DEVELOPMENT OF AN OPTIMUM SAMPLING PLAN FOR THE POPULATION OF APHIS-GOSSYPII (HOM, APHIDIDAE) ON COTTON IN GREECE

A sampling plan was developed for the study of the population dynamics of Aphis gossypii, Glover on cotton. The distribution of the populati on of A. gossypii within the plant is not randomly distributed in rela tion to height. At the beginning (end of May-June) and at the end (Oct ober) of the season, the upper part of the plant is more heavily infes ted than the lower but this relationship changes in mid-summer (July-A ugust). Therefore, the samples for estimating mean population density should be taken proportionally in respect to the height of the plant. The optimum size of the sampling unit was found to be 10 leaves but at the beginning of the season, when the number of leaves per plant is l ow, a smaller sampling unit could be adopted or even the whole plant c an be considered as the sampling unit. The number of samples required to attain a certain precision (i.e. a 10% standard error of the mean) was found to be a strong function of density, fewer samples are requir ed at the higher densities. This is the result of the relationship bet ween the mean and the variance of the population density as it is expr essed by the value of the coefficient b in the expression of Taylor's power law. At the low densities, the adoption of the above level of pr ecision leads to a very high number of samples and, therefore, a lower level of precision should be accepted.