Mortality of the cotton boll weevil in drip and sprinkler irrigated cotton crops

Abstract The cotton boll weevil, Anthonomus grandis grandis Boheman (Coleoptera: Curculionidae), is a key cotton crop pest in Brazil. Adverse climatic factors, such as high temperatures and low soil moisture, dehydrate oviposited cotton squares (bud flowers) on the ground and cause high mortality of its offspring within these plant structures. The objective of this research was to evaluate the mortality of the cotton boll weevil in drip and sprinkler irrigated cotton crops. The experimental design was in randomized blocks with two treatments: drip (T1) and sprinkler (T2, control) irrigated cotton crops with sixteen replications. Each parcel had one emergence cage, installed between two cotton rows per irrigation system, with 37 cotton squares with opened oviposition punctures and yellowish bracts, to capture adult cotton boll weevils. The average number of boll weevils that emerged from the cotton squares and the causes of mortality at different development stages were determined per treatment. Third-generation life tables of the boll weevil were prepared using the natural mortality data in drip and sprinkler irrigation treatments and plus actual, apparent and indispensable mortality rates and the lethality of each mortality cause. We conclude that the application of water directly to the root zone of the plants in a targeted manner, using the drip irrigation system, can cause high mortality of the cotton boll weevil immature stages inside cotton squares fallen on the ground. This is because the cotton squares fallen on the drier and hotter soil between the rows of drip-irrigated cotton dehydrates causing the boll weevils to die. This is important because it can reduce its population density of the pest and, consequently, the number of applications of chemical insecticides for its control. Thus, contributing to increase the viability of cotton production, mainly in areas of the Brazilian semiarid region where the cotton is cultivated in organic system.

the production costs of cotton crops (Bélot et al., 2016). In Brazil, 10% of the production costs per harvest are necessary to control the boll weevil (CEPEA, 2019) because its immature stages develop within the cotton reproductive structures (squares and small bolls) reducing the effect of the insecticides Barros et al., 2019). The dehydration of the cotton squares on the soil by adverse climatic factors, such as high temperatures and low soil moisture, increases mortality of boll weevil larvae and pupae inside these plant structures Ramalho and Wanderley, 1996).
The hypothesize of this study is that the drip irrigation can keep the soil between cotton rows drier and with high temperature, which will increase the mortality of the immature stages of the cotton boll weevil by its desiccation inside cotton squares. Natural mortality studies are important to develop management strategies for insect pests including the boll weevil populations, and to understand its mortality causes Ramalho and Wanderley, 1996). The construction of ecological life tables allows assessing qualitative and quantitative mortality factors (Pereira et al., 2007). The objective of this research was to evaluate the key mortality factors of the CBW in drip and sprinkler irrigation systems, using ecological life tables.

Field sites and plant material
The study was carried out at the Embrapa Cotton Experimental Station in Barbalha, state of Ceará, Brazil (7° 18' 18" S latitude and 39° 18' 7" W longitude) and at the Embrapa Algodão Entomology Laboratory in Campina Grande, state of Paraíba, Brazil (7° 13' 50" S latitude and 35° 52' 52" W longitude) between January and December, 2019.
The physiological conditions of the cotton plants and the weather of the cultivation area where the damaged cotton squares by CBW were collected were assumed not influencing the results. The abscission of punctured cotton squares by boll weevil oviposition depends on the enzyme pectinase produced by its second and third instar larvae (King, 1973). The cotton squares were obtained from cotton plants (genotype BRS 286), grown in a 480 m 2 experimental area (12 x 40 m) at Embrapa Algodão in Campina Grande in the first half of 2019. These plants were grown spaced 0.90 x 0.10 m leaving one per hole after thinning. Seventy days after germination, 1,184 cotton squares with boll weevil oviposition punctures plus yellow and open bracts under the plant canopies of and between the rows, easily detached from the plant, were randomly collected . A total of 37 cotton squares were placed per paper

Introduction
Brazil is one of the largest cotton producers in the world, with an estimated cultivated area of 1.1 million hectares (IBGE, 2020) mainly in the Cerrado Biome with rainfed conditions (Ferreira et al., 2020). However, this culture is expanding with areas irrigated, mainly small properties in the semi-arid region of the Northeast Brazil . The irrigation is essential in this region due to water deficiency during the crop cycle caused by low and poor distributed rainfall rates (Zonta et al., 2015). The Northeast irrigated areas accounts for, approximately 22% of those in Brazil (Paulino et al., 2011).
Water delivery and the environmental conditions created can affect insect pests in production units (Silva et al., 2019). Drip irrigation, for example, keeps the soil dry and warm between the rows of cultivated plants, which can harm insects that need environments with mild temperatures and high humidity to develop and reproduce (Amini et al., 2012;Rendon and Walton, 2019). Drip-irrigated coffee plants have fewer leaf miners than non-irrigated ones (Assis et al., 2012). The low water availability and low relative humidity increased the death rates of Drosophila suzukii (Matsumura) (Diptera: Drosophilidae) and its egg laying and increased the P. vindemmiae feeding on pupae of this pest. In addition, drip irrigation can turn soil humidity conditions unfavorable to soil pathogens (Panth et al., 2020) and can be used to apply fertilizers and pesticides, including biopesticides (Ghidiu et al., 2012;Jiang et al., 2019Jiang et al., , 2020. However, the high volumes of drip irrigation in vines can increase lateseason leafhopper populations (Daane and Williams, 2003).
Sprinkler irrigation, in turn, can remove pests such as aphids, caterpillars, mites and thrips from plants (Medeiros et al., 2011;Runia et al., 2014), but it also increases the risk of leaf diseases favoring phytopathogen spore germination (Lopes et al., 2006).
The cotton boll weevil (CBW) causes losses, such as high abscission of cotton squares (bud flowers) and abnormal opening of its capsules in all Brazilian cotton-producing regions due to its eating habits and oviposition in the plant reproductive structures (squares and small bolls) (Silva et al., , 2017. Cotton squares abscission occurs especially in those formed between 40 and 60 days old, depending on the cotton variety (Ribeiro et al., 2010;Pimenta et al., 2016).

Determination of the causes of the cotton boll weevil mortality
CBW survival per development stage was determined in an experimental field of the Embrapa Cotton in Barbalha using the cotton cultivar BRS 286, from July to September, 2019 with low precipitation (~ 7 mm) and average daily temperatures of ~ 25 ºC (INMET, 2009). Because the climatic conditions registered during the execution of the experiment were adjusted to the average of the last 30 years climatological data observed in the municipality of Barbalha, we assumed that it is unnecessary to repeat the bioassay for another year.
The experimental area was 560 m 2 (10 × 56 m) and the plants spaced 0.90 × 0.10 m, leaving one per role after thinning. The experiment started with the presence of cotton squares on the plants, 65 days after their germination.
The experimental design was in a randomized block with two treatments: drip (T1) and sprinkler (T2, control) irrigated cotton crops with 16 replications. The experimental unit consisted of a cylindrical emergence cage (the boll weevil adults alive emerged from the cotton squares) with a frame made with PVC pipe (20 cm high by 20 cm in diameter) surrounded by white nylon plastic screen, 0.5 mm mesh and closed at the top by transparent plastic film. The cages were installed between two of five 1.5 m long cotton rows. Thirty-seven cotton squares with an average size of 1.45 mm were collected and placed on the soil surface per cage. The total was 1,184 cotton squares damaged by the weevils. The soil between the cotton rows was kept clear during the experiment.
The irrigation regimes were based on the water storage capacity of the soil and the water consumption by the cotton plants (Dağdelen et al., 2009). Irrigation was carried out whenever the soil moisture was around 50% in the root zone at 0.60 m depth in both irrigation systems. Water consumption was estimated as a function of the reference evapotranspiration (IETo), calculated from the daily data obtained at the INMET weather station in Barbalha, by the cultivation coefficient (Kc) using the Penman-Monteith method and determined experimentally for the cotton culture (Bezerra et al., 2012).
In the sprinkler irrigation system, the water was uniformly applied before planting, with four pluviometers spaced 3m x 3m in the area with the micro-sprinklers (Andrade et al., 2017). Sprinkler irrigation was performed with 75% efficiency, using micro-sprinklers (Araujo-Revolution, Model HI-25), spaced 6m x 6m, with two lateral rows of 22.9 m long and a service pressure of 13.36 mca with an average precipitation of 12.06 mm h -1 . Drip irrigation consisted of two subunits, each with five lateral rows 12.9 m long, spaced 0.90 m (one lateral line per cotton row) with an outlet every 0.30 m and 95% efficiency with pressure service at the entrance of the lateral row of 21.09 mca and average flow of 0.756 l h -1 .
Soil temperature and humidity, in 16 and eight sampling points, respectively, per irrigation system were determined at 9, 11, 13 and 15 hours, of each day, for 10 days, corresponding to the period between the installation of the cages and the emergence of the cotton boll weevils. The correlation between the numbers of dead CBW and irrigation systems was calculated with the averages of these parameters. The sampling points were equidistant from each other, between two cotton rows, in the same line and the cages.
The soil temperature was determined with a digital thermometer (4 in 1 Soil Survey Instrument) and the relative humidity with an analog soil hygrometer (pH meter and soil moisture, Model PH-3000), daily, until the total emergence of the CBW adults from the cotton squares. These adults were removed from the cages and the cotton squares collected, packed in labeled paper bags and taken to the Laboratory for dissection and to determine the natural mortality causes (desiccation, unviable eggs, disease, parasitism and predation).
The dissection was performed by removing the calyx and petals of cotton squares with the aid of a stylet under stereomicroscope.
Unviable eggs due to desiccation were determined by subtracting the 13.5% unviable ones due to cell proliferation and/or formation of gelatinous substance in the cotton squares (Hinds, 1906) from the percentage of unviable eggs due to all mortality agents.
The instar of dead boll weevil larvae within the cotton squares was determined by its head capsule width, measured with a digital caliper. The widths of cephalic capsules of the first, second and third instar larvae of A. grandis grandis are 0.41, 0.62 and 0.99 mm, respectively (Parrott et al., 1970).

Cotton boll weevil life table
Two ecological life tables, from the third generations, were constructed with data on the natural mortality of the boll weevil within the cotton squares in drip and sprinkler irrigation treatments . These tables allow determining and quantifying the lethality of each one for the real, apparent and indispensable mortality of the boll weevil. The first column (X) represents the weevil developmental stage: egg, larva, pupa, adult pre-emerged (PA) (dead inside the cotton squares) and emerged adult (alive from the cotton squares) (EA). The second column (Lx) is the number of live insects in stage X.
The elaboration of the life tables started with the insect eggs in 585 and 588 cotton squares with oviposition punctures by the CBW. The value of Lx, for the X development stage, was obtained by subtracting the number of dead insects from those alive in the previous stage, until the total emerged boll weevils was obtained.
Mortality C is the factor responsible for the insect death in the stage X. This factors are unviable egg, desiccation, parasitism and predation. Unviable eggs include all embryo mortality causes (infertility, disease, desiccation, the proliferation of cellular tissues of cotton squares and/or formation of gelatinous substances responsible for crushing the egg inside these plant structures) (Hinds, 1906;. Dxc is the number of insects dead per stage X due to cause C. The percentage of actual mortality (% Dxc) is the number of dead individuals per stage calculated as a function of the number of insects at the beginning of each stage with the equation: % Dxc= Dxc/L(egg) + Dxc/ L(egg) x 100, where Dxc is the number of dead insects in stage X and L (egg) that of viable eggs at the beginning of the generation. The equation calculates the percentage of actual mortality, per generation, from egg to pre-emerged boll weevil adult.
The percentage of apparent mortality (100Qxc) is the number of dead insects per stage. It is calculated by the function of the number of live insects at the beginning of each of that of the boll weevil, with the equation: 100Qxc= Dxc/Lx (Dxc/Lx) x 100, where Dxc is the number of dead insects in the stage X due to the action of C, and Lx the number of live insects in stage X.
The mortality/survival ratio (MSxc) was estimated with the equation MSxc= Dxc/Lx (Dxc/Lx). The indispensable mortality (MIxc), for the development stage X and mortality cause C, was obtained by multiplying the number of adults emerged by the percentage of mortality (MSxc).

Data analysis
The number of adult CBW emerged from the cotton squares and the mortality data, at different development stages of this insect per treatment, were subjected to the Liliefors normality test and transformed into √ x + 0.5 when necessary, to meet the pre-requirements for the analysis of variance (ANOVA). The data were then subjected to analysis of variance and the means compared by Fisher's test (LSD) at 5% probability using the System of Statistical and Genetic Analysis (SAEG) of the Federal University of Viçosa.

Results and Discussion
The soil texture in the experimental area is sandy (93.8% sand, 2.5% silt and 3.7% clay) with a field capacity of 0.1 atm, an overall density of 1.5 kg .cm -3 , particles of 2.7 kg.cm -3 and total porosity of 44.8%. The uniformity coefficient, before irrigation, was 94.3%, close to the maximum desirable (Prado, 2016), characterizing a good water distribution in the irrigation systems. A total of 527.0 mm of water was supplied during the phenological cycle of the cotton culture by the sprinkler and drip irrigation systems, twice a week according to the capacity of soil water retention and the crop evapotranspiration.

Cotton boll weevil mortality causes
The number of adult CBW emerged from the cotton squares was higher in the sprinkler (0.35 ± 0.11) and lower (0.00 ± 0.00) in the drip treatment per emergency cage (F 1.15 = 9, 00; P <0.01). This is probably due to the higher mortality of the immature boll weevils in the drip irrigation system due to the higher temperature (41 ºC) (F 1 . 15 = 69.73; P <0.01) and lower relative humidity of the soil (9.7%) (F 1.15 = 29, 97; P <0.01) compared to temperature (37º C) and relative humidity (16%) of the soil in the sprinkler irrigation system (Figure 1). This was observed for the immature stages of the boll weevil, in the cotton squares with oviposition punctures submitted to temperatures of 35º C and 40º C and relative humidity of 89.1 and 96.2% (Greenberg et al., 2005). The water and nutrient losses in insects increase at temperatures above 50º C causing depletion, decreasing motility, tissue damage and rapid death (Harrison et al., 2012). Drip irrigation keeps the cotton rows clear in semi-arid areas what can reduce boll weevil populations by dehydrating the cotton squares with oviposition punctures fallen to the ground (dry and hot) .The number of adult boll weevils in the sprinkler treatment was relatively low, but it cannot be neglected, considering the high biotic potential of this pest (Pires et al., 2017).
The number of unviable eggs (F 1.15 = 4.99; P = 0.04) and post-embryonic immature stages of the boll weevil dead due to desiccation (F 1.15 = 4.34; P= 0.05) inside cotton squares with oviposition punctures on the soil surface differed between the irrigation systems ( Figure 2A) with interaction between them and the causes of unviable eggs (F 1.45 = 7.42; P < 0.01) (Figure 3). However, CBW killed by parasitism (F 1.15 = 1.00; P= 0.33) or predation (F 1.15 = 0.13; P > 0.05) was similar in drip or spray irrigated cotton. The number of unviable eggs due to desiccation was higher in the drip irrigation system (F 1.45 = 7.33; P < 0.01), similar to that observed for cell proliferation and/or formation of gelatinous substance in dripping and sprinkling (F 1.45 = 5.46; P = 0.02) systems, respectively. The number of unviable eggs due to cell proliferation and/or gelatinous substance formation was similar in both irrigation systems. Egg unviability due to the boll weevil infertility or disease is, normally, not significant, contrary to this due to desiccation or cell proliferation and/or gelatinous substance formation (Coelho et al., Figure 1. The number of cotton boll weevils (CBW) emerged per cotton square cage on the soil surface, and average daily temperature and relative humidity of the soil in drip and sprinkler irrigated cotton. Means followed by the same letter per column between treatments do not differ by Fisher's LSD test at 5%. Means of cotton boll weevil emerged from cotton squares per cage transformed into √x + 0.5 for statistical analysis. Original means are shown. 2016). Cell proliferation in the cotton square is quite fast, with new cells formed crushing and destroying the egg deposited by the boll weevil female inside the cotton square Hilker and Meiners, 2002).
The gelatinous substance is sticky and it is a product from the physiological reaction of the cotton squares to a foreign body, which can kill the eggs or prevent the A. grandis grandis larvae emergence (Hilker and Meiners, 2002). The term egg unviability, in this research, refers to agents of egg mortality besides unviable eggs due to desiccation. This condition was pre-established, because it is difficult to finding very small size eggs inside the completely dry cotton square and to determine is specific death cause. The numbers of unviable eggs (9.94 ± 0.63) and those killed by desiccation (5.05 ± 0.63) was higher in the drip than that of unviable (7.63 ± 0.71 individuals) or dried (2.73 ± 0.53 individuals) eggs in the sprinkling water system (Figures 2 and 3). On the other hand, the number of dead boll weevils killed by desiccation in the post-embryonic immature stages was higher in the sprinkling (28.63 ± 0.66 individuals) than in dripping (26.25 ± 0.74 individuals) systems. This is probably due to the higher temperature and lower relative humidity of the soil in the dripping system, which accelerates the egg drying process and mortality at the initial development stages of the boll weevil inside the cotton squares in this treatment.
The high mortality of larvae, pupae and pre-emerged adults of CBW by desiccation is the main cause of natural mortality of this insect pest followed by parasitism and predation (Ramalho and Wanderley, 1996;Ramalho et al., 2007;Santos et al., 2013). Arthropod mortality due to desiccation generally results from water loss because these organisms have a high surface area to body volume ratio, storing proportionally low amounts of fat and having high metabolic rate (Gibbs et al., 2003;Bujan et al., 2016). The tolerance threshold for water losses varies between insect species, even between those of the same gender, geographic location and frequency and duration of exposure to drought (Baudier et al., 2015;Kaspari et al., 2015).  The lowest mortality of the boll weevil mortality by natural causes in cotton plants irrigated by drip and sprinkler may be due to the physical barrier exerted by the cages preventing and/or hindering the access of the parasitoids Bracon vulgaris Ashmead (Hymenoptera: Braconidae) and Catolaccus grandis (Burks) (Hymenoptera: Pteromalidae) to cotton squares and reducing the number of predatory ants, which avoid cotton squares with immature boll weevils dead from desiccation (Hinds, 1907). The cotton squares with irregular holes and ripped edges, indicate predation by ants, but the ant genus responsible for the weevil predation was not determined . Ants preying larvae, pupae and pre-emerged adult of the boll weevils are mainly of the genera Solenopsis group germinata sp., Pheidole sp., and Crematogaster sp. .

Cotton boll weevil life table
The life tables, of CBW three generations, with natural mortality data showed the emergence of live adults of this insect in six and zero of the 585 and 588 cotton squares with oviposition punctures in the sprinkler and drip irrigation systems, respectively (Table 1). The higher emergence of the boll weevil in this last treatment is related to light, temperature, and humidity intensity .
The real (% Dxc) and apparent (100Qxc) mortality of the CBW, within cotton squares on the soil surface, was 100 and 99% in the areas with irrigation by drip and sprinkler systems, respectively (Table 1), mainly due to larvae desiccation, followed by egg desiccation. This is probably due to the greater vulnerability of these immature boll weevil stages as reported for desiccation, an important physical agent with deleterious effect of high temperatures and low soil humidity on this insect (Ramalho and Wanderley, 1996;Greenberg et al., 2004;Showler, 2007).
The indispensable mortality (Mlxc) of the immature boll weevil stages was completely estimated only in the cotton crop irrigated by sprinkling, because all immature stages of this insect died in the drip system. The lack of knowledge of all the mortality causes per boll weevil development stage in the drip system makes it impossible to estimate the relative importance of these factors in the subsequent population density of this insect (Southwood and Henderson, 2000;Naranjo and Ellsworth, 2017).
The lethality (Kxc) of desiccation was the greatest mortality per development stage of the boll weevil (Table 1). The reduction in the lethality of parasitism and predation of the boll weevil larvae may be related to the difficulty of these natural enemies to access the immature boll weevil stages and/or the lower preference of predators for these immature stages killed by desiccation (Hinds, 1907;. The absence of CBW, emerging from the cotton squares on the soil surface and between the rows of cotton plants in the drip system, indicates that this irrigation system can reduce their populations by desiccation because the water is applied directly to the root zone of plants. This reduces soil humidity and increases the temperatures, mainly between the cotton rows where the majority of cotton squares punctured for oviposition by the boll weevil fall . The drip irrigation system is more advantageous than the sprinkler irrigation because it maximize yields from cotton cultivation on small rural properties in the semiarid region (Bhatti and Patel, 2015;Zulfiqar et al., 2017), and can reduce or prevent the emergence of the CBW adults from infested cotton squares fallen on the ground. This reduction in the emergence of cotton boll weevils occurs because the cotton squares dehydrate on the drier and warmer soil between the rows of cotton drip-irrigated causing the death of immature stages of this pest by desiccation. These are interesting findings because the environment control through irrigation has never been reported as an important cultural practice to cause high mortality in cotton boll weevil inside cotton squares fallen on the ground. Therefore, this is important because it can reduce its population density and, consequently, the number of applications of chemical insecticides to control this pest. Thus contributing to increase the viability of cotton production, mainly in areas of the Brazilian semiarid region where the cotton is cultivated in organic system. pre-emergent adult) inside cotton squares on the soil in relation to the total of dead boll weevils within these plant structures in drip and sprinkler irrigated cotton. Means followed by the same letter in the columns, due to mortality, do not differ by Fisher's LSD test at 5% probability. Means transformed into √x + 0.5 for statistical analysis.