MATERIALS AND METHODS

Study area

The Laguna El Llano (ELEL) estuary is located at 19°40'05" N and 96° 23'54" W, in the municipality of Actopan, Veracruz (Figure 1). It is a 226 ha body of water and the depth at the internal margins is less than 1 m deep, but in the central channel, it reaches a depth interval that varies seasonally from 2 to 4 m.

Figure 1 Location of the Laguna El Llano estuary (ELEL), Actopan municipality, Veracruz, Mexico. The numbers indicate the sampling sites (SM1-SM6) in May 2013. 

The climate is warm humid with rains in summer (Aw2) (^{García, 2004}), the average annual precipitation is 1286 mm and the average annual temperature is 24 °C. Three climatic seasons are different in the region: the warm one runs between April and June, the rainy season between July and October, and the cold season occurs from November to March (^{Morán-Silva et al., 2005}).

Precisely at this season, a sandy bar forms at the mouth of the estuary due to coastal sedimentary processes. For this geomorphological feature, the definition of the estuary by ^{Potter et al. (2010)} is used here. That alludes to the temporary isolation of the estuary and the potential that this event causes hyperhalinity conditions.

The specimens of P. mexicana were only collected in May 2013. This was the initial sampling of an investigation on the ELEL fish community that was carried out between that month and May 2014. The collections were made in six sampling sites (SM) near the banks of the red mangrove Rhizophora mangle and the black mangrove Avicennia germinans, characterized by muddy bottoms and close to reefs of the Crassostrea virginica oyster. Sampling was restricted to depths less than 1.2 m; it did not extend to the southern part of the estuary due to the shallowness of the water column (less than 30 cm) and the muddy consistency (> 1 m) of the upper horizon of the substratum. The SMs were located with a Garmin 10X GPS. Dissolved oxygen (mg/L) and water temperature (°C) was also recorded with an Oakton DO 300 oximeter with 0.1 °C accuracy; salinity (UPS) with a Vista A366ATC refractometer with 0.1 UPS accuracy; turbidity with a La Motte 2020 (UNT) turbidimeter with 0.1 UNT accuracy; pH with an Oakton pH 110 Meter with 0.1 pH unit accuracy, and the depth (cm) was recorded with an Echotest II Plastimo probe with a precision of 0.1 cm. The salinity categories were defined based on the proposal of the Venice System (^{Ito, 1959}).

Specimens were captured using a 25 m long by 2 m high seine net with a 6.5 mm mesh span. It was tried that the collection area in each SM was 300 m²; only in SM 2 it was 125 m². All the collected specimens were placed on ice and then in bottles with 70 % ethanol. In the laboratory they were washed, fixed with 5 % formalin and stored in 70 % ethanol. The specimens were identified with ^{Álvarez del Villar (1970)} and ^{Miller et al. (2009)}. Each fish was measured with a standard length scale (LP) up to 0.01 mm. The individual weight (P), the gonads (PG), and the hepatopancreas (PH) were measured with a Cole-Parmer Simmetry electronic balance with a precision of 0.001 g. The difference of P minus PG was taken as the somatic weight (PS).

Statistical analysis

The size structure was established by separating the organisms in 10 mm length intervals. The sizes were defined using the frequency analysis method of ^{Pope et al. (2010)}. Relative growth was evaluated based on the biometric relationship LP vs. P, where LP was the independent variable. This biometric relationship was adjusted using a potential function (Y = aX^b), where a is the ordinate to the origin (initial growth coefficient) and b is the slope (growth coefficient or also called coefficient of allometry). The LP and P values were transformed into logarithms and the result of the linear relationship was adjusted with a least-squares regression using P as the dependent variable (Le Creen, 1951; ^{Ricker, 1975}). The degree of association between the variables was established utilizing the determination coefficient (r²). The LP vs. P ratios were calculated for the entire fish sample and each sex.

To test the isometric growth hypothesis, a student's t-test was used, testing Ho: b = 3 against H1: b ≠ 3 with n-1 degrees of freedom and a significance level of P = 0.05. In the same way, to determine if differences occurred between the values of b of males and females, Ho was tested: b_males - b_females = 0, indicating differences between the slopes, against H1: b_males - b_females 4 0, indicating differences in the value of b between the groups. Also, an ANCOVA test was used to compare the differences between the biometric relationships of LP vs. P between sexes, using LP as a covariate. All statistical procedures were calculated with the PAST program (^{Hammer et al., 2001}) with a significance of P = 0.05.

The Fulton K condition factor was estimated, which was used to compare the welfare of the fish with the assumption that a heavier fish at a given length has a better condition (Le Creen, 1951), according to the equation:

Where,

P = weight of each fish in g

LP = individual standard length in mm

b was used as the relative growth constant and factor 100 is used to scale K to unity. This model was applied separately to the group of males and females, using the respective value of the allometric coefficient b.

The individual values of K of males and females were correlated to their respective values of LP and P as independent variables. The Gonadosomatic Index (IGS, ^{De Vlaming et al., 1982}) of males and females was also correlated to LP, P, PS, PG and PH (^{Vazzoler, 1996}). In all cases the Pearson's correlation test was used with P = 0.05.

Gonadic maturity and stages of embryonic development were defined based on the criteria of ^{Riesch et al. (2011)}. Oocytes and embryos were counted in the gonads of 50 females. With these records, fertility was estimated.

The composition of the diet was determined with the numerical grid method (^{Trujillo-Jiménez and Toledo, 2007}). Food types were identified at the most precise taxonomic level possible. Diatom species were identified with ^{Prescott (1978)} and ^{Bellinger and Sigee (2010)}. Diet composition between sizes was compared with a one-way ANOVA test with P = 0.05.

RESULTS

The hydrological records for May 2013 corresponded to warm, slightly alkaline, well-oxygenated and hyperhaline water (Table 1). In June 2013 the sandy bar disintegrated and the exchange of water between the estuary and the marine platform resumed.

Table 1 Hydrological records in ELEL for May 2013 during the closed estuarine inlet phase. UNT: Nephelometric Turbidity Units, UPS: Practical Salinity Units, DS: Standard Deviation. 

Poecilia mexicana was only collected in May 2013 and was the most abundant species. Three other species were also obtained: Mugil cephalus, Centropomus undecimalis, and Lutjanus griseus. Of the 170 individuals, 128 were females in an LP range of 4.28-59.21 mm and 0.04-5.48 g P and 42 were males that presented in an interval of 15.73-51.41 mm LP and 0.091-2.97 g P. The size classes of 10-19 and 20-29 mm LP were the most numerous. The organisms from these ranges grouped 90.5 % of the abundance of males and 79.7 % of the abundance of females (Tabla 2).

Table 2 Size classes and abundance of P. mexicana in ELEL in May 2013. 

Females of P. mexicana presented negative allometric growth (P < 0.001) while males tended towards isometric growth (P < 0.001) (Table 3). The weight increases were not different between males and females (P > 0.05).

Table 3 Parameters of the relative growth model (Weight P vs. Standard Length LP) for females, males, and all individuals of P. mexicana in ELEL. 

The condition factor K of the females was recorded in a range of 0.036-4.31 (mean ± SD: 0.33 ± 0.488) and for males in a range of 0.86-2.64 (mean ± SD: 0.405 ± 0.464). K was positively correlated to LP and P of females and males (P < 0.001) (Figure 2a and Figure 2b).

Figure 2 Condition factor (K) by sex of P. mexicana in ELEL. a) condition factor K correlated to LP as an independent variable; b) condition factor K correlated to P as an independent variable. 

The female LP showed positive relationships with PS, PG, and IGS; on the other hand, weight (P) had a high correlation with PS and a positive correlation with PG, which in turn showed a positive correlation with IGS. These results indicate that the increases of LP and P of the females did not influence the gonad development. All the correlations were significant (P < 0.001). In the case of males, positive correlations were found between LP and P with respect to PS, PG, and PH. The correlations with the gonadosomatic index were higher between LP and P with respect to PG and IGS. All the correlations were also significant (P < 0.001) (Table 4).

Table 4 Correlations between somatic variables and gonadic variables of P. mexicana under hyperhalinity conditions in ELEL. 

Of 50 females reviewed, 11 did not show gonad development. Females with some level of gonad maturity were in intervals of 13.7-34.7 mm LP and 0.07-1.31 g P. Ten females between 36.3-50.7 mm LP loaded reproductive products such as eggs and embryos. Of the latter, in four embryos were found at different levels of development. Finally, a 50.69 mm LP female presented oocytes, eggs, and embryos. Fertility was recorded between 12-179 oocytes/female (average 31.3 oocytes/female), 8-162 eggs/female (average 54.3 eggs/female), and 29-72 embryos/female (average 51.2 embryos/female). Only positive correlation was found between female LP with the number of oocytes (n = 7, r = 0.85, P < 0.05).

The diet of P. mexicana in ELEL was mainly composed of detritus (97.7 %). In the remaining portion diatoms (1.5 %), unidentified algae, foraminifera, and crustacean remains (0.26 % each item) were found. The diatoms observed correspond to Navicula radiosa, N. chryptocephala, Diatoma hemiale, Surirella robusta var. splendida, Cymbella sp., and Nitzschia sp. A comparison of diet between length classes showed no difference (P < 0.05); this food composition identifies P. mexicana as a detritivorous consumer.

DISCUSSION

This is the first record of P. mexicana in hyperhaline estuarine waters. Regarding its tolerance to salinity, ^{Miller (1983)} published its occurrence in estuaries of the Gulf of Mexico. In later writings, it has been recorded in salinities less than 15 UPS (^{Castro-Aguirre and Mora-Pérez, 1984}; ^{Rodríguez-Varela et al., 2010}; ^{Aguirre-León et al., 2014}; ^{Chávez-López et al., 2015}).

The ability of P. mexicana to withstand high salinities can be explained by biochemical and morphological processes described in other species, such as the secretion of carbohydrates in the intestine (^{Laverty and Skadhauge, 2015}), which causes calcium precipitation and increases the ability to absorb the sodium chloride from hypersaline waters (^{Whittamore, 2012}). Other investigations consider the reduction of branchial permeability by regulating the levels of aquaporin proteins in the branchial epithelia (Tipsmarck et al., 2010; Verkman, 2012), which can be related to changes in the water ingestion rates that increase in fish that colonize saline waters (^{Nordlie, 2006}). Other evidences indicate that it decreases the ionic permeability of the body surface and in the gill ion-secreting cells, that are essential to eliminate large amounts of salts in marine and hyperhaline habitats (^{González, 2012}).

The hyperhalinity of the water did not limit the reproduction of P. mexicana. The evidence provided by the females when presenting oocytes and embryos demonstrates the use of energy for reproductive processes, although it is proposed that under conditions of environmental stress the energy budget should be reoriented towards basal physiological maintenance (^{Stearns, 1992}). Even the averages of the number of oocytes (31.3 oocytes/female) and eggs (54.3 eggs/female) were similar to those found in an oligohaline coastal lagoon in the region (^{Chávez-López et al., 2015}). Furthermore, both the condition factor and the gonadal weight were positively correlated to the length and weight increases. The intervals of these two parameters were similar to those found in females collected in superficial aquatic and cave habitats in southern Mexico (^{Riesch et al., 2006}).

The average of embryos (51.2 embryos/female) was similar to that found for females of the Alvarado Lagoon System (average 58.9 embryos/female). These values contrast with the interval of embryos carried by females that inhabit sulfurous waters in caverns (1-14 embryos) (^{Tobler and Plath, 2011}). Although a low number of eggs and embryos would have been expected in ELEL females due to the influence of hyperhalinity, the intervals recorded in ELEL are more approximate to those recorded in oligohaline habitats in the region, which would indicate a minimal effect of salinity in the reproductive effort of females.

In other cyprinodontiforms, this response has been found to imply high energy investments for reproduction, but the level of body condition decreases as in Gambusia holbrooki (^{Alcaraz and García-Berthou, 2007}). Increases in salinity create sub-lethal stress levels that lead to increased energy expenditure due to additional osmoregulation, reducing the energy destined for reproduction, maintenance of body condition, and lecithotrophic provisioning (^{Martin et al., 2009}).

Diet analysis showed that detritus was the main food for P. mexicana, which is why it is recognized as a detritivorous consumer (^{Miller, 2009}). This specie is capable of modifying its diet under different hydrological conditions based on the availability of other foods; for example, different species of algae were found as the main food in specimens from the Lagunar System of Alvarado, Veracruz (^{Chávez-López et al., 2015}). Under extreme conditions in caves with sulphurous waters in Tabasco, Mexico, aquatic arthropods were mainly consumed, but in caves with freshwater currents, combined consumption of detritus with bat guano was found (^{Tobler, 2008}). With this evidence, the detritivorous diet of P. mexicana stands out as a great advantage to survive in hyperhaline waters. Due to the vast availability of detritus in the estuary, food is no longer a limiting environmental factor.

Based on this information, P. mexicana is an exception to the reasons proposed by ^{Whitfield (2015)} to explain why freshwater fish species are underrepresented in estuaries. The first reason indicates its inability to withstand physiologically wide salinity intervals. The information presented here revealed the tolerance levels of this species. The second reason refers to adaptation to altered food resources in estuaries compared to freshwater systems. In the case of P. mexicana, its detritivorous habits facilitate this condition. The third reason indicates that freshwater species are not skillful competitors with respect to marine and estuarine species in estuaries. Here, the contrary is demonstrated by the greater abundance registered for collected marine species, as well as by the maintenance of their reproductive processes. The fourth reason refers to their possible combination in assemblages of fish of marine origin and the last reason suggests that freshwater species are difficult to overcome environmental obstacles between estuaries and freshwater streams. The distribution of P. mexicana in this region of the Gulf of Mexico and its occurrence in hyperhalinity indicates a great capacity to colonize estuarine habitats in extreme environmental conditions.