A method, based on samples taken at random is described. It is stressed that the obtained frequency scores directly correlate with cover percentages. When used in gradient and/or mozaic vegetations the results obtained are comparable to those of homogenous vegetations.