Relative Net Precision Research Articles

Spatial Dispersion and Sequential Sampling Plan of the Southern Green Stink Bug (Hemiptera: Pentatomidae) on Fresh Market Tomatoes

The spatial dispersion and sampling methodology of the southern green stink bug (SGSB), Nezara viridula (L.), were studied on fresh market tomatoes, Lycopersicum esculentum Mill. A visual count was used to sample SGSB populations from horizontal plant strata. More SGSB individuals were sampled from fruit clusters than from vegetative structures and samples from fruit clusters had the lowest coefficients of variation (CV) and the highest mean SGSB counts. Values of mean SGSB, CV, relative variation, and relative net precision indicated that the penultimate fruit cluster was the most efficient sampling unit among plant strata. Time consumed in counting SGSB on each stratum ranged from 3 to 10 s and showed little variation in mean or CV values. Regression equations estimating wholeplant populations of SGSB from individual fruit clusters were developed. The between-plant dispersion of SGSB fitted a negative binomial distribution with a moderate aggregation. A sequential sampling plan using the penultimate fruit cluster as a sample unit was developed to classify SGSB populations into light, moderate, and severe damage levels.

Sampling Procedures for Tobacco Flea Beetles (Coleoptera: Chrysomelidae) in Flue-cured Tobacco

Methods for estimating population densities of larvae, pupae, and emerging adult tobacco flea beetles, Epitrix hirtipennis (Melsheimer), were examined. Soil cores (10 and 20 cm in diameter) were used to sample larvae. Relative net precision analysis indicated that soil cores of 10 cm in diameter were relatively more precise than cores of 20 cm in diameter as a unit for sampling larvae. Optimum sample number for a coefficient of variability equal to 0.3 was ca. 50 when larval density was five per core, a common density. Soil cores 10 cm in diameter were used to sample pupae. At a pupal density of about two per core, a common density, optimum sample number needed for a coefficient of variability of 0.3 was ca. 12 cores. This represents a major reduction in collecting soil cores compared with sampling larvae. However, processing soil cores can be time-consuming, ranging from 10 to 40 min per core. An emergence trap was designed and used for collecting emerging adults. At a catch of ca. 0.75 beetles per trap, ca. 30 traps were needed to reach a coefficient of variability of 0.3. A method was developed to estimate density of emerging adults from the proportion of traps containing at least one beetle.

Comparison of Sampling Methods for Insect Pests of Cantaloupe

Sampling of cantaloupe fields was conducted during three growing seasons to determine insect pests present on melons in the lower Rio Grande Valley. Comparisons were made between two sampling methods (visual observations and a combined suction/visual technique) and three sample area sizes (1, 2, and 3 m2) during two of the seasons to determine sampling precision and efficiency. Five species and one species complex of insects considered as pests were monitored. Means, relative variation (RV), and relative net precision (RNP) for sampling techniques and sample area sizes were compared. Suction sampling followed by immediate visual examination of the plant material vacuumed yielded lowest RV values and highest RNP values. The largest-size area sampled resulted in lowest RV and RNP values for both techniques. The required number of samples associated with differing degrees of precision for both methods and all sample areas was very large.

Evaluation of Sampling Methods for Dingy Cutworm Larvae (Lepidoptera: Noctuidae)1

Sack traps, pitfall traps, and soil sieving were evaluated as sampling tools for larvae of Feltia jaculifera (Guenée). The largest number of larvae was collected in pitfall traps. Similar trap catches were found in unbaited traps, and in traps baited with either sunflower seedlings or a mixture of molasses and bran. Sack trapping was the most effective method tested, as measured by relative net precision (RNP). Soil sieving required considerably more time to obtain samples and had the lowest RNP value.

Evaluation of Sampling Techniques for Three Biological Control Agents of Carduus Thistles

Suction, visual, and pan sampling were evaluated in relation to their efficiency for population estimates of Rhinocyllus conicus, Trichosirocalus horridus, and Cassida rubiginosa. The duration of suction did not influence their removal from musk thistle rosettes. Significantly fewer R. conicus adults were removed from thistle rosettes in a simulated grazed habitat than in an ungrazed habitat. No differences occurred in removal by suction sampling between habitats for either T. horridus or C. rubiginosa. Removal by suction sampling was not dependent on density. Suction sampling and visual searches provided for equal numbers of R. conicus for short ( 50 cm) thistles. More T. horridus were found by suction sampling than by visual counts for all thistle height categories. Pan sampling yielded nearly equal numbers of T. horridus and slightly fewer R. conicus and C. rubiginosa when compared with visual counts. The pan technique was less variable. Relative net precision estimates indicated that the pan technique was superior in overall efficiency to visual counts for each insect.

Lepidopterous Pests of Cowpeas: Seasonal Occurrence, Distribution Pattern, and Comparison of Ground Cloth and Sweep-Net Sampling for Larvae

Larval populations of Anticarsia gemmatalis Hubner, Heliothis zea (Boddie). and Pseudoplusia includens (Walker) were monitored for seasonal abundance in cowpeas. Counts of H. zea , and P. includens were too low to reveal population fluctuations. Most A. gemmatalis appeared during September and October, when plants were producing pods. Larvae of A. gemmatalis were usually distributed based on the negative binomial model. H. zea and P. includens larval distribution generally satisfied both the negative binomial and Poisson expectations. A comparison of sweet-net and ground cloth-shake sampling methods showed that both methods produced <10% relative variation and gave unbiased estimates of small (<2.5 cm) and large larvae; however, based on the relative net precision values, the sweep-net method was preferred.

Estimating Populations of the Hessian Fly 123

Estimates of populations of Race B Mayetiola destructor (Say) produced by sampling 3 Indiana wheat fields during 1970–72 ranged from about 0.1–1.5 million pupae (flaxseeds) per acre. The spatial distribution of the counts of the flaxseed stage was determined to be consistent with the negative binomial distribution. The relative variability (RV) of the estimates of the flaxseeds and of plant culms per acre were below 10, well within a level of acceptable variability. As few as 15 and a maximum of 30 samples per field would be required to yield an estimate of absolute population density of the insect with an error of 10% or less of the mean. Costs and relative net precision were determined based on effort expended and the level of accuracy obtained.

Green Cloverworm Populations in Iowa Soybean with Special Reference to Sampling Procedure12

Larval populations of Plathypeana scabra (F.) were sampled in Iowa soybean during the summers of 1968-70. Intensive (small plots) and extensive (broad area) sampling programs allowed comparison of fumigation-cage, shake, sweep-net, vacuum-net, and pitfall collection techniques. Samples from the intensive program near Ames, Iowa, showed an earlier peak in 1968, an outbreak year, than in 1969, a non outbreak year. The 1968 peak occurred during the critical period of soybean pod development and bean fill which created a greater potential for economic loss than in other years. Samples from the extensive program indicated higher populations across Iowa in 1969 than in 1970, and population peaks occurred late in the season and offered little potential for economic loss. Population trends were generally concordant throughout the State. Comparison of sampling procedures in the intensive program showed that the cage technique was most precise but that the sweep-net technique gave greater relative net precision (based on precision and cost). Sweep-net samples demonstrated great fidelity to cage samples (r = 0.92), but pitfall samples did not (r = 0.(9). None of the sampling techniques used in the Ames intensie program showed larval age-class bias. Shake samples would require more study for proper evaluation. Sweep-net and vacuum-net technique comparisons in the extensive program indicated that the sweep-net technique again was superior. Although sweep-net and vacuum-net samples were highly correlated (r = 0.94), the vacuum-net technique was less precise, more expensive, and showed age-class bias.