CITRUS RUST MITE (ACARI, ERIOPHYIDAE) COUNTS ON FRUIT AND THE NEGATIVE BINOMIAL-DISTRIBUTION

Count data for the number of citrus rust mites per cm(2) on fruit acro ss a 4-ha (10-acre) area of orange trees followed the negative binomia l probability distribution 79% of the time based on chi-square tests. A correlation of r = 0.993 was found between observed counts and count s projected based on the distribution. A common k of 0.149 was compute d but generally appeared more suitable for mean densities of 3.0 to 55 .0 than 0.5 to 3.0 citrus rust mites per cm(2). For mean densities of from 0.5 to 55 citrus rust mites per cm(2), the parameter k of the neg ative binomial was related to the mean density ((x) over bar): k = 0.0 81 + 0.1139(log(10)(x) over bar). Estimated h-values were used to dra w expected count data profiles for several mean densities ranging from 1 to 40 citrus rust mites per cm(2). Due to the skewness of the count data, the number of mites per cm(2) expected in most individual sampl es was always considerably smaller than the average density. Based on the negative binomial, mean rust mite densities could be estimated fro m the percentage of samples with at least one mite. Results of the stu dy provide a means to predict the relative frequency histogram of dens ities associated with a mean density of citrus rust mites per cm(2) ac ross an area of trees.