SPATIAL-DISTRIBUTION AND SAMPLING PROCEDURES FOR FRANKLINIELLA SPP (THYSANOPTERA, THRIPIDAE) IN STAKED TOMATO

Phytophagous thrips were sampled in commercial tomato, Lycopersicum esculentum L., fields in western North Carolina during a 3-yr period form 1990 to 1992. Tomato flowers were collected to sample thrips and evaluate 3 different sampling units (2,5, and 10 flowers) for their relative net precision. Flower thrips, Frankliniella tritici (Fitch), and western flower thrips, Frankliniella occidentalis (Pergande), were the dominant species collected, accounting for 96% of all specimens. Relative net precision values indicated that a 10-flower unit was more precise and cost-efficient than either the 2-or 5-flower unit. Frequency data collected from the 10-flower unit were described by the negative binomial distribution. Both F.occidentalis and F. tritici exhibited an aggregated distribution in tomato fields. Taylor power law did not differ among thrips species or total thrips. A fixed-precision-level sequential sampling plan was developed using the parameters of Taylor power law generated from total number of thrips in a 10-flower sampling unit. The resulting sampling plan was tested with sequential sampling bootstrap simulations (n=500) using 11 independent data sets for the validation. Bootstrap simulation over a range of densities demonstrated that actual D-0 values at D-0 = 0.30 averaged < 0.26 in every case.