Life history of Gymnotus refugio (Gymnotiformes;Gymnotidae): an endangered speciesof weakly electric fish

Aline Salvador Vanin & Julia Giora &

Clarice Bernhardt Fialho

Received: 16 June 2016 /Accepted: 21 November 2016 /Published online: 28 November 2016# Springer Science+Business Media Dordrecht 2016

Abstract The present study describes the life history ofGymnotus refugio, a species classified as Endangered inthe last published list of threatened species of theBrazilian fauna. The study was conducted at a conser-vation unity that protect one of the last remainingsemideciduous forests in the region. The reproductiveperiod was estimated as occurring from the end ofwinter to the last summer months. Gymnotus refugioexibited fractional spawning, the lowest relative fecun-dity registered among the Gymnotifomes species stud-ied at the present, and male parental care behavior. Theanalyses showed a seasonal pattern on the species diet,associating different food categories to winter, autumn,and spring. According to food items analysis and esti-mated intestinal quotient, G. refugio was classified asinvertivorous, feeding mainly on autochthonous insects.The results obtained herein suggest that the position ofG. refugio as an Endangered species might be influ-enced by its territoriality, habitat specificity, parentalcare behavior, and low fecundity, reinforcing the impor-tance of swampy forest environment conservation as theonly means of the species maintenance.

Keywords Electric-fish . Threatened species .

Reproduction . Feeding

Introduction

The order Gymnotiformes is composed by freshwaterweakly electric fishes, which are able to produce anddetect electric fields with the function of localization andintra and interspecific communication (Crampton andHopkins 2005). They can be found from southernMexico to Argentina, and in the Caribbean island ofTrinidad (Albert and Crampton 2003). The genusGymnotus is the representative with the largest distribu-tion among the order, including 40 valid species(Eschmeyer and Fong 2016) occurring over their entiredistribution area (Mago-Leccia 1994). Environmentalconditions (e.g., water flow, temperature, dissolved ox-ygen, conductivity and vegetation) are important factorsinfluencing the distribution of electric fishes (Crampton1998). These fishes also have several strategies whichfacilitate adaptations to environments with a large rangeof abiotic and biotic factors (Hopkins 1974; Kirschbaum1975, 1979, 2000; Crampton 1998; Albert andCrampton 2005), including air breathing adaptations tocope with prolonged periods of hypoxia or anoxia(Hopkins 1974; Crampton 1998; Kirschbaum 1975,1979, 2000; Albert and Crampton 2005).

The recently described species, Gymnotus refugio(Giora and Malabarba 2016), belongs to the speciesgroup Gymnotus pantherinus currently represented by15 species. Gymnotus refugio was formerly

Environ Biol Fish (2017) 100:69–84DOI 10.1007/s10641-016-0556-z

Electronic supplementary material The online version of thisarticle (doi:10.1007/s10641-016-0556-z) contains supplementarymaterial, which is available to authorized users.

A. S. Vanin : J. Giora (*) : C. B. FialhoLaboratório de Ictiologia, Departamento de Zoologia, Instituto deBiociências, Universidade Federal do Rio Grande do Sul, AvenidaBento Gonçalves, 9500, Prédio 43435, Porto Alegre, RS CEP91501-970, Brazile-mail: jugiora@hotmail.com

misidentified as either the species G. pantherinus(Corrêa et al. 2015) or G. aff. pantherinus (Malabarbaet al. 2013) in the state of Rio Grande do Sul.Gymnotusrefugio is found inhabiting the interior of wetlands andedges of swampy forests composed by dense and float-ing riparian vegetation (Ferrer et al. 2015), and is dis-tributed from the coastal rivers of the southern RioGrande do Sul state to the southern Santa Catarina state(Giora and Malabarba 2016). According to the lastpublished list of threatened fauna species of the RioGrande do Sul state, G. refugio (quoted as Gymnotuspantherinus) is labelled as EN (Endangered) accordingto the UCN criteria (FZB 2014). This classification isbased on the species distribution in the state, whichcovers an estimated area of only 24 km2, due its habitatspecificity, and restriction to streams and fragmentedareas exposed to anthropic actions. Up until now,G. refugio is easily found and noted in populations withhigher number of specimens only in two conservationareas: Refúgio da Vida Silvestre Banhado dos Pachecos(RVSBP) and Parque Estadual de Itapeva, being rareand scarce in all the other places where it has beenregistered (Giora and Malabarba 2016).

Characters related to measures of habitat condi-tions, behavior, and trophic requirements havebeen useful for relating the distribution of fresh-water fishes to environmental variables (Oldenet al. 2006). On the other hand, life history traitsare strong predictors of the vulnerability of fishpopulations as they determine how resilient a spe-cies is to disturbances such as habitat loss and theadverse effects of invasive species (Hamidan andBritton 2015). Reproduction represents one of themost important aspects of the biology of a species,the maintenance of viable populations dependingon its success (Suzuki and Agostinho 1997).Feeding ecology is thoroughly linked to populationdynamics and contributes to the understanding ofsubjects such as resource partitioning, habitat pref-erence, prey selection, predation, competition, tro-phic ecology, and evolution (Braga et al. 2012).Since a disturbed ecosystem can affect directly thedynamic, seasonality, and behaviour of species,developing conservation efforts can require com-prehension of patterns that shape the biologicalcycles and the ecological interaction between theseorganisms and their habitat. On account of theconservation status of G. refugio populations, con-servation measures are needed and the design of

effective ones requires knowledge on the species’life history traits. Therefore, the present study aimsto understand the particular biological predictorsrelated to the low distribution and abundance ofG . refugio , which lead the species to theEndangered category in southern Brazil. The re-sults concerning reproductive biology, gonadalmaturation, and feeding habits unveiled subsidizingfuture conservation actions focusing the speciesand its habitat.

Material and method

Study area

The sampling area is located in the Refugio da VidaSilvestre Banhado dos Pachecos (RVSBP) (30°09′57″S;50°84′99 W) in Viamão municipality, state of RioGrande do Sul, Brazil. The RVSBP is a conservationunit created in 2002, covering 2.560 ha inserted in theGravataí river basin, where the medium and low seg-ments are extremely impacted by human activities(IBGE 2010). According to Oliveira et al. (2005), thefragments of swampy forest surrounding the Gravataíriver are the last remaining of semideciduous forestpermanently submitted to the fluvial influence of theriver basin. The stream whereG. refugiowas sampled ischaracterized by murky water, abundant floating vege-tation, muddy substrate, and dense riparian vegetationknown as Mata Paludosa (= swampy forest).

Sampling

The specimens of G. refugio were monthly sampledfrom March/2011 to February/2012 between 9:00 and17:00 h. Fishes were collected using a dip net underfloating vegetation and an electric fish finder (Cramptonet al. 2007). In the field, the specimens sampled wereeuthanized by immersion in 10% eugenol solution andthen fixed in 10% formalin solution. Water and airtemperature, water pH, dissolved oxygen, and conduc-tivity were recorded at the time and place of sampling.The monthly amounts of rainfall (in millimeters) andtime of sunshine (in hours) were obtained at the 8thDistrict of Meteorology of Porto Alegre (available atMeteorological databank for Education and Research ofthe National Institute of Meteorology, Brazil, station83967).

70 Environ Biol Fish (2017) 100:69–84

Data analysis

In the laboratory, fishes were transferred to 70% ethanolsolution and, afterward, the total length in millimeterand total weight in grams were measured. Individualswere dissected to register gonad and stomach weight,and intestine length. Stomach repletion index (RI) andgonadosomatic index (GSI) were estimated followingthe formula adapted from Santos (1978). These indexesrepresent the percentage organ weight related to fishtotal weight: RI = Ws ×100/Wt and GSI = Wg ×100/Wt.Ws corresponds to stomach weight, Wg to gonadweight, and Wt to total weight. The intestinal quotient(IQ) represents the ratio of the intestine length related tothe fish total length: IQ = Li/Lt. Li corresponds to theintestine length and Lt to the total length.

The reproductive period for males and females wasestablished through the analyses of monthly variation ofthe mean GSI values. The analysis of variance(ANOVA) with Tukey’s post-test was applied to verifypossible differences between the monthly values of bothGSI and RI of males and females separately. A simplelinear regression was applied to verify possible correla-tion between the RI and GSI of males and females. Amultiple linear regression was applied to verify possiblecorrelation between the GSI and RI of females andmales with the abiotic factors (photoperiod, rainfall,water and air temperature, water conductivity, pH, anddissolved O2).

To corroborate the macroscopic characterization anddefine the gonadal maturation phases, 30 male and 36female gonads were selected for histological analysisbeing dehydrated in ethanol series, and infiltrated andembedded in glycolmethacrylate resin. Sections of 3 μmwere performed on a Leica RM2245 microtome withglass knifes, and were stained with Toluidine blue. Theslides were photographed under a Nikon AZ 100 mmmicroscope. The phases of gonadal maturation wereclassified according to Brown-Peterson et al. (2011).For monthly variation of the gonadal maturation phasesof females and males the developing subphases were notincluded. Absolute fecundity was estimated by countingall vitellogenic oocytes present in the ovaries of sevenfemales with the highest GSI values recorded. The rel-ative fecundity was determined by the number ofvitellogenic oocytes counted per female milligram oftotal weight (Adebisi 1987). The same gonads selectedfor fecundity analysis were used for the determination ofthe spawning type. A sub-sample of 150 oocytes was

randomly removed from each selected gonad and thelargest possible oocyte diameter was obtained by exam-ination under a stereomicroscope equipped with amillimetered ocular (Vazzoler 1996).

The sex ratio was determined by the distribution ofmale and female frequency during the sample period.The χ2 test (α <0.05) was applied to determine theexistence of significant differences between the propor-tion of males and females in the population. The sametest was also applied to the distribution of relative fre-quencies of males and females in different total lengthclasses in order to test sexual dimorphism related to totallength.

Stomach content analysis was performed with thehelp of a stereomicroscope and the items of stomachcontents were identified to the lowest possible taxonom-ic level (Mugnai et al. 2010). The alimentary items wereanalyzed by the frequency of occurrence method(Hyslop 1980), including all food items, where thenumber of times that each item has occurred was treatedas the percent of total occurrence number of all items.For statistical analyses, the food items were grouped ineight broad categories according to their ecologicalcharacteristics and origin: allochthonous insects (Al),autochthonous insects (Au), digested organic matter(DOM), fish (Fis), crustacea (Crus), plant material(PM), sediment (Sed), and others (Other). For the PMcategory, all items of vegetal origin such as aquaticplants (leaves and roots) and algae were considered;whereas Sed included substrate components such asrock fragments, earth, sand and mud.

The factors sex, seasonality, and length classes weretested on the diet composition of the species through thePermutational Multivariate Analysis of Variance(PERMANOVA) (p ≤ 0.05) (Anderson 2001) with theBray-Curtis dissimilarity matrix (Borcard et al. 2011). Inorder to test significant variations in the diet of G.refugio, the months of the year were grouped into sea-sons: autumn from March to May; winter from June toAugust; spring from September to November; summerfrom December to February. The length classes weredetermined according to the Sturges rule (Vieira 1991).The Principal Coordinate’s Analysis (PCoA) was ap-plied in order to visualize possible variations in thecategories consumed due to the factors considered.The Indicator Value Index (IndVal) was applied to testwhich alimentary categories might be the most influ-enced by seasonality, sex, and length class. The IndValresults were also expressed in terms of specificity (A)

Environ Biol Fish (2017) 100:69–84 71

and fidelity (B), both ranging [0–1] and representingrespectively the abundance of a category over allgroups, and the presence or absence of a category withinthe site group. All analyses were performed through thesoftware R Project for Statistical Computing 3.0.1.

Results

Reproduction

A total of 123 specimens of G. refugio, 62 females(55.8 mm – 200.8 mm) and 61 males (63 mm –240.7 mm), were sampled. The estimated reproduc-tive period lasted from August to March, with GSIpeak occurring in October for females and August andOctober for males (Fig. 1). According to the analysis

of variance (ANOVA) with Tukey’s post-test, maleand female mean GSI values differ significantly bet-ween the months of sampling (F = 5.34, p < 0.05 formales; F = 3.75, p < 0.05 for females). For females,mean GSI value obtained in October differed fromApril to July and from January. For males, mean GSIvalue obtained in October differed significantly fromMay to July and December differed from August,October and November. The simple linear regressionindicated a positive relationship between the GSI ofboth females (F = 1.33; p < 0.05) and males (F = 11.1;p < 0.05) and their respective RI. Although themonthly data for dissolved oxygen, pH, and watertemperature (Table 1) were not completely recordedfor each month, the multiple linear regression indica-ted no correlation between these abiotic factors andthe GSI of both females (F = 0.62; p = 0.72) and males

Fig. 1 Monthly variation ofmean gonadosomatic index (GSI)of Gymnotus refugio females (a)and males (b). Vertical barsrepresent the standard deviation.Numbers above the barscorrespond to the numbers ofspecimens included in theanalysis

72 Environ Biol Fish (2017) 100:69–84

(F = 0.81; p = 0.61), and the RI of both females(F = 1.7; p = 0.31) and males (F = 1.01; p = 0.52).

The monthly variation of gonadal maturation phasesof females and males (Fig. 2) and the histological anal-ysis of female gonads (Figs. 3 and 4) demonstrated allgonadal maturation phases as proposed by Brown-Peterson et al. (2011). For females, immature individ-uals occurred only in October, ovaries showing predom-inance of oogonia and primary growth oocytes (Fig. 3a).Females in the developing phase were captured duringthe whole sampling period except byMarch, being morepredominant from May to September. During this peri-od, the female gonads presented primary growth oo-cytes, more abundant in the early developing subphase(Fig. 3b), with primary and secondary vitellogenic oo-cytes increasing in number as the gonad maturationreached the mid and late developing subphase (Fig. 3cand 3d, respectivelly). The spawning capable phase wasobserved in August, and from October to February(Fig. 4a); whereas actively spawning females were col-lected in October, November, and March (Fig. 4b).These two phases differed from each other due thepresence of postovulatory follicle complex in the active-ly spawning phase indicating oocytes released.Regressing individuals were sampled in April, January,and February, when the histological analysis demon-strated unorganized gonads with abundant blood ves-sels, postovulatory follicle complexes, and enlargedovarian lamellae (Fig. 4c). The regenerating phase ob-served in April, October, and February was

characterized by the presence of blood vessels, oogo-nias, and primary growth oocytes (Fig. 4d), indicating areorganization in the structure for a new beginning ofthe gonad maturation process.

The histological analysis of male gonads showedthe following phases of gonadal maturation: devel-oping, spawning capable, regressing and regenerating(Fig. 5). Males in the developing phase (Fig. 5a, b)were captured throughout most of the samplingperiod, except by November and February.Histologically, these gonads showed active spermato-genesis and presence of great amount of primary andsecondary spermatocytes, crescent number of sperma-tids and spermatozoa, and scarce spermatogonia.Spawning capable individuals were sampled inAugust and from October to February showing highamount of spermatozoa in lumen of sperm ducts, theother spermatogenic cells being poorly scatter in thegonad (Fig. 5c). Regressing individuals occurred inAugust and December, showing depleted stores ofspermatozoa in sperm ducts and the lumen of thelobules, cessation of spermatogenesis, and an increas-ing number of spermatogonias (Fig. 5d). Fishes re-main in the regressing phase for a relatively shorttime, moving to the regenerating phase, which wasobserved in April, July, and September. This phaseis characterized by the massive presence of sper-matogonia and residual spermatozoa in sperm ducts(Fig. 5e). Fish in regenerating phase are sexuallymature but reproductively inactive. There are no

Table 1 Monthly variation of the water conductivity (μS/cm),dissolved oxygen (mg/l), pH, water and air temperature (°C),photoperiod (hours), and rainfall (mm) from March/2011 to

February/2012 in the Refugio da Vida Silvestre Banhado dosPachecos, Rio Grande do Sul, Brazil

Conductivity Diss. O2 pH Temp. (H20) Temp. (air) Photoperiod Rainfall

Mar 43.5 23 215 83.1

Apr 36.4 4.63 23.7 25.5 162.7 172.7

May 19.7 4.1 4.96 22.6 24 122.4 50.1

Jun 18.7 5.6 4.9 14.7 17.5 107.5 109.6

Jul 26.5 3.5 4.8 18.6 24 101.3 225.7

Aug 18.5 3.5 5.4 19.3 29 118.6 182.3

Sep 41.1 3.4 5.4 19.2 16.7 180.3 53

Oct 43.5 2.8 5.25 25.5 27 200 123.7

Nov 22.1 23.5 34 262 13.7

Dec 48.6 3.6 31.3 22.8 233.2 52.1

Jan 47.2 2.3 4.68 23 26 298.4 166

Feb 28 5.7 4.63 27.4 226.5 139.5

Environ Biol Fish (2017) 100:69–84 73

results for males in March/2011 and January/2012,since only females were sampled in these months.

The absolute fecundity of G. refugio had an averagevalue of 419 oocytes ranging from 0.15 mm −2.85 mmof diameter for females with total length varying from121.3 mm to 196.8 mm (Table 2). Average relativefecundity was estimated as 0.05 oocytes per mg totalweight. Analyses of absolute frequency distribution ofoocytes diameters show that the specie has oocyte de-velopment synchronic in more than two groups,iteroparity, and fractional spawning (Fig. 6).

The chi-square test (p < 0.05) did not show signifi-cant differences in total length related to sexual dimor-phism forG. refugio however, males reached the largest

length class established (Fig. 7). According to the sametest (p < 0.05), the sex ratio established for G. refugiowas 1:1 in each sampled month, and also in the totalsample.

Feeding

A total of 22 food items were identified in the 118stomach analysed. Five empty stomachs were regis-tered, one stomach in April, June, and January, andtwo stomachs in September. Empty stomachs were ex-cluded from the analysis. The most frequent items ob-served were larvae of diptera, odonata and trichoptera,plant material, and digested organic matter (Table 3).

Fig. 2 Monthly variation of thegonadal maturation phases offemales (a) and males (b) ofGymnotus refugio histologicallyanalysed

74 Environ Biol Fish (2017) 100:69–84

The distribution of mean values of RI (Fig. 8) showsignificant differences throughout the sampling periodonly for males (F = 10.35; p < 0.05), December differingof all the other months except of May, and Februarydiffering of August and September. There was no sig-nificant difference between the mean RI values of malesand females and the abiotic factors analyzed.

The PERMANOVA results indicate a seasonal patternin the species diet (F = 16.5 p = 0.001). The foodcategories were associated with neither ontogeny(F = 0.72 p = 0.49) nor sex (F = 2.05 p = 0.12). ThePCoA demonstrate a seasonal pattern in the distributionof alimentary categories from autumn to spring (Fig. 9).In addition, the IndVal values show significant ingestionof autochthonous insects during winter (Stat: 0.55p = 0.038), allochthonous insects during autumn (Stat:0.39 p = 0.025), and ingestion of digested organic matterduring spring (Stat: 0.57 p = 0.001). Summer was not

associated to any alimentary category in particular. Inaddition, the terms A and B (IndVal) indicate the occur-rence of autochthonous insects in the stomachs of allspecimens sampled on winter, although the category isnot season exclusive (A = 0.3; B = 1). The same is shownfor digested organic matter found in almost all stomachsin the spring (A = 0.36; B = 0.9). Allochthonous insectswere consumed mainly in autumn, not occurring in allstomachs analysed in the season (A = 0.76 B = 0.2). Theintestinal quotient for G. refugio was estimated as 0.29,with a standard deviation of 0.05.

Discussion

Similarly to other species of gymnotiforms inhabitingthe southern South America (Cognato and Fialho 2006;Giora and Fialho 2009; Schaan et al. 2009; Giora et al.

Fig. 3 Histological sections of Gymnotus refugio ovaries in dif-ferent gonadal maturation phases. (a) Immature. (b) Early devel-oping. (c) Mid developing. (d) Late developing. * oogonias, pg

primary growth oocytes, ca cortical alveolar oocytes, vtg1 primaryvitellogenic oocytes, vtg2 secundary vitellogenic oocytes, vtg3tertiray vitellogenic oocytes

Environ Biol Fish (2017) 100:69–84 75

2014), G. refugio presented a long reproductive period.In addition, the higher means of males GSI in the firstmonth of the established reproductive period suggeststhat males were ready for reproduction before females.The increasing size of gonads during the reproductiveperiod occupies larger space in the body cavity and mayreduce the space available for other organs such as thestomach, interfering on the feeding behavior (Vazzoler1996). However, both males and females of G. refugioexhibited a positive relation between the GSI and RIvalues, suggesting that the species might increase feed-ing activities during the reproductive cycle in order tocope with the energetic demand. Although no correla-tion was observed between the species GSI and theabiotic factors, the reproductive period matched monthswith the highest photoperiod and water temperatureregistered (INMET 2015). Similar results were

described for other gymnotiforms inhabiting southernBrazil and Uruguay, which shown reproduction posi-tively associated with both temperature and photoperiod(Silva et al. 1999; Silva et al. 2002; Ardanaz et al. 2001;Giora et al. 2012; Giora et al. 2014); therefore,G. refugio fits the pattern proposed for all gymnotiformsfrom high latitudes (Giora et al. 2014).

The histological analysis of both males and femalesgonads indicated individuals capable of spawning inthe same months defined as the species reproductiveperiod. Even though the reproductive period was welldefined for males, the histology of the testes demon-strated a fast and continuous spermatogenesis processthroughout the sampling period, impeding the differ-entiation of the actively spawning subphase integratingthe spawning capable phase (Brown-Peterson et al.2011). This fast and continuous spermatogenesis can

Fig. 4 Histological sections of Gymnotus refugio ovaries in dif-ferent gonadal maturation phases. (a) Spawning capable. (b) Ac-tively spawning. (c) Regressing. (d) Regenerating. * oogonias, pgprimary growth oocytes, ca cortical alveolar oocytes, vtg1 primary

vitellogenic oocytes, vtg2 secundary vitellogenic oocytes, vtg3tertiray vitellogenic oocytes, dt ovarian duct, bv blood vessels, afatresia; pof postovulatory follicle complexes

76 Environ Biol Fish (2017) 100:69–84

be proven by the presence of spermatozoa in almost allphases of the maturation cycle. The smallest male(63 mm) and female (59 mm) collected were both indifferent phases of gonadal development, with thefemale being immature and the male in its mid-development phase. This observation suggests thatthe process of gonadal maturation for G. refugio maystart when individuals reach lengths shorter than

60 mm, pointing to an earlier development of the malejuveniles. Although the samples were performed withspecific equipment for the capture of electric fishes,larvae agglomeration and juveniles were neither ob-served nor sampled in this study. Therefore, it ispossible to presume that the species uses the floodedinterior of the riparian forest, inaccessible for sam-pling, as a nursery for breeding and early development

Fig. 5 Histological sections of Gymnotus refugio testes in differ-ent gonadal maturation phases. (a) Mid developing. (b) Latedeveloping. (c) Sapawning Capable. (d) Regressing. (e)

Regenerating. g spermatogonia, c1 primary spermatocytes, c2secondary spermatocytes, t spermatids, ez spermatozoas, bv bloodvessels

Environ Biol Fish (2017) 100:69–84 77

of the juveniles. Although the production of spermdemands some energetic investment (Nakatsuru andKramer 1982) that would be much lower than femaleenergetic investment to produce oocytes. In that case,the results suggest that G. refugio males benefits of thecontinuous spermatogenesis and early maturation, be-ing ready for mate spawning capable females through-out a long period of time.

According to the obtained data, G. refugio has verylow relative fecundity. Available informationconcerning relative fecundity of gymnotiforms pointsout the species low fecundity value, Brachyhypopomusgauderio reaching 0.2 (Giora et al . 2014) ,Brachyhypopomus draco 0.17 (Schaan et al. 2009),Brachyhypopomus bombilla 0.21 (Giora et al. 2012),and Eigenmannia virescens 0.27 (Kirschbaum 1979), aswell as other Gymnotus species such as Gymnotus aff.carapo reaching 0.20 oocytes/mg body weight(Cognato and Fialho 2006). However, G. refugio shows

relative fecundity much lower than all these valuesregistered for gymnotiforms up until now. Accordingto Vazzoler (1996), fecundity depends on the availabil-ity of female body cavity and on the size of oocytesproduced. In line with these suggestions, the maximumdiameter of a vitellogenic oocyte of G. refugio wasestimated as 2.85 mm, whereas Cognato and Fialho(2006) estimated for Gymnotus aff. carapo a maximumdiameter of vitellogenic oocytes of 1.7 mm.

Parental care is a behavior that improves offspringfitness and is often related to low fecundity values, sincegreater parental investment in individual progeny im-prove juvenile survivorship (Winemiller 1987; Menezesand Vazzoler 1992). This behavior is largely describedamong gymnotiform species, with larvae guarding andnest building by males documented for different species(Assunção and Schwassmann 1995; Vazzoler 1996;Kirschbaum and Schugardt 2002; Quintana et al.2004; Crampton and Hopkins 2005; Cognato andFialho 2006; Giora and Fialho 2009). Although directevidence of parental care (i.e detection of larval agglom-eration protected by an adult, and/or nests) were notrecord in the field, the species low fecundity indicatethat the same behavior might be performed byG. refugio. Moreover, as similarly associated byCognato and Fialho (2006) for G. aff. carapo, the de-crease in the RI ofG. refugiomales after periods of highGSI, might be another indicator of the existence ofparental caring.

The species spawning analysis demonstrated the ex-istence of an actively spawning subphase in spawningcapable females, classifying the species as fractionalspawner. Moreover, the distribution of vitellogenic

Table 2 Total length (Lt), total weight (Wt), gonadosomatic index(GSI), absolute fecundity (Fa), and relative fecundity (Fr) of 7females of Gymnotus refugio

Lt (mm) Wt (g) IGS Fa Fr

121.32 4.64 11.1 286 0.06

135.33 5.56 12.4 330 0.05

148.91 8.06 11.4 409 0.05

164.84 9.87 8.2 452 0.04

165.74 8.16 9 238 0.02

170.36 11.06 7.7 518 0.04

196.81 9.13 10.9 698 0.07

Mean 157.62 8.07 10.1 419 0.05

Fig. 6 Distribution of relative frequency of 150 oocyte diametersof the seven females of Gymnotus refugio with the highest regis-tered GSI values

Fig. 7 Distribution of relative frequency (%) ofGymnotus refugiofemales and male for total length classes (mm). Numbers abovethe column represent the number of specimens in each length class

78 Environ Biol Fish (2017) 100:69–84

oocytes in different maturation classes indicates thatoocytes are released in different moments until the endof reproductive period, whereas store oocytes (primarygrowth oocytes) and vitellogenic oocyte were continu-ously observed through the maturation phases (Brown-Peterson et al. 2011). The same spawning strategy isdescribed for all gymnotiform species studied at thepresent time (Barbieri and Barbieri 1982; Assunçãoand Schwassmann 1995; Kirschbaum and Schugardt2002; Quintana et al. 2004; Crampton and Hopkins2005; Cognato and Fialho 2006; Giora and Fialho2009; Schaan et al. 2009; Giora et al. 2012; Gioraet al. 2014). Although fractional spawning can requirea higher reproductive effort than a single spawning (Burtet al. 1988), species with this strategy tend to be betteradapted to unfavorable environmental conditions and

may solve competition issues for spawning sites bet-ween the females of the same population (Nikolskii1969). Since the swampy forest is a habitat extremelyunstable, with seasonal great variance in water deep andconsequently micro-habitats availability, the fractionalspawning can cope with G. refugio population mainte-nance. Additionally, gymnotiform species have greatlyreduced coelomatic cavity, and fractional spawningstrategy can help to enhance fecundity under a spacerestriction. Therefore, the quoted strategies (i.e fraction-al spawning, long reproductive period, parental care andlow fecundity) agree with the life history characters of‘K-strategists’ species as formerly proposed by Pianka(1970). The sex ratio 1:1 defined for G. refugio is alsocommonly observed in natural population of fishes andmay diverge when one of the sexes has particular

Table 3 Frequency of occurrence (%) of alimentary items identified for Gymnotus refugio

Alimentary items Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb

Allochthonous insect 100 11.1 30 0 0 0 25 0 7.1 30 20 9

Acari 0 0 0 0 0 0 12.5 0 14.3 10 0 27.3

Coleopetra (adult) 0 11.1 20 0 0 0 0 0 0 10 0 0

Hymenoptera (adult) 100 0 0 0 0 0 0 0 7 0 0 9

Unidentified parts of insect 0 0 0 0 0 0 25 0 0 0 20 0

Autochthous insect 100 100 100 100 100 100 100 93.7 100 90 100 81.8

Coleoptera (larvae) 0 0 10 0 0 0 37.5 0 0 0 70 0

Diptera (larvae) 100 88.8 100 100 100 100 87.5 93.7 100 100 100 63.6

Diptera (pupae) 0 0 20 11 20 50 0 12.5 7 0 20 0

Ephemeroptera (larvae) 0 44.5 10 0 0 0 25 6.25 35.7 20 0 27.3

Hemiptera (larvae) 0 0 0 0 0 0 12.5 0 0 0 10 0

Odonata (larvae) 100 66.6 70 88.8 60 70 37.5 50 57 30 10 18.2

Plecoptera (larvae) 0 11 0 11 0 0 0 0 0 0 0 0

Trichoptera (larvae) 100 33.3 80 66.6 60 40 25 12.5 7 30 70 45.5

Unidentified parts of insect 0 0 0 0 0 0 0 0 0 10 57 18.2

Crustaceae 0 0 10 11 20 10 12.5 18.8 14.3 0 0 18.2

Amphipoda 0 0 0 11 20 10 11 18.8 0 0 0 18.2

Cladocera 0 0 0 0 0 0 0 0 14.3 0 0 0

Copepoda 0 0 10 0 0 0 0 0 0 0 0 0

DOM 100 44.4 30 22.2 0 90 87 81.3 100 70 70 91

Fish 0 0 0 0 0 0 0 0 0 10 0 0

Other 0 0 10 20 0 0 12.5 6.3 14.3 0 10 27.3

Fish egg 0 0 0 0 0 0 12.5 6.3 0 0 10 0

Insect egg 0 0 10 20 0 0 0 0 0 0 0 0

Plant material 0 55.5 20 22.2 30 60 75 56.3 43 50 20 9

Sediment 0 11 0 0 0 0 0 0 0 20 0 0

Total of stomachs 1 9 10 9 10 10 8 16 14 10 10 11

Environ Biol Fish (2017) 100:69–84 79

advantage (Reay 1984). On the other hand, when com-paring the body length of G. refugiomales and females,the highest length classes were reached by males.Similarly, species of different Gymnotiformes genus,such as Egeinmannia (Kirschbaum 1979; Giora andFialho 2009) and Brachyhypopomus (Giora et al.2012; Giora et al. 2014) had the same pattern of sexratio and total length distribution described for bothmales and females.

Even though the direct influence of abiotic fac-tors on the species diet was not significant, theseasonal variations on the diet of G. refugio mightbe a result of environmental changes defining eachseason. For example, the increasing RI of bothmales and females from May to August, matchesthe period in which the rainfall tend to be more

intense in the studied region (Nimer 1990).Consequently, the water level might rise facilitat-ing the predation on the most available items,which may include Diptera larvae and other im-mature invertebrates highly abundant in streams(Lowe-McConnel 1987; Higuti and Takeda 2002;Motta and Uieda 2004; Hahn and Fugi 2007;Uieda and Motta 2007). Moreover, the continuousdistribution of riparian forests surrounding thestreams contributes indirectly with the allochtho-nous items in fishes diet, serving as shelter andfood resource for the invertebrates preyed (Soares1979). Further, the frequent plant material found inthe stomachs could integrate the species diet, but itis also possible that this could be an item eventu-ally consumed in association with the capture of

Fig. 8 Monthly variation ofmean repletion index (RI) forGymnotus refugio females (a) andmales (b). Vertical bars representthe standard deviation. Numbersabove the bars correspond to thenumbers of specimens included inthe analysis

80 Environ Biol Fish (2017) 100:69–84

other preys. Nonetheless, the dryness in the sam-pling area during summer months might have hin-dered the species locomotion through the watercolumn affecting the foraging on specific preys,which might explain the association of summerwith no particular food category.

Gymnotiform fishes are characterized by a crepuscu-lar and nocturnal behavior, hunting actively at night(Mago-Leccia 1994; Albert and Crampton 2005).Given all sampling occurred at the daylight, part of theitems was already digested at the moment the specimenswere fixed, compromising the identification of fooditems, and explaining the high frequency of digestedorganic matter observed. Furthermore, although previ-ous studies (Albert and Crampton 2003) claim thatspecies of the genusGymnotus can prey on small fishes,there is no direct evidence of this behavior in studiesregarding the feeding biology of the species of the genus(Winemiller and Adite 1997; Mérigoux and Ponton1998; Penczak et al. 2000; Giora et al. 2005; Luz-Agostinho et al. 2006). For G. refugio, the only sign ofpiscivory observed is related to a single specimenamong all the individuals sampled, suggesting an op-portunistic predation on this alimentary item.

The low intestinal quotient estimated forG. refugio isin line with that proposed for species consuming ali-mentary resources of animal origin according to thescale proposed by Fryer and Iles (1972) to compare

trophic category and intestine length. According to theresults obtained concerning the diet composition of G.refugio, the species can be classified as invertivorouswith tendency to insectivory. In comparison to otherstudies regarding feeding habits of gymnotiform species(Winemiller and Adite 1997; Penczak et al. 2000; Gioraet al. 2005; Giora et al. 2012, 2014; Luz-Agostinho et al.2006) it is possible to conclude thatG. refugio integratesthe order pattern of diet as it was proposed by Gioraet al. (2014).

Anthropic activities, such as pollution, habitatdegradation, and the expansion of urban areas havecontributed to the depletion of fish populationsworldwide (Olden et al. 2007; Olden et al. 2010;Vitule et al. 2009; Abilhoa et al. 2011), increasingthe need to discover our biodiversity in order topreserve it (Braga et al. 2012). Rapid expansion ofagriculture over the past century has left an indeliblemark on the world as croplands, and livestock pas-tures now cover an area larger than many of Earth’snatural biomes (Foley et al. 2005). Due to its pres-ence restricted to areas of hydromorphic soil, theswampy forests are considered to be ecosystemsnaturally fragmented, with a blotch distributionmerged to grasslands and other forest formation(Teixeira and Assis 2009). Although gymnotiformspecies show diverse characteristics to cope withdifferent environmental conditions, the results obtain-ed herein are pointing to a very strong associationbetween G. refugio population and the swampy for-est habitat. The surrounding riparian forest can di-rectly influence the species ecology, providing ener-getic resources and serving, in its flooded interior, asa nursery and shelter during dry periods. The aquaticenvironment profile, with excessive aquatic vegeta-tion and low water deep also prevent the occurrenceof great size carnivorous fish (pers. obs.), benefitingK-strategist species as G. refugio. The largest num-ber of threatened freshwater species in Rio Grandedo Sul State (27 species; 67.5%) belong to thefamily Rivulidae (Bertaco et al. 2016), which includethe species known as annual fishes and live intemporary pools and swamps. These species are allconsidered threatened due to their restricted area ofoccurrence and advanced degree of loss and frag-mentation of habitats (Reis et al. 2003; FZB 2014).Therefore, the extreme dependence of G. refugio onits specific habitat can explain the status ofEndangered [EN] of the species, and can ensure that

Fig. 9 Diagram from principal coordinate analysis (PCoA) for theseasonal distribution of the following food categories identified forGymnotus refugio. Al allochthonous insects, Au autochthonousinsects, Crus Crustacea, DOM digested organic matter, Fis fish,Other other items, PM plant material, Sed sediment

Environ Biol Fish (2017) 100:69–84 81

conservation actions focused on the reduction ofanthropisation of its fragmented habitats in southernBrazil are of extreme importance for maintenance ofthe fish fauna.

Acknowledgements We are grateful to André Osorio Rosa,director of Refúgio da Vida Silvestre Banho dos Pachecos conser-vation unity for supporting the collect expedition in the area. Thisproject was funded by Programa Nacional de PósDoutorado—PNPD-CAPES (process 2282/09), and ConselhoNacional de Desenvolvimento Científico e Tecnológico—CNPq(process 476821/2003-7; 478002/2006-8).

Compliance with ethical standards Field work and samplingwere executed according to the Authorization for Scientific Activ-ities (number 25542–1) concede by the Sistema de Autorização eInformação em Biodiversidade − SISBIO of the Instituto ChicoMendes de Conservação da Biodiversidade − ICMBio. ICMBioAuthorization for Scientific Activities is obligatory for any re-search project that includes field work in the Brazilian territory.Ethical approval was obtained through the Comissão de Ética noUso de Animais − CEUA of the Universidade Federal do RioGrande do Sul − UFRGS, where the study was conducted.

References

Abilhoa V, Braga RR, Bornatowski H, Vitule JRS (2011) Fishes ofthe Atlantic rain Forest streams: ecological patterns andconservation. In: Grillo O (ed) Changing diversity in chang-ing environment. InTech, Rijeka, pp. 259–282

Adebisi AA (1987) The relationships between fecundities,gonadossomatics indices and egg sizes of some fishes ofOgun River, Nigéria. Archiv fuer Hydrobiology 111(1):151–156

Albert JS, Crampton WGR (2003) Seven new species of theNeotropical elect r ic f ish Gymnotus (Teleostei :Gymnotiformes) with redecoration of G. carapo(Linnaeus). Zootaxa 287:1–54

Albert JS, Crampton WGR (2005) Diversity and phylogeny ofNeotropical eletric fishes (Gymnotiformes). In: Bullock TE,Hopkins CD, Popper AN, Fay FR (eds) Electroreception.Springer, New York, pp. 360–409

Anderson MJ (2001) A new method for non-parametric multivar-iate analysis of variance. Austral Ecol 26(1):32–46

Ardanaz JL, Silva A, Macadar O (2001) Temperature sensitivity ofthe electric organ discharge waveform in Gymnotus carapo. JComp Physiol A 187:853–864. doi:10.1007/s0035900102568

AssunçãoMIS, SchwassmannHO (1995) Reproduction and larvaldevelopment of Electrophorus electricus on Marajó Island(Pará, Brazil). Ichth Expl Freshw 6:175–184

Barbieri G, Barbieri MC (1982) Fecundidade e tipo de desova deGymnotus carapo (Linnaeus, 1758), na represa do Lobo,Estado de São Paulo (Pisces, Gymnotidae). Spectr J BrasCiênc 2:25–29

Bertaco VA, Ferrer J, Carvalho FR, Malabarba LR (2016)Inventory of the freshwater fishes from a densely collected

area in South America —a case study of the current knowl-edge of Neotropical fish diversity. Zootaxa 4138(3):401–440

Borcard D, Gillet F, Legendre P (2011) Numerical ecology with R.Springer, New York

Braga RR, Bornatowski H, Vitule JRS (2012) Feeding ecology offishes: an overview of worldwide publications. Rev Fish BiolFish 22(4):915–929

Brown-Peterson NJ, Wyanski DM, Saborido-Rey F, MacewiczBJ, Lowerre-Barbieri SK (2011) A standardized terminologyfor describing reproductive development in fishes. Marineand Coastal Fisheries 3(1):52–70

Burt A, Kramer DL, Nakatsuru K, Spry C (1988) The tempo ofreproduction in Hyphessobrycon pulchripinnis (Characidae),with a discussion on the biology of ‘multiple spawning’infishes. Environ Biol Fish 22(1):15–27

Cognato DP, Fialho CB (2006) Reproductive biology of a popu-lation of Gymnotus Aff. carapo (Teleostei: Gymnotidae)from southern Brazil. Neotropical Ichthyology 4(3):339–348

Corrêa F, Oliveira EF, Tuchtenhagen T, Pouey J, Piedras S (2015)Ichthyofauna of the hydrographic basin of the Chasqueirostream (Mirim lagoon system, southern Brazil): generatingsubsidies for conservation and management. BiotaNeotropica 15(4):e0006. doi:10.1590/1676–0611–bn–2015–0006

Crampton WGR (1998) Effects of anoxia on the distribution,respiratory strategies and electric diversity of gymnotiformfishes. J Fish Biol 53:307–330. doi:10.1111/j.1095–8649.1998.tb01034.x

CramptonWGR, Hopkins CD (2005) Nesting and paternal care inthe weakly electric fish Gymnotus (Gymnotiformes:Gymnotidae) with descriptions of larval and adult electricorgan discharges of two species. Copeia 2005:48–60

CramptonWGR,Wells JK, Smyth C,Walz SA (2007) Design andconstruction of an electric fish finder. Neotrop Ichth 5:425–428. doi:10.1590/s167962252007000300022

Eschmeyer WN, Fong JD (2016) Catalog of Fishes. Availablefrom http://researcharchive.calacademy.org/research/ichthyology/catalog/SpeciesByFamily.as. Accessed 05May 2016

Ferrer J, Azevedo MA, Giora J, Cavalheiro LW, Wingert JM,Aguiar AR, Vargas ND, Santos JPM, Vanin AS, Fialho CB,Hartz SM, Malabarba LR (2015) Refúgio de Vida SilvestreBanhado dos Pachecos: um verdadeiro abrigo para aictiofauna relictual de mata paludosa na região metropolitanade Porto Alegre, RS, Brasil. Boletim, Sociedade Brasileira deIctiologia 1808–1436:8–15

Foley JA, DeFries R, Asner GP, Barford C, Bonan G, CarpenterSR (2005) Global consequences of landuse. Science309(5734):570–574

Fryer G, Iles TD (1972) The cichlid fishes of the great lakes ofAfrica. Oliver & Boyd, Edinburgh

FZB Fundação Zoobotânica do Rio Grande do Sul (2014)Avaliação do Estado de Conservação de Espécies da Fauna.Lista Vermelha da Fauna

Giora J, Fialho CB (2009) Reproductive biology of weakly electricfish Eigenmannia trilineata López & Castello, 1966(TELEOSTEI, Sternopygidae). Braz Arch Biol Technol52(3):617–628

Giora J, Malabarba LR (2016) Gymnotus refugio, a new andendangered species of electric fish of the Gymnotus

82 Environ Biol Fish (2017) 100:69–84

pantherinus species-group from southern Brazil(Gymnotiformes: Gymnotidae). Zootaxa 4066(5):581–590

Giora J, Fialho CB, Dufech APS (2005) Feeding habit ofEigenmannia trilineata Lopez & Castello, 1996 (Teleostei:Sternopygidae) of Parque Estadual de Itapuã, RS, Brazil.N e o t r o p I c h t h 3 : 2 9 1 – 2 9 8 . d o i : 1 0 . 1 5 9 0/s167962252005000200007

Giora J, Tarasconi HM, Fialho CB (2012) Reproduction andfeeding habits of the highly seasonal Brachyhypopomusbombilla (Gymnotiformes: Hypopomidae) from southernBrazil with evidence for a dormancy period. Environm BiolFish 94:649–662. doi:10.1007/s1064101199713

Giora J, Tarasconi HM, Fialho CB (2014) Reproduction andfeeding of the electric fish Brachyhypopomus gauderio(Gymnotiformes: Hypopomidae) and the discussion of a lifehistory pattern for gymnotiforms from high latitudes. PLoSOne 9(9):e106515. doi:10.1371/journal.pone.0106515

HahnNS, Fugi R (2007) A alimentação de peixes em reservatóriosbrasileiros: alterações e consequências nos estágios iniciaisdo represamento. Oecologia Brasiliensis 11(4):469–480

Hamidan N, Britton JR (2015) Reproductive ecology of Garraghorensis, a critically endangered fish in Jordan. EnvironBiol Fish 98:1399–1409

Higuti J, Takeda AM (2002) Spatial and temporal variation indensities of chironomid larvae (diptera) in two lagoons andtwo tributaries of the upper Paraná River floodplain, Brazil.Braz J Biol 62(4B):807–818

Hopkins CD (1974) Electric communication: functions in thesocial behavior of Eigenmannia virescens. Behaviour 50:270–305. doi:10.1163/156853974x00499

Hyslop EJ (1980) Stomach contents analisys; a review of methodsand their application. J Fish Biol 17:411–429. doi:10.1111/j.1095-8649.1980.tb02775.x

IBGE Instituto Brasileiro de Geografia e Estatística (2010)Indicadores de desenvolvimento sustentável. Estudos ePesquisas, Informação Geográfica n. 7

Instituto Nacional de Meteorologia- INMET [WWW Document](2015) Ministério da Agricultura, Pecuária e Abastecimento(Accessed 30.Oct.15)

Kirschbaum F (1975) Environmental factors control the periodicalreproduction of tropical electric fish. Experientia 31:1159–1160. doi:10.1007/bf02326767

Kirschbaum F (1979) Reproduction of the weakly electric fishEigenmannia virescens (Rhamphichtyidae, Teleostei) in cap-tivity. Behav Ecol Sociobiol 4:331–355. doi:10.1007/bf00303241

Kirschbaum F (2000) The breeding of tropical freshwater fishesthrough experimental variation of exogenous parameters.Breedin through simulation of high and lowwater conditions.Aquageografia 20:95–105

Kirschbaum F, Schugardt C (2002) Reproductive strategies anddevelopmental aspects in mormyrid and gymnotiform fishes.J Physiol Paris 96(2002):557–566. doi:10.1016/s09284257(03)000111

Lowe-McConnel RH (1987) Ecological studies in tropical fishcommunities. University Press, Cambridge

Luz-Agostinho KDG, Bini LM, Fugi R, Agostinho AA, Júlio HF Jr(2006) Food spectrum and trophic structure of the ichthyofau-na of Corumbá reservoir, Paraná river basin, Brazil. NeotropIchth 4(1):61–68. doi:10.1590/s167962252006000100005

Malabarba LR, Neto PC, Bertaco VA, Carvalho TP, Dos Santos JF,Artioli LGS (2013) Guia de identificação dos peixes da baciado rio Tramandaí. Via Sapiens, Porto Alegre

Mago-Leccia F (1994) Electric fishes of the continental waters ofAmérica. Caracas, Clemente editores, p207

Menezes NA, Vazzoler AEADM (1992) Reproductive character-istics of Characiformes. In: Hamlett WC (ed) Reproductivebiology of south American vertebrates. Springer, New York,pp. 60–70

Mérigoux S, Ponton D (1998) Body shape, diet and ontogeneticdiet shifts in young fish of the Sinnamary River, FrenchGuiana, South America. J Fish 52:556–569

Motta RL, Uieda VS (2004) Diet and trophic groups of an aquaticinsect community in a tropical stream. Brazil J Biol 64:809–817. doi:10.1590/S1519-69842004000500010

Mugnai R, Nessimian JL, Baptista DF (2010) Manual deidentificação de macroinvertebrados aquáticos do Estado doRio de Janeiro: para atividades técnicas, de ensino etreinamento em programas de avaliação da qualidadeecológica dos ecossistemas lóticos. Technical BooksEditora, Rio de Janeiro

Nakatsuru K, Kramer DL (1982) Is sperm cheap? Limited malefertility and female choice in the lemon tetra (Pisces,Characidae). Science 216(4547):753–755

Nikolskii GV (1969) Theory of fish population dynamics asthe biological background for rational exploitation andmanagement of fishery resources. Oliver & Boyd Ltda,Edinburgh, p. 323

Nimer E (1990) Clima. Geografia do Brasil: Região Sul.Portuguese. FIBGE, Rio de Janeiro

Olden JD, Poff NL, Bestgen KR (2006) Life-history strategiespredict fish invasions and extirpations in the Colorado Riverbasin. Ecol Monogr 76(1):25–40

Olden JD, Hogan ZS, Zanden MJV (2007) Small fish, bigfish, red fish, blue fish: size-biased extinction risk ofthe world’s freshwater and marine fishes. Glob EcolBiogeogr 16:694–701

Olden JD, Kennard MJ, Leprieur F, Tedesco PA, Winemiller KO,García-Berthou E (2010) Conservation biogeography offreshwater fishes: recent progress and future challenges.Divers Distrib 16:496–513

Oliveira MDLLA, Balbueno RA, Senna RM (2005)Levantamento florístico de fragmentos florestais na baciahidrográfica do rio Gravataí, Rio Grande do Sul, Brasil.Iheringia Série botânica 60(2):269–284

Penczak T, Agostinho AA, Hahn N (2000) Na ordination tech-nique for fis diet comparison. Bras Achiv Biol Techn 43(1):101–101

Pianka ER (1970) On r-and K-selection. Am Nat 104(940):592–597Quintana L, Silva A, Berois N, Macadar O (2004) Temperature

induces gonadal maturation and affects electrophysiologicalsexual maturity indicators in Brachyhypopomuspinnicaudatus from a temperate climate. J Exper Biol 207:1843–1853. doi:10.1242/jeb.00954

Reay PJ (1984) Reproductive tatics: a non-event in aquaculture?In: Potts GW,WootonMN (eds) Fish reproduction: strategiesand tatics. Academic, London, pp. 291–309

Reis RE, Lucena ZMS, Lucena CAS, Malabarba LR (2003)Peixes. In: Fontana CS, Bencke GA, Reis RE (eds) Livrovermelho da fauna ameaçada de extinção no Rio Grande doSul. Edipucrs, Porto Alegre, pp. 117–145

Environ Biol Fish (2017) 100:69–84 83

Santos EP (1978) Dinâmica de populações aplicada à pesca episcicultura. São Paulo, Edusp, p. p129

Schaan AB, Giora J, Fialho CB (2009) Reproductive biology ofthe Neotropical electric fish Brachyhypopomus draco(Teleostei: Hypopomidae) from southern Brazil.Neotropical Ichthyology 7(4):737–744

Silva A, Quintana L, GaleanoM, Errandonea P, Macadar O (1999)Water temperature sensitivity of EOD waveform inBrachyhypopomus pinnicaudatus. J Comp Physiol A 185:187–197. doi:10.1007/s003590050377

Silva A, Quintana L, Ardanaz JL, Macadar O (2002)Environmental and hormonal influences upon EOD wave-form in gymnotiform pulse fish. J Physiol 96:473–484.doi:10.1016/s09284257(03)000032

Soares MGM (1979) Aspectos ecológicos, alimentação ereprodução dos peixes do igarapé do Porto, Aripuanã, MT.Acta Amaz 9:325–352

Suzuki HI, Agostinho AA (1997) Reprodução de peixes doreservatório de Segredo. In: Agostinho AA, Gomes LC(eds) Reservatório de Segredo: Bases ecológicas para omanejo. Eduem, Maringá, pp. 163–181

Teixeira AP, Assis MA (2009) Relação entre heterogeneidadeambiental e distribuição de espécies em uma floresta

paludosa no Município de Cristais Paulista, SP, Brasil. ActaBotanica Brasilica 23:843–853

Uieda VS, Motta RL (2007) Trophic organization and food webstructure of southeastern Brazilian streams: a review. ActaLimnologica Brasiliensia 19(1):15–30

Vazzoler AEAM (1996) Biologia da reprodução de peixesteleósteos: teoria e prática. Editora da Universidade,Maringá, p. 169

Vieira S (1991) Introdução à Bioestatística, 2nd edn. Ed. Campus,Rio de Janeiro, p. 294

Vitule JRS, Freire CA, Simberloff D (2009) Introduction ofnon-native freshwater fish can certainly be bad. FishFish 10:98–108

Winemiller KO (1987) Tests of ecomorphological and communitylevel convergence among neotropical fish assemblages.Dissertation, University of Texas

Winemiller KO, Adite A (1997b) Convergent evolution ofweakly electric fishes from floodplain habitats inAfrica and South America. Environm Biol Fish 49:175–186

Winemiller KO, Adite A (1997a) Convergent evolution of weaklyelectric fishes from floodplain habitats in Africa and SouthAmerica. Environm Biol Fish 49:175–186

84 Environ Biol Fish (2017) 100:69–84