ACTA ZOOLOGICA BULGARICA Research Article Acta zool. bulg., Suppl. 9, 2017: 117-130 Variation in Life History and Feeding Ecology of the Invasive Eastern Mosquitoish, Gambusia holbrooki Girard, 1859 (Poeciliidae), in a Groundwater-dependent Wetland in Turkey Baran Yoğurtçuoğlu & F. Güler Ekmekçi Hydrobiology Section, Department of Biology, Faculty of Science, Hacettepe University, 06800 Ankara, Turkey; E-mail: yokbaran@gmail.com Abstract: We studied the variations in life history and feeding ecology of G. holbrooki in Acıgöl, a groundwaterdependent wetland and the only distribution area of the critically endangered Aphanius transgrediens. Fish were sampled from three sites with diferent ecological characteristics: Site 1 with stable conditions, small size and high ish density; Site 2 with stable conditions, large size and medium ish density; Site 3 with variable conditions, large size and low ish density. The scale readings showed two age groups in males and three in females. The reproduction period was from April to September. More embryos and heavier gonads were found in specimens at Site 3. The overall ratio of fertilised eggs and the absolute and relative fecundities were the lowest at Site 1. Two diferent trophic positions, plant-based omnivory and animal-based omnivory, were detected. The main food items at Site 2 and Site 3 were adults and larvae of Diptera, as well as zooplankton. At the same time, mosquitoish fed predominantly on plant sources and occasionally on the other ones at Site 1. Signiicant temporal and ontogenetic diferences in feeding were also determined. The highest diet diversity and broadest trophic niche were found at Site 2. Key words: Gambusia holbrooki, biodiversity, feeding strategy, Acıgöl, Aphanius transgrediens Introduction The way to understand the impacts of invasive species on native fauna, in a broad sense on biodiversity, is to understand their biology. The invasive eastern mosquitoish, Gambusia holbrooki Girard, 1859, has been successfully introduced into freshwater and brackish ecosystems throughout the world as a malaria control agent (Fernández-Delgado 1989, Pyke 2005). Since the 1990s, its reputation as a ‘successful mosquito control agent’ has been disreputed thanks to recent studies (Arthington 1989, Rupp 1996, Cabral & Marques 1999, Garcia-Berthou 1999), which have disagreed on its eiciency in mosquito control. Gambusia holbrooki was irst introduced in Turkey in the 1930s (Erençin 1978) for the same purpose and now has spread throughout the country (Ekmekçi et al. 2013). While its biology and ecology has been well documented in its native range (Krumholz 1948, Brown-Peterson & Peterson 1990, Meffe 1991, Haynes & Cashner 1995, Matthews & Matthews 2011) and in Europe (Fernández-Delgado 1989, Vargas & de Sostoa 1996, Garcia-Berthou 1999, Specziár 2004, Scalici et al. 2007, Gkenas et al. 2012, Zarev 2012), much fewer studies have been conducted about the introduced populations in Turkey (Öztürk & İkiz 2002, Ergüden 2013). In this study, we presented life history variations of G. holbrooki among habitats with diferent characteristics in Acıgöl, a shallow Mediterranean lake. In addition, we elaborate its feeding strategy with special emphasis on variations 117 Yoğurtçuoğlu B. & F. G. Ekmekçi in spatiotemporal, ontogenetic and sexual traits. Given the fact that Acıgöl is the only distribution area of the critically endangered Aphanius transgrediens (Ermin, 1946) (Freyhof 2014, Yoğurtçuoğlu & Ekmekçi 2014), the results of the current study may also provide insights into key elements of conservation actions. Materials and Methods Study area The study was carried out in Acıgöl, a shallow hypersaline (mostly Na2SO4 salinity) Mediterranean lake (N 37°48.53’; E 29°52.81’, Fig. 1). It is located in an endorheic basin, on a tectonic depression between two fault lines. These faults give rise to several freshwater springs lowing into the lake with various low rates. The area between discharge points of the springs and the lake margin constitutes the main wetland, where G. holbrooki and the critically endangered A. transgrediens coexist. The surface area of the lake together with the wetland is approximately 80 km2 with a maximum depth of about 2 m. However, the area of the main lake body varies greatly throughout the year. The emergent plant communities were dominated by Phragmites australis, Typha sp. and Juncus sp. with variable cover percentages depending on the season. The other most common species were the submerged Nasturtium oicinalis and loating Lemna minor beside the epilithic and ilamentous algal communities. We examined three sites, which were spatially close but ecologically distinct. The distinctions between the sampling sites were due to their variability of the physicochemical properties, size and density of G. holbrooki (Table 1). Physicochemical characteristics of the water (e.g. salinity, pH, dissolved oxygen) were stable in all of the spring outlets, except the one at the northern side (hereafter, Site 3). This spring is connected with the outlet channel of one of the sodium sulphate plant ponds. Therefore, unpredictable salinity increases occurred throughout the sampling period depending on the pond discharge. Data collection and statistical analysis From September 2013 to September 2014, 934 ish specimens were caught by a beach seine net (3 m x 1.5 m x 1.5 m, 4 mm mesh size) from the three sites. The monthly sampling was possible only at Site 1, because of the insuicient number of ish caught in some months at Site 2 and Site 3. The ish were immediately anesthetised with clove oil, then ixed 118 in 4% formaldehyde solution and transported to the laboratory, where they were measured by a digital calliper (total length, TL, to the nearest 0.01 mm) and weighed by a digital precision balance (to the nearest 0.001 g). Scales were used for age determination. The length-weight relationship was calculated, using the equation W = a x Lb, where W is the body weight, L the total length, a the intercept and b the slope. The parameter estimation was established using a linear regression analysis on log-transformed data (Froese 2006). Analysis of covariance (ANCOVA) was used for comparison of slopes between sites for each sex, with the total body weight as the dependent variable and total length as covariate. The parameters of the von Bertalanfy growth equation (L∞ – asymptotic length; k – growth coeicient; t0 – hypothetical age at which length is 0) and the growth performance index (f’ = 2log (L∞) + log (k)) were calculated using ELEFAN I method (Gayanilo et al. 1988). The sex was determined by the anal in morphology and by examining the gonad structure (usually in individuals < 20 mm TL). The gonads were dissected and weighed and the gonadosomatic index (GSI = gonadal weight/ total weight) was calculated for both sexes. Diferences in GSI throughout the reproduction period were tested by one-way ANOVA for Site 1. Embryonic development and fecundity were studied over 102 females for Site 1, 20 females for Site 2 and 18 females for Site 3. The ovarian and embryonic developmental stages were modiied from Haynes (1995) as follows: (1) small, granular, white in colour and opaque; (2) opaque, yellow or orange, yolked eggs; (3) recently fertilised ovum, embryo appears as a small white cap on the surface of the yolk; (4) embryo appears as a thin white streak, covering one-half of the yolk diameter; (5) optic vesicles begin to pigment; (6) eyes enlarged but not full sized, pigmentation process almost completed, yolk sac is partially absorbed; (7) eye pigmentation process fully completed, body begins to pigment; (8) caudal in elongated enough to cover the head, paired ins appeared, yolk sac absorption almost completed; (9) yolk sac absorption fully completed, paired ins elongated, scales present, embryo resembles small adult. The absolute (total number of embryos) and relative (number of embryos per unit of body weight) fecundity was estimated from the gravid females, which possess fertilised eggs (from stage 3 and older stages). The diference in size of the gravid females during the reproductive period was tested by one-way ANOVA, followed by Tukey test. Linear regression analyses were used to describe the relationship between fecundity and total length. Variation in Life History and Feeding Ecology of the Invasive Eastern Mosquitoish, Gambusia holbrooki... Fig. 1. Location of the sampling sites in Acıgöl, Turkey Table 1. Some characteristics of the sampling sites, as measured and observed from September 2013 to September 2014. x – mean; min – minimum; max – maximum; Sd – standard deviation; D – depth; S – size; IsFD – intraspecific Gambusia density; AVD – aquatic vegetation density; *** – high; ** – medium; * – low Salinity (ppt) x (min-max) Temperature (°C) Sd x (min-max) Physical Features Sd ~D (cm) Density ~S (m2) IsGD AVD Site 1 0.87 (0.72-0.95) 0.06 19.8 (19.2-20.7) 0.5 50-200 60 *** ** Site 2 0.71 (0.60-0.79) 0.05 19.9 (19.4-21.1) 0.4 50-100 2500 ** *** Site 3 10.17 (0.54-58.56) 17.35 18.6 (10.9-26.8) 5.4 100-200 1200 * * The diference between the slopes of the regression equations was tested by ANCOVA. For analysis of the feeding ecology, 175 stomach samples were selected from coinciding sampling months in order to compare feeding habits among sampling sites. The entire gut was removed and dissected, and the gut contents were identiied and counted under a dissecting microscope for macroscopic food items. A Sedgewick rafter chamber was used to count and identify smaller organisms. The volume of the gut contents and food items were estimated using two methods. Firstly, the individual gut contents and/or inseparable unique categories (e.g. detritus) were squashed on a plate to a uniform depth (here we used the Sedgewick rafter) and the area of the squash was measured (Hellawell & Abel 1971). Secondly, the volume of speciic food items was estimated through calculation of a volume of geometric shape closest to it (Sun & Liu 2003). This procedure was carried out by a digital image 119 Yoğurtçuoğlu B. & F. G. Ekmekçi analysis software – IMAGEJ (Schindelin et al. 2012). The identiiable parts of organisms, such as heads, were considered as individuals. The dietary importance of each food category was expressed by percent number (%N), percent volume (%V), and frequency of occurrence (%FO), as follows: %N – number of prey i/total number of prey x 100; %V – volume of prey i/total volume of prey x 100; %FO – number of the guts containing prey i/total number of the guts containing prey x 100 (Hyslop 1980). The index of relative importance (IRI = (%N + %V) x %FO) was used to incorporate these indices (Pinkas et al. 1971). To test the intraspeciic (among size groups and between sexes) and seasonal diferences in the diet composition, permutational multivariate analysis of variance (PERMANOVA) was applied, using Bray-Curtis similarity resemblance matrices of square-root transformed volume data, with 9999 permutations. Three length classes (< 30 mm, 3040 mm, and > 40 mm) were assigned to assess the ontogenetic variation in feeding habits. Canonical Analysis of Principal coordinates (CAP) was performed to display a visual ordination of the dietary pattern among sampling sites by using IRI data. This method was preferred against traditional ordination methods since it uncovers patterns that are masked in an unconstrained scaling ordination (Anderson & Willis 2003). PERMANOVA and CAP analysis were performed by a PERMANOVA+ v1.0.1. PRIMER v6 software package (Anderson et al. 2008). The diet diversity was calculated using the Shannon index (H΄): , (H΄) where p is the proportion of the number of prey item ‘i’ to the total number of prey organisms. The niche breadth, an indication of trophic generalisation level, was quantiied by the Levins index (B) (Krebs 1989 ): where Pi is the proportion of each food category ‘i’ in the diet and n is the total number of food categories in the diet of G. holbrooki. The Levins index was also standardised by the equation (B-1) / (n-1) to represent the niche breadth on a scale from 0 (narrow niche) to 1 (broad niche). The fractional trophic level (TROPH) values were estimated using TrophLab (Stergiou & Karpouzi 2002). Finally, to visualise the feeding strategy, the Costello graphical method (Costello 1990), modiied by Amundsen et al. (1996) was used. 120 Here, the prey-speciic abundance (%Pi) was plotted against the frequency of occurrence (%FO). The %Pi is the percentage of prey category i to the total number of prey items in the stomachs, which contained the prey category i. Results Life history traits Of the total ish captured, 622 were females and 371 were males (c2=62.9, P<0.001). This ratio difered between the sampling sites and the seasons with a decrease in female/male ratio in spring and summer (Table 2). Three age groups (0+, 1+ and 2+) in females and two (0+ and 1+) in males were determined. The dominant age group was 0+ in both sexes (42.4% in females and 58.0% in males). In fact, only two individuals of 1+ males were sampled from Site 3. The total length ranged between 14.0 mm and 61.4 mm in females and between 16.2 mm and 33.3 mm in males (Table 3). The analysis of covariance ANCOVA showed signiicant diferences in length-weight relationships (Table 3) between the sampling sites for each sex (females F=11.79, P=<0.001; males F=5.64, P=0.004). The females were longer and heavier than the males of the same age (Student t-test for 0+ age group t=5.74, P<0.001 and for 1+ age group t=16.91, P<0.001). The seasonal growth varied between sexes as depicted by the mean monthly lengths of individuals (Fig. 2). The null growth in females initiated in February and followed by a notable secondary stage just before the beginning of summer (March-April). The initial growth stage in males occurred less obviously in October. The growth performance index was higher in females than in males, whereas the males reached to their asymptotic length at a faster rate (Table 4). The smallest gravid female with mature embryos was with 28.6 mm TL, the smallest male with fully developed gonopodium was with 20.0 mm TL. The average GSI value started to increase from April and reached its maximum in June for both sexes at Site 3, where the monthly sampling was achieved completely (Fig. 3). The highest GSI value was recorded as 40.9% for a female captured from Site 3 in August. When assessing the GSI together with the frequency of gravid females (Fig. 4), the reproductive period of G. holbrooki in Acıgöl could be seen to commence in April and end in September. The absolute and relative fecundity values and fertilisation success that was estimated for the females caught at the three sampling sites (Table 5) revealed the highest average absolute fecundity Variation in Life History and Feeding Ecology of the Invasive Eastern Mosquitoish, Gambusia holbrooki... Table 2. Seasonal sex ratio of Gambusia holbrooki caught in Acıgöl from September 2013 to September 2014. Signiicant diferences from parity were tested by chi-square test Season Number of individuals Location Females Autumn Winter Spring Summer Ratio Females:Males c2 P-value Males Site 1 109 69 1.58:1 8.55 < 0.05 Site 2 89 44 2.02:1 14.56 < 0.05 Site 3 100 44 2.27:1 21.01 < 0.05 Site 1 50 17 2.94:1 15.28 < 0.05 Site 2 0 0 – – – Site 3 0 0 – – – Site 1 70 46 1.52:1 4.56 < 0.05 Site 2 34 35 0.97:1 0.01 0.95 Site 3 20 12 1.67:1 1.53 0.216 Site 1 116 69 1.68:1 11.44 < 0.05 Site 2 0 0 – – – Site 3 34 35 0.97:1 0.01 0.95 Table 3. Estimated length-weight relationship parameters for Gambusia holbrooki males and females in Acıgöl. F – females; M – males; C – sexes combined; n – sample size; TL – total length; W – total body weight; a and b – parameters of the equations (see Materials and Methods); Cl – conidence limits; r2 – coeicient of determination TL (cm) W (g) Regression Parameters Sites Sex n Min Max Min Max a b b Cl %95 r2 Site 1 F 322 1.85 5.14 0.06 1.80 0.0064 3.48 3.44-3.53 0.988 M 186 2.01 3.33 0.07 0.38 0.0071 3.27 3.14-3.39 0.936 C 508 1.85 5.14 0.06 1.80 0.0055 3.57 3.53-3.61 0.985 Site 2 Site 3 F 128 1.40 5.09 0.03 1.80 0.0077 3.25 3.17-3.32 0.983 M 84 1.62 3.20 0.04 0.30 0.0094 2.95 2.85-3.10 0.974 C 212 1.40 5.09 0.03 1.80 0.0074 3.26 3.20-3.33 0.981 F 134 1.84 6.14 0.06 3.07 0.0071 3.33 3.24-3.42 0.976 M 80 1.76 3.23 0.05 0.36 0.0083 3.09 2.88-3.29 0.918 C 214 1.76 6.14 0.05 3.07 0.0067 3.36 3.28-3.43 0.974 Table 4. Estimated von Bertalanfy parameters and growth performance indices of Gambusia holbrooki in Acıgöl. L∞ – asymptotic length; k – growth coeicient; t0 – hypothetical age at which length is 0 (year); f’ – growth performance index Parameters Site 1 Site 2 Site 3 Females Males Females Males Females Males L∞ (mm) 55.57 34.10 58.57 35.63 64.29 38.38 k (yr-1) 0.61 0.89 0.36 0.69 0.36 0.76 t0 (yr) -0.29 -0.31 -0.63 -0.19 -0.37 -0.19 1.28 1.02 1.09 0.94 1.17 1.05 f’ at Site 3. However, the average relative fecundity was the highest at Site 2. The fertilisation success was higher at Site 3 with the highest percentage of malformed eggs. The size of the gravid females was signiicantly diferent during the reproductive period (ANOVA, F=3.87, P=0.023; Site 1 = Site 2, Site 1 = Site 3 and Site 2 < Site 3, according to the Tukey test). The equations of the relationships between the fecundity (F) and total length (TL) of G. holbrooki at the three sites were: FSite1 = -36.41 + 1.51 TL (r2=22.9%; F=29.16, P<0.001), FSite2 = -109.4 + 3.93 TL (r2=64.3%; F=32.54, P<0.001) and FSite3 = -92.70 + 3.39 TL (r2=78.1%; F=57.19, P<0.001) 121 Yoğurtçuoğlu B. & F. G. Ekmekçi Fig. 2. Monthly growth pattern of Gambusia holbrooki at Site 1 in Acıgöl. Average total lengths with bars representing standard deviation Fig. 3. Variation in mean (± standard deviation) gonadosomatic indices of male and female Gambusia holbrooki, sampled in Acıgöl, between September 2013 and September 2014 122 Variation in Life History and Feeding Ecology of the Invasive Eastern Mosquitoish, Gambusia holbrooki... Fig. 4. Changes in the percentage of female Gambusia holbrooki bearing eggs/ embryos in each developmental stage as deined in Materials and Methods with the egg diameter change during the reproductive period. np – non-pregnant Table 5. Absolute and relative fecundity and fertilisation success of Gambusia holbrooki in Acıgöl. AF – absolute fecundity; RF – relative fecundity; Nf – non-fertilised eggs; M – malformed eggs; n – number of females examined for fecundity; n’ – number of eggs examined for fertilisation success Fecundity Fertilisation success Sites n AF (Mean±sd) Min-max RF (Mean±sd) Min-max n’ Nf (%) M (%) Site 1 102 26.3±15.2 1-76 27.1±14.5 1.4-85.5 2689 5.60 0.80 Site 2 20 43.1±32.2 10-125 54.2±21.7 32.1-127.7 862 3.55 0.78 Site 3 18 57.7±41.3 6-130 38.9±13.3 7.1-68.7 1039 1.93 2.48 The diference in the slopes of the regression equations was statistically signiicant (ANCOVA, F=29.01, P<0.001), due to the signiicance of diference between Site 1 and the other two sites (Tukey test). P<0.001). The most important food component was the adult nematocerans in spring (Fig. 5). The diet shifted towards the plant sources and detritus in summer, followed by the dominance of nematoceran larvae in autumn. Overall diet description and seasonal variation None of the 175 stomach samples examined was completely empty. The dietary spectrum was summarised under 26 food categories (Table 6). Regarding the occurrence in the diet and the percent volume, the most common food items were plant materials (especially ilamentous algae) and detritus at Site 1 and insect groups at Site 2 (mostly nematoceran Diptera) and Site 3 (mostly brachyceran Diptera). The most important contribution to the overall food components in terms of the percent number was made by invertebrate eggs at Site 1 and by zooplankton at Site 2. The demonstration of the full seasonal variation in feeding was possible only for Site 1. IRI was signiicantly diferent between the seasons at this sampling site (PERMANOVA, pseudo-F = 4.17, Feeding ecology Two diferent trophic positions were represented by G. holbrooki in Acıgöl: omnivory with a preference for plants at Site 1 and omnivory with a preference for animals at Sites 2 and 3. The main food items explaining the variation among dietary composition of the individuals are displayed in Fig. 6, after performing PERMANOVA to detect signiicant diferences in the diet between the sites (pseudo-F = 11.56, P<0.001). Despite the spring feeding pattern seen at Site 1 (Fig. 5), most of the variation was contributed by ilamentous algae, detritus and diatoms. The diet composition was dominated by Hydracinidae, Cladocera and Copepoda in small individuals, whereas shifted towards diferent taxa, e.g., insects (at Site 2 and Site 3) and ilamentous 123 Yoğurtçuoğlu B. & F. G. Ekmekçi Table 6. Overall diet composition of Gambusia holbrooki in Acıgöl: frequency of occurrence, % number and % volume of the main food categories. Estimated S-W diversity index and niche breadth with conidence intervals and trophic level ± standard error Prey Categories Site 1 Site 2 Site 3 % FO %N %V % FO %N %V % FO %N %V 54.22 27.06 36.72 78.13 32.43 88.94 94.23 84.89 89.71 Coleoptera 2.41 0.66 3.55 12.50 3.75 46.05 19.23 6.58 8.80 Hemiptera 1.21 0.17 1.14 3.13 0.27 1.60 7.69 2.19 17.41 Homoptera 1.21 0.17 0.09 0.00 0.00 0.00 1.92 0.44 0.44 Brachycera 8.43 2.31 10.17 6.25 0.80 10.27 40.39 19.30 49.01 Nematocera (adult) 31.36 9.74 13.03 71.88 13.94 16.98 53.85 33.77 6.50 Nematocera (larvae) 18.07 3.47 7.02 46.88 10.99 6.68 5.77 2.87 0.67 Insecta Trichoptera (larvae) 3.61 0.83 0.96 0.00 0.00 0.00 0.00 0.00 0.00 Ephemeroptera (larvae) 0.00 0.00 0.00 9.38 1.88 6.79 1.92 1.32 5.47 Other (unidentiied) 8.43 9.74 0.77 6.25 0.80 0.57 11.54 18.42 1.41 Zooplankton 18.07 8.09 0.03 53.13 41.56 2.23 9.62 7.02 0.10 Cladocera 14.46 6.77 0.03 37.50 25.74 0.75 5.77 3.51 0.01 Copepoda 0.00 0.00 0.00 31.25 15.55 1.42 7.70 2.63 0.08 Ostracoda 0.00 0.00 0.00 3.13 0.27 0.07 1.92 0.44 0.02 3.61 1.32 <0.00 0.00 0.00 0.00 1.92 0.44 <0.00 13.25 59.08 0.17 31.25 21.72 2.58 23.08 6.14 7.90 Rotifera Other Invertebrates Amphipoda 0.00 0.00 0.00 6.25 0.54 1.13 7.69 2.63 7.84 Gastropoda 1.21 0.17 0.07 6.25 0.54 1.10 0.00 0.00 0.00 Hydrachnidia 0.00 0.00 0.00 15.63 1.34 0.07 13.46 3.51 0.06 Invertebrate eggs 8.43 54.95 0.02 6.25 18.50 0.03 0.00 0.00 0.00 Other (unidentiied) Vertebrata Aphanius scales 1.21 3.96 0.08 3.13 0.80 0.25 0.00 0.00 0.00 6.02 0.99 0.69 18.75 3.79 0.05 1.92 0.44 <0.00 3.61 0.66 0.01 12.50 3.22 0.04 1.92 0.44 <0.00 Aphanius adult 1.21 0.17 0.68 0.00 0.00 0.00 0.00 0.00 0.00 Gambusia scales 1.21 0.17 <0.00 3.13 0.57 0.10 0.00 0.00 0.00 Plant-Detritus 89.16 4.79 62.40 18.75 0.54 6.21 19.23 1.51 2.28 Detritus 39.76 – 11.96 9.38 – 1.07 5.77 – 0.07 Filamentous Algae 75.90 – 47.88 12.50 – 3.83 17.31 – 0.16 Diatoms 53.01 – 2.03 6.25 – 1.31 1.92 – <0.00 Macrophytes 6.02 3.96 0.52 3.13 0.27 0.01 1.92 1.51 2.06 Pollen 2.41 0.83 <0.00 0.00 0.00 0.00 0.00 – 0.00 Shannon-Wiener (H’) 2.14 (2.01-2.24) 2.59 (2.51-2.66) 2.18 (2.07-2.28) Levin’s Niche Breadth (B) 6.39 (5.42-7.24) 10.65 (9.43-11.67) 6.53 (5.57-7.45) Std. Niche Breadth (BA) 0.23 (0.19-0.27) 0.42 (0.37-0.46) 0.24 (0.20-0.28) 2.42±0.25 3.04±0.39 3.09±0.41 Trophic Level ± SE algae (at Site 1) in larger individuals. This ontogenetic pattern was evident between the >40mm and <30mm individuals, but not between these length classes and the 30-40 mm length class. Indeed, the factorial test with the length and sex as ixed factors indicated a signiicant efect of length rather than of sex or length x sex interaction on the overall diet of the species (Table 7). 124 The feeding strategy diagrams show similar feeding strategy patterns at all examined sites (Fig. 7). At all sites, a mixed-feeding strategy, with alternating levels of specialisation and generalisation on diferent prey types was presented. The percent of individual predators specialised on diferent types of prey, which indicates a broad niche, was relatively higher at Site 2. This pattern Variation in Life History and Feeding Ecology of the Invasive Eastern Mosquitoish, Gambusia holbrooki... Fig. 5. Seasonal change in the percent relative importance index of Gambusia holbrooki at Site 1 in Acıgöl. IRI – index of relative importance; Zoo=zooplankton; NDl=Nematorecan Diptera (larvae); UI=unidentiied Insecta; NDa=Nematorecan Diptera (adults); Op=other plant material; Fa=ilamentous algae; Det=detritus Fig. 6. Two-dimensional ordination plots resulting from the diet composition data of Gambusia holbrooki at all sites in Acıgöl. Amp=Amphipoda; BD=Brachyceran Diptera; Cla=Cladocera; Cop=Copepoda; Det=detritus; Dia=diatoms; Epl=Ephemeroptera (larvae); Fa=ilamentous algae; Hyd=Hydracinidia; NDl=Nematorecan Diptera (larvae); NDa=Nematorecan Diptera (adults) was also evident in the Shannon diversity function and Levins niche breadth estimations (Table 6). The trophic levels of the food items that dominated the population were diferent between the sites (ilamentous algae and diatoms at Site 1 and Diptera groups at Sites 2 and 3). Discussion The most prominent contrasting factors between the studied sites were the water salinity and temperature, which greatly luctuated at Site 3, and were highly stable throughout the year at Sites 1 and 2 (Table 1). 125 Yoğurtçuoğlu B. & F. G. Ekmekçi Fig. 7. Graphical analysis of Gambusia holbrooki feeding strategy in Acıgöl, using the modiied Costello method. %Pi – prey-speciic abundance; %FO – frequency of occurrence; Amp=Amphipoda; Ata=A. transgrediens (adult); Ats=A. transgrediens (scale); BD=Brachyceran Diptera; Cla=Cladocera; Col=Coleoptera; Cop=Copepoda; Det=detritus; Dia=diatoms; Epl=Ephemeroptera (larvae); Fa=ilamentous algae; Gas=Gastropoda; Ghs=G. holbrooki (scale); Hem=Hemiptera; Hy=Hydracinidia; Inp=insect part; Mac=macrophytes; NDl=Nematorecan Diptera (larvae); NDa=Nematorecan Diptera (adults); Os=Ostracoda; Tri=Trichoptera (larvae) 126 Variation in Life History and Feeding Ecology of the Invasive Eastern Mosquitoish, Gambusia holbrooki... Table 7. The factorial PERMANOVA test with length and sex of Gambusia holbrooki as ixed factors. df – degrees of freedom; MS – mean square; Pseudo-F – permutational F-statistics; p(perm) – permutational p value Source df MS Pseudo-F P(perm) Total length 2 9990 2.8083 0.0008 Sex 1 4853.2 1.3643 0.1868 1 6349.2 1.7848 0.0714 139 3557.3 Total length x Sex Residuals Therefore, the individuals of G. holbrooki at Site 3 were exposed to a higher degree of environmental stress than those at the other sites. The lack of individuals at this site in most of the season was related to these conditions – mostly to the extreme salinity luctuations. Considering the overall sampling in the Acıgöl Lake, the females of G. holbrooki markedly dominated the sample with lower proportion in spring and summer. The sex ratio of Gambusia spp., as it is known, is 1:1 at birth (Krumholz 1948). The later stage imparity can often be linked to selective mortality, sampling bias and diferent habitat preferences of males and females (Fernández-Delgado 1989, Vargas & de Sostoa 1996, Fernández-Delgado & Rossomanno 1997). Approaching the 1:1 ratio in summer and spring in this study can be attributed to a combination of one or more factors: (i) mating aggregation; (ii) acting of selective mortality by predators at the expense of females due to pregnancy (Vargas & de Sostoa 1996, Cabral & Marques 1999); and (iii) the new generation represented in the sample (Haynes & Cashner 1995). Many studies on Gambusia spp. have shown that females have higher survival rate and larger longevity than males (Pérez-Bote & López 2005, Patimar et al. 2011, Ruiz-Navarro et al. 2011). While females survived after the reproduction period in Acıgöl, a great proportion of males from the parental cohort did not survive the winter, especially those sampled at Site 3. The somatic costs of reproduction against unpredictable conditions, which can be related to high adult mortality, may explain this situation, considering that only two individuals of 1+ males have been caught. Making a comprehensive comparison on the seasonal growth between the sampling sites was impossible, since we could not ind ish in some months, except at Site 1. However, according to the estimated von Bertalanfy parameters, diferent asymptotic length values were estimated among the sampling sites, being the highest at Site 3, where the water temperature reached 10°C. The efect of temperature was more apparent on the asymptotic length than on the growth coeicient (k) and L∞ tended to be greater as temperature declined (Basoline et al. 2004). The seasonal growth pattern illustrated for Site 1 is similar to other Mediterranean populations (Cabral & Marques 1999, Pérez-Bote & López 2005). It has been suggested that the timing of the reproductive cycling in Gambusia spp. is mainly related to photoperiod rather than temperature (Hubbs 1971, Milton & Arthington 1983, Cech et al. 1992). This was also proven for Acıgöl, as the gravidity pattern and GSI were similar between the sites even though the temperature was stable throughout the year at Site 1 and Site 2, but not at Site 3. The reproductive season started at the beginning of April and ended at the end of September as observed in many other populations (Vargas & de Sostoa 1996, Specziár 2004, Pérez-Bote & López 2005, Zarev 2012). Given that the pregnancy period suggested for mosquitoish is about four weeks (Krumholz 1948, Milton & Arthington 1983, Meffe 1990), it means that G. holbrooki females could release up to six broods in Acıgöl between April and September. Like many other ish species, the fecundity has been known to be positively correlated with female length in Gambusia spp. (Krumholz 1948). The strongest correlation in the fecundity-length relationship was estimated at Site 3, with larger females and more embryos. Many studies have shown that the reproductive investment increases (e.g. more embryos, heavier gonads) with salinity (Brown-Peterson & Peterson 1990, Vargas & de Sostoa, 1996, Alcaraz et al. 2008, Swenton & Kodric-Brown 2012). Therefore, the highest absolute fecundity and the maximum number of embryos found in a female were obtained at Site 3, with the cost of the highest percent of malformed eggs. Although G. holbrooki is known to prefer animal prey (Arthington 1989, Arthington & Marshall 1999, Garcia-Berthou 1999), a few studies suggest that the species feeds more on plant material (Gophen et al. 1998, Arthington & Marshall 1999, Specziár, 2004). Prey shifting in response to availability of sources is common in the mosquitoish (Pyke 2005). Most likely the plant-weighed feeding habit demonstrated at Site 1, but not at other sites, was associated with the high intraspeciic density. It is diicult to characterise the feeding of G. holbrooki. Even though some studies reveal that its 127 Yoğurtçuoğlu B. & F. G. Ekmekçi feeding is based mainly on zooplankton (Bence & Murdoch 1986, Kramer et al. 1987, Singh & Gupta 2010), or on larvae of Diptera (Hoy et al. 1972, Peck & Walton 2008), it has a highly lexible diet depending on the available sources (Pyke 2005). Similarly, in Acıgöl, three distinct compositions for three spatially close sampling sites were identiied (Table 6). The ontogenetic shift in diet has been reported in previous studies (Miura et al. 1979, Farley 1980, Garcia-Berthou 1999). In Acıgöl, the diet composition of the small individuals was dominated by Hydracinidae, Cladocera and Copepoda, and shifted towards larger taxa speciic to each sampling site. Although there was remarkable sexual dimorphism in terms of body size, the efect of the length on the overall diet variation was more signiicant than that of the sex. This may support the indings of Crivelli & Boy (1987) who hypothesised that the ontogenetic resource partitioning could involve diferent microhabitat use of diferent size classes. The modiied Costello method outlined a mixed-feeding strategy for G. holbrooki in Acıgöl, which could be unclear in the summary table of dietary composition. Relatively broader trophic niches with higher rate of specialised individuals were illustrated for Site 2, which was also evident from the diversity indices. This was not surprising, when considering the mild environmental conditions with a larger habitat size at Site 2. Several studies have demonstrated the ability of G. holbrooki to outcompete native ishes (Pyke 2005, Alcaraz & Garcia-Berthou 2007, Alcaraz et al. 2008). Even though there is evidence that G. holbrooki and the critically endangered A. transgrediens (the only native ish in the lake) do not compete notably for food in Acıgöl (unpublished data), Aphanius scales were observed in the stomach of G. holbrooki, which indicates its aggressive pressure on A. transgrediens. This behaviour seems to be species-speciic having in mind the lower occurrence of Gambusia scales in the examined stomachs. Acknowledgement: This study is part of a PhD thesis of the irst author and was supported by the Ruford Small Grants (Projects 16079-2 and 12983-1). References Alcaraz C & García-Berthou E. 2007. Life history variation of invasive mosquitoish (Gambusia holbrooki) along a salinity gradient. Biological Conservation 139: 83-92. Alcaraz C., Bisazza A. & Garcia-Berthou E. 2008. Salinity mediates the competitive interactions between invasive mosquitoish and an endangered ish. Oecologia 155: 205-213. Amundsen P. A., Gabler H. M. & Staldvik F. J. 1996. A new approach to graphical analysis of feeding strategy from stomach contents data – modiication of the Costello (1990) method. Journal of Fish Biology 48: 607-614. Anderson M. J. & Willis T. J. 2003. Canonical analysis of principal coordinates: a useful method of constrained ordination for ecology. Ecology 84: 511-525. Anderson M.J., Gorley R. N. & Clarke K. R. 2008. PERMANOVA+ for PRIMER: Guide to software and statistical methods. Plymouth, UK: PRIMER-E Ltd., 214 p. Arthington A. H. 1989. Diet of Gambusia ainis holbrooki, Xiphophorus helleri, X. maculatus, and Poecilia reticulata (Pisces: Poeciliidae) in streams of southeastern Queensland, Australia. Asian Fisheries Science 2: 193-212. Arthington A. H. & Marshall C. J. 1999. Diet of the exotic mosquitoish Gambusia holbrooki in an Australian lake and potential for competition with indigenous ish species. Asian Fisheries Science 12: 1-16. Basoline G., Guisande C., Patti B., Mazzola S., Cuttitta A., Bonanno A. & Kallianiotis A. 2004. Linking habitat conditions and growth in European anchovy (Engraulis encrasicolus). Fisheries Research 68: 9-19. Bence J. R. & Murdoch W. W. 1986. Prey size selection by the mosquitoish: relation to optimal diet theory. Ecology 67 (2): 324-336. 128 Brown-Peterson N. & Peterson M. S. 1990. Comparative life history of female mosquitoish Gambusia ainis in tidal freshwater and oligohaline habitats. Environmental Biology of Fishes 27: 33-41. Cabral J. A. & Marques J. C. 1999. Life history, population dynamics and production of eastern mosquitofish, Gambusia holbrooki (Pisces, Poeciliidae), in rice ields of the Lower Mondego River Valley, West Portugal. Acta Oecologica 20: 607-620. Cech J. J. Jr., Schwab R. G., Coles W. C. & Bridges B. B. 1992. Mosquitoish reproduction: efects of photoperiod and nutrition. Aquaculture 101: 361-369. Costello M. J. 1990. Predator feeding strategy and prey importance: a new graphical analysis. Journal of Fish Biology 36: 261-263. Crivelli A. J. & Boy V. 1987. The diet of the mosquitoish (Gambusia affinis (Baird and Girard) Poeciliidae) in Mediterranean France. Annual Review of Ecology and Systematics 42: 421-435. Ekmekçi F. G., Kirankaya Ş. G., Gençoğlu L. & Yoğurtçuoğlu B. 2013. Present status of invasive ishes in inland waters of Turkey and assessment of the effects of invasion. Istanbul University Journal of Fisheries & Aquatic Sciences 28: 105-140. Erençin Z. 1978. Opinions on malaria (mosquitoes) control and the ish farming activities. Journal of the Faculty of Veterinary Medicine, Ankara University 25: 760-766. (in Turkish) Ergüden S. A. 2013. Age, growth, sex ratio and diet of eastern mosquitoish Gambusia holbrooki Girard, 1859 in Seyhan Dam Lake (Adana/Turkey). Iranian Journal of Fisheries Sciences 12 (1): 204-218. Variation in Life History and Feeding Ecology of the Invasive Eastern Mosquitoish, Gambusia holbrooki... Farley D. G. 1980. Prey selection by the mosquitoish, Gambusia ainis (Baird and Girard) on selected non-target organisms in Fresno County rice ields. Proceedings of the California Mosquito and Vector Control Association 48: 51-54. Fernández-Delgado C. 1989. Life-history patterns of the mosquitofish, Gambusia afinis, in the estuary of the Guadalquivir river of south-west Spain. Freshwater Biology 22 (3): 395-404. Fernández-Delgado C. & Rossomanno S. 1997. Reproductive biology of the mosquitofish in a permanent natural lagoon in south-west Spain: two tactics for one species. Journal of Fish Biology 51: 80-92. Freyhof J. 2014. Aphanius transgrediens. The IUCN Red List of Threatened Species. http://dx.doi.org/10.2305/IUCN. UK.2014-1.RLTS.T19384476A19848952.en Froese R. 2006. Cube law, condition factor and weight-length relationships: history, meta-analysis and recommendations. Journal of Applied Ichthyology 22: 241-253. Garcia-Berthou E. 1999. Food of introduced mosquitoish: ontogenetic diet shift and prey selection. Journal of Fish Biology 55: 135-147. Gayanilo F. C., Jr., Soriano M. & Pauly D. 1988. A draft guide to the Compleat ELEFAN. ICLARM Software 2. Manila, Philippines: International Center for Living Aquatic Resources Management, 65 p. Gkenas C., Oikonomou A., Economou A., Kiosse F. & Leonardos I. 2012. Life history pattern and feeding habits of the invasive mosquitoish, Gambusia holbrooki, in Lake Pamvotis (NW Greece). Journal of Biological ResearchThessaloniki 17: 121-136. Gophen M., Yehuda Y., Malinkov A. & Degani G. 1998. Food composition of the ish community in Lake Agmon. Hydrobiologia 380: 49-57. Haynes J. L. 1995. Standardized classification of Poeciliid development for life history studies. Copeia 1995: 147154. Haynes J. L. & Cashner R. C. 1995. Life history and population dynamics of the western mosquitoish: a comparison of natural and introduced populations. Journal of Fish Biology 46: 1026-1041. Hellawell J. M. & Abel R. 1971. A rapid volumetric method for the analysis of the food of ishes. Journal of Fish Biology 3: 29-37. Hoy J. B., Kaufmann E. E. & O’Berg A. G. 1972. A large-scale ield test of Gambusia ainis and chlorpyrifos for mosquito control. Mosquito News 32: 161-171. Hubbs C. 1971. Competition and isolation mechanisms in the Gambusia ainis x G. heterochir hybrid swarm. Texas Memorial Museum Bulletin 19: 1-47. Hyslop E. J. 1980. Stomach content analysis: a review of methods and their applications. Journal of Fish Biology 17 (4): 411-429. Kramer V. L., Garcia R. & Colwell A. E. 1987. An evaluation of the mosquitoish, Gambusia ainis, and the inland silverside, Menidia beryllina, as mosquito control agents in California wild rice ields. Journal of the American Mosquito Control Association 3: 626-32. Krebs C. J. 1989. Ecological Methodology. New York, USA: Harper & Row, 634 p. Krumholz L. A. 1948. Reproduction in the western mosquitoish Gambusia ainis ainis (Baird & Girard) and its use in mosquito control. Ecological Monograph 18: 1-43. Matthews W. J. & Matthews E. M. 2011. An invasive ish species within its native range: community efects and population dynamics of Gambusia ainis in the central United States. Freshwater Biology 56: 2609-2619. Meffe G. K. 1990. Offspring size variation in eastern mosquitoish (Gambusia holbrooki: Poeciliidae) from contrasting thermal environments. Copeia 1990 (1): 10-18. Meffe G. K. 1991. Life history changes in eastern mosquitoish (Gambusia holbrooki) induced by thermal elevation. Canadian Journal of Fisheries and Aquatic Sciences 48: 60-66. Milton D. A. & Arthington A. H. 1983. Reproductive biology of Gambusia ainisholbrooki (Baird and Girard), Xiphophorus helleri (Günther) and X. maculatus (Heckel) (Pisces; Poeciliidae) in Queensland, Australia. Journal of Fish Biology 23: 23-41. Miura T., Takahashi R. M. & Stewart R. J. 1979. Habitat and food selection by the mosquitoish, Gambusia ainis. Proceedings of the California Mosquito and Vector Control Association 47: 46-50. Öztürk S. & İkiz R. 2002. Some biological properties of mosquitoish populations (Gambusia ainis) living in inland waters of the Western Mediterranean region of Turkey. Turkish Journal of Veterinary and Animal Science 28: 355-361. Patimar R., Ghorbani M., Gol-Mohammadi A. & Azimi Glugahi H. 2011. Life history pattern of mosquitoish Gambusia holbrooki (Girard, 1859) in the Tajan River (Southern Caspian Sea to Iran). Chinese Journal of Oceanology and Limnology 29 (1): 167-173. Peck G. W. & Walton W. E. 2008. Efect of mosquitoish (Gambusia ainis) and sestonic food abundance on the invertebrate community within a constructed treatment wetland. Freshwater Biology 53 (11): 2220-2233. Pérez-Bote J. L. & López M. T. 2005. Life-history pattern of the introduced eastern mosquitoish, Gambusia holbrooki (Baird & Girard, 1854), in a Mediterranean-type river: The River Guadiana (SW Iberian Peninsula). Italian Journal of Zoology 72 (3): 241-248. Pinkas L., Oliphant M. S. & Iverson I. L. K. 1971. Food habits of albacore, bluein tuna, and bonito in California waters. State of California the Resources Agency Department of Fish and Game, Fish Bulletin 152, 105 p. Pyke G. H. 2005. A review of the biology of Gambusia ainis and G. holbrooki. Reviews in Fish Biology and Fisheries 15: 339-365. Ruiz-Navarro A., Moreno-Valcárcel R., Torralva M. & Oliva-Paterna F. J. 2011. Life-history traits of the invasive ish Gambusia holbrooki in saline streams (SE Iberian Peninsula): does salinity limit its invasive success? Aquatic Biology 13: 149-161. Rupp H. R. 1996. Adverse assessments of Gambusia ainis: an alternate view for mosquito control practitioners. Journal of the American Mosquito Control Association 12: 155-166. Scalici M., Avetrani P. & Gibertini G. 2007. Mosquitoish life history in a Mediterranean wetland. Journal of Natural History 41: 887-900. Schindelin J., Arganda-Carreras I. & Frise E. 2012. Fiji: an open-source platform for biological-image analysis. Nature methods 9 (7): 676-682. Singh N. & Gupta P. K. 2010. Food and feeding habits of an 129 Yoğurtçuoğlu B. & F. G. Ekmekçi introduced mosquitoish, Gambusia holbrooki (Girard) (Poeciliidae) in a subtropical lake, Lake Nainital, India. Asian Fisheries Science 23 (3): 355-366. Specziár A. 2004. Life history pattern and feeding ecology of the introduced eastern mosquitoish, Gambusia holbrooki, in a thermal spa under temperate climate, of Lake Hévíz, Hungary. Hydrobiologia 522: 249-260. Stergiou Κ. Ι. & Karpouzi V. S. 2002. Feeding habits and trophic levels of Mediterranean ish. Reviews in Fish Biology and Fisheries 11: 217-254. Sun J. & Liu D. 2003. Geometric models for calculating cell biovolume and surface area for phytoplankton. Journal of Plankton Research 25 (11): 1331-1346. 130 Swenton D. M. & Kodric-Brown A. 2012. Habitat and life history diferences between two species of Gambusia. Environmental Biology of Fishes 94 (4): 669-680. Vargas M. J. & De Sostoa A. 1996. Life history of Gambusia holbrooki (Pisces, Poeciliidae) in the Ebro delta (NE Iberian Peninsula). Hydrobiologia 341: 215-224. Yoğurtçuoğlu B. & Ekmekçi F. G. 2014. Threatened ishes of the world: Aphanius transgrediens Ermin, 1946 (Cyprinodontidae). Croatian Journal of Fisheries 72 (4): 186-187. Zarev V. Y. 2012. Some life-history traits of Gambusia holbrooki (Pisces: Poeciliidae) from Bulgaria. Acta Zoologica Bulgarica 64 (3): 263-272.