification of ecological formations, and of plant succession. Ashby (1936) has discussed some of the difficulties of the problem, and has reviewed the statistical methods which have been applied. Blackman (1935) has studied the distribution of a number of plants in grassland, and has attempted to fit the Poisson cutve to the distribution of those plants which could be scored as discrete individuals. By the use Of x2 test he has demonstrated that the common plants conform well with the random (Poisson) distribution, while occasional species do not. In at least one instance he attributes the lack of randomness, or correlation in occurrence of the plants, to the fact that vegetative reproduction occurs. Clapham (1936) has analyzed data collected by Steiger (1930) on the distribution of pl-ants on high and low prairie, using relative variance as an index of overor under-dispersion. He has found that the majority of the plants studied by Steiger show marked aggregation or over-dispersion, and has pointed out that, as a result of the over-dispersion of many plants, their mean density does not have much ecological meaning. Ashby (1935) has proposed using a quadrat consisting of a lattice of 16 or 25 squares, and counting the number of plants in the quadrat and the number of empty squares. If the distribution is random, the expected number of empty squares can be calculated from the number of plants in the quadrat. When the distribution departs significantly from random, a correction factor may be applied to make the observed number of empty squares agree with the calculated. The correction factor is a measure of the departure from random.