Factors regulating growth pattern and condition factor of an amphibious fish Periophthalmus gracilis living in the Mekong Delta

Study site and fish analysis

This study lasted for 12 months from April 2020 to March 2021 at four sites in the estuarine and coastal regions in MD, including Duyen Hai - Tra Vinh (TV, 9°40′29.5″N 106°34′49.5″E); Tran De - Soc Trang (ST, 9°26′19.7″N 105°10′48.1″E); Dong Hai - Bac Lieu (BL, 9° 05′50.5″N 105°29′54.7″E) and Dam Doi - Ca Mau (CM, 8°58′10.4″N 105°22′58.9″E) (Fig. 1). There are two seasons in these sites, including the dry season from January to May (with no rain) and the wet season from June to December (with heavy rain) (Le et al., 2006). The pH ranged 7.6−8.0, and the salinity varied widely from 12.3 to 23.5‰. The pH change depends on site but not season, whereas the salinity variation gives the opposite result (Dinh et al., 2021a). The primary vegetation at TV and ST is Sonneratia caseolaris and Rhizophora apiculata, while R. apiculata is the predominant vegetation in BM and CM, according to our observation.

An area of 120 m² (6-m width × 20-m length) in each site was chosen to monthly catch fish species at night as fish was so active in the daytime, from April 2020 to March 2021. Fish samples were collected monthly by hand-catching for 4-hours continuously during low tide. Semi-tides represent the site samples, and during low tide, fish gather in large numbers on the mudflats. Fish specimens were easily distinguished from congeners as P. gracilis was covered by many irregular blackish dots, whereas P. chrysospilos and P. variabilis were surrounded by tiny orange spots and greyish brown (Murdy & Jaafar, 2017). MS222 (30 mg/l) dissolved with water taken from the sampling site was used to anaesthetize the fish specimens before being preserved in formalin buffer 5% (dilute from formalin with the ratio of 1 formalin: 9 water taken from the sampling site) and shipped to the laboratory. Fish was sexing using genital papilla, which was triangle males and oval shape in females. Then, fish total length (TL) was measured using a ruler to the nearest 0.1 cm, and fish weight (W) was weighted using an electric scale to the nearest 0.01 g. The Council for Science and Education, School of Education, Can Tho University approves the fish use in this study (Animal Welfare Assessment number: BQ2020-03/KSP).

Data analysis

The length-weight relationship of fish was defined as W=a ×TL^b (W: fish weight, TL: fish total length, a: intercept parameter and b: slope parameter) (Ricker, 1973). The condition factor (CF), according to Le Cren (1951), was calculated as CF=W/ (a × TL^b). T-test verified if the b value obtained from the LWR s was ≈3. Fish showed positive allometry (b >  3), negative allometry (b < 3) and isometry (b ≈3) (Martin, 1949). T-test qualified whether CF was regulated by sex, size and season (dry season and wet season), while one-way ANOVA verified if CF changed with month and site (Mahmood et al., 2012). T-test confirmed if CF was ≈1, whereas General Linear Model qualified gender × season, gender × site and season × site affecting CF (Dinh, 2016a). Fish was divided into the immature group when TL <L_m and the mature group if TL ≥L_m. Fish length at first maturity (L_m) of males and females at each site was calculated from the formula: P = 1/(1 + exp[ − r × (TL − L_m)]) (P: proportion of mature individuals in a length class; TL: fish total length; and r: model parameter) (Zar, 1999). L_m was length at first maturity of female and male which was 5.0 cm and 5.7 cm in TV; 4.6 cm and 5.8 cm in ST; 4.9 cm and 5.2 cm in BL; and 6.2 cm and 5.9 cm in CM, respectively (Dinh et al., in press). Before weighting to the nearest 0.01 mg, ovarian and testicular development stages were classified into six developmental stages according to the methods of Dinh et al. (2020). Data analysis was performed using SPSS v.21, and all tests were set at p < 0.05. To lessen the Type I error of all tests, the Benjamini–Hochberg procedure was performed (Benjamini & Hochberg, 1995; McDonald, 2014).

Growth pattern

The total length and weight of 486 individuals collected at four sites from Tra Vinh to Ca Mau were 2.9–5.9 cm and 0.13–1.66 g, respectively (Table 1). The LWRs of P. gracilis in different fish sexes, sizes, seasons, sites and months were presented in Figs. 2–4. The growth pattern of P. gracilis was obtained from the slope value (b) of the length-weight relationship. Specifically, as a slope b (2.69 ± 0.06 SE) got from LWR was significantly less than the threshold of 3 (n = 486, df = 484, p <  0.01, t =  − 5.46), P. gracilis belonged to a negative allometric growth pattern.

In terms of sex, although the slope b value of females (2.80 ± 0.08 SE, n = 236) was higher than that of males (2.50 ± 0.07 SE, n = 250), both males and females displayed negative allometry as these values were <3 (df_females = 234, t_females = −2.40, p_females = 0.02; df_males = 248, t_males = −6.99, p_males<0.01, Figs. 2A–2B). Regardings fish size, the b value of mature fish (2.80 ± 0.37 SE, n = 63, Fig. 2D) was higher than that of immature fish (2.62 ± 0.06 SE, n = 423, Fig. 2C). The immature fish showed negative allometry due to b<3 (df = 421, t =  − 6.18, p <  0.01, Fig. 2C), whereas the mature fish displayed isometry because of b ≈3 (n = 63, df = 61, t =  − 1.99, p = 0.05, Fig. 2D). Similar to size, the growth pattern of P. gracilis changed with the season variable as it showed negative allometry in the dry season (b = 2.51 ± 0.08, <3, df = 195, t =  − 6.25, p <  0.01, Fig. 2E) but isometry in the wet season (b = 2.87 ± 0.08, ≈3, p = 0.08, t =  − 1.79, Fig. 2F). Growth pattern of this mudskipper varied with site as it showed negative allometry at TV, ST and BL but isometry in CM. Indeed, b values of this fish at TV (2.79 ± 0.10 SE, n = 139, Fig. 3A), ST (2.11 ± 0.10 SE, n = 95, Fig. 3B) and BL (2.68 ± 0.09, n = 129, Fig. 3C) were <3 (df_TV = 137, t_TV = −2.05, p_TV = 0.04; df_ST = 93, t_ST = −8.83, p_ST<0.1; df_BL = 127, t_BL = −3.43, p_BL<0.01). By contrast, this value in CM (b = 2.96 ± 0.11, n = 123, Fig. 3D) was ≈3 (df = 121, t =  − 0.41, p = 0.69). The lowest value of b-value in ST could suggest that most of fish caught from ST belonged to immature fish.

The growth pattern of this fish was also changed with the month variable as it showed negative allometry to isometry to positive allometry. Spefically, as b value in February (2.39 ± 0.16 SE, n = 33), April (2.35 ± 0.27 SE, n = 34), May (2.56 ± 0.19SE, n = 46), June (2.43 ± 0.23SE, n = 37), August (2.07 ± 0.20 SE, n = 51) was signicantly less than 3 (df_April = 32, t_April = −2.42, p_April = 0.02; df_May = 44, t_May = −2.31, p_May = 0.03; df_June = 35, t_June = −2.50, p_June = 0.02; df_August = 49, t_August = −4.71, p_August < 0.01; df_February = 31, t_February = −3.80, p_February < 0.01), this fish showed negative allometry. Only in September, the species displayed positive allometry since its b value (b = 3.47 ±.01, n = 42) was >3 (df = 40, t = 3.47, p < 0.01). By contrast, this fish showed isometric growth as b value of this fish in the remaning months, e.g., in January (2.67 ± 0.17, n = 35), March (2.84 ± 0.16, n = 49), July (2.49 ± 0.26, n = 43), October (3.03 ± 0.17, n = 43), November (2.93 ± 0.13, n = 38) and December (3.14 ± 0.25, n = 35), equivalented to 3 (df_July = 44, t_July = −1.98, p_July = 0.03; df_October = 41, t_October = 0.18, p_October = 0.86; df_November = 38, t_November = −0.52, p_November = 0.60; df_December = 35, t_December = 0.56, p_December = 0.58; df_January = 33, t_January = −1.99, p_January = 0.05; df_March = 41, t_March =  − 1.00, p_March = 0.32) (Fig. 4).

Condition factor

Condition factor (CF) of female P. gracilis (1.09 ± 0.02 SE, n = 236) was higher than that of males (0.96 ± 0.01, n = 250) (n = 486, df = 484, t = 5.94, p < 0.01, CI 95% = [0.12–0.05]). The CF of both sexes were significantly larger than 1 (n_females = 236, df_females = 235, t_females = 4.92, p_females < 0.01, CI 95%_females = [0.17 to −0.07]; dfn_males = 250, _males = 249, tmales = −3.33, p_males = 0.01, CI 95%_males = [−0.01 to −0.06]). Likewise, the CF of immature fish group (1.01 ± 0.01 SE, n = 423) was significantly less than that of mature one (1.12 ± 0.03 SE, n = 63) (df = 484, t =  − 3.19, p < 0.01, CI 95% = [(−0.04)–(−0.17)]). The CF of immature fish was ≈1 (n = 423, df = 422, t = 0.90, p = 0.36, CI 95% = [0.03–(−0.01)]); this value of mature fish was >1 (n = 63, df = 61, t = 4.50, p <  0.01, CI 95% = [0.17–0.06]). The variation of CF was also found in season variable as this value in the dry season (0.99 ± 0.01 SE, n = 197) was lower than in the wet season (1.05 ± 0.02 SE, n = 289) (df = 484, t =  − 2.74, p < 0.01, CI 95% = [(−0.02)–(−0.10)]. In the dry season, the CF was ≈1 (n = 197, df = 196, t =  − 0.88, p = 0.38, CI 95% [0.15–(−0.03)]); however, in the wet season the CF increases and was >1 (n = 289, df = 288, t = 3.09, p <  0.01, CI 95% = [0.08–0.01]).

The CF value of P. gracilis varied with site (one-way ANOVA, n = 486, df = 3, F_2,3 = 4.56, p <  0.01). This value at BL (0.94 ± 0.02 SE, n = 129) was significantly lower than that at the remaining 3 sites, comprising TV (1.09 ± 0.02 SE, n = 139), ST (1.04 ± 0.03 SE, n = 95) and CM (1.03 ± 0.02 SE, n = 123) (Fig. 5). At ST and CM, the CF of this fish was equivalent to 1 (n_ST = 95, df_ST = 94, t_ST = 1.19, p_ST = 2.35, CI 95%_ST = [0.09 to −0.02]; n_CM = 123, df_CM = 122, t_CM = 1.74, p_CM = 0.08, CI 95%_CM = 0.06 − (−0.01)). Meanwhile at BL, the CF was less than 1 (n = 129, df = 128, t =  − 3.09, p <  0.01, CI 95%= [−0.02 to −0.09]). and at TV this value was higher than 1 (n = 139, df = 138, t = 4.37, p <  0.01, CI 95% [0.12–0.04]). The CF of this fish fluctuated during the 12-month study (n = 486, df = 11, F_2,11 = 3.43, p <  0.01), reaching the highest value in March and August-October (1.07 ± 0.04 SE to 1.10 ± 0.06 SE) and the lowest value in February (0.89 ± 0.02 SE, n = 33) (Fig. 6).

The CF was also changed by the interaction of sex × season (GML, n = 486, F_2,3 = 4.18, p = 0.04) (Fig. 7), but sex × site (n = 486, F_2,3 = 0.91, p = 0.44) (Fig. 8) and season × site (n = 486, F_2,3 = 0.31, p = 0.82) (Fig. 9). In overall, CF value of this mudskipper (1.05 ± 0.02, n = 495) was significantly higher than an ideal threshold value of 1 (t = 2.21, n = 486, df = 485, p = 0.03, CI 95% = [0.05–0.01]).