**2. Materials and methods**

The study was carried out in the 2016 harvest at the Gredos Agricultural Farm, of the Martin and Martin Group, located 12 km from the city of Pedro Juan Caballero, in the Vice President Sánchez neighbourhood, on route 5 "General Bernardino Caballero", Department of Amambay, georeferenced at 22°36′17″S 55°48′46″W. The property is located at approximately 640 m above sea level, the topography is flat and the area is used for extensive crops under a direct sowing system in rotation with soybeans, second-harvest corn, wheat and chia.

The region is characterized by having a transitional climate between a Mediterranean type and a frankly humid climate with an average annual temperature of 22°C and an average annual rainfall of between 1000 and 1200 mm per year, with the rainiest months being December and January, the less rainy months June, July and August. The soil of the place corresponds to an alfisol (Soil Taxonomy), with a clayey loam texture and shows a prolonged agricultural use of more than twenty-seven years of sustained production, with twelve years of production of cereals and other grains under the direct sowing system.

The study area was made up of eight areas randomly distributed over a 700 m2 surface, the dimensions of the areas being 20 m wide and 35 m long with five rows of corn separated 0.45 m by 10 m long, with an average population of 108.2 plants. For comparison purposes, they were classified into plants attacked and not attacked by *D. melacanthus*.

The criteria adopted to classify the attacked plants consisted of the presence of three to four holes in the corn leaf blade that generally appear in three to four rows and are characteristic damages that form on the leaves after the puncture-puncture performed by the stylets from bed bugs. Meanwhile, for the plants not attacked, those maize plants that did not show the characteristic damage described were considered.

The number of Bt maize plants attacked and not attacked by *D. melacanthus* and the number of punctures per plant, height of the plant in the V6 phenological stage and the weight of grains of the attacked and non-attacked Bt maize plants have been quantified by *D. melacanthus*. The estimation of attacked and unattacked plants, number of punctures and height of the plant, was carried out 33 days after sowing, and the grain weight of the attacked and unattacked plants was carried out after harvest.

The calculation of the percentage of plants attacked and/or attacked was carried out through the analysis of relative frequency (hi) = fi/N, where fi corresponds to the number of times that a piece of data is repeated within the set and N, the total number of data from the set.

The unattacked corn plants were marked with a red tape, in all the areas attacked plants were found, that is, approximately 85% of the population was attacked by green-bellied bugs, even after making two applications of the insecticide Imidacloprid to control the pest, even stinkbugs continued to appear in the areas, so a third application of the product was made. In total, 866 Bt maize plants were evaluated.

In the demarcated areas, the number of punctures caused by the green-bellied bug was quantified, which was recorded by direct observation of the continuous

perforations in each leaf and noted on a spreadsheet. The height of the plant was measured with a tape measure in the phenological stage V6 of maize, at that stage the plants presented a high attack by stinkbugs.

Once the maize reached commercial maturity, the ears were harvested manually. For this, the ears of corn were pulled out and placed in bags, which received an identification. Subsequently, the ears were exposed to the sun for 12 hours to reduce their humidity, and then the shelling and cleaning were carried out manually to eliminate impurities. The production was packed in the bags and identified according to their respective areas. Finally, the corn grains were weighed on a Profield brand four-digit precision scale.

The estimation of the corn yield was made by multiplying the average weight of the corn grains by the number of existing plants in the evaluated area, later, by the simple rule of three, the value of the corn yield per hectare was estimated. The calculation of the loss of income was made by multiplying the average weight loss of grams of corn per hectare by the commercial sale value of the product.

The data were arranged into two groups, a population of plants attacked by *D. melacanthus* and another population of plants not attacked by *D. melacanthus*. To verify the normal distribution, the data referring to the number of attacked and/or attacked plants as well as the height of the plants were subjected to the Shapiro Wilk test at 5% significance, the analysis showed that the data were homogeneous and met the requirements for the distribution. Use of parametric statistics, by and analyzed, using the Student's T-test at 5% significance for comparison of two independent groups and verification of the existence of significant differences between treatments. The statistical program used was BioStat 5.3.
