http://dx.doi.org/10.15446/abc.v19n1.38265

FOOD WEB OF A TROPICAL HIGH MOUNTAIN STREAM: EFFECTS OF NUTRIENT ADDITION

Red trófica de un arroyo de montaña tropical: efectos de la adición de nutrientes

MARÍA ISABEL CASTRO-REBOLLEDO1, Ph. D.; ISABEL MUÑOZ-GRACIA, Ph. D.2; JOHN CHARLES DONATO-RONDÓN, Ph. D.1

1 Laboratorio Diatomeas. Universidad Nacional de Colombia. Facultad de Ciencias, Departamento de Biología. micastror@unal.edu.co. jcdonator@unal.edu.co

2 Departamento de Ecología. Universidad de Barcelona. imunoz@ub.edu. Send correspondence for: micastror@unal.edu.co

Received 27 May 2013, first decision 21 October 2013, accepted 18 November 2013.
Citation / Citar este artículo como: CASTRO-REBOLLEDO MI, MUÑOZ-GRACIA I, DONATO-RONDÓN JC. Food web of a tropical high mountain stream:
effects of nutrient addition. Acta biol. Colomb. 2014. 19(1):33-42.

In order to define the effect of nutrient enrichment on trophic webs in an Andean mountain stream we performed an experiment using stable isotope ratios (δ15N and δ13C) to analyze different trophic compartments: 1) basal level: CPOM and biofilm; 2) primary consumers - macroinvertebrates: collector-gatherers (Heterelmis sp., Thraulodes sp. and Trichorythodes sp.), and collectorfilterers (Simulium sp.); 3) predators - fish (Oncorhynchus mykiss and Trichomycterus bogotensis). The average fractionation of nitrogen among the primary consumers with respect to CPOM was 4.7 ‰, and 1.7 ‰ with respect to biofilm. Predators incremented their δ15N signal by 5.9 % with respect to primary consumers. A depletion of δ15N was observed in Impact with respect to Control reach after fertilization in different compartments (biofilm, Heterelmis, Simulium andTricorythodes), while depletion was not significant for top predators. In most cases, the δ13C signal of biofilm overlapped with that of primary consumers, but a clear enrichment was observed with respect to CPOM. Macroinvertebrate gut contents showed fine detritus to be their most abundant food, and that in general there were no changes in diet as a consequence of nutrient enrichment. The only exception was Heterelmis, who increased its consumption in the Impact reach.

Keywords: food webs, gut content, isotope ratios, nutrient enrichment, trophic compartments.

RESUMEN

Con el fin de determinar el efecto del incremento de nutrientes sobre la red trófica de un río de montaña andino, se realizó un experimento en donde se analizaron las proporciones de isotopos estables (δ15N y δ13C) para analizar los siguientes compartimientos tróficos: 1) Nivel basal: CPMO y Biofilm; 2) Consumidores primarios - macroinvertebrados: colectores - recolectores (Heterelmis sp., Thraulodes sp., yTrichorythodes sp.) y colectores - filtradores (Simulium sp.); 3) Depredadores - peces (Oncorhynchus mykiss y Trichomycterus bogotensis). La fracción promedio de nitrógeno entre los consumidores primarios con respecto a CPOM fue de 4.7 ‰, y de 1.7 ‰ con respecto al biofilm. Los depredadores incrementaron en un 5.9 % la señal δ15N con respecto a los consumidores primarios. Después de la fertilización, se observó en diferentes compartimientos (biofilm, Heterelmis, Simulium yTricorythodes) del tramo impacto un agotamiento de δ15N con respecto al control, mientras que el agotamiento no fue significativo para los depredadores superiores. En la mayoría de los casos la señal δ13C del biofilm se sobrepuso con la de los consumidores primarios pero un claro enriquecimiento fue observado con respecto a CPOM. Los macroinvertebrados referidos fueron seleccionados para analizar su contenido estomacal y los resultados nos mostraron que el detritus fino es el alimento más abundante para los invertebrados, y únicamente, Heterelmis mostró, después de la adición de nutrientes, diferencias significativas en el detritus fino y el material vegetal entre el tramo control y el impactado.

Palabras clave: compartimientos tróficos, contenido estomacal, enriquecimiento de nutrientes, proporción de isotopos, redes tróficas.

INTRODUCTION

Food webs are complex trophic connections among interacting organisms in ecosystems (Elser and Hessen, 2005), and their structure influences population dynamics, community structure and ecosystem function (Polis et al., 1997). Knowledge of the food web in freshwater systems is essential to integrate the dynamics of organic matter with organism interactions. The maximum food-chain length is an important food-web property that is correlated with resource availability, ecosystem size, environmental stability and colonization history (Doi, 2011). Some of these correlations may result from environmental effects on predator-prey mass ratios (Jennings, 2005). In streams, the structure of food webs is affected by numerous factors, such as biogeography, stream order, disturbance, temperature, resource type and anthropogenic activities (Hershey et al., 2007). Light availability and nutrient levels are often regarded as the most important factors influencing primary production in streams. Moreover, nutrient excess is one of the most common disturbances affecting river ecosystems, through "bottom-up" effects to the whole community structure (Biggs and Smith, 2002).

Resources in streams are usually represented by detritus and primary producers that are always associated with fungi, bacteria and micro- and meio-fauna. Invertebrates can thus be both primary and secondary consumers. Predators are usually omnivorous with a mixed diet of prey, detritus and algae. These characteristics produce confusing results that make it difficult to clearly determinate the trophic position of animals.

Stable isotope analysis has proved to be a useful tool in reconstructing diets, elucidating patterns of resource allocation, characterizing trophic relationships and constructing food webs (Boecklen et al., 2011), thus providing a measurement of trophic position that integrates the assimilation of energy or mass flow through all the different trophic pathways leading to an organism (Post, 2002). Carbon and nitrogen stable isotopes are frequently used to study energy sources and food web structure in ecosystems (Bergfur et al., 2009), as well as to show which processes or components are more sensitive to perturbation (Peterson and Fry, 1987). Nitrogen isotopes are useful for differentiating trophic levels and food-web dynamics (Bergfur et al., 2009) because δ15N isotopic fractioning increases with each trophic level (Finlay, 2001) and a consumer is typically enriched by 3-4 ‰ relative to its diet (Jardine et al., 2012). In contrast, the ratio of carbon isotope (δ13C) changes little (0.3-0.5 ‰ on average) as carbon moves through food webs (Peterson and Fry, 1987), what makes it is an effective diet tracer (Finlay et al., 2002). Analysis of δ13C signature has an advantage over gut-content analysis because it measures the amount of carbon assimilated from each food source as opposed to that ingested (March and Pringle, 2003). However, as mentioned, due to high overlapping in the diets of stream organisms, gut contents provide basic and complementary information about food sources.

MATERIALS AND METHODS

Experimental Design

The two study reaches were located in the Tota stream, within the area of the municipality of Cuítiva, department of Boyacá (Colombia).

In order to define the effect of nutrient enrichment on trophic webs, we chose two 50 m reaches that were geo-morphologically and hydrologically similar: Control (C) and Impact (I). Both reaches were studied for 12 months prior to the enrichment and ten months after. The C reach was located 700 m upstream from the I reach, where the nutrients were added. The continuous addition of nutrients was performed using a drop system over a 10-month period. Two commercial grain fertilizers (Nitron 26 (26-0-0) and Abocol (NPK) (10-30-10)) were diluted in a 500 L tank in order to at least double the average basal (natural) phosphate and ammonium concentrations in the stream. Nutrient addition was adjusted bi-weekly, and natural N:P proportions were maintained as well.

Hydrological, Physical and Chemical Variables

Measurements of hydrological, physical and chemical variables were taken bi-monthly. Current velocity and flow (Q) were taken with a digital flow-meter SCHILTKNECHT (MiniAir 20). Temperature and dissolved oxygen (were registered with a HACH LDO HQ30d oxygen sensor). Conductivity was measured with a YSI model 556 MPS multi-parametric probe. The pH was measured with a SCHOTT pH 11/SET sensor. The ammonium, nitrate, nitrite and phosphate were all measured spectrophotometrically by following the techniques described by APHA-AWWA-WEF (2005).

Biological Sampling

Two samplings were carried out in each stream reach (C, I), one before (B, April 2008) and one after (A, January 2009) the enrichment. Three replicates of each trophic compartment (Coarse Particulate Organic Matter (CPOM), biofilm, macroinvertebrates and fish) were taken on each occasion. Samples were collected and processed according to the indications of Muñoz et al. (2009) and Hershey et al. (2007). The samples were refrigerated between sampling and lab processing.

Coarse Particulate Organic Matter.We collected three fractions of accumulated leaves from the stream bed in the field using a surber net of 900 cm2.

Macroinvertebrates. Samples were taken in rock substrata using a Surber sampler with a 900 cm2 surface area and 500 μm net mesh size. In the lab we then sorted the animals and left them in filtered river water under temperature-regulated conditions for 12 hours to clean their stomach content. For isotope analysis, we selected the most abundant species of insecta: Heterelmis sp. (adult stage), Simulium sp.,Tricorythodes sp. and Thraulodes sp.

Fish. Were sampled using an electric fishing device (ELT 60II GI) Individuals from the only two species found in the river (Oncorhynchus mykiss and Trichomycterus bogotensis) were collected, from which we obtained a subsample of 1 g of muscle tissue.

Sample Analyses

The extract of samples for analysis of δ13C and δ15N were dried at 60 °C for three days and were then crushed with a mortar to obtain a homogeneous sample of 0.1 mm particle size. For liquid extract samples of biofilm, we added 1 ml of the extract concentrate onto the pre-weighed tin capsule, and then dried and reweighed it. All samples were subsequently packed into tin capsules and stored in dry conditions. Samples were analyzed in a Thermo Elemental Analyzer 1108 associated to a mass spectrometer. Standards specified by the International Atomic Energy Agency (IAEA)were used to calibrate the isotopic signal: sucrose, polyethylene and graphite for carbon; ammonium sulphate and potassium nitrate for nitrogen. The standard test was run repeatedly to ensure linearity. The results were compared with the isotopic composition of atmospheric nitrogen for nitrogen, and PeeDee belemnite carbonate rock (PDB) for carbon as reference.

We analyzed gut contents of invertebrates, not of fish. After organism collection, the animals were deposited in plastic bags before being frozen. Five individuals of each species were analyzed. The individuals were placed in vials containing rose Bengal for 24 hours. Afterwards, the digestive tract of each individual was extracted and the anterior part removed under a stereomicroscope. This material was then placed on a slide to be studied under a microscope (400x). Ten visual fields were selected at random and photographed. The photos were quantified using the Coral Point (CpCe 3.4) program. Five categories of food sources were identified: coarse detritus (CD), fine detritus (FD), diatom algae (DA), filamentous algae (FA) and plant matter (PM).

Isotopic signals were represented in a bi-plot figure using the SIGMAPLOT 10 program. Differences in isotopic signals and in category percentages in gut content were tested with Student's t-test to find differences between reaches and time separately.

RESULTS In general terms, physical and hydrological variables presented similar conditions in the different reaches (C, I) and enrichment time (B, A) (Table 1). However, during nutrient addition, maximum discharge values were lower and the minimum higher than during the period before (Fig. 1). Significant differences were found in BACI analysis for nutrient concentrations, with Impact reach being higher after enrichment for NH4 + (n = 26, F = 4.685, p = 0.042) and PO4 3- (n = 26. F = 6.638, p = 0.017) (Fig. 2).

The δ15N signal clearly established three trophic levels: 1) basal level with CPOM and biofilm; 2) primary consumers (macroinvertebrates: collector-gatherers (Heterelmis, Thraulodes and Trichorythodes) and collector-filterers (Simulium); 3) predatorsfish, (Onchorhynchus mykiss and Trichomyterus bogotensis, Fig. 3). The δ15N of consumers was enriched compared to primary sources - mainly CPOM - in different proportions, depending on the feeding habits of each taxa (Table 2). The CPOM presented similar isotopic values in all phases of the experiment and constituted an indicator of the base of the food chain. However, the biofilm showed a higher δ15N signal (between 3.5 and 4 ‰) than CPOM, except for the treatment reach after the nutrient addition where the δ15N was 0.10 ‰. The average fractionation of nitrogen of the primary consumers with respect to CPOM was 4.7 ‰ (range, 3.8 - 5.5 ‰) and 1.7 ‰ with respect to biofilm (ranging between 0.3 and 4.0 ‰). Values forTricorythodes δ15N in impact reach after enrichment have not been taken into account in this range due to their unusual low values). Predators increased their δ15N signal by 5.9 ‰ (from 4.3 to 10.17 ‰) with respect to primary consumers (Table 2).

In most cases, the δ13C signal of biofilm overlapped with that of primary consumers, but a clear enrichment was observed with respect to CPOM. Following the δ13C and the gut content results (see below), one would predict that almost all the invertebrates analyzed were actually feeding on CPOM and biofilm (Fig. 3). An increase in δ13C values of biofilm was observed in both reaches after the enrichment (n = 3, t = -5.259, p = 0.006 in control reach and t = -2.944, p = 0.05 in I reach), indicating an enrichment of 13C with respect to 12C, probably related to environmental changes (e.g. flow). Oncorhychus mykiss and Trichomycterus bogotensis fed on macroinvertebrates in both reaches, and in I reach Oncorhychus became a prey of Trichomycterus (δ15N fractionation: 2.14 ‰ and 2.32 ‰, before and after respectively).

A depletion in δ15N was observed in I with respect to C reach after fertilization in different compartments: biofilm (t = 13.453, p = 0.001), Heterelmis (t = 5.572, p = 0.01) Simulium (t = 4.019, p = 0.02) andTricorythodes (t = 17.42, p < 0.001). Biofilm reflected the use of inorganic N from fertilizer (δ15N = 0,1 ± 0 ‰) and invertebrate signals in its consumption. This depletion was not significant for top predators.

Fine detritus (FD) was the most abundant food in the four invertebrate species whose guts were analyzed. Proportions of algae (diatom and filamentous algae) were always 10 % lower in gut contents. Heterelmis showed significant differences in FD (n = 5 t = -5.159, p < 0,01) and PM (n = 5, t = 3.533, p = 0.001) between C and I reaches after nutrient addition. Simulium in PM (n = 5, t = -2.496, p = 0.017), Thraulodes in CD (n = 5, t = 6.760, p < 0.001) and FD (n = 40, t = 7.027, p < 0.001).Tricorythodes did not show significant differences in any of the food categories (Fig. 4).

DISCUSSION

In spite of progress in the study of isotopes in food webs, there is little information available regarding tropical high mountain systems. In our study, the Tota stream food web shows about three to four trophic levels according to the spatial work scale of the study (Post et al., 2000). This high mountain creek is hydrologically very dynamic, and shows nutrient limitations (Rivera and Donato, 2008) that lead to low productivity (Abuhatab, 2011). All these could be important factors for limiting connectivity between species (Schmid- Araya et al., 2002; Jardine et al., 2012).

δ13C values obtained for biofilm (-25.33 to -21.82) and CPOM (-27.41 to -26.85) in the Tota stream are similar to those of other tropical streams, although slightly higher than those reported for leaves in other small tropical streams (March and Pringle, 2003; Dudgeon et al., 2010). In addition, our values correspond to those given by Peterson and Fry (1987) for C3 plants. Several factors can affect algal fractionation of C (Finlay et al., 2002), and result in a broad range of values. Fine detritus is derived from both algal and detrital components and has intermediate δ13C values (Hershey et al., 2007). It is an important source for consumer diet, it was common in the guts of Tota invertebrates, but, unfortunately, we did not analyze this compartment for stable isotopes.

In general it is very difficult to obtain clean samples of periphyton from the field since algal cells grow into the biofilm matrix together with bacteria, microfauna and detritus, thus resulting in higher values of δ15N than other basal resources. This makes it more difficult to calculate nitrogen fractionation for primary consumers with respect to resources. Values in our system are near those predicted in the literature (an average of 3.4 ‰, Post, 2002) although with high variability (ranging between 0.3 and 4.0 ‰) when biofilm is used in calculations. Fish nitrogen fractionation values were an average of 5.9 ‰ and clear δ13C enrichment was observed with respect to the invertebrates, showing their feeding dependence.

We found a strong relationship between collector-gatherers and biofilm in the two reaches (C, I) and periods (B, A), indicating strong reliance on algal carbon in this feeding group.Tricorythodes was the collector-gatherer that had the closest connection with the biofilm. A previous study (Donato- Rondon et al., 2010) has already shown that this species is strongly associated with periphyton resources, while Tomanova et al. (2006) indicated that most invertebrate collectors in tropical rivers are not food specialized and their dietary changes are related to the availability of resources.

An increase in δ13C values of biofilm was observed in both reaches after nutrient addition. This increase could be a product of discharge temporality (before samples were taken at the beginning of the high discharge period and the after ones were taken in the low discharge period). In this period a thicker periphyton, active photosynthesis, combined with diffusion-limited movement from the water to cells, is likely to cause the depletion of inorganic carbon within the periphyton matrix and higher values of δ13C (Hill y Middleton, 2006). On the contrary, Hladyz et al. (2011) raise the point that maximum enrichment in cobble biofilm δ13C signature occurred following periods of high discharge, while maximum depletion occurred during the low discharge period. These findings highlight the fact that trophic links between basal resources and primary consumers can be altered profoundly and that changes in hydrology can alter food chains and energy fluxes to the higher trophic levels (Perkins et al., 2010). As we hypothesized, nutrient addition, partly in nitrogen form, reduced the δ15N signature of πbiofilm and of most of the primary consumers, although no evidence of higher consumption was observed in gut contents. Conversely, detritus entering from riparian forest was not affected. Moreover, the proportion of fine detritus in gut contents significantly increased in Heterelmis and Thraulodes after addition. Both results may indicate a higher consumption of fine detritus and biofilm in the impacted reach, even though no clear significant differences were found in their quality as a consequence of fertilization. However, this depletion was not reflected in fish. Their high mobility along the river (Jardine et al., 2012) and long life cycle would lead to different patterns for these top predators.

The natural distribution of isotopic abundance in Tota stream showed the links between a resource or prey and predator in the two reaches and periods studied, and the nitrogen addition works as a tracer approach confirming those links. Complementary to this, the gut-contents data help to decide which link is the most correct and to discern dietary changes due to nutrient addition.

AKNOWLEDGMENTS

The authors thank the financing of the Universidad Nacional de Colombia, Colciencias and the BBVA Foundation, in the frame work of the project: "Global changes in fluvial systems: effects on the trophic web biodiversity and the functional aspects" (GLOBRIO) and thetechnical services of the University of Barcelona for the analysis of of isotops and CNP samples.

REFERENCES

Abuhatab Y. Actividad metabólica diaria del biofilm en el sector medio de un río de alta montaña (río Tota, Boyacá - Colombia). [tesis de maestría]. Bogotá: Departamento de Biología, Facultad de Ciencias, Universidad Nacional de Colombia; 2011. p. 34-63.         [ Links ]

Bergfurg J, Johnson RK, Sandin L, Goedkoop W. Effects of nutrient enrichment on C and N stable isotope ratios of invertebrates, fish and their food resources in boreal streams. Hydrobiologia. 2009;628:(1)67-79. DOI:10.1007/ s10750-009-9746-4.         [ Links ]

Biggs BJF, Smith RA. Taxonomic richness of stream benthic algae: Effect of flood disturbance and nutrients. Limnol Oceanogr. 2002;47(4):1175-1186.         [ Links ]

Boecklen JW, Yarnes CT, Cook BA, James AC. On the use of stable isotopes in trophic ecology. Annu Rev Ecol Evol. 2011;42:411-440. DOI: 10.1146/annurev-ecolsys-102 209-144726.         [ Links ]

Chará JD, Baird DJ, Telfer TC, Rubio EA. Feeding ecology and habitat preferences of the catfish genus Trichomycterus in low-order streams of the Colombian Andes. J Fish Biol. 2006;68(4):1026-1040. DOI: 10.1111/j.0022-1112. 2006.00984.x.         [ Links ]

Doi H. Resource productivity and availability impacts for food-chain length. Ecol Res. 2011;27(3):521-527. DOI: 10.1007/s11284-012-0941-9.         [ Links ]

Donato-Rondon JC, Morales-Duarte SJ, Castro-Rebolledo MI. Effects of eutrophication on the interaction between algae and grazers in an Andean stream. Hydrobiologia. 2010;657(1):159-166. DOI: 10.1007/s10750-010-0194-y.         [ Links ]

Dudgeon D, Cheung FKW, Mantel SK. Food web structure in small streams: do we need different models for the tropics?. J North Am Benthol Soc. 2010;29(2):395-412. DOI: http://dx.doi.org/10.1899/09-058.1.         [ Links ]

Elser JJ, HESSEN DO. Biosimplicity via stoichiometry: the evolution of food-web structure and process. En: Belgrano A, Scharler UM, Dunne J y Ulanowicz RE, editors. Aquatic Food Webs: An ecosystem approach. Oxford: Oxford University Press. 2005. p. 7-18.         [ Links ]

Finlay J. Stable-Carbon-Isotope Ratios of River Biota: Implications for Energy Flow in Lotic Food. Ecology. 2001; 82(4):1052-1064.         [ Links ]

Finlay J, Khandwala S, Power ME. Spatial Scales of Carbon Flow in a River Food. Ecology. 2002;83(7):1845-1859.         [ Links ]

Habit E, Victoriano P, Campos H. Ecología trófica y aspectos reproductivos de Trichomycterus areolatus (Pisces, Trichomycteridae) en ambientes lóticos artificiales Rev Bio Trop. 2005;53(1-2):195-210.         [ Links ]

Hershey A, Fortino K, Peterson Bj, Ulseth AJ. Stream Food Webs. En: Hauer FRR, Lamberti GA, editores. Methods in Stream Ecology. 2nd ed. San Diego: Elsevier. 2007. p. 637-659.         [ Links ]

Hill WR, Middleton RG. Changes in Carbon stable isotope ratios during periphyton development. Limnol Oceanogr. 2006;51(5):2360-2369.         [ Links ]

Hladyz S, Cook RA, Petrie R, Nielsen DL. Influence of substratum on the variability of benthic biofilm stable isotope signatures: implications for energy flow to a primary consumer. Hydrobiologia. 2011;664:(1)135- 146. DOI: 10.1007/s10750-010-0593-0.         [ Links ]

Jardine T, Pettit NE, Warfe DM, Pusey BJ, Ward DP, Douglas MM, et al. Consumer-resource coupling in wet-dry tropical rivers. J Anim Ecol. 2012;81(2):310-322. DOI: 10.1111/j.1365-2656.2011.01925.x.         [ Links ]

Jennings S. Size-based analyses of aquatic food webs. En: Belgrano A, Scharler UM, Dunne J, Ulanowicz RE, editores. Aquatic Food Webs an ecosystem approach. Oxford: Oxford University Press. 2005. p. 86-97.         [ Links ]

March JG, Pringle CM. Food Web Structure and Basal Resource Utilization along a Tropical Island Stream Continuum, Puerto Rico. Biotropica. 2003;35(1):84-93. DOI: 10.1111/j.1744-7429.2003.tb00265.x.         [ Links ]

Muñoz, I, Romaní AM, Rodriguez-Capitulo A, García-Berthou E. Relaciones tróficas en ecología fluvial. En: Elosegi A, Sabater S, editores. Conceptos y técnicas en ecología fluvial. Bilbao: Fundación BBVA. 2009. p. 347-366.         [ Links ]

Perkins D, Reiss J, Yvon-Durocher G, Woodward G. Global change and food webs in running waters. Hydrobiologia. 2010;657(1):181-198. DOI: 10.1007/s10750-009-0080-7.         [ Links ] Peterson B, Fry B. Stable isotopes in ecosystem studies. Annu Rev Ecol Syst.1987;18:293-320. DOI: 10.1146/annurev.es. 18.110187.001453.         [ Links ]

Polis G, Anderson WB, Holt RD. Toward an integration of landscape and food web ecology: the dynamics of spatially subsidized food. Annu Rev Ecol Syst. 1997; 28:289-316. DOI: 10.1146/annurev.ecolsys.28.1.289.         [ Links ]

Post DM, Pace ML, Hairston NG Jr. Ecosystem size determines food-chain length in lakes. Nature. 2000; 405:1047-1049.DOI: 10.1038/35016565.         [ Links ]

Post D. Using Stable Isotopes to Estimate Trophic Position: Models, Methods and Assumptions. Ecology. 2002;83 (3):703-718.         [ Links ]

Rivera C, Donato JC. Influencia de las variaciones hidrológicas y químicas sobre la diversidad de diatomeas bénticas. En: Donato J, editor. Ecología de un río de montaña de los Andes colombianos (río Tota, Boyacá). Universidad Nacional de Colombia. Facultad de ciencias. Bogotá. 2008. p. 83-101.         [ Links ]

Roman-Valencia C. Trophic and reproductive ecology of Trichomycterus caliense and Astroblepus cyclopus (Pisces: Siluriformes) in Quindío river, Colombia. Rev Biol Trop. 2001;49(4):657-666. DOI: 10.1111/j.0022-1112.2006. 00984.x        [ Links ]

Schmid-Araya J, Schmid PE, Robertson A, Winterbottom J, Gjerløv C, Hildrew AG. Connectance in Stream Food Webs. J Anim Ecol. 2002;71(6):1056-1062. DOI: 10. 1046/j.1365-2656.2002.00668.x.         [ Links ]

Tomanova S, Goitia E, HeleÅ¡ic J. Trophic levels and functional feeding groups of macroinvertebrates in neotropical streams. Hydrobiologia. 2006;556(1):251-264. DOI: 10. 1007/s10750-005-1255-5.         [ Links ]