ABSTRACT Estimates of diversity indices provide information on distribution, richness and abundance of species within a community. The aim of the current study was to estimate faunistic indices (richness, constancy, dominance and abundance) of the species collected by beat cloth and entomological net sampling methods in different soybean crops and to determine the sample size necessary to estimate faunistic indices. Fauna density from 100 soybeans crops located in nine municipalities in Rio Grande do Sul, Brazil, were evaluated. Thirty points, 20 meters apart from one another, were marked in the crops. At each point, insect pests and natural enemies were collected using the two sampling methods. Faunistic indices, including the average and interval estimates carried out by the resampling method, were calculated for each soybean crop. In the samplings using the beat cloth, richness > dominance = abundance > constancy, and using the entomological nets, richness > dominance > abundance > constancy. All faunistic indices other than dominance were greater when using the beat cloth method than those estimated by the entomological net method. The sample sizes required to estimate richness, constancy, dominance and abundance indices were 38, 78, 47 and more than 300, respectively, with an estimation error of 20 % of the average (p = 0.05). Higher values were observed when entomological net was used.

RESUMO: O objetivo deste trabalho foi estimar o tamanho de amostra para medir a densidade populacional de espécies de percevejos, variando métodos de coleta, altitude e cultivares de soja. Foram utilizadas 100 lavouras de soja, distribuídas em nove municípios da região central do Rio Grande do Sul, em três safras agrícolas (2010/2011, 2011/2012, 2012/2013). Em cada lavoura, foram demarcados 30 pontos distantes em 20 metros entre si. Em cada ponto, foram coletados percevejos (adultos + ninfas) das espécies Dichelops sp., Piezodorus guildinii e Euchistus heros, por meio dos métodos pano-de-batida largo e rede entomológica, totalizando 6.000 coletas. Para cada lavoura, método de coleta e espéciede percevejo, foram estimadas a média da densidade populacional e o tamanho de amostra por meio de reamostragem. O tamanho de amostra (número de pontos) para determinar a densidade populacional de percevejos varia com o método de coleta e a densidade populacional. Usando o pano-de-batida largo para a coleta de percevejos em soja, para um erro de estimação (amplitude do intervalo de credibilidade) igual a dois percevejos e densidade populacional na classe de 1,5 a 2,0 percevejos m-2, o tamanho de amostra é 13, 77 e 15, respectivamente, para as espécies Dichelops sp., P. guildinii e E. heros.

ABSTRACT: The objective was to estimate the sample size to evaluate the population density of bugs species, varying collection methods, altitude and soybean cultivars. There were used 100 soybean fields, distributed in nine districts of the central region of Rio Grande do Sul on three growing seasons (2010/2011, 2011/2012, 2012/2013). In each field, 30 equidistant points were marked at 20 meters. Bugs (adults and nymphs) of Dichelops sp., Piezodorus guildinii and Euchistus heros species were collected at each point using beat cloth and sweep net, totaling 6,000 collections. For each field, collection method and bug speciesit was estimated average population density, and sample size by resampling. The sample size (number of points) to estimate the population density of bugs vary with the species, altitude, soybean cultivar and the collection method. Using the beat cloth method to collect bugs in soybean, for an estimate error equal to two bugs and population density at class of 1.5 to 2.0 bugs m-2, the number of sampling points are 13, 77 and 15, respectively, for Dichelops sp., P. guildinii and E. heros.