Convergent responses of fish belonging to different feeding guilds to sewage pollution

Marina Rodrigues do Prado1, Débora Reis de Carvalho1, Carlos Bernardo Mascarenhas Alves2, Marcelo Zacharias Moreira3 and Paulo Santos Pompeu1

PDF: EN   XML: EN  | Supplementary: S1 S2 S3 S4 | Cite this article



This study aimed to evaluate if the presence of pollutants promotes changes in feeding habits of fish species from different trophic guilds: the detritivorous species, Hypostomus francisci, and the piscivorous, Hoplias intermedius. Both species were sampled at 12 sites (with different degrees of pollution) in the Rio das Velhas basin, which is heavily polluted by domestic and industrial sewage from the Metropolitan Region of Belo Horizonte (MRBH). Stable isotope analyses of carbon (δ13C) and nitrogen (δ15N) of fish tissue and the main food resources were performed. Fishes from both trophic guilds altered their diets in degraded environments, but the detritivorous species showed greater trophic plasticity. The isotopic niche of both trophic guilds was broadest in unpolluted sites and more δ15N enriched in polluted regions. The detritivorous species presented high niche-breadth in unpolluted sites, probably due to the greater variety of resources consumed. In addition, the δ15N of the detritivorous was more enriched than the piscivorous species in polluted sites. In conclusion, fishes from both trophic guilds presented similar isotopic responses to environmental pollution. However, the detritivorous species was more sensitive to these alterations and therefore, is likely a better indicator of environmental condition than the piscivorous.

Keywords: Detritivorous; Hoplias intermedius; Hypostomus francisci; Piscivorous; Stable isotopes


Este estudo teve como objetivo avaliar se a presença de poluentes promove mudanças nos hábitos alimentares de espécies de peixes de diferentes guildas tróficas: a espécie detritívora, Hypostomus francisci, e a piscívora, Hoplias intermedius. Ambas espécies foram amostradas em 12 locais (com diferentes níveis de poluição) na bacia do Rio das Velhas, que é altamente poluída por esgoto doméstico e industrial da Região Metropolitana de Belo Horizonte (RMBH). Foram realizadas análises de isótopos estáveis de carbono (δ13C) e nitrogênio (δ15N) dos tecidos dos peixes e dos principais recursos alimentares. Espécies de ambas guildas tróficas alteraram suas dietas em ambientes degradados, mas a espécie detritívora apresentou maior plasticidade trófica. O nicho isotópico de ambas as espécies foi mais amplo em locais menos perturbados e mais enriquecido em δ15N em regiões poluídas. A espécie detritívora apresentou grande amplitude em seu nicho isotópico em locais menos perturbados, provavelmente devido à maior variedade de recursos consumidos. Além disso, o δ15N da espécie detritívora foi mais enriquecido que a espécie piscívora em locais poluídos. Em conclusão, ambas as espécies apresentaram respostas isotópicas semelhantes à poluição ambiental. No entanto, a espécie detritívora foi mais sensível a essas alterações e, portanto, é provavelmente uma melhor indicadora de condição ambiental do que a espécie piscívora.

Palavras-chave: Detritívoros; Hoplias intermedius; Hypostomus francisci; Isótopos estáveis; Piscívoros


Fish are good indicators of environmental quality and are used to assess the integrity of aquatic environments (Karr, 1981; Pompeu et al., 2005; de Carvalho et al., 2017b). The selection of these organisms for environmental biomonitoring occurs because changes in habitat can be reflected in multiple dimensions among ichthyofauna, for example, by changes in feeding habit (Carvalho et al., 2015; de Carvalho et al., 2017a), reproduction (Schulz, Martins-Junior, 2001), and community composition (Fausch et al., 1990; Cunico et al., 2006). In addition, fishes are represented at several trophic levels in aquatic food webs (Freitas, Siqueira-Souza, 2009), and many species have long-life cycles (Karr, 1981; Smith et al., 1997). Therefore, assessment of ichthyofauna to understand the effects of anthropogenic impacts on aquatic ecosystems is very important.

The discharge of domestic and industrial sewage without proper treatment into the rivers is considered as one of the main anthropogenic impacts in aquatic systems (Camargo et al., 1995; Soares et al., 2016). One of the main consequences of sewage pollution is the increased nutrient load in aquatic ecosystems, which can lead to a rapid decrease in oxygen levels through eutrophication and have catastrophic impacts on fish diversity (Smith, 2003; Silva, Fonseca, 2016).

More subtle impacts, such as changes in the food webs of aquatic ecosystems, can also be observed in polluted environments (Goulart, Callisto, 2003), related to changes in primary productivity (Delitti, 1995). In response, fish species may seek alternative resources to complement their diets in modified habitats and improve survival probability, a behaviour known as trophic plasticity (Abelha et al., 2001). Such changes in the feeding preferences are related to both seasonal and spatial differences of the supply of food (Uieda, Pinto, 2011). Although morphology may set limits to patterns of resource use, since ecomorphological traits are specific to each trophic guild (Albouy et al., 2011), these limits are broad enough to allow fishes changing their choice of prey resources to respond to local biotic and/or abiotic conditions (Ibañez et al., 2007). Detritivorous for instance, are expected to consume more algae in polluted sites and more periphyton in undisturbed sites, as the presence and abundance of these resources are related to water quality (Moschini-Carlos, 1999).

A common underlying assumption is that environmental degradation leads to the decline of the proportion of trophic specialists and carnivores, while the proportion of omnivores increases (Fausch et al., 1990). Several studies have also suggested that certain anthropogenic activities may alter the trophic niche of aquatic organisms (e.g., Crook et al., 2015; Castro et al., 2016; de Carvalho et al., 2017a; Alonso et al., 2019). Large ranges of carbon sources exploited by benthic macroinvertebrates and fishes have been reported, for instance, in streams under influence of pasture (Castro et al., 2016; de Carvalho et al., 2017a). However, little is known about the impact of sewage pollution on the trophic niche of fish species. Furthermore, it is not known whether species of different guilds respond in the same way to stressors, such as pollutants.

The stable isotopes of carbon (δ13C) and nitrogen (δ15N) are key tools for accessing information on feeding habits and trophic structure, since the isotopic composition of the consumers reflects the isotopic composition of their diet, providing long-term feeding information (Manetta, Benedito-Cecilio, 2003). Isotopic niche has been used as a proxy of the trophic niche, and represented with isotopic values (δ15N and δ13C) as coordinates, because an animal’s chemical composition is directly influenced by what it consumes as well as the habitat in which it lives (Newsome et al., 2007). A positive factor in using isotopes is that, unlike stomach contents, it is possible to analyse the main food sources of consumers over weeks (Sacramento et al., 2016; Winter et al., 2019), not just what they fed on in the moments prior to capture. Therefore, from the isotopic composition of consumers and its main food sources it is possible to evaluate the diet of captured organisms, energy flow, and the trophic dynamics of aquatic environments (Manetta, Benedito-Cecilio, 2003).

Considering that pollution can affect biota in different ways, the main objective of this work was to evaluate if the presence of sewage in aquatic environments alters the feeding habits and isotopic niche of fish species from different trophic guilds. For this, the carbon and nitrogen isotopic composition of two species – the piscivorous Hoplias intermedius (Günther, 1864) and the detritivorous Hypostomus francisci (Lütken, 1874) – were evaluated in different regions of a highly polluted Brazilian river basin. These species were selected because they represent different trophic guilds with resident habit (non-migratory) and because they were widely distributed and abundant throughout the studied basin. Hypostomus francisci (Siluriformes, Loricariidae), inhabit rocky or sandy bottoms in places with running water, and are considered detritivorous, as they ingest large amounts of organic matter from sediments (Cardone et al., 2006). As a consequence, these fish are import for the recycling of nutrients in aquatic environments (Pereira, Resende, 1998; Flecker, Taylor, 2004). Hoplias intermedius (Characiformes, Erythrinidae) are of predatory habit, feeding preferentially on other fishes (Carvalho et al., 2002). In addition, this species is economically important and a source of food for the local community. The following hypotheses were tested: 1) Species of different trophic guilds respond similarly to the presence of pollution in the aquatic environment; 2) piscivorous species are more sensitive to pollution, since negative effects strengthen across trophic levels; and 3) both trophic guilds will present larger isotopic niches in unpolluted habitats due to a greater availability of different resources.

Material and methods

Study area. The study was carried out in the main channel and main tributaries of the Rio das Velhas sub-basin, that belongs to the Rio São Francisco basin, located in the Brazilian state of Minas Gerais (Fig. 1). The headwaters of the rio das Velhas are located in the municipality of Ouro Preto, and has an overall extension of approximately 801 km, making it the largest tributary of the Rio São Francisco (Machado et al., 2008). The Rio das Velhas basin encompasses 51 municipalities of Minas Gerais, totalling almost five million inhabitants (IBGE, 2000). The river is responsible for most of the water supply in this region, presenting significant economic and social importance. Annual average temperature varies between 19 and 23 ºC, with precipitation levels between 900 and 2000 mm, and well-defined seasons – dry (winter) and rainy (summer) (PDRH, 2015). The Cerrado (Brazilian Savanna) is the predominant natural vegetation in the watershed, however; approximately 90% has already been modified by human activities. Due to the diverse human activities developed in the region, this basin is in an advanced state of degradation. Main sources of pollution along the river’s course include waste products from iron ore mining, and the discharge of sewage and other pollutants from the Metropolitan Region of Belo Horizonte (MRBH) in its middle portion. Despite this scenario, it is still possible to find relatively unpolluted tributaries, such as the Rio Cipó, which has its headwaters in the Serra do Cipó National Park.

FIGURE 1| Sampling network in the rio das Velhas basin, Minas Gerais, Brazil. Sampling sites at rio das Velhas main stem (RV-01 to RV-05), rio Taquaraçu (TQ); rio Jaboticatubas (JB); rio Cipó (CP1 and CP2); rio da Onça (ON); rio Bicudo (BI); rio Curimataí (CU); and Sewage Treatment Plants (STP’s).

Sampling design. Samples were taken at five sites along the main channel of the Rio das Velhas (RV), which were later grouped into three regions: one sampling site in the upper section of Rio das Velhas – “Upper RV” (region with the best water quality); two sample sites in the middle section of Rio das Velhas – “Middle RV” (most impacted) and two sampling sites in the lower section of Rio das Velhas – “Lower RV” (region farthest from the MRBH and near the river’s mouth, presents a partial improvement in the water quality as the distance from MRBH increases). In addition, seven sites were sampled in six of the main tributaries of the Rio das Velhas basin, which were considered as controls due to the absence of human disturbance (Rio Jabuticatubas, Rio Taquaraçu, Rio da Onça, Rio Bicudo, Rio Curimataí and two sites in Rio Cipó). Two sewage treatment plants (STP), Arrudas and Onça, were also sampled to obtain complementary samples of the suspended material to determine the isotope signature of sewage (Tab. 1; Fig. 1). Fish collections were carried out over three campaigns in the main channel (two in the dry season and one in the rainy season) and two campaigns in tributaries (one in the dry season and one during the rainy season) (Tab. 1).

TABLE 1 | Geographic location (in degrees/minutes/seconds and UTM, date, altitude and municipality) and water quality of the sampling sites sampled in the main channel and tributaries of rio das Velhas. Cond.= Condutivity (ìS/cm); D.O.= Dissolved oxygen (mg/l); Am. Nitr. = Ammoniacal nitrogen (mg/l); Phosp. = Total phosphorus (mg/l); Tox. Contam. = Toxic contamination; Deg. Level = degradation level ranging from I (undisturbed/unpolluted) to IV (degraded/polluted) (Feio et al., 2015). *Sites with hypereutrophic condition according IGAM.

Characteristics of sampling sites

Water Quality


Sampling sites

Date of sampling


Altitude (m)




Am. Nitr.


Tox. Contam.

Deg. Level


Upper RV


19/08/2015 20/01/2016 9/06/2016

20˚01’10.7"S 43˚49’45.4"W 23K 622454 E 7785916  W


Nova Lima








Middle RV


10/08/2015 11/01/2016 31/05/2016

18˚48’19.2"S 44˚09’09.2"W 23K 589298 E 7920498 W







Arsenic and total ammoniacal nitrogen


Gold metallurgy and discharge of domestic sewage

Middle RV


11/08/2015 12/01/2016 1/06/2016

18˚25’33.2"S 44˚11’10.9"W 23K 585926 E 7962502 W










Lower RV


13/08/2015 13/01/2016 3/06/2016

17˚51’55.4"S 44˚32’57.4"W 23K 547752 E 8024649 W









Discharge of domestic sewage and agriculture (sugar cane)

Lower RV


12/08/2015 14/01/2016 2/06/2016

17˚12’25.9"S 44˚48’49.8"W 23K 519793 E 8097515 W


Várzea da Palma







Discharge of domestic sewage


Rio Taquaraçu (TQ)

19/06/2015 23/10/2015

19º36’03.1”S 43º44’09.9”W 23K 632551 E 7832196 N


Taquaraçu de Minas








Rio Jaboticatubas (JB)

16/06/2015 22/10/2015

19º27’51.4”S 43º54’18.0”W 23K 614931 E 7847433 N











Rio Cipó


15/06/2015 21/10/2015

19º20’01.1”S 43º39’17.0”W 23K 641315 E 7861707 N


Santana do Riacho









Rio da Onça (ON)

11/07/2015 18/10/2015

19º02’48.1”S 44º15’15.9”W 23K 578454 E 7893841 N











Rio Cipó (CP2)

10/07/2015 13/10/2015

18º41’07.0”S 43º59’48.4”W 23K 605789 E 7933699 N


Presidente Juscelino









Rio Bicudo (BI)

06/07/2015 17/10/2015

18º21’30.7”S 44º35’30.8”W 23K 543117 E 7970099 N











Rio Curimataí (CU)

08/07/2015 15/10/2015

17º59’33.4”S 44º10’47.5”W 23K 586829 E 8010438 N


Augusto de Lima








Sewage – MRBH

STP Arrudas

20/07/2016 25/01/2017











Sewage – MRBH

STP Onça

20/07/2016 18/01/2017



Belo Horizonte









Information about the level of degradation in sampling sites (excluding Rio Taquaraçu that did not have a corresponding point) was obtained from previously published data (Feio et al., 2015). Degradation levels ranged from I (undisturbed/unpolluted) to IV (degraded/polluted) (Tab. 1).

Data on water quality, hypereutrophic condition, toxic contamination and pressure factors acting in studies sites were accessed through Instituto Mineiro de Gestão das Águas website (IGAM, 2018), which publishes quarterly monitoring reports from several points across the Rio das Velhas basin. Values of conductivity, dissolved oxygen, ammoniacal nitrogen and total phosphorous presented in this study correspond to average values obtained from the IGAM measurements during 2015 and 2016. The hypereutrophic condition, toxic contamination and pressure factors acting in sampling sites were obtained from the report of the year 2017. Based on their geographic proximity, the IGAM monitoring sites BV010; BV136; BV139, BV141, BV144, BV147, BV149; BV150; BV151; BV162; SC33 were considered as the correspondents of CP-01, JB-01, RV-01, RV-02, ON-01, BI-01, RV-05, RV-03, RV-04, CP-02 and CU-01, respectively (Tab. 1).

Fish sampling. Trophic guilds were determined from stomach content analysis previously conducted by the authors (Tab. S1, available only in the online version). The specimens of H. intermedius and H. francisci were collected with gillnets (20 m long, with 3-16 cm stretch measure mesh), seines (5 m long, 1 mm mesh), cast nets (3 cm stretch measure mesh), and mosquito nets (80 cm in diameter, 1 mm mesh). Gill nets were deployed in the water column for 14 h overnight. Seines were used in shallow areas or littoral zones, mosquito nets were used in near-shore aquatic macrophytes (both shorelines), undercut banks and in riffles, and cast nets were used in habitats too deep to wade. The three latter methods were employed for 1-3 h. In the field, muscle samples from adult specimens were removed and were kept frozen until laboratory processing. In laboratory, these samples were lyophilized for 24 hours and ground to a fine and homogeneous powder using a pestle and mortar and subsequently stored in eppendorf tubes.

Collected species were deposited in the Coleção Ictiológica da Universidade Federal de Lavras (CI-UFLA) and the Museu de Zoologia da Universidade de São Paulo (MZUSP), with the following catalogue numbers: H. francisci (CI-UFLA 1029, MZUSP 73724); H. intermedius (MZUSP 73655, MZUSP 73735, MZUSP 73839).

Resource sampling. Food resources usually consumed by H. francisci and H. intermedius were identified by analysing the stomach contents of individuals of both species (see Tab. S1, available only in the online version) and through information in the literature (e.g., Carvalho et al., 2002; Cardone et al., 2006). We collected five replicates (at each site) of the following resources: periphyton, filamentous algae, grasses, riparian vegetation, suspended material (raw sewage) and fish (due to piscivorous habit of H. intermedius) (see Tab. S2, available only in the online version).

Samples of algae and vegetation (grasses and riparian vegetation) were collected randomly at all sites where present, stored in plastic pots and kept frozen until laboratory processing. Periphyton were collected by scraping stones and collected material stored in plastic bottles with distilled water. To obtain the isotope signature of sewage, suspended matter in the water was sampled using a phytoplankton net (45 µm mesh) for a period of three minutes in the sewage treatment plants (STPs) Arrudas and Onça. After collection, the liquid samples (periphyton and suspended matter) were immediately frozen for preservation of the material. In the laboratory, samples were filtered using a filtration apparatus attached to a vacuum pump using calcined quartz fiber filters (Whatman® QMA quartz filters). All basal resource samples were then dried in an oven at 60°C for a minimum period of 48 hours. Afterwards, dried samples were ground to a fine and homogeneous powder using pestle and mortar and stored in eppendorf tubes.

Other fish species (see Tab. S2, available only in the online version), which make up the assembly of each sample point and are the primary food source for H. intermedius, were collected with the same fishing equipment and sampling effort mentioned in the previous section. Sample processing was also done in the same manner; however, small fish individuals (with less than 3cm) were processed as a whole, without the digestive tract. To avoid contamination, all equipment and material used were washed in distilled water along the processing.

Isotopic analysis. After laboratory processing, a total of 129 fish samples (66 H. francisci and 63 H. intermedius) and 1725 resources samples (including fish from other species that served as potential resources for H. intermedius) were sent to the Centre for Nuclear Energy in Agriculture (CENA) at the University of São Paulo for isotopic analysis. Approximately 2-5 mg of dry material from animal tissue, and about 5-10 mg of basal resources were selected for analysis. To determine the isotopic ratio, a mass spectrometer system in the Continuous-flow (CF-IRMS) mode was used with a Carlo Erba elemental analyzer (CHN 1110) coupled to a Delta Plus mass spectrometer (Thermo Scientific). Results were expressed as the relative difference of international reference standards (air nitrogen and PeeDee Belemnite), in the delta notation (δ‰), and calculated using the following formula:

δX = [(Rsample /Rstandard) -1] x 103

where “X” is 13C or 15N and “R” represents the isotopic ratio 13C/12C or 15N/14N (Barrie, Prosser, 1996).

Statistical Analysis. Differences in the isotopic ratios of δ13C and δ15N of consumers and resources between the four regions were tested using one-way analyses of variance (ANOVAs) where assumptions of normality and homoscedasticity were met. The nonparametric Kruskal-Wallis test was used for data with non-normal distributions. When significant differences (p<0.05) were observed, the means were compared using the post-hoc Tukey’s test. These analyses were performed in the software Statistica 6.0 (Statsoft, 2004).

To evaluate the trophic structure of the detritivorous and piscivorous populations, the individuals of two species were plotted in the bi-plot space according to the isotopic values of carbon (x-axis) and nitrogen (y-axis) in each region. Source contributions to the detritivorous and piscivorous diet were estimated for the four regions (Upper RV, Middle RV, Lower RV, and Control) based on stable isotope data analysed through Bayesian stable isotope mixed models (Moore, Semmens, 2008; Parnell et al., 2010), using the MixSIAR package in R (Stock, Semmens, 2016b). A Markov chain Monte Carlo sampling was conducted based on the following parameters: number of chains = 3; chain length = 100,000; burn in = 50,000; thin = 50 and model 4 (Resid*Process) error structure (Stock, Semmens, 2016a). Diagnostic tests (Gelmin-Rubin, Heidelberger-Welch and Geweke) and trace plots were used to examine models. The fractionation values used for consumers were 0.4 ± 1.3‰ for C and 2.54 ± 1.27‰ for N (Post, 2002; Vanderklift, Ponsard, 2003). Both the graphical representation and the partition analysis were done using the MixSIAR package using the R software.

The isotopic niches of the detritivorous and piscivorous in both regions (Upper RV, Middle RV, Lower RV and Control) were quantified based on standard ellipse areas (SEA – expressed in ‰2) through use of the Stable Isotope Bayesian Ellipses package in R – SIBER (Jackson et al., 2011). The standard ellipse area (SEA) represents the core isotopic niche space and it is a proxy of the richness and evenness of resources consumed by the population (Bearhop et al., 2004).


Individuals of the detritivorous (and/or algivorous) Hypostomus francisci and the piscivorous Hoplias intermedius were collected at all sampling sites. Both the detritivorous and piscivorous showed significant variation in their carbon and nitrogen isotopic composition between study regions. For both trophic guilds, the δ13C and δ15N were enriched in regions under the influence of heavy pollution (middle and lower RV) (Fig. 2). Basal resources also presented extensive variation in isotopic composition of δ13C and δ15N between study regions, except for riparian vegetation that did not vary in δ13C composition between the four sampled regions (Tab. 2).

FIGURE 2| Variation in the isotopic composition of carbon (A. and C.) and nitrogen (C. and D.) in the piscivorous species Hoplias intermedius (A. and C.) and the detritivorous species Hypostomus francisci (B. and D.) among the studied regions.

TABLE 2 | Variation in the carbon and nitrogen isotopic composition of the resources sampled in the four regions of the rio das Velhas basin, Minas Gerais, Brazil. Letters a, b, c and d indicate significant differences according to post-hoc Tukey’s test.






Riparian vegetation








































Upper RV

















Middle RV



















Lower RV










































Upper RV

















Middle RV



















Lower RV


























The two trophic guilds varied in their distribution in bi-plot space in unpolluted (upper RV and control) and polluted regions (middle and lower RV). In unpolluted regions, the detritivorous presented greater amplitude in δ13C and smaller amplitudes of δ15N when compared to the piscivorous (Figs. 3A, D). However, in polluted regions, the detritivorous presented more enriched δ15N values than the piscivorous (Figs. 3 B, C).

FIGURE 3| Distribution of the piscivorous species Hoplias intermedius (red points) and the detritivorous species Hypostomus francisci (blue points) in the bi-plot space by study regions: A. Upper RV, B. Middle RV, C. Lower RV and D. Control sites. Resources: GR = Grasses; RV= Riparian vegetation; PE = periphyton; SW = Sewage (before treatment); AL = filamentous algae and FS = Fish.

Partition analyses revealed that the proportions of different resources assimilated by fishes varied between sample regions. The most assimilated resource by the piscivorous was fish. However, in the upper section of the Rio das Velhas, such guild also assimilated basal resources (filamentous algae and periphyton), as well as allochthonous items from the forest in the surroundings (riparian vegetation and grasses) (Fig. 4, see Tab. S3, available only in the online version).

FIGURE 4| Mean proportion of each resource assimilated by A. Hoplias intermedius (piscivorous) and B. Hypostomus francisci (detritivorous) at each study region (MixSIAR results). AL = filamentous algae; SW = Sewage (before treatment); GR = Grasses; RV= Riparian vegetation; PE = periphyton and FS = Fish.

The highest variation in the proportion of assimilated resources was observed for the detritivorous. Filamentous algae were the main resource, except in lower RV sites, where the main source of carbon came from grasses. The importance of grasses as a carbon source was also observed in the upper section of the Rio das Velhas (approximately 22%). Periphyton were the second most important resource for the detritivorous in all regions (mainly in the middle section) (Fig. 4, Tab. S4, available only in the online version).

Comparing the isotopic niche occupied by these species in different regions, we observed that the two fish species, despite being members of different trophic guilds, presented similar responses to the presence of pollutants (Fig. 5). In regions under high influence of sewage (middle and lower RV), both trophic guilds presented isotopic niches with values more enriched in nitrogen. In addition to influencing nitrogen levels, the presence of pollutants also altered the type and range of assimilated resources, with fishes from both trophic guilds feeding on resources with depleted δ13C in unpolluted regions. Such variation in the assimilated resources was more visible for the detritivorous, as the amplitude of the x axis (carbon) was much wider in unpolluted environments.

FIGURE 5| Isotopic niche (measured by ellipse area with 95% confidence interval) occupied by A. Hoplias intermedius (piscivorous) B. Hypostomus francisci (detritivorous) in different regions of the rio das Velhas basin, Minas Gerais, Brazil.


Using isotopic analyses, we confirm the hypothesis that fish from different trophic guilds respond similarly to the presence of pollution from domestic and industrial sewage discharge in aquatic environments. It was also possible to confirm our third hypothesis that isotopic niches occupied by fish species are wider in unpolluted environments, probably due to the consumption of a greater variety of resources. However, the greater trophic plasticity and higher nitrogen enrichment of the detritivorous diets in polluted sites, suggests that such guild may be more sensitive to variation in environmental conditions and, therefore, a better bioindicator of water quality.

We found striking evidence of nitrogen enrichment in fish species and basal resources sampled in polluted regions. Such enrichment has also been observed in aquatic environments under different anthropogenic impacts (e.g., Carvalho et al., 2015; Loomer et al., 2015; Castro et al., 2016; Orlandi et al., 2017). Although δ15N values above 25‰ are rarely recorded, they have been consistently observed in fishes and resources from the Rio das Velhas basin (Alonso et al., 2019; Carvalho et al., 2019). The most extreme δ15N values were observed in the region where the high discharge of pollutants occurs (middle RV) and downstream of this region (lower RV). At lower RV, despite the high δ15N of fishes, the δ15N was slightly less enriched than in the region close to MRBH (middle RV) likely due the sewage dilution. Such high values are probably related to eutrophication processes (Silva, Fonseca, 2016), and were detected in all components of trophic webs, from autotrophic organisms (e.g., algae and periphyton) to the consumers.

In the most polluted regions (middle and lower RV), δ15N of detritivorous fish were more enriched than piscivorous fish. This pattern is not expected, since, by theory, the piscivorous fishes should occupy the top of the food web. The reason why the δ15N enrichment of piscivorous fishes is lower than that of detritivores is not clear. Both piscivorous and detritivorous species sampled in this study are resident, so it is unlikely piscivorous fishes are not feeding on fish from the same site where they were sampled. As detritivorous fishes usually presents bottom habits and feed in the resources located in this portion of the water bodies, probably they may be consuming δ15N enriched resources (not sampled in the present study) that are not being consumed by fishes from other guilds. In addition, the detritivorous from the genus Hypostomus are barely predated due its morphology and adoption of defence strategies (Bruton, 1996; Kirchheim, Goulart, 2010), what could explain why the δ15N enrichment is not propagating to piscivorous fishes. Therefore, the consume of δ15N enriched unsampled resources together with the lack of predation of detritivorous fishes by piscivorous fishes (otherwise the piscivorous should be more enriched than the detritivorous) seems to be the main hypothesis to explain why detritivorous fishes are more δ15N enriched than the piscivorous fishes.

The different carbon isotopic compositions between control sites and the polluted ones, suggests that fishes from both trophic guilds are modifying their feeding habits in response to changes in the aquatic environment (Abelha et al., 2001), despite differences in their trophic plasticity. Such changes were reflected in the isotopic niches occupied by the consumers, and similar patterns of variation in niche amplitude, even though these species belong to different trophic guilds. Moreover, a clear homogenization in the isotopic composition of available resources in polluted areas may also explain the narrower isotopic niches of consumers in such regions (Abelha et al., 2001; Pusey, Arthington, 2003).

Fish were the most consumed resources by the piscivorous, as expected for species of the genus Hoplias (Pompeu, Godinho, 2001; Carvalho et al., 2002; Montenegro et al., 2013). However, other items were also assimilated by this species, which probably influenced the variation of the isotopic composition of this species between the regions. Vegetable remnants and algae/periphyton, may have been ingested together with the prey during predation events, as they are not normally part of the diet of this species (Corrêa, Piedras, 2009; Beliene et al., 2014). However, in some regions (Upper and Middle RV), this accounted for over 20% of assimilated resources, which weakens the hypothesis of accidental consumption. Another hypothesis is that these resources may have been indirectly assimilated through the ingestion of aquatic insects and their larvae, which may be herbivores/detritivorous (feeding on small algae and organic matter), or small particulate filter-feeders suspended in water. As approximately two thirds of sampled individuals were juveniles reinforces this hypothesis, as juveniles predominantly feed on aquatic invertebrates, and adults may do so also in presence of other piscivorous species (Pompeu, Godinho, 2001).

The detritivorous species assimilated a range of basal resources, including algae, periphyton and C4 grasses among the sampled regions, which consequently was reflected in the variation of the δ13C isotopic composition. The detritivore/ herbivore habit has been described to representatives of the genus Hypostomus (Pessoa et al., 2013) and was confirmed through stomach content analysis, which explains the consumption of allochthonous items, such as plant remnants (grasses). The high consumption of algae and periphyton was expected, as this genus is also considered algivorous (Cardone et al., 2006). In addition, as this species adheres to the substrate during foraging, some items outside their normal diet can be ingested (Ross, 1986; Villares-Junior et al., 2016), such as insect larvae and pupae that are buried in sediments. Therefore, some items reflected in the isotopic composition of the detritivorous may also have been assimilated indirectly, through the accidental consumption of insects and other aquatic organisms.

Greater algal growth is expected in more eutrophic sites, where there is sewage discharge (Tundisi, Matsumura, 2008). On the other hand, periphyton are more associated with undisturbed environments, where water transparency (Cetto et al., 2004), and the presence of free surfaces, such as rocks and/or submerged vegetation, favours their development (Tundisi, Matsumura, 2008). Nevertheless, the opposite was observed in the control region, with algae and periphyton contributing approximately 86% and 10% of assimilated resources, respectively. This could reflect the specific characteristics of the studied tributaries, which may have favoured the proliferation of algae rather than periphyton. However, the abundance of each resource in each sampled region was not evaluated in the present study. In addition, detritivorous may exhibit selectivity and preference for algae over periphyton, even when both resources are abundant (Bozza, Hahn, 2010).

It is interesting to note that carbon from grasses contributed 22% of the isotopic composition of detritivorous in the upper RV and 51% in the lower RV. This high assimilation of grasses was not expected, as it is not considered as a primary energy source for detritivorous fish (Araujo-Lima et al., 1986). In addition, C4 grasses are generally consumed by fish in small quantities because they contain fewer nutrients and are difficult to digest. However, recent studies have suggested that the contribution of C4 plants can be substantial to aquatic communities (Hoeinghaus et al., 2007; Ferreira et al., 2012), and herbivorous fish species can feed on this resource in impacted environments (Carvalho et al., 2015). In addition, we also need to consider the indirect assimilation of C4 sources through debris consume (Garcia et al., 2016), since it is composed by a mix of autochthonous (e.g., algae, macrophytes and periphyton) and allochthonous material (e.g., leaves from trees and grasses) provided by river banks.

Through the results it was possible to observe that the detritivorous species showed great variation in the proportion of assimilated resources in different regions and high enrichment in the isotopic composition of nitrogen in polluted sites. Detritivorous species are important for the functioning of aquatic ecosystems (Pereira, Resende, 1998) and usually comprise a significant portion of the total community (Vasconcelos-Filho et al., 2009). Therefore, environmental degradation, and subsequent alterations in resources abundance and in the feeding habits of this trophic group, may have implications for ecosystem functioning.

We can conclude that the use of carbon and nitrogen isotopes contributes greatly to our understanding of the trophic ecology of fish species in impacted aquatic environments. We observed that fishes from both trophic guilds exhibited high food plasticity, varying the proportion of resources consumed among regions. Moreover, pollution promoted nitrogen enrichment in resources and consumers which, consequently, was reflected in the isotopic niches occupied by them. Such changes may have major impacts on ecosystem functioning in aquatic systems, especially when important changes are observed at different levels of trophic webs (e.g., top predators and detritivorous). In addition, for future biomonitoring and assessment of the biotic integrity, we suggest the use of first-level consumers rather than species that occupy higher trophic levels (predators), as they could be more sensitive to changes in environmental conditions, especially where domestic sewage is the main stressor. Finally, it is important to emphasise that our results are based on the evaluation of only two species in a large area. Therefore, whole fish community evaluations should be encouraged.


We thank the Agência Peixe Vivo and the Comitê de Bacia Hidrográfica do Rio das Velhas (CBH – Rio das Velhas) for the project financial support and the Projeto Manuelzão for the logistical support. Thanks to Aline J. Grossi (Federal University of Lavras/ UFLA) and Luiza Hoehne (Federal University of Minas Gerais/ UFMG) for the support on processing of samples, and Lucas Grossi (Projeto Manuelzão) for the map elaboration. Thanks also to the Benthos Ecology Laboratory (UFMG) and the Laboratory of Fish Ecology (UFLA) that allowed processing of samples and infrastructure, and to the Centre for Nuclear Energy in Agriculture (CENA) for their support and partnership in the isotopic analysis. Paulo S. Pompeu received a research grant and a research fellowship from the Conselho Nacional de Desenvolvimento Científico e Tecnológico (303548/2017-7) and from the Fundação de Amparo à Pesquisa do Estado de Minas Gerais (00608/2015). Débora R. Carvalho was financed in part by the Coordenação de Aperfeiçoamento Pessoal de Nível Superior – Brasil (CAPES) – Finance code 001.


Abelha MCF, Agostinho AA, Goulart E. Plasticidade trófica em peixes de água doce. Acta Sci Biol Sci. 2001; 23(2):425–34.

Albouy C, Guilhaumon F, Villéger S, Mouchet M, Mercier L, Culioli JM, Tomasini JA, Le Loc’h F, Mouillot D. Predicting trophic guild and diet overlap from functional traits: statistics, opportunities and limitations for marine ecology. Mar Ecol Prog Ser. 2011; 436:17–28.

Alonso MB, Carvalho DR, Alves CBM, Moreira MZ, Pompeu PS. Changes in trophic characteristics of two fish species of Astyanax (Teleostei: Characidae) in response to aquatic pollution. Zoologia. 2019; 36:e30445.

Araujo-Lima CARM, Forsberg BR, Victoria R, Martinelli L. Energy sources for detritivorous fish in the Amazon. Science. 1986; 234(4781):1256–58.

Barrie A, Prosser SJ. Automated analysis of light-element stable isotopes by isotope ratio mass spectrometry. In: Boutton TW, Yamasaki SI, editors. Mass spectrometry of soils. New York: Marcel Dekker; 1996. p.1–46.

Bearhop S, Adams CE, Waldron S, Fuller RA, Macleod H. Determining trophic niche width: a novel approach using stable isotope analysis. J Anim Ecol. 2004; 73(5):1007–12.

Beliene GH, Rocha ARM, Souza CMM. Parâmetros alimentares de Hoplias malabaricus, como ferramenta de análise ambiental na Lagoa Feia, RJ, Brasil. E&S Engineering and Science. 2014; 1(1):1–8.

Bozza AN, Hahn NS. Uso de recursos alimentares por peixes imaturos e adultos de espécies piscívoras em uma planície de inundação neotropical. Biota Neotrop. 2010; 10(3):217–26.

Bruton MN. Alternative life-history strategies of catfishes. Aquat Living Resour. 1996; 9:35–41.

Camargo AFM, Bini LM, Schiavetti A. Avaliação dos impactos provocados pelas descargas de esgotos orgânicos em alguns corpos d’água do município de Rio Claro. In: Esteves FA, editor.Volume I: Estrutura, funcionamento e manejo de ecossistemas brasileiros. Rio de Janeiro: Oecologia Brasiliensis; 1995. p.395–406.

Cardone IB, Lima-Junior SE, Goitein R. Diet and capture of Hypostomus strigaticeps (Siluriformes, Loricariidae) in a small Brazilian stream: Relationship with limnological aspects. Braz J Biol. 2006; 66(1A):25–33.

Castro DMP, Carvalho DR, Pompeu PS, Moreira MZ, Nardoto GB, Callisto M. Land use influences niche size and the assimilation of resources by benthic macroinvertebrates in Tropical headwater streams. PLoS One. 2016; 11(3):e0150527.

Cetto JM, Leandrini JA, Felisberto SA, Rodrigues L. Comunidade de algas perifíticas no reservatório de Irai, Estado do Paraná, Brasil. Acta Sci Biol Sci. 2004; 26(1):1–7.

Corrêa F, Piedras SRN. Alimentação de Hoplias aff. malabaricus (Bloch, 1794) e Oligosarcus robustus Menezes, 1969 em uma lagoa sob influência estuarina, Pelotas, RS. Biotemas. 2009; 22(3):121–28.

Crook DA, Lowe WH, Allendorf FW, Erõs T, Finn DS, Gillanders BM, et al. Human effects on ecological connectivity in aquatic ecosystems: Integrating scientific approaches to support management and mitigation. Sci Total Environ. 2015; 534:52–64.

Cunico AM, Agostinho AA, Latini JD. Influência da urbanização sobre as assembléias de peixes em três córregos de Maringá, Paraná. Rev Bras Zool. 2006; 23(4):1101–10.

de Carvalho DR, Castro D, Callisto M, Moreira MZ, Pompeu PS. Isotopic variation in five species of stream fishes under the influence of different land uses. J Fish Biol. 2015; 87(3):559–78.

de Carvalho DR, Castro DMP, Callisto M, Moreira MZ, Pompeu PS. The trophic structure of fish communities from streams in the Brazilian Cerrado under different land uses: an approach using stable isotopes. Hydrobiologia. 2017a; 795(1):199–217.

de Carvalho DR, Leal CG, Junqueira NT, Castro MA, Fagundes DC, Alves CBM, et al. A fish-based multimetric index for Brazilian savanna streams. Ecol Indic. 2017b; 77:386–96.

Delitti WBC. Estudos de ciclagem de nutrientes: Instrumentos para a análise funcional de ecossistemas terrestres. In: Esteves FA, editor. Volume I: Estrutura, funcionamento e manejo de ecossistemas brasileiros. Rio de Jeneiro: Oecologia Brasiliensis; 1995. p.469–86

Fausch KD, Lyons J, Karr JR, Angermeier PL. Fish communities as indicators of environmental degradation. Am Fish Soc Symp. 1990; 8:122–44.

Feio MJ, Ferreira WR, Macedo DR, Eller AP, Alves CBM, França JS, Callisto M. Defining and testing targets for the recovery of tropical streams based on macroinvertebrate communities and abiotic conditions. River Res Appl. 2015; 31(1):70–84.

Ferreira A, Cyrino JEP, Duarte-Neto PJ, Martinelli LA. Permeability of riparian forest strips in agricultural, small subtropical watersheds in south-eastern Brazil. Mar Freshw Res. 2012; 63(12):1272–82.

Flecker AS, Taylor BW. Tropical fishes as biological bulldozers: Density effects on resource heterogeneity and species diversity. Ecology. 2004; 85(8):2267–78.

Freitas CEC, Siqueira-Souza FK. O uso de peixes como bioindicador ambiental em áreas de várzea da bacia amazônica. Rev Agrogeoambiental. 2009; 1(2):39–45.

Garcia AM, Claudino MC, Mont’Alverne R, Pereyra PER, Copertino M, Vieira JP. Temporal variability in assimilation of basal food sources by an omnivorous fish at Patos Lagoon Estuary revealed by stable isotopes (2010–2014). Mar Biol Res. 2016; 13(1):98–107.

Goulart MDC, Callisto M. Bioindicadores de qualidade de água como ferramenta em estudos de impacto ambiental. Revista da FAPAM. 2003; 2(1):153–64.

Hoeinghaus DJ, Winemiller KO, Agostinho AA. Landscape-Scale hydrologic characteristics differentiate patterns of Carbon flow in large-river food webs. Ecosystems. 2007; 10(6):1019–33.

Ibañez C, Tedesco PA, Bigorne R, Hugueny B, Pouilly M, Zepita C, Zubieta J, Oberdorff T. Dietary-morphological relationships in fish assemblages of small forested streams in the Bolivian Amazon. Aquat Living Resour. 2007; 20(2):131–42.

IBGE, Instituto Brasileiro de Geografia e Estatística. Sinopse preliminar do censo demográfico. Rio de Janeiro, IBGE. 2000. 450p.

IGAM-Instituto Mineiro de Gestão das águas. Monitoramento de Qualidade das Águas. www Document. Monit. Qual. das Águas. 2018. URL. http://portalinfohidro.igam

Jackson AL, Inger R, Parnell AC, Bearhop S. Comparing isotopic niche widths among and within communities: SIBER – Stable Isotope Bayesian Ellipses in R. J Anim Ecol. 2011; 80(3):595–602.

Karr JR. Assessment of biotic integrity using fish communities. Fisheries. 1981; 6(6):21–27.;2

Kirchheim PD, Goulart E. Ecomorfologia de predação e antipredação em Siluriformes (Osteichthyes). Oecol Australis. 2010; 14(2):550–68.

Loomer HA, Oakes KD, Schiff SL, Taylor WD, Servos MR. Use of stable isotopes to trace municipal wastewater 3ffluents into food webs within a highly developed River system. River Res Appl. 2015; 31(9):1093–1100.

Machado ATM, Alves CBM, Callisto M. Projeto Manuelzão: Metodologia e resultados. In: Lisboa AH, Goulart EMA, Diniz LFM, organizers. Projeto Manuelzão: A história da mobilização que começou em torno de um rio. Belo Horizonte: Instituto Guaicuy; 2008. p.39–54.

Manetta GI, Benedito-Cecilio E. Aplicação da técnica de isótopos estáveis na estimativa da taxa de turnover em estudos ecológicos: uma síntese. Acta Sci Biol Sci. 2003; 25(1):121–29.

Montenegro AKA, Vieira ACB, Cardoso MML, Souza JERT, Crispim MC. Piscivory by Hoplias aff. malabaricus (Bloch, 1974): a question of prey availability? Acta Limnol Bras. 2013; 25(1):68–78.

Moore JW, Semmens BX. Incorporating uncertainty and prior information into stable isotope mixing models. Ecol Lett. 2008; 11(5):470–80.

Moschini-Carlos V. Importância, estrutura e dinâmica da comunidade perifítica nos ecossistemas aquáticos continentais. In: Pompêo MLM, editor. Perspectivas na Limnologia do Brasil. São Luís: Gráfica e Editora União; 1999. p.1–11.

Newsome SD, Martinez del Rio C, Bearhop S, Phillips DL. A niche for isotopic ecology. Front Ecol Environ. 2007; 5(8):429–36.

Orlandi L, Calizza E, Careddu G, Carlino P, Costantini ML, Rossi L. The effects of nitrogen pollutants on the isotopic signal (∆15N) of Ulva lactuca: Microcosm experiments. Mar Pollut Bull. 2017; 115(1–2):429–35.

Parnell AC, Inger R, Bearhop S, Jackson AL. Source partitioning using stable isotopes: coping with too much variation. PLoS One. 2010; 5(3):e9672.

PDRH – Plano Diretor de Recursos Hídricos do Rio das Velhas. Ecoplan, Belo Horizonte. 2015.

Pereira RAC, Resende EK. Peixes detritívoros da planície inundável do rio Miranda, Pantanal, Mato Grosso do Sul Brasil. Corumbá: Embrapa; 1998.

Pessoa EKR, Silva NB, Chellappa NT, Souza AA, Chellappa S. Morfologia comparativa do trato digestório dos peixesHoplias malabaricus e Hypostomus pusarum do açude Marechal Dutra, Rio Grande do Norte, Brasil. Biota Amazônia. 2013; 3(1):48–57.

Pompeu PS, Alves CBM, Callisto M. The effects of urbanization on biodiversity and water quality in the rio das Velhas basin, Brazil. In: Brown LR, Gray RH, Hughes RM, Meador M, editors Effects of urbanization on stream ecosystems. AFS Symposyum, 47. American Fisheries Society; 2005. p.11–22.

Pompeu PS, Godinho AL. Mudança na dieta da traíra Hoplias malabaricus. Rev Bras Zool. 2001; 18(4):1219–25.

Post DM. Using stable isotopes to estimate trophic position: Models, methods, and assumptions. Ecology. 2002; 83(3):703–18.[0703:USITET]2.0.CO;2

Pusey BJ, Arthington AH. Importance of the riparian zone to the conservation and management of freshwater fish: A review. Mar Freshw Res. 2003; 54(1):1–16.

Ross ST. Resource partitioning in fish assemblages: A review of field studies. Copeia. 1986; 1986(2):352–88.

Sacramento PA, Manetta GI, Benedito E. Diet‐tissue discrimination factors (∆13C and ∆15N) and turnover rate in somatic tissues of a neotropical detritivorous fish on C3 and C4 diets. J Fish Biol. 2016; 89(1):213–19.

Schulz UH, Martins-Junior H. Astyanax fasciatus as bioindicator of water pollution of rio dos Sinos, RS, Brazil. Braz J Biol. 2001; 61(4):615–22.

Silva AR, Fonseca ALDO. Eutrofização dos recursos hídricos como ferramenta para a compreensão das doenças de vinculação hídrica.Geosul. 2016; 31(62):247–70.

Smith VH. Eutrophication of freshwater and coastal marine ecosystems a global problem. Environ. Sci Pollut Res. 2003; 10(2):126–39.

Smith WS, Barrella W, Cetra M. Comunidades de peixes como indicadoras de poluição ambiental. Rev Bras Ecol. 1997; 1:67–71.

Soares AFS, Araújo AS, Rodrigues NUA, Cunha NC. Compensação ambiental devido à falta de tratamento de esgotos domésticos no município de Campanha (MG). Ambiência. 2016; 12(Edição Especial):831–39.

Stock B, Semmens B. MixSIAR GUI user manual. Version 3.1 [Internet]. 2016b.

Stock BC, Semmens BX. Unifying error structures in commonly used biotracer mixing models. Ecology. 2016a; 97(10): 2562-–69.

Tundisi G, Matsumura T. Limnologia. Oficina de Textos; 2008.

Uieda VS, Pinto TLF. Feeding selectivity of ichthyofauna in a tropical stream: space-time variations in trophic plasticity. Community Ecol. 2011; 12(1):31–39.

Vanderklift MA, Ponsard S. Sources of variation in consumer-diet delta 15N enrichment: a meta-analysis. Oecologia. 2003; 136(2):169–82.

Vasconcelos-Filho AL, Neumann-Leitão S, Eskinaqzi-Lessa E, Oliveira MAS, Porto-Neto FF. Hábitos alimentares de consumidores primários da ictiofauna do sistema estuarino de Itamaracá (Pernambuco – Brasil). Rev Bras Enga Pesca. 2009; 4(1):21–31.

Villares-Junior GA, Cardone IB, Goitein R. Comparative feeding ecology of four syntopic Hypostomus species in a Brazilian southeastern river. Braz J Biol. 2016; 76(3):692–99.

Winter ER, Nolan ET, Busst GMA, Britton JR. Estimating stable isotope turnover rates of epidermal mucus and dorsal muscle for an omnivorous fish using a diet-switch experiment. Hydrobiologia. 2019; 828(1):245–58.


Marina Rodrigues do Prado1, Débora Reis de Carvalho1, Carlos Bernardo Mascarenhas Alves2, Marcelo Zacharias Moreira3 and Paulo Santos Pompeu1

[1]    Laboratório de Ecologia de Peixes, Setor de Ecologia, Departamento de Biologia; Universidade Federal de Lavras, CâmpusUniversitário, Caixa Postal 3037, 37200-000 Lavras, MG, Brazil. (MRP); (DRC); (PSP)

[2]    Laboratório Nuvelhas, Projeto Manuelzão, Universidade Federal de Minas Gerais, Av. Antônio Carlos, 6627, Pampulha, 31270- 901 Belo Horizonte, MG, Brazil.

[3]    Laboratório de Ecologia Isotópica, Centro de Energia Nuclear na Agricultura – CENA, Universidade de São Paulo, Av. Centenário, 303, São Dimas, 13416-000 Piracicaba, SP, Brazil.

Authors’ Contribution

Marina Rodrigues do Prado: Formal Analysis, Investigation, Methodology, Writing (original draft).

Débora Reis de Carvalho: Conceptualization, Data curation, Formal Analysis, Writing (original draft).

Carlos Bernardo Mascarenhas Alves: Conceptualization, Formal Analysis, Methodology, Writing (original draft).

Marcelo Zacharias Moreira: Conceptualization, Formal Analysis, Writing (review & editing).

Paulo Santos Pompeu: Conceptualization, Formal Analysis, Funding acquisition, Investigation, Methodology, Project administration, Writing (original draft), Writing (review & editing).

Ethical Statement​

This project was conducted under the approval of the Animal Ethics Committees (CEUA – UFLA 070/15) and Collection License (ICMBio 10327-2).

Competing Interests

Not applicable.

How to cite this article

Prado MR, Carvalho DR, Alves CBM, Moreira MZ, Pompeu PS. Convergent responses of fish belonging to different feeding guilds to sewage pollution. Neotrop Ichthyol. 2020; 18(1):e190045.

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

Distributed under

Creative Commons CC-BY 4.0

© 2020 The Authors.

Diversity and Distributions Published by SBI

Accepted January 29, 2020 by David Hoeinghaus

Submitted May 9, 2019

Epub April 20, 2020