Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Ecologically relevant biomarkers reveal that chronic effects of nitrate depend on sex and life stage in the invasive fish Gambusia holbrooki

  • Oriol Cano-Rocabayera ,

    Roles Conceptualization, Formal analysis, Funding acquisition, Investigation, Methodology, Writing – original draft

    Affiliation Department of Evolutionary biology, Ecology and Environmental Sciences–Institute of Research in Biodiversity (IRBio-UB), Universitat de Barcelona, Spain

  • Adolfo de Sostoa,

    Roles Conceptualization, Funding acquisition, Project administration, Supervision, Writing – review & editing

    Affiliation Department of Evolutionary biology, Ecology and Environmental Sciences–Institute of Research in Biodiversity (IRBio-UB), Universitat de Barcelona, Spain

  • Francesc Padrós,

    Roles Methodology, Supervision, Validation, Writing – review & editing

    Affiliation Servei de Diagnòstic Patològic en Peixos, Facultat de Veterinària, Universitat Autònoma de Barcelona, Bellaterra, Spain

  • Lorena Cárdenas,

    Roles Formal analysis, Investigation, Methodology

    Affiliation Department of Evolutionary biology, Ecology and Environmental Sciences–Institute of Research in Biodiversity (IRBio-UB), Universitat de Barcelona, Spain

  • Alberto Maceda-Veiga

    Roles Conceptualization, Funding acquisition, Investigation, Supervision, Validation, Writing – review & editing

    Affiliation Department of Evolutionary biology, Ecology and Environmental Sciences–Institute of Research in Biodiversity (IRBio-UB), Universitat de Barcelona, Spain


Agricultural intensification and shifts in precipitation regimes due to global climate change are expected to increase nutrient concentrations in aquatic ecosystems. However, the direct effects of nutrients widely present in wastewaters, such as nitrate, are poorly studied. Here, we use multiple indicators of fish health to experimentally test the effects of three ecologically relevant nitrate concentrations (<10, 50 and 250 mg NO3-/l) on wild-collected mosquitofish (Gambusia holbrooki), a species widely introduced for mosquito biocontrol in often eutrophic waters. Overall, biomarkers (histopathology, feeding assays, growth and caloric content and stable isotopes as indicators of energy content) did not detect overt signs of serious disease in juveniles, males or females of mosquitofish. However, males reduced food intake at the highest nitrate concentration compared to the controls and females. Similarly, juveniles reduced energy reserves without significant changes in growth or food intake. Calorimetry was positively associated with the number of perivisceral fat cells in juveniles, and the growth rate of females was negatively associated with δ15N signature in muscle. This study shows that females are more tolerant to nitrate than males and juveniles and illustrates the advantages of combing short- and long-term biomarkers in environmental risk assessment, including when testing for the adequacy of legal thresholds for pollutants.


Nutrient pollution results in man-made eutrophication, which is amongst the most pernicious forms of global change affecting aquatic ecosystems around the world [1,2]. Human causes of eutrophication are the inefficient use of fertilizers, aquaculture and urban outflows and atmospheric nitrogen deposition from combustion [35]. The ecological effects of eutrophication are well-known, including toxic algal blooms and high mortality of animals due to dissolved oxygen depletion at night [68]. Water authorities attempt to mitigate eutrophication by establishing safe nutrient concentrations (e.g. OECD, 1982; Directive 91/676/ECC [9,10]). However, the direct toxicity of nutrients to wildlife under chronic exposure is still poorly studied [11,12]. Considering agricultural intensification continues unabated and water purification is costly [13], there is the pressing need to get better insight into the health effects that environmentally relevant nutrient concentrations have on wildlife.

Nitrate (NO3-) is a widely distributed nutrient that naturally occurs at a low environmental concentration [3]. However, it can reach up to 2000 mg NO3-/l in aquaculture tanks and 345 mg NO3-/l in surface waters in nitrate vulnerable zones [14,15]. From 2012 to 2015 the surface area vulnerable to nitrate pollution increased from 1951898 km2 to 2175861 km2 just in Europe, representing 61% of the total agricultural area [16]. Alongside surface waters, nitrate pollution degrades groundwater, with reported concentrations of more than 395 mg NO3-/l [17], which exceeds the legal thresholds for Europe (50 mg NO3-/l; Directive 91/676/ECC [10]) and U.S. (44 mg NO3-/l; USEPA SWDA [18]). Ground and surface waters are linked, buffering groundwater against shortages of surface water during drought [7]. Moreover, climate change may intensify the effects of nitrate pollution on temperate rivers if shifts in precipitation regimes increase agricultural run-off [2].

Nitrate toxicity has long attracted the attention of public health agencies after nitrate-induced oxidation of respiratory pigment (methemoglobinemia) was recorded in U.S. babies [19]. Studies have since reported diseases other than respiratory issues in humans and in laboratory and domestic animals after drinking nitrate-polluted water, including mortality, oxidative stress, hypertension, birth defects, diabetes, impaired thyroid function, spontaneous abortions or cancer [11,20,21]. For water-breathing animals, nitrate was generally considered of little concern, possibly because nitrate has low branchial permeability compared to the highly toxic ammonia and nitrite [22,23]. This view changed after experimental evidence showed methemoglobinemia and alterations in hormone levels, behaviour, growth or in vulnerability to diseases in aquatic taxa under chronic nitrate exposure [11,24,25]. However, these studies used eggs, juveniles or adults of one sex of different species, all of which are factors that may affect the toxic response [26]. Moreover, toxic responses can be delayed, so that the combined use of short- (e.g. feeding assays) and long-term biomarkers (e.g. growth) will provide a more holistic view of nitrate toxicity to wildlife than the often used single-type biomarker approach [27].

The eastern mosquitofish (Gambusia holbrooki) is one of the world’s worst piscine invaders, which has been introduced in many temperate regions due to a misguided strategy for mosquito control [28]. Although extensively used in ecotoxicology, nitrate toxicity to mosquitofish has not been examined in detail. Reduced sperm counts and increased testicular weight in male mosquitofish were associated with concentrations of up to 22 mg NO3-/l in U.S. streams [29]. However, there is no experimental evidence for other nitrate-induced alterations in mosquitofish. The effects of nitrate on other fish species are negative [11,24,30], almost neutral [31] and even protective against a disease [32]. Nevertheless, these studies were mostly conducted on captive-reared species, which may have a higher tolerance to nitrate than wild fish because nitrate accumulates in aquaculture tanks and nitrate pre-exposure increases tolerance [33]. This rationale may apply to wild taxa if pre-exposure to the many pollutants occurring in natural waters induces co-tolerance [34].

The present experimental study monitored over 8 weeks the effects of three ecologically relevant nitrate concentrations on wild males, females and juveniles of mosquitofish using endpoints associated with their ecological impact. If males are the sicker sex [35,36,37] and juveniles are more vulnerable than adults to pollution [26,38], then we expected female mosquitofish to be the most tolerant to nitrate pollution. If the effects of nitrate pollution are subtle, then we expected nitrate effects to be more apparent in short- than in long-term biomarkers. Finally, responses in nitrate treatments should be comparable to those in controls if mosquitofish can cope with nitrate toxicity. Given the ecological relevance of the biomarkers used, our work will explore whether the ecological impact of the mosquitofish can be modulated by changing water-nutrient concentrations.

Materials and methods

Fish origin and general fish maintenance

The male and female mosquitofish used in this study were captured with dip nets in November 2012 in channels draining an agricultural area in the Llobregat river, Barcelona, Spain (41°16’52”N, 2°02’04”E). Fish were brought to the University of Barcelona in opaque plastic tanks provided with air-pumps and were acclimatised for one week to the laboratory conditions in two mixed sex stock 500 L tanks provided with an external filter, artificial plants and flowerpots for refugee. Fish were maintained in acclimation and experimental conditions as follows. A malaquite green/formaline bath was applied at a prophylactic dose upon arrival (see [39]). Water was then fully renewed by using dechlorinated tap water as we did to maintain the experimental environmental conditions (see section 2.3). Water properties in the laboratory tap were: pH = 7.7, mg/l, ammonia <0.5 mg/l, nitrite <0.03 mg/l, nitrate = 7.4 mg/l, sulphates = 81.2 mg/l, chloride = 130 mg/l, bicarbonate = 221 mg/l and conductivity = 784 μS/cm. Pregnant females (N = 15) with overt signs of giving birth soon were introduced in batches of three in 100 L tanks provided with nets to collect recently newborn juveniles (N = 165). Both adults and newborn were kept under 22±1°C and 12 h light:12 h dark cycle and fed daily with crushed commercial Sera Vipan flakes and weekly with frozen bloodworms for adults and live Artemia nauplii for newborns. Fish were fed once daily until satiety and uneaten food and faeces were removed daily with a dipnet. Each tank had a biological filter to prevent metabolic waste built-up (NH4+ and NO2-) and ensure water oxygenation.

Ethics statement

The experimental procedure was authorised by the Natural Environment and Biodiversity Division at the Catalan Department of Agriculture and Fisheries (Num. DAAM 8290). Fish capture and maintenance were approved by the Committee for an Ethical use of Experimental Animals at the University of Barcelona (Num. 87/15). All fish were humanely euthanized on the termination of the experiment in compliance with Spanish legislation for the management of invasive species (Real Decreto 1628/2011 [40]).

Experimental nitrate concentrations and exposure conditions

Sodium nitrate (NaNO3, CAS Number: 7631-99-4) was used to make two nitrate solutions (50 and 250 mg NO3-/l, equivalent to 11.5 and 57 mg NO3N/l, respectively) using dechlorinated tap water, which was also used in the control treatment (<10 mg NO3-/l). The lowest nitrate concentration is the safety nitrate threshold for European waters (Directive 91/676/ECC [10]) and the highest level is within the range reported in aquaculture [24] and in rivers draining nitrate vulnerable zones in Europe [41] and tropical countries [42]. Experimental nitrate concentrations represented a 0, 5- and 25-fold increase, respectively, for mosquitofish in relation to the nitrate concentration at the collection site (9.9 ± 3.0 mg NO3-/l, based on quarterly water analyses over one year).

Male and female mosquitofish were visually size-matched per sex (total length, female: TL = 37.6 ± 0.4 and male: 25.9 ± 0.2 mm) and exposed for 8 weeks to the experimental nitrate solutions in 20 L aquaria (N = 5 tanks per treatment and sex) with six males or females in each replicate. For juveniles, the same experimental setting was used but these were exposed in batches of 11 siblings per tank. The exposure started by increasing nitrate concentrations in each aquarium drop by drop (~3 h) via a 5 mm ø tube connected to a tank with clean water or one of the two nitrate solutions. This drop-by-drop system was used to refill each tank with fresh water from each experimental condition after 50% of water was changed every three days. Water samples randomly analysed from the different treatments using the colorimetric kit VISOCOLOR indicated that the water quality conditions remained constant through the experiment at 24 h of the next water change (Table 1).

Table 1. Water quality properties measured in the experimental tanks.

Overview of the biomarkers measured to appraise nitrate toxicity

The effects of chronic nitrate exposure on males, females and juveniles of mosquitofish were examined using 13 variables (Table 2). All of them are indicators of fish health, but calorimetry and stable isotope signatures inform the quality of fish for piscivores (see section 2.4.2). Alterations in the quantity of food eaten or in the reaction time to a stimulus show how nitrate may alter mosquitofish performance in ecosystems. Indicators related to fish growth, body condition based on mass-length relationships, energetic content and histopathology were recorded on the termination of the experiment (8 weeks). However, the feeding behaviour of mosquitofish was monitored at 0, 4, 6 and 8 weeks. All fish were euthanized at 8 weeks using an overdose of the anaesthetic MS-222.

Table 2. List of the biomarkers used to assess the effects of nitrate on mosquitofish (Gambusia holbrooki).

Growth and body condition based on mass-length relationships.

Male and female mosquitofish were anesthetised with MS-222 (0.02%), measured (TL, mm) and weighted (0.001g) at Time 0 and at 8 weeks. Newborns other than those used in the experiment were used to estimate the size of experiment juveniles at Time 0 (TL = 8.9 ± 1.2 mm) to avoid compromising the health of the tiny tested individuals due to handling.

Fish size measures were used to calculate the specific fish growth rate (G) using the equation G = (ln Lt−ln L0) / tn [43], where ln Lt is the natural logarithm of fish length at 8 weeks, ln L0 is that of fish length at Time 0 and tn is the duration of the experiment (8 weeks). We ranked fish in each tank by body length at Time 0 and tn to identify fish individuals and be able to calculate G because our tagging equipment (e.g. elastomer) is not suitable for such small fish. Moreover, we calculated the Scaled Mass Index (SMI) as an index of body condition: SMI = Wi (Lo/Li)bSMA [44], where Wi and Li are the weight and length of each fish individual, respectively, L0 is the arithmetic mean length of all the tested mosquitofish and bSMA is the slope of a standardised major axis regression of the mass-length relationship. The SMI is regarded as a correlate of energy and fitness measures [44].

Energetic reserves.

Changes in fish δ13C and δ15N stable isotope signatures and calorimetry were used as two complementary measures of energetic reserves. Stable isotopes are widely used in studies of trophic ecology because the isotopic composition of predator tissue is naturally altered by the type and amount of food assimilated [45]. Moreover, tissue isotopic differences are due to changes in metabolism, including increased lipid storage [46,47], because lipids are about 6–7% depleted in 13C relative to protein [48]. We used the δ13C and δ15N ratio as proxy for lipid content in white muscle because this is the tissue widely used in fisheries (e.g. [47,49]). However, lipid content varies amongst fish tissues [50], so that we used the caloric content of the whole fish as an additional measure of energy content.

For stable isotope analyses, we freeze-dried fish muscle samples from below the dorsal fin of 3 adult fish from each tank and we ground them to fine power. Two sub-samples of 0.30 mg each were placed into tin buckets and crimped for combustion to determine δ13C and δ15N using a Flash EA1112 and TC/EA coupled to a stable isotope mass spectrometer Delta C through a Conflo III interface (ThermoFinnigan). Analytical accuracy was controlled using replicate assays of certified standards indicating an analytical error of ±0.1‰ and ±0.3‰ for δ13C and δ15N, respectively. Isotope ratios are expressed conventionally as δ values in ppt (‰) according to the following equation: δX = ((Rsample/Rstandard)– 1) ·1000, where X (‰) is 13C, 15N, and R are the corresponding ratios 13C/12C 15N/14N, related to the standard values: R standard for 13C is Pee Dee Belemnite, for 15N is atmospheric nitrogen.

For calorimetry, we used three adults per tank and we pooled all juveniles from each tank to reach the detection limits of the IKA Calorimeter c7000 (Germany). Samples were oven-dried at 60ºC for 48 h, weighted and the caloric content was expressed as joules per gram (J/g).


A random sample of three fish from each tank was processed for histology. The head and viscera of adult fish were fixed individually in 10% buffered formalin, dehydrated in ethanol, cleared in xylene and embedded in paraffin wax [51]. Juveniles were processed as a whole due to small size.

Sagittal sections of 5 μm thick were cut in all fish at the same position and stained with conventional haematoxylin-eosin [51]. Observations were made under an Olympus BH light microscope at 400x magnification. We focused on liver and gills because nitrate uptake is through gills [52] and liver is the target of many toxicants [27]. Gill alterations and the number of mucous cells in the gills were recorded and expressed as a ratio out of the 50 secondary lamellae and spaces examined. Liver alterations and the number of melanomacrophague centres were expressed as a ratio out of the number of microscope fields examined. We took photographs with a Jenoptik ProgRes C3 camera and used the ImageJ software to quantify the area (μm2) of perivisceral fat in 3 sections of each juvenile fish as an additional measure of energy storage. Outcomes were expressed as average per juvenile.

Feeding behaviour.

Temporal changes in feeding behaviour were quantified in all experimental fish, which were not fed 24 h before the assay to ensure all fish were hungry. The assay was conducted in an aquarium with a plastic sheet in one side, where the tested fish were left for 3 minutes to acclimatise before the trial (Fig 1). The trial began when the sheet was gently removed and ten size-matched unfrozen Artemia adults in 200 ml of water were gently released into the experimental area on the opposed side of the tested fish. Live Artemia nauplii were used for juveniles. The assay was repeated individually for all fish and the same observer, sat in front of the aquarium, recorded: i) latency defined as the time spent to capture the first item, ii) voracity defined as the time needed to capture the first items (four in adults and ten in juveniles), and iii) satiety defined as the total number of items eaten without stopping > 2 minutes. If a fish ate all 10 prey items, then we added Artemia individuals on the water surface until satiety.

Fig 1. Experimental setting for the feeding behaviour assay of adults and juveniles of mosquitofish.

a) 10 prey items are offered to an isolated adult mosquitofish. b) Live brine shrimp nauplii are offered with a syringe to an isolated juvenile mosquitofish in a tray. After adding food (t0) we quantified: feeding latency time (time to capture the first prey, tLAT), voracity time (time to capture 4 preys in adults and 10 in juveniles, tVOR) and satiety.

Statistical analyses

All analyses were conducted using the R software [53] and the functions outlined below. Spearman rank correlation coefficients were used to examine associations amongst all biomarker responses. The effects of chronic nitrate exposure on mortality, fish growth, SMI, caloric content, the C/N ratio, histopathological and behavioural measures were assessed using generalized linear mixed models (GLMMs). Data on males, females and juveniles were analysed separately because there were significant differences in all response variables (S5 Table). To avoid pseudoreplication, aquarium ID was included as random intercept in all models to account for the fact that fish were exposed to nitrate in batches (6 adults or 11 juveniles). For juveniles, aquarium ID was nested within mother ID as random intercept to account for systematic differences amongst clutches. Nitrate was included as fixed effect in all models. The interaction between nitrate and time was included in the behaviour model to test whether the effects of nitrate on fish varied with exposure time. The distribution of all response variables was visually inspected and the error distribution in GLMMs was chosen accordingly (e.g. Gaussian for fish growth, Poisson for the number of prey eaten). Model assumptions were checked by inspecting diagnostic plots of residuals [54]. The function Anova within the package car [55] was used to assess significance at P ≤ 0.05.


Mosquitofish were evenly distributed by size and sex amongst treatments and fish did not differ in size amongst aquaria at Time 0, either for males (F = 0.54; P = 0.90) or females (F = 1.22; P = 0.28). However, females (Mean ± S.E. = 37.6 ± 0.42 mm) were significantly bigger than males (25.9 ± 0.18 mm; t = 25.8; P < 0.001). There was no mortality in males due to nitrate and only minor mortality was recorded for juveniles (3.6%) and females (3.3%). Nonetheless, mortality did not differ significantly between treatments (Nitrate: χ2 = 1.59, P = 0.66) or sexes (Sex: χ2 = 0.43, P = 0.93).

Growth, body condition and energetic reserves

An eight-week nitrate exposure did not alter significantly the growth rate or body condition, as defined by the SMI, of males, females and juveniles (Table 3). The caloric content of juveniles at 50 mg NO3-/l and in the controls at <10 mg NO3-/l was markedly higher than at 250 mg NO3-/l (Table 3). However, males significantly increased in caloric content at 50 mg NO3-/l compared to the other two concentrations. Females showed no significant differences amongst treatments (Table 3). Nitrate also did not affect the δ13C and δ15N measures in the white muscle of all tested fish (Table 3).

Table 3. Biomarkers used to appraise the health status (energy content, mass–length measures, histopathology) of males, females and juveniles of mosquitofish (Gambusia holbrooki) exposed to three experimental nitrate concentrations (<10 mg/l, 50 mg/l, 250 mg/l).


A detailed examination of all slides did not reveal overt clinical signs of disease, but some tissular changes were observed in liver and gills (Table 3, S2 and S3 Figs). Occasional telangiectasia and slight epithelial lifting were observed in secondary lamellae. However, we did not observe other tissue alterations such as hyperplasia, hypertrophy or increased number of mucous cells. All liver samples had regularly aligned cords of hepatocytes (S3 Fig). However, slight changes in the staining intensity of the cytoplasm were observed in liver tissue, probably related to glycogen deposits and occasional lipid droplets. The presence of macrophage aggregates was restricted to adult fish, with females having a larger number than males, but without significant changes due to nitrate (Table 3). The amount of perivisceral adipose tissue in juveniles at the highest nitrate concentration was lower than in juveniles in the other treatments (Table 3, S4 Table).

Feeding behaviour

Males, females and juveniles showed differences in latency, voracity and satiety, but without significant changes due to nitrate apart from males from the sixth week onwards (Fig 2, S1 Table). Males at 50 and 250 mg NO3-/l exhibited lower satiety values and lower voracity (i.e. > time to capture prey) than those in the controls at <10 mg NO3-/l (Fig 2). Overall, females and juveniles had a higher voracity and satiety than males, and juveniles had greater latency times than adults (Fig 2). Females and juveniles also tended to increase voracity and satiety and to reduce latency times throughout the experiment compared to males (Fig 2).

Fig 2. Feeding behaviour variables along the experiment.

Symbols and bars represent means ± 95% confidence intervals for each variable and assigned for each treatment (black: control, red: 50 mg NO3-/l, blue: 250 mg NO3-/l). a) Latency is the time spent to capture the first food item. b) Voracity is the time to capture a given number of food items. c) Satiety is the total number of eaten food items.

Pair-wise correlations amongst biomarkers

Spearman rank correlation coefficients were generally low amongst all biomarkers measured in juveniles, males and females of mosquitofish. There was a strong positive correlation between calorimetry and growth (ρ = 0.73, P = 0.002) and between calorimetry and the amount of perivisceral fat cells in juveniles (ρ = 0.72, P = 0.003, S6 Table). However, a marked negative correlation was found between satiety and latency time of males (ρ = -0.78, P <0.001, S7 Table). A strong negative association was observed in females for d15N and growth (ρ = -0.74, P = 0.002), even though the relationship was positive between C/N and growth (ρ = 0.77, P <0.001, S8 Table).


This is the first comprehensive study examining the chronic effects of nitrate on a widely introduced fish species, as exemplified by the eastern mosquitofish (Gambusia holbrooki) [28]. Moreover, this is one of the few ecotoxicological studies using short- and long-term biomarkers (e.g. growth, histopathology, feeding assays) in females, males and juveniles of the same species. Overall, we did not find overt clinical signs of disease, which supports the prevailing idea that many invasive species, including mosquitofish, have wide tolerance to changes in water quality [28,56]. However, the fact that nitrate altered food intake or energetic reserves in males and juveniles suggests that concentrations >50 mg NO3-/l cannot be considered completely safe [11,21].

Many studies have shown that males are more likely to acquire diseases than females, including fish [37,5759]. We did not find gross pathological alterations in any fish, but the more marked effects of nitrate on males provide some support for males being the sicker sex [35,36]. The weaker response of the tested fish to nitrate is unlikely to be much attributed to pre-acclimation to nitrate at the collection site (9.9 ± 3.0 mg NO3-/l), as reported for amphibians [33]. However, this outcome does not exclude the possibility of nitrate tolerance being increased due to pre-exposure to other ions, including those of water hardness (see [60]), which is high at the collection site and in the laboratory dechlorinated tap water due to a calcareous geology. The effects of metals and chlorine compounds on fish in the animal facility probably were negligible because tap water is filtered through active charcoal and the aquarium product Sera Aquatan is used to further guarantee water is free of metals and chlorine (see [39]). The fact that juveniles had a greater tolerance to nitrate than males was unexpected because young fish are generally more sensitive to chronic pollution than adults [26,38]. However, 96-h LC50 tests revealed that susceptibility to nitrate increases with body size in the Siberian sturgeon Acipenser baeri [61]. Although we cannot reveal the mechanisms for the mild effects of nitrate on juveniles because nitrate metabolites in tissues were not measured (e.g. nitric oxide [62]), differences in space and behaviour between adults and juveniles might partially explain outcomes. Adults were kept at lower densities (6 fish per tank) than juveniles (11 fish per tank) and not surprisingly, we observed more agonistic interactions among adults due to confinement.

Mosquitofish populations are often confined in small water bodies and female-biased [63,64], including in the collection site of the studied fish (authors pers. observ.). The biased sex-ratio has been attributed to the high life-span of females compared to males [65]. We built on this knowledge by showing that females may dominate in number because they are more tolerant than males to polluted waters, where mosquitofish often occur (e.g. [66]). Female mosquitofish had higher feeding rates than males regardless of the nitrate treatment, which is consistent with previous data in clean water [64]. Given that higher nitrogen excretion rates have been reported in females [64], it is possible that tolerance to environmental nitrate can be predicted from nitrogen excretion rates in fish. Nonetheless, sensitivity to nitrate probably depends on many factors in wild fish, including temperature, predation, and the fact that a parasite with more severe effects on males than females [37] is more sensitive to nitrate than the fish host [32]. In contrast to mainstream literature in which external factors other than pollutants are often not included in toxicological assays [27], our study accounted for intraspecific interactions; that is, several individuals were exposed to nitrate in the same tank instead of fish being exposed individually. Agonistic interactions probably are amongst the most important factors to explain G. holbrooki performance alongside sex because females are more aggressive towards conspecifics than males [67].This might explain why more females died during the experiment than males, although mortality did not differ significantly amongst treatments.

Feeding traits were more affected by nitrate than other biomarkers measured in male mosquitofish, which supports that food intake is amongst the most sensitive biomarkers in ecotoxicology [68,69]. Although we cannot identify the mechanisms, it might be a response-mediated by stress hormones (e.g. cortisol) because high levels of these hormones often reduce appetite in fish and other animals [70]. However, cortisol levels remained stable in females of Siberian sturgeon (Acipenser baeri) after 30-day exposure to 250 mg NO3-/l, as opposed to the reproductive hormones testosterone and estradiol [24]. Moreover, there is correlative evidence for reduced sperm count in male mosquitofish at < 22 mg NO3-/l [29]. These studies illustrate that nitrate is an endocrine disruptor through in-vivo conversion to nitric oxide, which is involved in many metabolic pathways [71], suggesting that it is possible that the effects of nitrate on fish probably would have been stronger than observed if we had used biochemical biomarkers. Nevertheless, many biochemical alterations often do not have far reaching impacts on individuals, reason for which they are considered of less ecological relevance than behavioural assays, including the feeding traits we measured [27].

Even though fish differed in food intake amongst nitrate treatments, we did not observe overt signs of disease, including reduced fish growth. Reduced food ingestion in males may be attributed to fatigue because nitrate forms methaemoglobin, which transports oxygen worse than haemoglobin [19]. However, fish can cope with moderate methaemoglobinemia [72], especially in hard water, such as ours in the laboratory, which may have mitigated nitrate adverse effects [60]. Growth was expected to decrease in mosquitofish because iodine uptake, which is needed for thyroid functions and animal development, is altered by nitrate, but concentrations up to 11 mg NO3-/l did not impair the thyroid function in perch (Perca fluviatilis) and Crucian carp (Carassius carassius) [73]. The neutral effect of nitrate we saw on mosquitofish growth agrees with Freitag et al. [74], who found that concentrations up to 450 mg NO3-/l had no effect on the thyroid hormone levels in Atlantic salmon (Salmo salar). Our outcome is also consistent with studies in other freshwater taxa showing that nitrate effects on growth and survival occur at > 500 mg NO3-/l (e.g. [31,7577]). However, the neutral effect of nitrate on mosquitofish does not exclude the possibility that fish exposed to nitrate may reduce their ability to cope with other pollutants if nitrate alters the internal ionic composition of fish at an osmoregulation cost and probably impairs important enzymatic complexes, such as those involved in detoxification [62].

Histopathological analyses revealed no relevant tissue alterations because slight epithelial lifting and other alterations we saw are not pathological but tissue processing artifacts (see [78]). Changes in caloric content only matched with histological data for juveniles at 250 mg NO3-/l, which reduced energy reserves as also exemplified by peripheral fat content, but no major changes in δ13C and δ15N measures of muscle occurred in fish from any treatment combination. These findings suggest that no single tissue can be a good proxy of overall fish energy reserves because they vary greatly amongst tissues [50]. Moreover, our findings confirmed that no biomarker, including mass-length relationship indices such as the SMI, can be assumed to accurately reflect ‘true condition’ without analysing body composition [27,44]. The lack of response of the SMI may be attributed to the fact that nitrate did not markedly change fish weight, possibly because although fish reduced food intake, fish were fed daily until satiety. However, the biochemical composition of fish tissues might have changed due to nitrate because pollutants often alter tissue stoichiometry [52] with potential far reaching impacts for fish predators. In this regard, juvenile mosquitofish altered energy content in tissues, but food intake or growth were not affected, which suggests that growth is prioritised over lipid storage, probably to reduce size-dependent predation mortality [79].

In our study, energy costs are likely to be mostly attributed to intraspecific interactions and osmoregulation due to nitrate. Osmoregulation cost probably was caused mainly by the anion nitrate (NO3-, see [80]) and, to a minor degree, by the cation sodium (Na+) of the salt (NaNO3). The sodium concentration in the highest nitrate concentration was below 1 parts per thousand (ppt) and there is no experimental evidence for major changes in mosquitofish metabolism at 20 ppt [81] or in mosquitofish plasma osmotic concentration at 10 ppt [82]. Surprisingly, we found that the caloric content of males was higher at 50 mg NO3-/l than at 250 mg NO3-/l and in the controls at <10 mg NO3-/l. Reduced caloric content in males at 250 mg NO3-/l compared to 50 mg NO3-/l can be due to osmoregulation cost increasing nitrate concentration. However, this rationale does not explain why males at 250 mg NO3-/l had a similar caloric content to those in the controls at <10 mg NO3-/l, although the latter had the highest feeding rates in the study. Less energy stored implies that males at <10 mg NO3-/l had an additional cost than those at 50 mg NO3-/l, which may be the courtship display. Control males were reproductive active and we observed copulation attempts with other males, a behaviour that often occurs in male poeciliids in the absence of females [83]. Courtship display has an energetic cost [84,85], which probably was reduced at 50 mg NO3-/l because nitrate, even at lower concentrations, reduces testosterone [62,86] and this hormone promotes the sexual characteristics of males.


Our study shows that females of the invasive fish G. holbrooki are more tolerant to nitrate pollution than males and juveniles, but that there are all weak effects combining short- and long-term biomarkers. Therefore, the ecological impact of this invasive fish seems not to be much affected by nitrate pollution, especially if populations are female-biased. However, our study cannot inform the indirect effects that nitrate may have on G. holbrooki through the alteration of aquatic food-webs, including a possible reduction in prey numbers accompanied by impaired food intake in males. There is the pressing need for an–omics screening (e.g. transcriptomics) to identify simultaneously all the metabolic pathways that are altered in fish exposed to nitrate in order to improve the mechanistic understanding of the effects of this widely distributed subsidy and pollutant in aquatic ecosystems.

Supporting information

S1 Table. Mixed models analysis of variance of satiety, latency to eat and voracity of juveniles, males and females along the experiment.


S2 Table. Mixed models analysis of variance of calorimetry and stable isotopes of juveniles, males and females at the end of the experiment.


S3 Table. Mixed models analysis of variance of growth and body condition of juveniles, males and females at the end of the experiment.


S4 Table. Mixed models analysis of variance of histopathology variables of juveniles, males and females at the end of the experiment.


S5 Table. Mixed models analysis of variance of all biomarkers with sex and length as principal explicative variables.


S6 Table. Spearman rank correlation coefficients examining the associations amongst all biomarkers in juveniles.


S7 Table. Spearman rank correlation coefficients examining the associations amongst all biomarkers in males.


S8 Table. Spearman rank correlation coefficients examining the associations amongst all biomarkers in females.


S9 Table. Mean (±Standard Error) of the variables used to appraise the feeding rates of males, females and juveniles of mosquitofish.


S1 Fig. Histological samples of the abdominal cavity.


S2 Fig. Histological samples of the gill tissue examined.


S3 Fig. Histological samples of the hepatic tissue.


S1 File. Compressed file including separate datasets for juveniles, females and males data.



We thank the staff at the animal house facilities, especially Jordi Guinea. We also thank lab technician Cèlia Arcarons for histology sample processing, as well as Mercè Durfort and Jordi Correas for histology methods guidance. Bomb calorimetry analysis was done with the guidance of Xavier Remesar and Rosa M. Marimon. We also thank Pili Rubio for stable isotopes processing, Helena Satorras and Mònica Utjés for image editing and Mario Monroy for support in fish collection and lab procedures. We would also like to thank the two anonymous reviewers for their suggestions on this text.


  1. 1. Smith VH, Tilman GD, Nekola JC. Eutrophication: impacts of excess nutrient inputs on freshwater, marine, and terrestrial ecosystems. Environmental Pollution 1999; 100(1–3): 179–196. pmid:15093117
  2. 2. Sinha E, Michalak AM, Balaji V. Eutrophication will increase during the 21st century as a result of precipitation changes. Science 2017; 357(6349): 405–408. pmid:28751610
  3. 3. Galloway JN, Schlesinger WH, Levy H, Michaels A, Schnoor JL. Nitrogen fixation: Anthropogenic enhancement‐environmental response. Global Biogeochemical Cycles 1995; 9(2): 235–252.
  4. 4. Puckett LJ. Identifying the major sources of nutrient water pollution. Environmental Science & Technology 1995; 29(9): 408A–414A.
  5. 5. Fields S. Global nitrogen: cycling out of control. Environmental Health Perspectives 2004; 112(10): A556–A563. pmid:15238298
  6. 6. Kolkwitz R, Marsson M. Ökologie der tierischen Saprobien. Beiträge zur Lehre von der biologischen Gewässerbeurteilung. Internationale Revue der Gesamten Hydrobiologie und Hydrographie 1909; 2(1–2): 126–152.
  7. 7. Margalef R. Limnología. Barcelona: Ediciones Omega; 1983.
  8. 8. Smith VH, Schindler DW. Eutrophication science: where do we go from here? Trends in ecology & evolution 2009; 24(4): 201–207. pmid:19246117
  9. 9. OECD. Eutrophication of waters: monitoring, assessment and control. Final report. Paris, France: Organisation for Economic Cooperation and Development; 1982.
  10. 10. European Commission. Council Directive 91/676/EEC of 12 December 1991 concerning the protection of waters against pollution caused by nitrates from agricultural sources. Official Journal of the European Communities 1991; 375, 1–8.
  11. 11. Camargo JA, Alonso Á. Ecological and toxicological effects of inorganic nitrogen pollution in aquatic ecosystems: a global assessment. Environment international 2006; 32(6): 831–849. pmid:16781774
  12. 12. Moore AP, Bringolf RB. Effects of nitrate on freshwater mussel glochidia attachment and metamorphosis success to the juvenile stage. Environmental Pollution 2018; 242: 807–813. pmid:30032077
  13. 13. Lee DR, Barrett CB. Tradeoffs or synergies? Agricultural intensification, economic development and the environment. CABI; 2001.
  14. 14. Honda H, Watanabe Y, Kikuchi K, Iwata N, Takeda S, Uemoto H, et al. High density rearing of Japanese flounder, Paralichthys Olivaceus with a closed seawater recirculation system equipped with a denitrification unit. Aquaculture Science 1993; 41(1): 19–26.
  15. 15. Agència Catalana de l’Aigua. 2014 Sep 20. In: Catalan Water Agency. Interactive applications [Internet]. Available from:
  16. 16. European Commission. 2018 Sep 20. In: Report from the commission to the council and the European Parlament on the implementation of Council Directive 91/676/EEC concerning the protection of waters against pollution caused by nitrates from agricultural sources based on member states reports from the period 2012–2015 [Internet]. Available from:
  17. 17. Grup de Defensa del Ter. 2018 Sep 20. In: Nitrate concentration in springs of Osona [Internet]. Available from:
  18. 18. EPA US. The Safe Drinking Water Act. Public law 1974; 93–523.
  19. 19. Craun GF, Greathouse DG, Gunderson DH. Methaemoglobin levels in young children consuming high nitrate well water in the United States. International journal of epidemiology 1981; 10(4): 309–317. pmid:7327829
  20. 20. Rodríguez-Estival J, Martínez-Haro M, Martín-Hernando MP, Mateo R. Sub-chronic effects of nitrate in drinking water on red-legged partridge (Alectoris rufa): Oxidative stress and T-cell mediated immune function. Environmental Research 2010; 110(5): 469–475. pmid:20382380
  21. 21. Poulsen R, Cedergreen N, Hayes T, Hansen M. Nitrate: An Environmental Endocrine Disruptor? A Review of Evidence and Research Needs. Environmental science & technology 2018; 52(7): 3869–3887. pmid:29494771
  22. 22. Thurston RV, Russo RC, Vinogradov GA. Ammonia toxicity to fishes. Effect of pH on the toxicity of the unionized ammonia species. Environmental science & technology 1981; 15(7): 837–840.
  23. 23. Jensen FB. Uptake and effects of nitrite and nitrate in animals. Nitrogen metabolism and excretion. In: Walsh PJ, Wright P, editors. Nitrogen metabolism and excretion. Boca Raton: CRC Press; 1995. pp 289–303.
  24. 24. Hamlin HJ, Moore BC, Edwards TM, Larkin IL, Boggs A, High WJ, et al. Nitrate-induced elevations in circulating sex steroid concentrations in female Siberian sturgeon (Acipenser baeri) in commercial aquaculture. Aquaculture 2008; 281(1–4): 118–125.
  25. 25. Secondi J, Lepetz V, Cossard G, Sourice S. Nitrate affects courting and breathing but not escape performance in adult newts. Behavioral ecology and socio-biology 2013; 67(11): 1757–1765.
  26. 26. Férard JF, Blaise C. Encyclopedia of Aquatic Ecotoxicology. Dordrecht, Netherlands:Springer; 2013.
  27. 27. Colin N, Porte C, Fernandes D, Barata C, Padrós F, Carrassón M, et al. Ecological relevance of biomarkers in monitoring studies of macro-invertebrates and fish in Mediterranean rivers. Science of the Total Environment 2016; 540: 307–323. pmid:26148426
  28. 28. Pyke GH. Plague minnow or mosquito fish? A review of the biology and impacts of introduced Gambusia species. Annual review of ecology, evolution, and systematic 2008: 39(1): 171–191.
  29. 29. Edwards TM, Guillette LJ Jr. Reproductive characteristics of male mosquitofish (Gambusia holbrooki) from nitrate-contaminated springs in Florida. Aquatic Toxicology 2007; 85(1): 40–47. pmid:17767965
  30. 30. Hord NG, Beaver L, Axton E, Truong L, St. Marie L, Tanguay R, et al. Nitrate and nitrite exposure affect cognitive behavior and oxygen consumption during exercise in zebrafish. The FASEB Journal 2017; 31(1_supplement): lb277.
  31. 31. Pereira A, Carvalho AP, Cruz C, Saraiva A. Histopathological changes and zootechnical performance in juvenile zebrafish (Danio rerio) under chronic exposure to nitrate. Aquaculture 2017; 473: 197–205.
  32. 32. Smallbone W, Cable J, Maceda-Veiga A. Chronic nitrate enrichment decreases severity and induces protection against an infectious disease. Environment international 2016; 91: 265–270. pmid:26995268
  33. 33. Johansson M, Räsänen K, Merilä J. Comparison of nitrate tolerance between different populations of the common frog, Rana temporaria. Aquatic Toxicology 2001; 54(1–2): 1–14. pmid:11451421
  34. 34. Coutellec MA, Barata C. An introduction to evolutionary processes in ecotoxicology. Ecotoxicology 2011; 20(3): 493–496. pmid:21416110
  35. 35. Bateman AJ. Intra-sexual selection in Drosophila. Heredity 1948; 2: 349–368. pmid:18103134
  36. 36. Rolff J. Bateman’s principle and immunity. Proceedings of the Royal Society B: Biological Sciences 2002; 269(1493): 867–872. pmid:11958720
  37. 37. Stephenson JF, Kinsella C, Cable J, Van Oosterhout C. A further cost for the sicker sex? Evidence for male‐biased parasite‐induced vulnerability to predation. Ecology and Evolution 2016; 6(8): 2506–2515. pmid:27066240
  38. 38. McKim JM. Evaluation of tests with early life stages of fish for predicting long-term toxicity. Journal of the Fisheries Research Board of Canada 1977; 34(8): 1148–1154.
  39. 39. Noga EJ. Fish disease: diagnosis and treatment. John Wiley & Sons; 2011.
  40. 40. Gobierno de España. Real Decreto 1628/2011, de 14 de noviembre, por el que se regula el listado y catálogo español de especies exóticas invasoras. BOE 2011; 298.
  41. 41. Bouraoui F, Grizzetti B, Aloe A. Nutrient discharge from rivers to seas for year 2000. JRC Scientific and Technical Reports, European Commission; 2009.
  42. 42. WHO. Guidelines for drinking-water quality: first addendum to third edition. Vol. 1, Recommendations. World Health Organization; 2006.
  43. 43. Hopkins KD. Reporting fish growth: A review of the Basics. Journal of the world aquaculture society 1992; 23(3): 173–179.
  44. 44. Peig J, Green AJ. New perspectives for estimating body condition from mass/length data: the scaled mass index as an alternative method. Oikos 2009; 118(12): 1883–1891.
  45. 45. Post DM. Using stable isotopes to estimate trophic position: models, methods, and assumptions. Ecology 2002; 83(3): 703–718.
  46. 46. Gannes LZ, O’Brien DM, Del Rio CM. Stable isotopes in animal ecology: assumptions, caveats, and a call for more laboratory experiments. Ecology 1997; 78(4): 1271–1276.
  47. 47. Gaye-Siessegger J, Focken U, Muetzel S, Abel H, Becker K. Feeding level and individual metabolic rate affect δ13C and δ15N values in carp: implications for food web studies. Oecologia 2004; 138(2): 175–183. pmid:14608500
  48. 48. Focken U, Becker K. Metabolic fractionation of stable carbon isotopes: implications of different proximate compositions for studies of the aquatic food webs using δ13C data. Oecologia 1998; 115(3): 337–343. pmid:28308424
  49. 49. Cano-Rocabayera O, Maceda-Veiga A, de Sostoa A. Fish fins and scales as non-lethally sampled tissues for stable isotope analysis in five fish species of north–eastern Spain. Environmental biology of fishes 2015; 98(3): 925–932.
  50. 50. Wallace PD, Hulme TJ. The fat/water Relationship in the Mackerel, Scomber scombrus (L.), pilchard, Sardina pilchardus (Walbaum) and sprat, Sprattus sprattus (L.) and the seasonal variations in fat content by size and maturity. Directorate of Fisheries Research, Lowestoft 1977; 35: 1–10
  51. 51. Durfort M. Iniciació a les tècniques histològiques vegetals i animals. Barcelona: Institut d’Estudis Catalans; 2006.
  52. 52. Stormer J, Jensen FB, Rankin JC. Uptake of nitrite, nitrate, and bromide in rainbow trout,(Oncorhynchus mykiss): effects on ionic balance. Canadian Journal of Fisheries and Aquatic Sciences 1996; 53(9): 1943–1950.
  53. 53. R Core Team. 2018 Sep 20. In: R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria [software]. Available from:
  54. 54. Zuur AF, Ieno EN, Walker N, Saveliev AA, Smith GM. Mixed effects models and extensions in ecology with R. New York: Springer; 2009.
  55. 55. Fox J, Weisberg S. In: An R companion to applied regression, 2nd edition [software]. Thousand Oaks CA: Sage; 2011. Available from:
  56. 56. Marchetti MP, Moyle PB, Levine R. Alien fishes in California watersheds: characteristics of successful and failed invaders. Ecological Applications 2004; 14(2): 587–596.
  57. 57. Nakazawa T, Hamaguchi S, Kyono-Hamaguchi Y. Histochemistry of liver tumors induced by diethylnitrosamine and differential sex susceptibility to carcinogenesis in Oryzias latipes. Journal of the National Cancer Institute 1985; 75(3): 567–573. pmid:2993731
  58. 58. Perkins EJ, Griffin B, Hobbs M, Gollon J, Wolford L, Schlenk D. Sexual differences in mortality and sublethal stress in channel catfish following a 10 week exposure to copper sulfate. Aquatic toxicology 1997; 37(4): 327–339.
  59. 59. Klein SL, Flanagan KL. Sex differences in immune responses. Nature Reviews Immunology 2016; 16(10): 626–638. pmid:27546235
  60. 60. Baker JA, Gilron G, Chalmers BA, Elphick JR. Evaluation of the effect of water type on the toxicity of nitrate to aquatic organisms. Chemosphere 2017; 168: 435–440. pmid:27810544
  61. 61. Hamlin HJ. Nitrate toxicity in Siberian sturgeon (Acipenser baeri). Aquaculture 2006; 253: 688–693.
  62. 62. Edwards TM, Hamlin HJ. Reproductive endocrinology of environmental nitrate. General and Comparative Endocrinology 2018; 265: 31–40. pmid:29577898
  63. 63. Vargas MJ, de Sostoa A. Life history of Gambusia holbrooki (Pisces, Poeciliidae) in the Ebro delta (NE Iberian peninsula). Hydrobiologia 1996; 341(3): 215–224.
  64. 64. Fryxell DC, Arnett HA, Apgar TM, Kinnison MT, Palkovacs EP. Sex ratio variation shapes the ecological effects of a globally introduced freshwater fish. Proceedings of the Royal Society B: Biological Sciences 2015; 282(1817): 20151970. pmid:26490793
  65. 65. Haynes JL, Cashner RC. Life history and population dynamics of the western mosquitofish: a comparison of natural and introduced populations. Journal of Fish Biology 1995; 46(6): 1026–1041.
  66. 66. Maceda-Veiga A, Baselga A, Sousa R, Vilà M, Doadrio I, de Sostoa A. Fine-scale determinants of conservation value of river reaches in a hotspot of native and non-native species diversity. Science of the Total Environment 2017; 574: 455–466. pmid:27644023
  67. 67. Smith CC, Sargent RC. Female fitness declines with increasing female density but not male harassment in the western mosquitofish, Gambusia affinis. Animal Behaviour 2006; 71: 401–407.
  68. 68. Kestemont P, Baras E. Environmental factors and feed intake: mechanisms and interactions. In: Houlihan D, Boujard T, Jobling M, editors. Food intake in fish. Blackwell Science Ltd; 2001. pp. 131–156.
  69. 69. Lall SP. Disorders of nutrition and metabolism. In: Leatherland JF, Woo TF, editors. Fish diseases and disorders, Volume 2. University of Guelph, Canada: CABI; 2010. pp. 202–237.
  70. 70. Bernier NJ. Food intake regulation and disorders. In: Leatherland JF, Woo TF, editors. Fish diseases and disorders, Volume 2. University of Guelph, Canada: CABI; 2010. pp. 238–266.
  71. 71. Guillette LJ Jr, Edwards TM. Is Nitrate an Ecologically Relevant Endocrine Disruptor in Vertebrates? Integrative and Comparative Biology 2005; 45(1): 19–27. pmid:21676740
  72. 72. Eddy FB, Williams EM. Nitrite and freshwater fish. Chemistry and Ecology 1987; 3(1): 1–38.
  73. 73. Lahti E, Harri M, Lindqvist OV. Uptake and distribution of radioiodine, and the effect of ambient nitrate, in some fish species. Comparative Biochemistry and Physiology Part A: Physiology 1985; 80(3): 337–342.
  74. 74. Freitag AR, Thayer LR, Leonetti C, Stapleton HM, Hamlin HJ. Effects of elevated nitrate on endocrine function in Atlantic salmon, Salmo salar. Aquaculture 2015; 436: 8–12.
  75. 75. Hrubec TC, Smith SA, Robertson JL. Nitrate toxicity: a potential problem of recirculating systems. Northeast regional agricultural engineering service Ithaca—Publications–NRAES 1996; 1(98): 41–48.
  76. 76. Scott G, Crunkilton RL. Acute and chronic toxicity of nitrate to fathead minnows (Pimephales promelas), Ceriodaphnia dubia, and Daphnia magna. Environmental Toxicology and Chemistry 2000; 19(12): 2918–2922.
  77. 77. Shimura R, Ijiri K, Mizuno R, Nagaoka S. Aquatic animal research in space station and its issues—focus on support technology on nitrate toxicity-. Advances in space research 2002; 30(4): 803–808. pmid:12530377
  78. 78. Wolf JC, Baumgartner WA, Blazer VS, Camus AC, Engelhardt JA, Fournie JW, et al. Nonlesions, Misdiagnoses, Missed Diagnoses, and Other Interpretive Challenges in Fish Histopathology Studies: A Guide for Investigators, Authors, Reviewers, and Readers. Toxicologic Pathology 2015; 43: 297–325. pmid:25112278
  79. 79. Martin BT, Heintz R, Danner EM, Nisbet RM. Integrating lipid storage into general representations of fish energetic. Journal of Animal Ecology 2017; 86: 812–825. pmid:28326538
  80. 80. Jensen FB. Nitrite and red cell function in carp: control factors for nitrite entry, membrane potassium ion permeation, oxygen affinity and methaemoglobin formation. Journal of Experimental Biology 1990; 152(1): 149–166.
  81. 81. Akin S, Neill WH. Routine metabolism of mosquitofish (Gambusia affinis) at three different salinities. Texas Journal of Science 2003; 55: 255–226.
  82. 82. Nordlie FG, Mirandi A. Salinity relationships in a freshwater population of eastern mosquitofish. Journal of Fish Biology 1996; 49: 1226–1232.
  83. 83. Field KL, Waite TA. Absence of female conspecifics induces homosexual behaviour in male guppies. Animal Behaviour 2004; 68: 1381–1389.
  84. 84. Christiansen JS, Jorgensen EH, Jobling M. Oxygen consumption in relation to sustained exercise and social stress in Arctic Charr (Salvelinus alpinus L.). Journal of Experimental Zoology 1991; 260(2): 149–156.
  85. 85. Haller J. Biochemical background for an analysis of cost-benefit interrelations in aggression. Neuroseience and Biobehavioral Reviews 1995; 19(4): 599–604.
  86. 86. Panesar NS, Chan KW. Decreased steroid hormone synthesis from inorganic nitrite and nitrate: studies in vitro and in vivo. Toxicology and Applied Pharmacology 2000; 169: 222–230. pmid:11133344