Population structure and reproduction of Steindachneridion melanodermatum (Siluriformes: Pimelodidae), a large catfish endemic to Neotropical ecoregion

Keywords: Iguaçu River, reproductive biology, spawning period, surubim of Iguaçu.

Studies on fish reproduction generate valuable information that can be used to maintain populations in the natural environment (Cantanhêde et al. 2016), in addition to guaranteeing the survival and perpetuation of the species. From an ecological perspective, knowledge of the fish reproductive biology (reproduction, fertilisation and spawning) aids in the development of effective strategies for the conservation and management of species, especially for anthropized environments (Freitas and Montag 2019), and even technologies for fish farming.

The understanding of a species population structure, such as the size of the first gonadal maturation, may assist in the implementation of species-specific management and conservation plans (Wikelski and Cooke 2006; Rizzo and Bazzoli 2014; Pompei et al. 2017). The establishment of the closing period for fishing, or even permanent prohibition, is also a management measure based on the reproduction period and population aspects of a specific species. This type of approach is particularly necessary for endemic (Myers et al. 2000) and endangered species (Olden et al. 2006).

The extinction risk of endemic species is high because of their vulnerability to environmental change (Olden et al. 2008). Many stressed rivers around the world show evidence of endemic extinction of fish as a result of anthropogenic action (Brauer and Beheregaray 2020), such as the Iguaçu River in South America (Agostinho and Gomes 1997), which is being affected by the presence of successive hydroelectric dams along its course (Agostinho et al. 2007). The Iguaçu River is one of the largest tributaries of the Paraná River in southern Brazil and northern Argentina, and it is considered a distinct ecoregion (Abell et al. 2008). This is due to separation of the catchments by tectonic movements (Abell et al. 2008), which has isolated the Iguaçu River from Paraná River by the Great Iguaçu Falls (Brehm et al. 2016), and consequently, formed an area of high fish endemism (Garavello et al. 1997; Zawadzki et al. 1999; Abell et al. 2008). Nevertheless, the cumulative anthropogenic effects have contributed to the decline of fish endemism (decreased by 5.3% – species loss) in the Iguaçu River basin within a 13-year period, according to the records of Zawadzki et al. (1999) and Baumgartner et al. (2012).

The ichthyofauna of the Iguaçu River basin is predominantly characterised by small species (standard length <200 mm; Baumgartner et al. 2006, 2012). The catfish Steindachneridion melanodermatum Garavello 2005 is the only large species of the family Pimelodidae, with standard length >400 mm, found in this basin (Baumgartner et al. 2012). The genus Steindachneridion comprises six endemic species that are all restricted to South American waterways between southern and south-eastern Brazil (Lundberg and Littmann 2003; Garavello 2005). Two common characteristics shared among all species are their regional endemism and conservation status (endangered: S. melanodermatum, S. parahybae, S. punctatum and S. scriptum; and critically endangered: S. amblyurum, and S. doceanum; Rosa and Lima 2008; ICMBIO 2018).

Steindachneridion melanodermatum is known as surubim of Iguaçu or monjolo. It is possibly a migratory species (Agostinho and Gomes 1997; Ludwig et al. 2005; Brehm et al. 2016) that inhabits fast-flowing waters, rocky bottoms, and deep waters in stretches of rivers wherein natural flow is preserved (Garavello 2005). Currently, the population of this species in the wild is restricted to the Lower Iguaçu River, from downstream of the Salto Caxias Hydroelectric Dam to upstream of the Iguaçu Falls, on a river stretch ~190 km long (Assumpção et al. 2017), the last refuge of this endemic and endangered species. Habitat alteration and extirpation, as well as illegal and predatory fishing, pose imminent threats to this species (Brehm et al. 2016; Assumpção et al. 2017). Knowledge on species of the genus Steindachneridion is currently scarce, not only because of the reduced number of specimens in wild populations (Garavello 2005), but also due to difficulties in field sampling and the limited number of specimens that are typically captured relative to the necessity for biological studies (Garavello et al. 1997; Garavello 2005).

Rare and endemic freshwater fish species face many anthropogenic threats worldwide (Sutherland et al. 2009; Cooke et al. 2012; Teimori et al. 2016). Therefore, efforts should be focussed on basic natural history research (Cooke et al. 2012), together with the creation of conservation strategies for endemic species, especially for those with restricted distribution and a higher level of threat (Nogueira et al. 2010). Thus, considering that the lack of knowledge on the biological aspects of the S. melanodermatum population in the wild has hindered the planning and implementation of specific conservation and management strategies for this species, the present study characterised the population structure and reproductive biology of S. melanodermatum, including length structure, sex ratio, growth type, reproductive period and sexual maturity, as well as the relationship between environmental variables and reproduction. Information on biology and reproduction of S. melanodermatum is essential for understanding its lifecycle, and will assist in the establishment of conservation measures and criteria for this endangered species.

The Iguaçu River basin covers an area of 72 000 km² and borders Brazil (Paraná state, 79%; and Santa Catarina state, 19%) and Argentina (2%; Eletrosul 1978; Júlio-Júnior et al. 1997). The Iguaçu River is one of the largest tributaries of the Paraná River in southern Brazil and northern Argentina, runs 1200 km from its source to its mouth in the Paraná River (Maack 1981) and is subdivided into the following three hydrographic units according to the geomorphological characteristics: upper (1st plateau), medium (2nd plateau), and lower (3rd plateau; Baumgartner et al. 2012).

The Lower Iguaçu region belongs to the third plateau and is marked by the presence of numerous rapids and waterfalls such as Salto Santiago (40 m), Salto Osorio (30 m), Salto Caxias (67 m) and the Iguazu Falls (72 m; Maack 1981; Júlio-Júnior et al. 1997). These high-level differences allowed the formation of five cascade hydroelectric reservoirs in the main channel of the Iguaçu River (Agostinho and Gomes 1997; Agostinho et al. 2007). The sixth hydroelectric plant, the Baixo Iguaçu Hydroelectric Power Plant, is located between the cities of Capitão Leônidas Marques and Capanema, Paraná state, downstream of the Governador José Richa Hydroelectric Power Plant (known as Salto Caxias Hydroelectric Power Plant). The Baixo Iguaçu Dam is ~800 m upstream of the Gonçalves Dias River mouth (Fig. 1), which is bordered by the Iguaçu National Park Conservation Unit, PARNA of Iguaçu.

Fig. 1.  Steindachneridion melanodermatum sampling sites in the stretch of the Lower Iguaçu River from downstream of the Salto Caxias Hydroelectric Power Plant to Iguaçu Falls, in the State of Paraná, Brazil. Sampling sites P1, P6, and P7 were located in the main channel of the Iguaçu River, whereas P2 and P3 were located in Capanema River, P4 in Gonçalves Dias River and P5 in Santo Antônio River. The gray area corresponds to the Iguaçu National Park. Red dots indicate sampling sites where no specimens of S. melanodermatum were captured. Numbered sites indicate locations where specimens were captured during the study period.

The study area comprised a stretch of the Iguaçu River of ~190 km, extending from downstream of the Salto Caxias Dam to the Iguaçu Falls (25°32′52.61′′S, 53°31′29.96′′W to 25°35′51.85′′S, 54°23′29.89′′W; Fig. 1). In this stretch, the Iguaçu River is embedded with a rocky bed consisting of basaltic rock originated by the Mesozoic magnetism and presents remnants of the Atlantic Forest biome.

Specimens of S. melanodermatum evaluated were sampled (70 samplings) during January 2010 to January 2011 (13 samplings, 25 sites) and September 2012 to December 2016 (57 samplings, 27 sites). The main purpose of these samples was to assess and monitor the fish assemblages in the region. All sampling sites are shown in Fig. 1, and Sites P1– P7 correspond to those where the specimens were caught during the study period.

Sampling sites P1, P2, P3 and P5 were located outside the limits of the Iguaçu National Park Conservation Unit (PARNA of Iguaçu), whereas Sites P4, P6 and P7 were located within the PARNA of Iguaçu. Sampling sites P1, P6, and P7 were located in the main channel of the Iguaçu River, whereas Sites P2 and P3 were located in Capanema River, Site P4 in Gonçalves Dias River, and Site P5 in Santo Antônio River, tributaries of Iguaçu River. The Baixo Iguaçu Hydroelectric Power Plant (Fig. 1), located ~0.7 km from the upper limit of the PARNA of Iguaçu, was under construction when the study was conducted.

Several fishing gears were used to sample fish: gillnets (with 2.5–18 cm mesh size between opposing nodes), three layers of fishing nets (with 6, 7, and 8 cm mesh size between opposing nodes), and longlines (composed of 20 hooks, size 9/0) using fish and bovine heart pieces as bait. The gill-nets and longlines were set on the right bank and installed at 1500 hours and monitored at 6-h intervals, i.e. at 2100 hours, 0300 hours and at 0900 hours in the morning of the next day. The fishing effort was standardised at all sampling sites throughout the study period, by keeping the time and amount of fishing gear constant.

Specifically, at Sites P6 and P7, located in Iguaçu National Park, the authorisation to sacrifice captured specimens (limited to a few specimens) was granted only from April 2013 by the environmental agency, Instituto Chico Mendes de Conservação da Biodiversidade (ICMBio), because of the species’ endangered status (Assumpção et al. 2017), so as to avoid removing a large number of specimens from the environment. Thus, before that, several captured specimens that remained alive after being measured and weighed were released at the capture sites according to ICMBio authorisation. Other specimens used for analyses in the laboratory included those that died during the capture process and alive specimens sampled after April 2013 that were killed with benzocaine (250 mg L^–1).

The procedures applied in this study were approved by the Ethics Committee on Animal Use of the Universidade Estadual do Oeste do Paraná (UNIOESTE), according to protocol number 62/09, and certificate protocol on 10 December 2013. Collection licences were granted by the ICMBio and authorised by the Sistema de Autorização e Informação em Biodiversidade (SISBIO; nos 25 648-3 and 25 648-4), Instituto Ambiental do Paraná (IAP; nos 37 788 and 43 394), and ICMBio (no. 003/2014 and SEI Letter no. 63/2016-DIBIO/ICMBio). Voucher specimens were deposited at the Museu de Zoologia of the Universidade Estadual de Londrina (MZUEL; voucher numbers: 17 091, 17 092, 17 093, 15 702, 15 703, 15 704, 15 705).

The morphometric data for all specimens, including total length (TL, in cm) and bodyweight (BW, in g) were recorded. Subsequently, after ventral section, the sex and gonadal development of several specimens were determined macroscopically, according to Vazzoler (1996) and Brown-Peterson et al. (2011), and some gonads were removed, weighed and stored for histological processing.

Microscopic analysis of the gonads was performed to evaluate spawning and gonadal development, as well as to confirm and correct the maturation stage of the ovaries and testes that were initially macroscopically determined. In total, 62 gonads (30 ovaries and 32 testes) were fixed for 24 h in 10% formalin neutralised with calcium carbonate and, subsequently, processed using routine histological techniques. The material was embedded in paraffin, cut in 7-μm-thick slices, and stained with hematoxylin–eosin (H&E; Tolosa et al. 2003). The ovarian and testicular maturity scale was elaborated on the basis of the descriptions given by Vazzoler (1996), Lowerre-Barbieri et al. (2011) and Brown-Peterson et al. (2011) as the following stages: immature (A), development (B), capable of spawning (C) and post-spawning (D). The spawned and resting stages for females, and the spermiate and resting stages for males were grouped in the post-spawning stage, because of the reduced number of specimens in these stages.

Abiotic data were collected at the same time as fish sampling was undertaken. The water temperature (°C), dissolved oxygen concentration (mg L^–1), pH and electrical conductivity (μS cm^–1) were obtained using multiparameter (YSI Professional Plus, Yellow Springs OH, USA). The turbidity (NTU) was sampled with turbidimeter (Policontrol AP 2000 IR, Diadema SP, Brazil) and water transparency (m) using Secchi disk. Additionally, daily precipitation and discharge data were provided by the Technological Institute of Paraná (SIMEPAR).

The structure of the S. melanodermatum population was characterised in length, sex ratio and growth rate of females and males.

The length structure was determined by the fish frequency distribution among the TL classes. Eight length classes were defined at 10.1-cm intervals according to Sturge’s postulate (Vieira 2003). The Kruskal–Wallis test was used to evaluate whether fish lengths and weights differed between the sexes. The sex ratio for each length class and month was determined on the basis of the absolute frequency of males and females. The chi-square test (χ²; significant at P < 0.05 if χ² > 3.84) was used to determine whether the sex ratio significantly differed from 1 : 1 (Vazzoler 1996). The Kolmogorov–Smirnov test was applied to evaluate possible differences in the distribution of individuals within size classes (α = 0.05).

The characterisation of the population structure in length and weight employed all collected specimens, considering that some of them were returned to the environment after data collection of these variables. The type of growth exhibited by females and males was determined on the basis of the weight–length ratio, using the following equation: BW = a × TL^b, where a is the linear growth coefficient, and b the angular growth coefficient of the potential model (Le Cren 1951; Braga et al. 2007).

The maturation curves of females and males and the frequency of distribution of gonadal developmental stages (based on macroscopic and microscopic analysis) in the population were used to characterise the reproductive cycle and determine the reproductive period of the species. Specimens that were captured alive and released at capture sites (63 fish) were not included in the cycle and reproductive-period analyses, as well as four other fish that died and were not dissected (they were intended for scientific collections at the UNIOESTE, Iguaçu National Park and MZUEL).

The maturation curve was constructed from the monthly mean values of gonadosomatic index (GSI), which were initially calculated individually, as expressed by the following formula: GSI = (GW / BW)  × 100, where GW and BW correspond to the gonad weight and total bodyweight of the specimen respectively.

The length at first maturation (L₅₀) was determined from the curve plotting the length-class midpoints versus the relative frequency of adult fish in each class, using the expression , where Fr is the relative frequency of adult, e is the Euler number, and a is a linear coefficient and b is an angular coefficient estimated by the least-squares method after transforming the variables involved according to the procedures described in Fávaro et al. (2003) and Oliveira and Favaro (2011). Fish with immature gonads were considered juveniles, while those in all other stages of gonadal development were considered adults.

To summarise the dimensionality of the set of abiotic variables, principal component analysis (PCA) was applied in the abiotic data-matrix values (log-transformed, except the pH) measured monthly (seasonality). The retention and interpretation of the axes generated by the PCA followed the Kaiser–Guttman criterion (Jackson 1993). According to this empirical rule, only components having eigenvalues >1.0 (λi > 1) summarise shared variation and should be retained. Abiotic variables with eigenvectors (correlations) >0.30 were considered biologically important (Hair et al. 2009).

Relationships between abiotic variables (axes retained of PCA) and the reproduction (based on mean GSI values) of S. melanodermatum were investigated for females and males by using the Spearman correlation. The PCA axes retained for interpretation and the monthly GSI were tested by the Kruskal–Wallis test to verify the effect of seasonality on the reproductive activity of S. melanodermatum.

In this study, seasonality was determined on the basis of the collection date as follows: autumn was defined as the months between March and May, winter between June and August, spring between September and November, and summer between December and February.

In total, 182 S. melanodermatum specimens were analysed, of which 51 were females, 64 were males, 63 were fish that remained alive after capture and were released at the capture sites, and four were fish that died and were intended for scientific collections.

The analysis considering both sexes together showed that their TL and bodyweight varied between 21 and 102 cm and 82.8 and 15670 g respectively. Separate analyses of females and males showed that females exhibited higher TL and weight values compared with males (Table 1). However, the Kruskal–Wallis test did not show significant differences in length and weight between sexes.

Table 1.  Length and weight structure of the Steindachneridion melanodermatum population captured in the Lower Iguaçu River
Total length (TL) and bodyweight (BW), maximum, minimum, mean and standard deviation (s.d.) for each analysed category (both sexes combined, females and males) are shown. The analysis of the grouped genders included the specimens that were released back into the environment after collection

Analyses of females and males divided by length classes showed that the intermediate length classes occurred more frequently in both sexes. The slight predominance of males over females was observed in practically all length classes, with the exception of the 77-cm intermediate length class; nevertheless, there was no significant difference in the proportion of each length class comprising each sex (χ² < 3.84, P > 0.05). The smallest and largest length classes contained only females (Fig. 2).

Fig. 2.  Distribution of the absolute frequency of occurrence per length class of Steindachneridion melanodermatum females and males. The numbers above the bars indicate the χ²-value for the comparison of the frequency of females versus males for each length class (significant if χ² > 3.84 at P < 0.05).

The analysis of the monthly sex ratios showed no significant differences between the sexes, except in the month of April, when females predominated significantly. Moreover, the visual examination of the plotted data showed that the highest absolute occurrences of fish of both sexes were in the months of June, July and August (Fig. 3).

Fig. 3.  Distribution of the absolute monthly frequency of Steindachneridion melanodermatum female and male occurrences in the Lower Iguaçu River. The numbers above the bars indicate the χ²-value of the comparison between the frequency of females versus males for each month. Asterisk indicates a significant difference at the 5% level (χ² > 3.84, P < 0.05).

The determination of the type of growth using weight–length ratio data showed a positive allometric-type growth (b > 3) for all males and females, which indicates greater weight gain than length increase as the fish grow (Fig. 4).

Fig. 4.  Weight–length ratios for Steindachneridion melanodermatum (a) females and (b) males captured in the Lower Iguaçu River. Dashed lines indicate the curves fitted to the data for each sex on the basis of its regression coefficient values. BW, bodyweight; and TL, total length.

The microscopic analysis of ovary and testis histology allowed us to establish a microscopic maturity scale for both sexes of S. melanodermatum. Four stages of gonadal development were described in both females (Table 2, Fig. 5) and males (Table 2, Fig. 6).

Table 2.  Steindachneridion melanodermatum microscopic ovarian and testicular maturity scale
Adapted from Vazzoler (1996), Brown-Peterson et al. (2011) and Lowerre-Barbieri et al. (2011)

Fig. 5.  Stages of ovarian development in Steindachneridion melanodermatum: (a) Stage A, immature, predominance of oocytes in primary growth phase (PG); (b) Stage B, developing, PG, cortical alveolus (CA), and primary vitellogenesis (Vtg1) phases; (c) Stage C, capable of spawning, predominance of developed oocytes, tertiary vitellogenesis (Vtg3) phases; (d) Stage D, capable of spawning, predominance of developed oocytes, tertiary vitellogenesis (Vtg3) phases and post-ovulatory follicles (POF); (e) Stage E, post-spawning, PG and atretic follicles. Hematoxylin–eosin (HandE) staining. Scale bars: 100 μm.

Fig. 6.  Stages of testicular development in Steindachneridion melanodermatum: (a) Stage A, immature, partially closed seminiferous tubules with a predominance of spermatogonia (SG), seminiferous tubes still closed, consisting of spermatogonia and delimited by the dashed line; (b) Stage B, developing, seminiferous tubules exhibit different sperm cells, SG, primary spermatocytes (SC1), secondary spermatocytes (SC2), spermatids (ST) and spermatozoa (SZ); (c) Stage C, capable of spermiating, seminiferous tubules full of spermatozoa; (d) Stage D, post-spawning, seminiferous tubules present spaces (arrows) and reduced amounts of spermatozoa. Hematoxylin–eosin (HandE) staining. Scale bars: 50 μm.

Considering the GSI values in the maturation curve (Fig. 7), the highest values occurred from June to September, and may have extended until November, because in October and November there was no capture of females. In addition, a peak of GSI can be observed in August for males and in September for females.

Fig. 7.  Maturity curves of Steindachneridion melanodermatum females and males captured in the Lower Iguaçu River, Paraná, Brazil.

The reproductive season observed by the maturation curve was corroborated by the frequency distribution of the gonadal development stages in females and males, with a higher frequency of those capable of spawning (ovaries) or spermiating (testes) over this same period (Fig. 8). Between October and February (corresponding to lower mean GSI values), there was an increase in the frequency of spawned ovaries and depleted testes. The occurrence of mature individuals was recorded in winter and spring, whereas post-reproductive individuals were observed more frequently in summer, and immature ones occurred in autumn.

Fig. 8.  Distribution of the percentage monthly frequency of different gonadal development stages in Steindachneridion melanodermatum (a) females and (b) males sampled in the Lower Iguaçu River, Paraná, Brazil. Stages included immature (A), developing (B), capable of spawning or capable of spermiating (C) and post-spawning (D). Numbers inner the bars represent the number of individuals examined.

The length at first maturation (L₅₀) determined for S. melanodermatum was 39.5 cm for females and 43.9 cm for males (Fig. 9a, b), which represented 39.9% and 48.0% respectively, of the maximum TL observed in this study. The total length at which all adult individuals were mature (L₁₀₀) and, thus, capable of spawning was 87.0 cm for females and 70.0 cm for males (Fig. 9a, b).

Fig. 9.  Length at first maturation (L₅₀) for Steindachneridion melanodermatum (a) females and (b) males sampled in the Lower Iguaçu River, Paraná, Brazil.

Most of the analysed specimens were adults (95%), mainly sampled in the main channel of the Iguaçu River, at Sites P6 and P7 near the Iguaçu Falls. Juveniles accounted for 5%; the smallest ones occurred in two tributaries (Site P4, Gonçalves Dias; and Site P5, Santo Antônio) and the largest ones (almost adults) were registered in the Iguaçu River main channel, at Sites P6 and P7 (Table 3).

Table 3.  Number of Steindachneridion melanodermatum captured at sampled sites in the Lower Iguaçu River, frequency of occurrence (%) and total length of juveniles and adults, and number of specimens capable of spawning (C) and post-spawning (D)

Considering the macroscopic and microscopic analyses of female and male gonads, most specimens were in the capable of spawning (Stage C; 36.5%) and post-spawning (Stage D; 33.0%) stages, and they were mainly recorded in the main channel of the Iguaçu River, at Sites P6 and P7 (Table 3).

Mean GSI values were higher for both sexes in months with lower temperatures (winter and early spring), showing that the reproduction was more intense during this period of lower temperatures (14.78–22.60°C), high discharge (10 77.3–11 464.2 m³ s^–1) and high dissolved oxygen concentrations (5.94–9.75 mg L^–1; Table 4).

Table 4.  Mean (M), minimum (Min) and maximum (Max) monthly values for the abiotic variables measured during the sampling period in the Iguaçu River, Paraná, Brazil
DI, discharge (m³ s^–1); DO, dissolved oxygen concentration (mg L^–1); EC, electrical conductivity (μS cm^–1); pH, power of the hydrogen ion concentration; PR, precipitation (mm) ; TR, transparency (m); TU, water turbidity (NTU); WT, water temperature (°C)

The first two axes of the PCA were retained for the interpretation of the results. The PC1 and PC2 axes explained 52.24% of the Lower Iguaçu River abiotic gradient (32.70% and 19.54% respectively; Table 5, Fig. 10). The variables that positively influenced PC1 were water temperature and transparency, whereas dissolved oxygen concentration and discharge were negatively correlated with this axis. The PC1 evidenced the seasonality of the abiotic gradient, because it spatially separated the months (autumn and summer) with high temperatures and transparencies from the months with higher concentration of dissolved oxygen and discharge (winter and spring). Transparency was positively correlated with PC2, whereas precipitation, pH, turbidity and discharge variables were negatively correlated. The PC2 was influenced by monthly singularities, reporting occasional variations of the monthly abiotic gradient.

Table 5.  Results of the principal component analysis (PCA) for the axes retained for interpretation (PC1 and PC2)
Eigenvectors for each abiotic variable (values greater than 0.30 are in bold, Hair et al. 2009), eigenvalues and percentages of explanation are given

Fig. 10.  Principal component analysis (PCA) applied to the abiotic variables measured during the collection in the Lower Iguaçu River. DI, discharge; DO, dissolved oxygen concentration; EC, electrical conductivity; pH, hydrogen potential; PR, precipitation; TR, transparency; TU, turbidity; WT, water temperature. Aut, autumn; Win, winter; Spr, spring; Sum, summer.

The Spearman correlation showed that the PC1 was negatively correlated with female GSI (r_sp = –0.60; P < 0.05), which may indicate seasonal influence on reproductive activity. However, male GSI was positively correlated (r_sp = 0.83; P < 0.05) with female GSI, indicating that males follow the same trend of female reproductive activity. These results suggest that S. melanodermatum has a higher GSI in months with low water temperature and transparency.

Despite the high GSI averages in the winter months (June, July and August) for both sexes, the Kruskal–Wallis test values were not significant.

Our study is the first to investigate populational structure and reproductive biology of wild populations of the surubim of Iguaçu, S. melanodermatum, an endemic species in the Neotropical region. These biological aspects are necessary for the development of effective measures and specific criteria for the conservation and protection of this endangered species. So, like some large tropical catfish around the world, S. melanodermatum is rare and difficult to sample because it occurs in deep pools (Assumpção et al. 2017), and the knowledge gap so far has hindered the planning of conservation strategies.

In our study, the length structure of the wild S. melanodermatum population from the Lower Iguaçu River showed that this species’ maximum sizes and weights are higher than those recorded by Feiden et al. (2006) in captivity. Additionally, the absence of significant differences in length between S. melanodermatum males and females indicated no sexual dimorphism in size. This is an uncommon pattern in fishes, where studies generally report evidence that females have larger body sizes, which favour increased fertility and fecundity (Shine 1990).

The positive allometric growth (b > 3) observed in this study for S. melanodermatum is similar to the large catfish Pseudoplatystoma corruscans (Mateus and Penha 2007); however, it differs from that of the congeneric species S. scriptum, which exhibits negative allometric growth (Agostinho et al. 2008). Differences in the type of growth among species reflect their ways of using and allocating energetic resources to achieve better performance and condition in diverse environments, which exert different types and intensities of selection pressures (Silva et al. 2010). However, the population of the species studied exhibited a sex ratio of 1 : 1, as is usually expected in the natural environment (Nikolsky 1969).

The use of the histological techniques applied to ovaries and testes was fundamental for establishing a microscopic maturity scale for the gametes and gonads of S. melanodermatum females and males. It enabled the gonadal development dynamics and reproductive season of this species to be better understood and characterised (Manorama and Ramanujam 2017). The development of ovarian follicles followed the same development patterns as described for other teleosts. However, even though oocyte growth patterns are similar among teleosts, reproductive strategies and tactics vary considerably among species (Jobling 1995; Tyler and Sumpter 1996; Brooks et al. 1997).

Post-ovulatory follicles characterised histologically for S. melanodermatum females may be indicative of spawning (Arantes et al. 2010; Viana et al. 2018) of this species in the study area. Perini et al. (2013) mentioned the reproductive success of a migratory fish species in a three-river system from the Paraná River basin, associating the spawning site to some reproduction parameters such as a higher percentage of spawned females and a lower follicular atresia index.

Monthly variations in the GSI and the proportion of gonadal developmental stages are used to determine the reproductive season of a species for an ideal management measure (Mian et al. 2017). The highest GSI values and higher percentages of specimens (males and females) in the stage capable of spawning indicate that S. melanodermatum spawn mainly from June to September (winter and early spring); however, spawning may extend until November. Post-reproductive individuals were observed more frequently in the summer, whereas the small number of immature individuals occurred in the autumn, suggesting it as the recruitment period of this species. However, the occurrence in January of males in the stage capable of spawning and females in the stage of developing may indicate late-spawning specimens.

In contrast, different from specimens in the wild, a recent study on reproductive indicators in males of S. melanodermatum born and maintained in captivity (Tessaro et al. 2019) observed reproductive period from September to November (spring). Probably, environmental variables influence the gonadal development and reproductive activity of S. melanodermatum. In our study, the reproductive activity was more intense in months with lower temperatures and higher discharges, namely, winter to early spring. These results are similar to those previously obtained for S. scriptum (Sartori 2003; Reynalte-Tataje and Zaniboni-Filho 2008; Weingartner et al. 2012), namely, low temperatures in the early spring, and S. amblyurum (Vono and Garavello 2008), namely, high discharges. However, the reproductive period for S. melanodermatum differed from the temporal pattern previously observed for most fish species in the Paraná River basin (Vazzoler 1996), i.e. spring and summer, including the congeneric species S. parahybae (Caneppele et al. 2009; Honji et al. 2017), other large migratory catfish (Agostinho et al. 2003; Campos 2010), and other endemic pimelodid species of the Iguaçu River, such as Pimelodus britskii and P. ortmanni (Baumgartner et al. 2012). It is also noteworthy that the reproduction and recruitment period for fish species in temperate and subtropical climates is usually determined by the increase in temperature associated with the increased food availability that occurs during the spring and summer (Jobling 1995; Lowe-McConnell 1999). Thus, the reproductive period of the analysed species must be associated with the specific needs of its offspring and can be considered a tactic to achieve reproductive success.

Considering the different analyses applied in this study to evaluate reproductive biology of S. melanodermatum, the spawning is possibly total. Similarly, total spawning had been suggested during the induced reproduction of this species performed in the Experimental Station of Ichthyology Studies of the Ney Braga Hydroelectric Power Plant (Ludwig et al. 2005), and was also indicated by Tessaro et al. (2019). Furthermore, evidence of the same type of spawning was found for S. scriptum, a long-distance migratory species (Suzuki et al. 2009).

Tropical migratory fishes inhabit lotic environments (Sato et al. 2003) and generally exhibit total spawning (Vazzoler 1996); however, no studies have been conducted on the migratory behaviour of S. melanodermatum. Reproductive aggregation is expected in migratory species and reflects the occurrence of higher concentrations of individuals during the reproductive period (Godoy 1975). The highest abundance of S. melanodermatum specimens occurred in the months that preceded reproduction, corroborating what was previously observed for other pimelodid species (Brito and Bazzoli 2003; Holzbach et al. 2009).

The determination of the length at first sexual maturation (L₅₀) is important in the understanding a fish’s life history, because it can be used to establish fishing resource-conservation criteria (Barbieri and Hartz 1995) and plan the management of exploitation measures. Moreover, the minimum size at maturation is a reproductive parameter that depends on water temperature and varies among species, as well as whether the species are found in the natural environment or in captivity (Crepaldi et al. 2006) and may be related to the fishing activities that are performed. In many cases, the lack of this knowledge may lead to decreases in the size of the reproductive stock (Peixer et al. 2006; Mateus and Penha 2007). The L₅₀ value determined for Steindachneridion melanodermatum was smaller than that for S. scriptum (Agostinho et al. 2008). Also, females of S. melanodermatum reached maturity in smaller lengths than did males, that is, females participate in reproduction earlier. This information can assist in the management of fisheries resources at the species level by implementing minimum catch sizes (through the L₅₀) or establishing fishery closure period (through the reproductive period).

The correct identification of the gonadal development stages and the determination of the proportion of individuals reproducing constitute a few of the fundamental parameters needed to understand how a species uses its environment (Veregue and Orsi 2003). The reproductive tactics used by a species depend on its interactions with the environment, which are associated with the genetic, physiological, behavioural and ecological responses of individuals (Potts and Wootton 1984). Tactical variations determine the reproductive strategy used by the species to achieve reproductive success (Vazzoler 1996). It is evident that S. melanodermatum exhibits several reproductive tactics, including total spawning, reproductive activity more intense in the winter and early spring, and small size at first maturation, mostly compared with other large pimelodids (Agostinho et al. 2003). Moreover, the fact that this species reproduces in a different period in relation to other species in the Iguaçu River may allow its post-larvae to obtain appropriate food in the mouth-opening phase, perhaps by feeding on other species’ eggs and larvae.

Steindachneridion melanodermatum inhabits rapid waters in stretches of the river where the natural flow is still preserved, and high-depth environments (Garavello 2005), called deep pools, constitute the preferred habitat of this species (Assumpção et al. 2017). Our study observed specimens capable of spawning and post-spawning especially in the main channel of the Iguaçu River, mainly at P6 and P7 sites, a region called Poço Preto, which is a deep pool considered a sanctuary for this species (Assumpção et al. 2017). Additionally, juveniles also occurred at these sites and in the tributaries such as Capanema, Gonçalves Dias, Floriano and Santo Antonio rivers. The occurrence of juveniles at the mouth of the Nandu River, a tributary of the Iguaçu River on the Argentina riverbank near the Iguaçu Falls, was also mentioned by Casciotta et al. (2016). Studies have highlighted the relevance of dam-free tributaries for the conservation of fish diversity, especially of large migrants, in neotropical reservoirs (Perini et al. 2013; da Silva et al. 2015, Silva et al. 2019; Marques et al. 2018; Makrakis et al. 2019), and a remnant of the floodplain in the upper Paraná River (Silva et al. 2017). The tributaries support populations of endangered and threatened species and provide a variety of environmental conditions, access to spawning habitat, refugia for early life stages (Silva et al. 2019). For this reason, it is vital to maintain dam-free tributaries to safeguard this species. Tributaries of the Iguaçu River, such as Capanema, Gonçalves Dias, Floriano and Santo Antonio rivers, may play an important role for the development and growth of S. melanodermatum, and the habitat and spawning site of this species still needs to be investigated through the occurrence of eggs and early larvae.

One attribute that indicates whether a species is favoured and adapted to abiotic conditions is population abundance. In this sense, preservation of the Sites P6 and P7 located in the Iguaçu Conservation Unit promotes the high abundance of adults. Notably, the capture outside the limits of the Iguaçu National Park was low and may reflect the poor status of conservation of the watersheds, as shown by Celestino et al. (2019). The ability for parks to serve as protected areas depends on activities upstream from their boundaries (Lawrence et al. 2011). The maintenance of existing protected areas and the creation of new ones are urgent as a strategy for the conservation of freshwater biodiversity in Brazil, which is rapidly being degraded and could be lost if not properly protected (Azevedo-Santos et al. 2019). The Iguaçu National Park has great potential as a protected area for fish diversity and adequately preserving the tributaries and the free-flowing Iguaçu River stretches will contribute for the conservation of this large endemic pimelodid.

In conclusion, our study showed that the reproduction of wild S. melanodermatum population was more intense in the winter and early spring, a period associated with low temperatures, high dissolved oxygen concentrations and discharges in the Lower Iguaçu River. This study has provided hitherto unpublished valuable information on the biology and reproduction of wild populations of S. melanodermatum to assist in the planning, development and evaluation of conservation measures. Furthermore, crucial and urgent actions are mandatory to safeguard this endemic and endangered species from extinction such as the habitat preservation (deep pools) and intensive supervision to avoid illegal fishing of S. melanodermatum along the river stretches where the species still occurs.

The authors declare that they have no conflicts of interest.

We thank the Consórcio Empreendedor Baixo Iguaçu and the Macuco Safari for financial support. Coordination for the Improvement of Higher Education Personnel (CAPES) provided a doctoral scholarship to L. de Assumpção, and National Council for Scientific and Technological Development (CNPq) provided a Productivity Grant in Technological Development and Innovative Extension (DT) to S. Makrakis.