Fluctuating asymmetry in larvae of Brachythemis sp. (Odonate: Libellulidae) in relation to habitat conditions in four small Rivers in Tropical area (Côte d'Ivoire, West Africa)

J. Bio. & Env. Sci. 20 23 124 | Nakouma et al. RE RERE RESEARCH SEARCHSEARCH SEARCH PAPER PAPERPAPER PAPER OPEN ACCESS OPEN ACCESSOPEN ACCESS OPEN ACCESS Fluctuating asymmetry in larvae of Brachythemis sp. (Odonate: Libellulidae) in relation to habitat conditions in four small Rivers in Tropical area (Côte d’Ivoire, West Africa) Konaté Nakouma * , Konan K . Mexmin, Koné Y . Wayrawele, Edia O . Edia Laboratoire d’Environnement et de Biologie Aquatique, Université Nangui Abrogoua , Abidjan, Côte d’Ivoire Article published on September 11, 2023 Key words: Côte d’Ivoire, Fluctuating asymmetry, Brachythemis sp. Streams, Water quality Abstract Random and subtle deviations from bilateral symmetry (fluctuating asymmetry) have long been of interest to biologists who wish to study the susceptibility of organisms to changes in environmental quality. We examined fluctuating asymmetry (FA) in larvae of the genus Brachythemis sp. (Odonata: Libellulidae) from four small, relatively disturbed rivers (Banco, Anguédédou, Dougodou and Bou) in Côte d'Ivoire as a water quality assessment tool. In situ measurements of water pH, temperature, conductivity and dissolved oxygen were made at two sampling points along each river. Water was sampled and analyzed in the laboratory for nitrate, nitrite, phosphate, manganese, and ammonium concentrations. The left and right sides of 11 segments were photographed separately using a portable digital microscope (Celestron 2.0). Thus, the perimeter, the surface area of both sides of the labial palp, the ocular diameter, and the length of the different appendages (femur, tibia, and tarsus) of the 3 legs were measured using the imageJ measurement software. This study revealed that the water quality of each river affects the developmental stability of Brachythemis sp. We concluded that the length of the first femur (LF1) and the length of the third tibia (LT3) are useful bioindicators for this taxon because the impact of fluctuating asymmetry on these traits was strongly associated with five physicochemical parameters of each River. * Corresponding Author: Konaté Nakouma  konatenakou[email protected]om Journal of Biodiversity and Environmental Sciences (JBES) ISSN: 2220-6663 (Print) 2222-3045 (Online) Vol. 23, No. 3, p. 124-138, 2023 http://www.innspub.net J. Bio. & Env. Sci. 20 23 125 | Nakouma et al. Introduction The order Odonata, commonly known as dragonflies and damselflies, form a group of about 7000 species (Kalkman et al., 2008). They have long aroused the interest of scientists, because of the importance of their roles in aquatic and terrestrial ecosystems (Corbet, 1999). At the larval stage, they can be associated with aquatic plants or the bottom substrates of watercourses (Corbet, 1999; Juen et al., 2007). Numerous studies have indicated that odonates are sensitive to physical disturbance or habitat change, but tolerate a wide range of water quality conditions (Dolný et al., 2012; Júnior et al., 2015). Thus, morphological analysis plays an important role in many such studies, as it can be used to answer various research questions (Bell & Foster, 1994). A promising morphological analysis is fluctuating asymmetry (FA), which measures deviations from perfect morphological symmetry. Indeed, FA can offer an early warning tool to determine the effects of environmental stresses on organisms before critical changes in population and community structures (Clarke, 1993). Also, the developmental stability of any organism is indicated by its ability to maintain a normal form under the effects of certain conditions (Kozlov et al., 2002). Therefore, any deviation in the shape and/or size of any morphological trait is considered a disruption of developmental mechanisms due to genetic (Clarke, 1993) or environmental (Al-Shami et al., 2011) factors. In Côte d'Ivoire, water pollution caused by anthropogenic chemical residues has become a concern in recent years. Water quality studies generally detect polluting particles in water resources (Koffi et al., 2020; Ouattara et al., 2021) although they are generally in low quantities. Little is known about the effect of lightly polluted water on aquatic organisms. There are many methods to assess the level of pollutant in freshwater ecosystems and its effects on aquatic organisms (Bahroun et al., 2022; Charvet et al., 1998; Knoben et al., 1995; Ouattara et al., 2021) such as biological, physical and chemical indices. Despite the fact that chemical indices are widely used, biological indices have extraordinary advantages over chemical indices (Marneffe et al., 1996; Wallace et al., 1996) Although several studies focus on the impact of human activities related to aquatic organisms in the various rivers of the Ivory Coast (Camara et al., 2014; Edia et al., 2013; Jean-renaud et al., 2022; Stevens et al., 2022), none has been devoted to the use of fluctuating asymmetry in these organisms as a bioindicator of environmental stress. Thus, the present study aims to study the degree of AF in Brachythemis sp. (Odonata), relate them to water quality indices in four relatively disturbed small rivers in Côte d'Ivoire with the aim of confirming their relevance as a tool for monitoring water quality. Materials and methods Sampling site The data were collected in four rivers; namely Banco, Anguédédou, Dougoudou and Bou. Banco stream is located in the Banco National Park (5°23′40″ N and 4°03′07″ W) in the city of Abidjan (Fig.1). This stream receives wastewater discharges from the municipality of Abobo. The Anguédédou (5°23'13" N and 4°8'52" W) is located on the northwestern outskirts of Abidjan city. In addition to runoff water from surrounding municipalities and agricultural land, it also receives wastewater from a brewery. The Dougoudou River is located in Lauzoua area (5°19' W and 5°12' N) in the department of Guitry (Lôh Djiboua region). This river crosses an industrial manganese mining area (Nangah et al., 2012). Bou is a tributary of Bandama River that crosses a diamondiferous area located in Tortiya (Department of Katiola, Hambol region). The geographic coordinates of this locality are 8°46'0" N and 5°40'60" W. Two stations were selected on each of these hydrosystems (Fig. 1). For the Banco stream, station BAN1, close to the forestry school, receives runoff from the municipality of Abobo and station BAN2, located downstream just outside the park. The latter receives a significant amount of runoff from Abobo J. Bio. & Env. Sci. 20 23 126 | Nakouma et al. and the forestry school. As for the Anguédédou, ANG1 is located near a road, and bordered by a cassava field, and ANG2 which does not receive any domestic discharge. At the Dougoudou River, the both stations (DOU1 and DOU2) were chosen downstream manganese mining and bridges. In addition, DOU1 was closed to a village, cocoa plantation and rice fields while DOU2 was surrounded by fallow and cocoa plantation. On Bou River, BOU1 (upstream station) was located 5 km from the town of Tortiya. The main anthropogenic activities at this site are cashew plantation, fishing and gold panning. BOU2 (downstream station) was chosen near Tortiya, where Diamond mining, fishing, washing cars, washing and watering of oxen were observed. Fig. 1. Map of the study area. Water sampling and Analysis methods Sampling was carried out monthly from July 2020 to June 2021, with the exception of October 2020 on the Banco and Anguédédou streams. Parameters such as temperature, pH, electrical conductivity (EC) and dissolved oxygen (DO) were measured in situ using a WTW3110 multiparameter. The chemical analyses carried out in the laboratory concerned the NH 4+ , PO 43- , SO 42and NO 3ions using the analysis methods recommended by the (AFNOR, 1997) standards. As for the mining sites (Bou and Dougoudou rivers), these data come from a mining project carried out in 2015 entitled ‘Mapping and Assessing the Environmental and Health Impacts of Abandoned Mines in Sub-Saharan African Countries’ (Ani et al., 2020; Séniva et al., 2022). Odonata larvae sampling and morphometric measurement Odonata larvae were collected from all eight stations with the kick net using the kick sampling technique (Merritt & Cummins, 1996). The contents of the fillet are then preserved in labelled plastic jars and fixed with alcohol at 76°. In the laboratory, Odonata larvae were sorted and preserved in 76° ethanol for taxonomic identification and subsequent measurement. All Odonata specimens were identified using the taxonomic keys of (Stals, 2007; Tachet et al., 2010). In the present study, the anisopteran Brachythemis sp. (Libellulidae) was found to be the most dominant across all study sites, so it was selected for fluctuating asymmetry investigation. J. Bio. & Env. Sci. 20 23 127 | Nakouma et al. Late instar larvae were selected in this study. A total of 11 bilateral morphological traits (Table 1, Fig. 2) were selected to test the presence of FA and to relate it to environmental variables. The pairs of legs and the labial palps were dissected in a Petri dish containing glycerine to facilitate the correct orientation of each piece to be photographed. The left and right sides of each segment were photographed separately, using a portable digital microscope (Celestron 2.0) equipped with the imaging system. For all the photos, the microscope objective remained fixed and located 2 cm above each preparation. The perimeter, the area on both sides of the labial palp, the ocular diameter and the length of the different appendages (femur, tibia and tarsi) of the 3 legs were measured using the image measurement software. For all traits, measurements were taken twice following general procedures as recommended by (Palmer, 1994) and (Palmer & Strobeck, 2003). Table 1. Different traits measured on Brachythemis sp. collected in four Rivers of Côte d’Ivoire. Character measured on one side Abbreviation First femur length LF1 First tibia length LT1 First tarsus length Lta1 Second femur length LF2 Second tibia length LT2 Second tarsus length Lta2 Third femur length LF3 Third tibia length LT3 Third tarsus length Lta3 Perimeter of the labial palp PLP Area of the labial palp ALP Fig. 2. Morphological characteristics measured on dragonfly larvae (WfHC, 2004): Tarsal length (1-2, 56, 9-10, 13-14, 19-20, 23-24); Tibial length (2-3, 6-7, 10-11, 14-15, 18-19, 22-23); Femoral length; (3-4, 7-8, 11-12, 15-16, 17-18, 21-22); Eyepiece diameter (25-26, 27-28); R: Right labial palp; L: Left labial palp. Data analysis The Kruskal-Wallis test is a non-parametric alternative to the one-way ANOVA. It was used to test the significance of differences between the medians of the abiotic variables of the stream studied. Where the analysis revealed significant differences, the MannWhitney U test was used to test the significance of pairwise differences. For Calculation of the Water Quality Index (WQI), nine important parameters (pH, Dissolved Oxygen, Water Conductivity, Temperature, sulphate ion, phosphate ion, ammonium ion and nitrate ion were selected. This index is a water quality classification technique based on the comparison of water quality parameters with international or national Ivorian standards as part of this study. In this study, the WQI index is applied to estimate the influence of natural and anthropogenic factors on the basis of several key parameters of the surface waters of Côte d'Ivoire. This index will be calculated using the weighted arithmetic index method (Yidana & Yidana, 2010; Tyagi et al., 2013; Talhaoui et al., 2020). In this approach, a numerical value called relative weight (Wi), specific to each physico-chemical parameter, will be calculated according to the following formula: Wi =   (1) Where: k: constant of proportionality and can also be calculated using the following equation: k =  ∑    (2) n: number of parameters Si: maximum value of the WHO surface water standard (WHO, 2011) of each parameter. Then, a quality assessment scale (Qi) is calculated for each parameter according to the following formula: Q i =  x 100 (3) Qi: quality assessment scale for each parameter. Ci: the concentration of each parameter in mg/l J. Bio. & Env. Sci. 20 23 128 | Nakouma et al. Finally, the overall water quality index is calculated by the following equation: WQI = ∑ ×   ∑   (4) The analysis of the FA assessment must meet precise statistical assumptions, the departure of which could weaken the interpretation of the results. Before comparing the samples by FA, it will be worth first identifying the outliers of the differences between the two sides of the body (R-L) by visual inspection of the point clouds. If one or more values were suspected to be extreme, the use of Grubb's test as recommended by (Palmer & Strobeck, 2003) will reduce subjectivity. Next, the two-way ANOVA test used to test measurement error. A Kolmogorov-Smirnov test was used to test the normality of the values (R-L) in order to show the absence of anti-symmetry. The absence of directional asymmetry was tested with a t-test comparing the mean (R-L) to zero (Palmer, 1994; Palmer & Strobeck, 1986). Biased FA level results could be found when these levels depend on the size of traits within and between samples (Palmer, 1994). Finally, the Pearson correlation was performed between FA [| R - L |] and character size [(R + L) / 2] for each sample. In this study, FA levels were calculated as the average of the difference in absolute value between the measurements of the right and left sides |R-L| for each trait (FA1=mean |R-L|) (Palmer, 1994). To test for differences in FA between samples, multivariate tests were performed using one-way ANOVA and non-parametric one-way ANOVA (Almeida et al., 2008; Chang et al., 2009; Fessehaye et al., 2007). The results were corrected using the Bonferroni adjustment (Chang et al., 2009). To verify the extent of the influence of disturbance on the asymmetry, a linear regression between the AF1 index and the chemical variables was carried out. All these statistical analyses were performed using R software (RStudio Team, 2020) with a level of significance p = 5%. Results Physico-chemical analyzes The Kruskal-Wallis test makes it possible to differentiate the physico-chemical variables of the eight sites. Thus, all the parameters vary significantly from one station to another with the exception of dissolved oxygen and phosphate (Table 2). Table 2. Values of physico-chemical parameters (median ± SD) at the eight sampling sites (median values ± standard deviation. Different letters in the same line indicate significant differences (p < 0.05, according to the Kruskal-Wallis test)). Parameters DOU1 DOU2 BOU1 BOU2 ANG1 ANG2 BAN1 BAN2 median ± SD median ± SD median ± SD median ± SD median ± SD median ± SD median ± SD median ± SD T (°C) 27.20ab ± 1.17 26.9 ab ± 2.00 30.50a ± 2.60 28.20 ab ± 2.16 26.16 ab ± 0.72 26.2 ab ± 0.93 25.75b ± 0.55 25.65b ± 0.45 CND (µS/cm) 192.10a ± 113.10 176.60a ± 88.90 150.02ac ± 34.10 107.3ac ± 113.65 25.26b ± 2.90 26.25b ± 6.64 41.00bc ± 26.74 40.00bc ± 16.24 pH 6.93a ± 0.47 6.84a ± 0.54 6.82a ± 0.30 6.68ac ± 0.34 5.47b ± 0.65 5.63bc ± 0.54 5.86bc ± 0.41 5.76b ± 0.29 DO (mg/L) 4.20a ± 2.24 4.31a ± 2.31 4.84a ± 2.30 4.26a ± 1.98 2.9a ± 1.49 2.83a ± 1.80 4.21a ± 1.75 3.40a ± 1.86 Turb (NTU) 39.35acd ± 23.09 10.55a ± 6.25 51.31ab ± 20.2 22.21abc ± 38.25 4.68d ± 5.51 8.04abcd ± 27.96 58.60c ± 27.05 42.60ac ± 23.39 TDS 95.97a ± 57.46 64.35a ± 22.88 71.00a ± 16.90 71.15a ± 54.54 7.66b ± 5.46 13.25b ± 3.03 16.00b ± 14.17 16.70b ± 8.04 Nitra (mg/L) 3.5a ± 13.02 6.85 ± 11.48 4.25 ± 7.60 5.05 ± 16.34 5.00 ± 1.96 7.00 ± 3.11 8.20 ± 2.57 12.10b ± 3.36 Nitri (mg/L) 0.03a ± 0.02 0.02 ± 0.05 0.02 ± 0.00 0.02 ± 0.05 0.01b ± 0.01 0.01b ± 0.01 0.02 ± 0.00 0.02 ± 0.01 Am (mg/L) 0.54a ± 0.24 0.42a ± 0.25 0.41a ± 2.10 1.18a ± 2.25 0.07b ± 1.61 0.07b± 0.16 0.11 ± 0.53 0.17 ± 0.56 Phos (mg/L) 0.30 ± 0.87 0.22 ± 0.29 0.21 ± 1.20 0.30 ± 0.64 0.05 ± 0.18 0.14 ± 0.13 0.09 ± 0.10 0.10 ± 1.10 J. Bio. & Env. Sci. 20 23 129 | Nakouma et al. Indeed, the temperature varies from 25.65 ± 0.45 (BAN2) to 30.50 ± 2.60 (BOU1). This temperature changes significantly between BOU1 and the both sites of Banco stream. Turbidity varies significantly between ANG1 (4.68 ± 5.51) and BAN1 (58.60 ± 27.05). Nitrate levels are low in DOU1 (3.5 ± 13.02) and high levels are recorded in BAN2 (12.10 ± 3.36). Nitrite varies significantly between 0.01 ± 0.01 (ANG1, ANG2) and 0.03 ± 0.02 (DOU1). For conductivity, TDS, ammonium, iron and manganese, the results of the statistical tests indicate a significant difference between the stations in the mining areas (BOU1, BOU2, DOU1, DOU2) and the stations located in the peri-urban areas of Abidjan (BAN1, BAN2, ANG1, ANG2). Water Quality Index (WQI) After calculating the overall quality index of the IQE using the results of physico-chemical analyzes and the standard values of the WHO standard for drinking water. The water quality class is determined for the 8 sampling stations. Thus, the good quality class identifies the station Anguédédou2 (ANG2); the poorquality class for the Dougoudou2 (DOU2), Angueledou1 (ANG1), Banco1 (BAN1) and 2 (BAN1) stations; very poor quality and undrinkable classes are recorded respectively at the Dougoudou1 (DOU1), Bou1 and Bou2 stations (Table 3). Table 3. Results of the WQI index and water quality class of four different rivers in Côte d’Ivoire. Station WQI Quality classes DOU1 87.50 Very poor quality DOU2 70.67 poor quality BOU1 184.66 Non - drinking water BOU2 279.93 Non - drinking water BAN1 61.09 poor quality BAN2 62.67 poor quality ANG1 57.11 poor quality ANG2 27.07 Good quality Asymmetry and measurement error models A total of 374 specimens of Odonate larvae are examined for the presence of fluctuating asymmetry. Of the data set, 8.51% are outliers. After removing outliers, the effect of measurement errors on asymmetry was assessed for each trait and for each site using a two-way ANOVA. The sides*individuals interaction revealed that there was no significant difference (p > 0.05) between the sides for each trait, indicating an absence of directional asymmetry (DA) (Table 4). Table 4. The results of the two-way (error) ANOVA performed for each selected trait on larvae of the genus Brachythemis sp. collected in four rivers of Côte d’Ivoire. Trait BOU1 BOU2 DOU1 DOU2 F p F p F p F p LF1 0.002 0.967 0.023 0.880 0.003 0.959 0.001 0.973 LT1 0.011 0.916 0.001 0.970 0.007 0.934 0.005 0.945 Lta1 0.003 0.955 0.012 0.912 0.095 0.757 0.036 0.848 LF2 0.000 0.996 0.003 0.957 0.092 0.762 0.050 0.823 LT2 0.002 0.968 0.008 0.931 0.001 0.973 0.016 0.899 Lta2 0.072 0.787 0.076 0.783 0.018 0.893 0.055 0.814 LF3 0.002 0.961 0.000 0.993 0.007 0.933 0.030 0.862 LT3 0.004 0.949 0.011 0.918 0.060 0.806 0.000 0.994 Lta3 0.000 0.995 0.005 0.943 0.006 0.936 0.012 0.914 DO 0.003 0.952 0.001 0.977 0.204 0.652 0.000 0.991 PLP 0.011 0.918 0.002 0.966 0.019 0.891 0.000 0.992 ALP 0.000 0.988 0.001 0.978 0.001 0.971 0.001 0.977 ANG1 ANG2 BAN1 BAN2 F p F p F p F p LF1 0.002 0.968 0.078 0.780 0.000 0.991 0.001 0.973 LT1 0.008 0.927 0.009 0.924 0.001 0.973 0.001 0.980 Lta1 0.000 0.987 0.010 0.919 0.958 0.958 0.051 0.821 LF2 0.012 0.914 0.004 0.945 0.000 0.978 0.003 0.955 LT2 0.004 0.95 0.004 0.948 0.040 0.841 0.002 0.959 Lta2 0.022 0.881 0.008 0.927 0.002 0.964 0.000 0.993 LF3 0.001 0.972 0.009 0.921 0.002 0.964 0.001 0.976 LT3 0.000 0.986 0.003 0.951 0.037 0.848 0.000 0.989 Lta3 0.002 0.965 0.225 0.635 0.039 0.842 0.001 0.979 DO 0.000 0.996 0.004 0.945 0.005 0.940 0.000 0.981 PLP 0.040 0.841 0.031 0.861 0.000 0.985 0.000 0.986 ALP 0.010 0.919 0.052 0.820 0.000 0.998 0.000 0.990 J. Bio. & Env. Sci. 20 23 130 | Nakouma et al. For each trait and for each station, all the measurements show a normal distribution and a zero mean, with the exception of LT1, Lta2 and DO (BOU1); Lta3 and DO (DOU1); LT1, LF3 and DO (DOU2); LT1, Lta1, Lta2, LF3, and DO (ANG1); Lta1; Lta2, LF3, Lta3 and DO (ANG2); LT1, Lta3 and DO (BAN1) and finally Lta1, Lta2, Lta3 and DO (BAN2). The assumption of normality being satisfactory for these characters, so that there is no evidence of antisymmetry and directional asymmetry (Table 5). Therefore, we can assume that these traits exhibit fluctuating asymmetry. Of these characters, six (LF1, LF2, LT2, LT3 PLP and ALP) exhibit FA at all stations. These characters will be used for the determination of FA in this study. Table 5. Distribution of normality using Kolmogorov-Smirnov d-test and zero-mean tests of R-L values using one-sample t-test for all variables (p = p value ; K–S: Kolmogorov-Smirnov test). Traits BOU1 BOU2 DOU1 DOU2 Test for mean = 0 Test for normality Test for mean = 0 Test for normality Test for mean = 0 Test for normality Test for mean = 0 Test for normality t test p d (K–S) p t test p d (K–S) p t test p d (K–S) p t test p d (K–S) p LF1 -0.680 0.501 0.175 0.280 1.544 0.133 0.178 0.292 -0.802 0.426 0.143 0.263 -0.235 0.815 0.161 0.353 LT1 2.040 0.049 0.202 0.143 -0.604 0.550 0.209 0.143 -0.188 0.851 0.145 0.253 1.073 0.291 0.413 3.5e-5 Lta1 0.462 0.647 0.181 0.258 0.864 0.394 0.175 0.331 -0.100 0.920 0.183 0.074 -1.200 0.238 0.158 0.378 LF2 -1.508 0.141 0.189 0.199 0.515 0.610 0.255 0.051 -0.905 0.369 0.167 0.126 -0.383 0.704 0.166 0.337 LT2 -1.071 0.292 0.171 0.300 -1.948 0.061 0.158 0.460 -0.343 0.733 0.194 0.057 0.250 0.803 0.138 0.570 Lta2 0.000 1.000 0.273 0.016 0.509 0.613 0.203 0.165 -0.589 0.558 0.173 0.099 -1.646 0.109 0.239 0.051 LF3 0.304 0.763 0.166 0.353 -0.787 0.437 0.198 0.190 1.069 0.290 0.172 0.106 -2.200 0.034 0.187 0.182 LT3 1.161 0.255 0.183 0.261 -1.201 0.239 0.179 0.286 0.475 0.636 0.150 0.209 0.436 0.665 0.173 0.259 Lta3 -0.229 0.820 0.214 0.105 -0.586 0.562 0.218 0.127 -0.368 0.714 0.254 0.003 0.155 0.877 0.148 0.477 DO 1.853 0.073 0.267 0.017 0.319 0.751 0.213 0.142 -0.274 0.784 0.250 0.004 -1.559 0.128 0.260 0.019 PLP 0.656 0.516 0.117 0.787 0.462 0.647 0.132 0.684 -0.985 0.329 0.117 0.522 0.468 0.642 0.124 0.682 ALP -1.861 0.072 0.113 0.836 0.230 0.819 0.109 0.867 0.035 0.971 0.092 0.806 0.347 0.730 0.074 0.991 Traits ANG1 ANG2 BAN1 BAN2 Test for mean = 0 Test for normality Test for mean = 0 Test for normality Test for mean = 0 Test for normality Test for mean = 0 Test for normality t test p d (K–S) p t test p d (K–S) p t test p d (K–S) p t test p d (K–S) p LF1 -0.188 0.851 0.136 0.441 -0.102 0.918 0.116 0.305 0.558 0.578 0.122 0.234 0.796 0.430 0.130 0.538 LT1 0.297 0.767 0.229 0.036 1.498 0.138 0.150 0.080 0.289 0.772 0.194 0.008 -0.466 0.643 0.214 0.060 Lta1 0.725 0.472 0.224 0.035 1.435 0.155 0.207 0.004 1.551 0.125 0.156 0.058 -2.255 0.030 0.186 0.143 LF2 -1.5e-9 1.000 0.169 0.214 -1.652 1.103 0.144 0.112 -0.144 0.885 0.144 0.111 -1.732 0.091 0.168 0.257 LT2 0.000 1.000 0.173 0.191 0.499 0.618 0.157 0.066 -0.437 0.663 0.145 0.094 -1.480 0.147 0.176 0.197 Lta2 -1.865 0.069 0.263 0.008 0.814 0.417 0.269 5.7e-5 -0.092 0.926 0.159 0.060 2.537 0.015 0.152 0.342 LF3 2.614 0.012 0.229 0.032 -0.608 0.544 0.134 0.153 0.279 0.781 0.133 0.157 1.957 0.057 0.161 0.261 LT3 1.138 0.262 0.162 0.257 -1.354 0.179 0.168 0.303 -0.959 0.340 0.120 0.267 1.362 0.181 0.141 0.452 Lta3 1.462 0.152 0.198 0.099 -0.242 0.809 0.193 0.009 -2.720 0.008 0.218 0.002 3.450 0.0014 0.192 0.139 DO 0.190 0.850 0.229 0.029 -2.837 0.005 0.248 0.0002 -4.330 4.8e-05 0.251 0.0002 -2.434 0.019 0.275 0.005 PLP -0.396 0.694 0.134 0.528 -0.510 0.611 0.075 0.813 -1.951 0.055 0.050 0.995 0.600 0.552 0.114 0.699 ALP 0.587 0.560 0.114 0.666 1.172 0.245 0.051 0.991 -1.887 0.063 0.083 0.691 1.057 0.297 0.152 0.353 J. Bio. & Env. Sci. 20 23 131 | Nakouma et al. AF Size dependency The results of the Pearson correlations of | R−L| (the absolute difference between the right and left measurements of a specific character) versus (R+L)/2 (an indicator of character size) are shown in Table 6. These results indicated that the asymmetry in APL at BOU1 (r = 0.373, p < 0.05), BOU2 (r=0.476, p<0.001), DOU1 (r=0.306, p<0.05), ANG1 (r=0.372, p<0.05), ANG2 (r=0.439, p<0.001), BAN1 (r=0.384, p<0.001), BAN2 (r=0.637, p<0.001), in PPL at BOU2 (r = 0.591, p < 0.001), ANG2 (r=0.241, p<0.05), in LT3 at the BAN1 level (r = 0.241, p < 0.05), BAN2 (r = 0.355, p < 0.05) and in LT2 at ANG1 (r = 0.362, p < 0.05) indicated significantly positive correlations of asymmetry with feature size. The fluctuating asymmetry index The results of the various fluctuating asymmetry indices, namely AF1 = mean |R-L| are given in Table 7. These results show that the indices of the different characters studied vary from one station to another. For the length of femur1 (LF1), the AF1 index varies significantly (p˂0.05) from 0.051 ± 0.041 (BOU1) to 0.105 ± 0.079 (BAN1). At the level of the length of the femur2 (LF2), this index varies significantly (p<0.05) from 0.053 ± 0.043 (BOU1) to 0.097 ± 0.061 (BAN1). In addition, for the perimeter of the labial palp (PPL) and the surface of the labial palp (APL), they vary very significantly (p˂0.001) respectively from 0.070 ± 0.051 (BAN1) to 0.267 ± 0.220 (ANG2) and 0.048 ± 0.044 (BAN2) to 0.103 ± 0.090 (BAN1). However, for tibia length2 (LT2) and tibia length3 (LT3), this index does not indicate a significant variation (p˃0.05) between the different stations. Table 6. Descriptive statistics of selected traits of Brachythemis sp. larvae collected in different rivers of Côte d’Ivoire for fluctuating asymmetry. ( R+L )/2 Pearson’s r value R - L Station Trait N Mean ± SD Mean ± SD Skew ness Kurtosis T p T p BOU1 LF1 32 3.251 ± 1.500 0.165 - 0.007 ± 0.065 - 0.098 0.803 2.200 0.207 LF2 32 4.203 ± 2.061 0.327 - 0.018 ± 0.066 0.078 0.832 2.251 0.269 LT2 32 3.639 ± 1.680 0.048 - 0.016 ±0.083 - 0.213 0.582 2.070 0.158 LT3 30 5.321 ± 2.474 0.100 0.017 ± 0.079 0.330 0.401 2.171 0.230 PLP 31 7.119 ± 3.054 0.121 0.022 ± 0.183 0.529 0.185 3.198 0.821 ALP 30 3.058 ± 2.553 0.373* 0.029 ± 0.082 0.705 0.070 2.533 0.542 BOU2 LF1 30 3.252 ± 1.501 0.191 0.030 ± 0.104 0.105 0.773 2.352 0.362 LF2 30 4.204± 2.061 - 0.147 - 0.002 ± 0.075 - 0.436 0.252 1.965 0.112 LT2 30 3.639 ± 1.688 0.005 - 0.032 ± 0.089 0.037 0.924 2.893 0.906 LT3 30 5.321 ± 2.474 0.174 - 0.019 ± 0.087 0.310 0.424 2.875 0.882 PLP 30 7.119 ± 3.054 0.591*** 0.009 ± 0.137 - 0.241 0.521 2.714 0.712 ALP 31 3.058 ± 2.553 0.476** 0.001 ± 0.073 0.625 0.101 3.452 0.542 DOU1 LF1 49 3.467 ± 0.721 0.257 - 0.012 ± 0.101 0.088 0.778 2.102 0.092 LF2 49 4.198 ± 0.805 0.170 - 0.008 ± 0.104 - 0.259 0.408 2.450 0.353 LT2 50 3.796 ± 0.669 0.196 0.000 ± 0.091 0.247 0.426 2.681 0.626 LT3 50 5.222 ± 0.985 - 0.076 0.006 ± 0.094 0.090 0.771 2.602 0.503 PLP 48 5.630 ± 1.376 0.195 - 0.025 ± 0.177 0.199 0.524 3.031 0.963 ALP 48 1.909 ± 1.150 0.306* 0.000 ± 0.069 0.612 0.063 3.530 0.388 DOU2 LF1 33 2.769 ± 0.622 - 0.013 - 0.004 ± 0.094 0.105 0.762 2.750 0.745 LF2 32 3.319 ± 0.810 0.244 - 0.008 ± 0.117 - 0.127 0.731 2.104 0.152 LT2 32 2.960 ± 0.602 0.232 0.006 ± 0.134 - 0.394 0.29 2.824 0.817 LT3 34 4.190 ± 0.989 0.113 - 0.019 ± 0.102 0.088 0.794 2.131 0.158 PLP 33 6.360 ± 1.935 0.087 0.012 ± 0.143 0.109 0.752 2.000 0.118 ALP 34 2.478 ± 1.809 - 0.102 0.004 ± 0.070 0.119 0.744 2.715 0.700 ANG1 LF1 40 3.070 ± 0.664 0.198 0.033 ± 0.101 - 0.482 0.170 2.476 0.426 LF2 39 3.718 ± 0.853 - 0.068 0.000 ± 0.102 1.9e - 9 1.000 1.957 0.073 LT2 39 3.35 ± 0.765 0.362* 0.000 ± 0.100 0.082 0.806 2.215 0.199 LT3 39 4.51 ± 1.113 0.104 0.019 ± 0.107 - 0.128 0.689 2.455 0.410 PLP 36 5.066 ± 1.120 0.001 - 0.011 ± 0.174 0.240 0.516 2.512 0.483 ALP 37 1.572 ± 0.684 0.372* 0.010 ± 0.068 0.195 0.591 2.225 0.224 ANG2 LF1 69 3.667 ± 0.719 0.212 - 0.002 ± 0.126 - 0.073 0.777 2.273 0.143 LF2 69 4.349 ± 0.924 0.189 - 0.015 ± 0.092 - 0.294 0.308 2.276 0.124 LT2 71 3.898 ± 0.840 0.158 - 0.026 ± 0.115 0.127 0.636 2.627 0.461 J. Bio. & Env. Sci. 20 23 132 | Nakouma et al. LT3 68 5.519 ± 1.350 0.122 - 0.049 ± 0.101 0.119 0.656 2.048 0.063 PLP 70 6.073 ± 1.174 0.241 - 0.022 ± 0.363 0.103 0.716 3.018 0.979 ALP 70 2.137 ± 0.822 0.439*** 0.018 ± 0.127 - 0.114 0.679 3.179 0.737 BAN1 LF1 72 3.859 ± 0.493 0.040 0.008 ± 0.120 - 0.230 0.385 2.534 0.371 LF2 69 4.783 ± 0.709 0.120 - 0.002 ± 0.105 - 0.243 0.393 3.090 0.865 LT2 72 4.355 ± 0.640 0.053 - 0.006 ± 0.119 - 0.100 0.697 2.949 0.927 LT3 69 6.520 ± 1.099 0.241 - 0.015 ± 0.127 0.387 0.171 3.215 0.685 PLP 67 6.888 ± 0.992 - 0.185 - 0.020 ± 0.085 0.234 0.405 2.545 0.382 ALP 72 2.913 ± 0.740 0.384*** - 0.030 ± 0.134 - 0.004 0.999 3.919 0.056 BAN2 LF1 38 4.770 ± 1.209 0.104 0.017 ± 0.129 0.065 0.842 2.944 0.942 LF2 37 5.523 ± 1.443 - 0.079 - 0.039 ± 0.109 0.060 0.843 2.817 0.814 LT2 37 5.228 ± 1.433 - 0.008 - 0.019 ± 0.077 - 0.063 0.859 2.298 0.268 LT3 37 6.839 ± 1.781 0.355* 0.024 ± 0.107 0.195 0.579 2.709 0.660 PLP 38 6.788 ± 1.894 0.171 0.015 ± 0.105 - 0.210 0.563 2.635 0.625 ALP 37 2.649 ± 1.370 0.637*** 0.011 ± 0.065 0.338 0.358 2.817 0.812 N: sample size; (D+G)/2: average character length (mm); D-G: difference between the right and left sides of the character considered (mm); SD: standard deviation; * p < 0.05; ** p < 0.01; *** p < 0.001 Relationship between AF1 indices and chemical parameters The results of the linear regression test between the AF1 index and the environmental parameters and the water quality index (WQI) are given in Table 8. These results reveal that out of the six (6) trait studied, two (2) are influenced by the physico-chemical parameters. These are the length of the first femur (LF1) and the length of the third tibia (LT3). Indeed, the length of the first femur (LF1) is influenced by temperature (84%), pH (58%), dissolved oxygen (62%), phosphorus (56%) and iron at 57%. For the length of the third tibia (LT3), in addition to the water quality index (76%), six (6) physico-chemical parameters have an influence on this character, namely, temperature (74%), pH (68%), dissolved oxygen (53%), ammonium (67%), phosphorus (67%) and iron (87%). Table 7. Fluctuating asymmetry indices for each trait (median ±SD) of Brachythemis sp. collected in four different rivers in Côte d’Ivoire. Station LF1 LF2 LT2 LT3 PPL APL median ± SD median ± SD median ± SD median ± SD median ± SD median ± SD BOU1 0.051 b ± 0.041 0.053 b ± 0.043 0.067 ± 0.050 0.063 ± 0.049 0.135 a ± 0.124 0.064 a ± 0.057 BOU2 0.089 ab ± 0.061 0.061 ab ± 0.042 0.069 ± 0.063 0.065 ± 0.058 0.105 a ± 0.086 0.053 a ± 0.048 DOU1 0.084 ab ± 0.057 0.097 a ± 0.061 0.070 ± 0.056 0.074 ± 0.056 0.141 a ± 0.107 0.052 a ± 0.045 DOU2 0.073 ab ± 0.057 0.096 a ± 0.057 0.100 ± 0.086 0.085 ± 0.055 0.120 a ± 0.075 0.053 a ± 0.043 ANG1 0.101 a ± 0.060 0.084 ab ± 0.055 0.081 ± 0.056 0.081 ± 0.071 0.137 a ± 0.104 0.055 a ± 0.041 ANG2 0.099 a ± 0.070 0.075 ab ± 0.056 0.057 ± 0.049 0.092 ± 0.064 0.267 b ± 0.220 0.101 b ± 0.079 BAN1 0.105 a ± 0.079 0.079 ab ± 0.069 0.088 ± 0.079 0.097 ± 0.082 0.070 c ± 0.051 0.103 b ± 0.090 BAN2 0.097 ab ± 0.085 0.087 ab ± 0.075 0.063 ± 0.047 0.082 ± 0.071 0.116 a ± 0.094 0.048 a ± 0.044 p 0.018 0.022 0.297 0.486 < 0.001 < 0.001 median values ± standard deviation. Different letters in the same column indicate significant differences (P <0.05, according to the Kruskal-Wallis test). SD : Standard Deviation Table 8. Results of the linear regression between the AF1 index and the water quality index (WQI) as well as the physico-chemical parameters. Parameters LF1 LF2 LT2 LT3 PPL APL p R 2 p R 2 p R 2 p R 2 p R 2 p R 2 IQE 0.273 0.194 0.059 0.473 0.673 0.031 0.004 0.764 0.414 0.113 0.397 0.121 T 0.001 0.847 0.089 0.405 0.638 0.039 0.005 0.746 0.926 0.001 0.540 0.065 CND 0.070 0.446 0.775 0.014 0.301 0.175 0.074 0.436 0.587 0.051 0.199 0.257 pH 0.027 0.581 0.792 0.0125 0.792 0.012 0.012 0.680 0.578 0.054 0.304 0.173 DO 0.019 0.626 0.820 0.009 0.376 0.132 0.039 0.532 0.087 0.409 0.359 0.140 Turb 0.600 0.046 0.515 0.073 0.710 0.024 0.455 0.096 0.334 0.155 0.487 0.083 Nitra 0.783 0.013 0.972 0.009 0.733 0.020 0.183 0.273 0.273 0.194 0.300 0.175 Nitri 0.336 0.153 0.740 0.019 0.392 0.123 0.108 0.372 0.496 0.080 0.256 0.207