Eduardo Cargnin-Ferreira1, Bruce Rider Forsberg

1Instituto Nacional de Pesquisas da Amazônia (INPA)-CPEC, Caixa Postal 478 - 69011-970, Manaus/AM, Brazil

1Endereço atual: Universidade Federal de Santa Caratina, CCB, Depto. de Ciências Morfológicas.

--------------------------------------------------------------------------------

ABSTRACT

The River Continuum Concept (RCC) provides a holistic theoretical framework for understanding the trophic structure and functioning of fluvial ecosystems. Macroinvertebrate functional feeding groups (FFGs) play a key role in these systems, processing specific sizes classes and types of organic resources in similar ways. The RCC makes specific predictions about the spatial distributions of FFGs along the river size-continuum. We present here the first test of these predictions in a neotropical fluvial ecosystem, the Jaú River. Distributions of aquatic macroinvertebrate FFGs, determined along a continuum of river size within the Jaú basin, were compared to those predicted by the RCC. Aquatic macroinvertebrates were collected in 1st, 2nd, 3rd, 4th and 7th order tributaries of the Jaú using an optimized sampling scheme on four distinct occasions between March and November of 1995. As expected by the RCC, the relative abundances of most FFGs varied significantly with stream size. The specific patterns encountered for several FFGs though were quite distinct from those predicted by the RCC. The communities in headwater streams, which according to the RCC should be dominated by shredders and benthic particle collectors, were dominated by scrapers and omnivores, apparently feeding on fungi. The relative abundance of predators, which is expected to be relatively constant in the RCC, varied considerably between tributaries in the Jaú, due primarily to differences in prey availability. Two distinct pathways for the processing of allochthonous particulate organic matter were identified in the Jaú river system: 1) a dominant pathway beginning with the colonization and consumption of leaf litter by fungi and 2) a minor pathway, initiating with the consumption of submerged leaves by shredders.

Key words: river, macroinvertebrate, functional feeding groups, Amazon.

--------------------------------------------------------------------------------

INTRODUCTION

Since its creation by Vannote et al. (25), the River Continuum Concept (RCC) has become one of the central paradigms for research on lotic macroinvertebrate communities. The RCC provides a holistic theoretical framework for understanding the trophic structure and functioning of fluvial ecosystems. Macroinvertebrate functional feeding groups (FFGs) (9) play a key role in these systems, processing specific sizes classes and types of organic resources in similar ways. The RCC makes

 

specific predictions about the spatial distributions of FFGs in river systems. The relative abundance of each FFG is expected to vary in a unique way in response to systematic changes in habitat structure and resource availability which occur between the headwaters and mouth of a river. Shredders, which utilize whole leaves and other large organic matter as a resource, are expected to be more numerous in small headwater streams where these materials tend to accumulate. Vannote et al. (25) defined collectors as scavengers which feed on small particles produced by shredders and other functional groups or processes and separated them schematically into benthic and planktonic particle feeders. Here when define these two sub-groups as collectors and filterers, respectively. Benthic particle feeders (collectors) are expected to reach peak densities in small to intermediate sized tributaries where fine benthic particulate matter is most abundant. Planktonic particle feeders (filterers) are expected to predominate in larger rivers where ultra-fine suspended particulate matter, produced by the progressive reduction of organic detritus by upstream communities, tends to accumulate. Scrapers, which feed on periphytic algae, are expected to be most abundant in intermediate sized streams where reduced shading and shallow waters produce optimal light levels for algal growth (21, 25). Predators, which feed on prey from all functional groups, are expected to occur uniformly throughout the river size continuum. The independent variation in FFG distributions with stream size is expected to produce a continuum of macroinvertebrate communities which processes allochthonous and autochthonous carbon inputs in an integrated sequential fashion, maximizing the efficiency of organic carbon metabolism in the ecosystem (25).The predictions of the RCC regarding lotic macroinvertebrate distributions have been tested in a number of north temperate river systems (15, 8, 4, 3). Their accuracy in neotropical systems though has yet to be evaluated. The Amazon, the largest neotropical river system in the world, provides an ideal environment for such a test. Much of the Amazon river system drains undisturbed tropical forest and still contains a full spectrum of natural lotic habitats. The invertebrate communities in these systems have been studied extensively (26, 27, 16, 28, 33). However the specific predictions of the RCC regarding FFG distributions have never been tested explicitly. We present here an analysis of macroinvertebrate community structure in the Jaú river, a 7th order blackwater tributary of the Amazon. The distributions of FFGs encountered in the Jaú river system are compared to those predicted by the RCC. The implications of the results for the understanding, conservation and management of neotropical fluvial ecosystems are discussed.

MATERIAL AND METHODS

Study site

The Jaú is a 7th order black water tributary which flows into the right bank of the Negro River approximately 250 km upstream from its confluence with the Amazon main channel near Manaus. The fluvial network of the Jaú drains approximately 10,000km2 of predominantly undisturbed lowland tropical forest which is located entirely within the Jaú River National Park. The Jaú main channel is a slow meandering river with an extensive but poorly developed floodplain covered by annually flooded forest. The 4th to 6th order tributaries of the Jaú also have fringing forested floodplains, although they are less extensive and subject to more frequent flooding. The meanders in these streams also tend to be more frequent and smaller. Large but infrequent log dams are encountered in forth order streams and large submerged litter banks frequently occur on the interior margins of their meanders. The 1st to 3rd order tributaries of the Jaú generally drain upland forests. They are characterized by narrow, poorly defined floodplains which are inundated only during major storms, by frequent small log and root dams and submerged litter banks. Meanders are more frequent and smaller than in larger order streams. The waters in the Jaú drainage system are

 

exceptionally acid and low in suspended particulates and dissolved salts. The pH varies between 3.1 and 5.6, conductivity ranges from 6.1 and 18.7m S.cm-1 and suspended sediment concentrations vary from 1-19mg l-1. Dissolved oxygen concentrations vary considerably in the system, ranging

 

from 7.7mg l-1 in small headwater streams to 1.4mg.l-1 in the Jaú main channel. Surface water temperatures also vary significantly, ranging from 23oC in small shaded streams to 26.7oC in the main channel. Some physical characteristics of tributaries sampled in this study are given in Table 1.

Table 1. Physical characteristics and sampling locations of tributaries sampled in the Jaú River system.

Sampling and Analysis

To investigate changes in community structure related to stream size, aquatic macroinvertebrates were sampled in 1st, 2nd, 3rd, 4th and 7th order tributaries of the Jaú on four distinct occasions between March and November of 1995. The geographic location of the sampling sites is indicated in Table 1. Two complete meanders were sampled in each tributary to assure that most common habitats were represented. Different methodologies were used to collect invertebrates, depending on tributary size and invertebrate type, in order to obtain the most efficient collection possible in all environments. A 1mm mesh hand net was used to collect benthic invertebrates along the complete study reach in 1st, 2nd and 3rd order streams. These streams were small and shallow enough so that all benthic environments present in the study reach could be efficiently sampled with this method. Sampling effort was optimized in these tributaries by plotting the cumulative number of families encountered during the first collection against sampling time (Fig 1). The optimal sampling interval was defined as the minimum sampling time needed to saturate the family curve. A standard sampling interval (> optimal interval) of 15 minutes was used for all three stream sizes. An Eckman dredge was employed to sample benthic invertebrates in the deeper 4th and 7th order tributaries where the use of a hand net was impractical. Due to the greater length of these study reaches (0.5m - 2km) it was not feasible to sample all existing habitats. Instead, samples were collected along equidistant transects which extended

 

between river margins (excluding wetlands) perpendicular to the channel. The number of transects sampled was optimized by plotting the cumulative number of families encountered during the first collection campaign against transect number. The minimum number of transects necessary to saturate the family curve was adopted as the standard sampling effort in each system. The standard sampling effort determined in this manner for 4th and 7th order streams was 7 and 5 transects per reach, respectively. Planktonic invertebrates (zooplankton) in all tributaries were collected with a 65m m mesh zooplankton net. A single 100 hundred liter sample, collected near the surface at mid channel

 

and mid reach, was filtered for each tributary on each date. All benthic invertebrates larger than 1mm were separated from detritus in the field and preserved in 70% ethyl alcohol. Zooplankton samples were preserved in 10% Formalin.

Figure 1. Cumulative number of families encountered with increasing sampling time and transects in tributaries of different size in the Jaú River system.

 

 

All macroinvertebrates were identified to the family level and classified into one of six recognized functional feeding groups (9): shredders, collectors, filterers, scrapers, omnivores or predators. The omnivore group was included to account for organisms with multiple feeding strategies which were not considered in the original RCC model. The relative abundances of FFGs in each stream were calculated from the total number of organisms in each group encountered during the four collections. The variation in relative abundance between collections was not considered.

RESULTS AND DISCUSSION

Functional Feeding Group (FFGs) Distributions

The taxonomic composition of the macroinvertebrate community in the Jaú river system (Table 2) was similar to that found in other streams around the World. All of the major invertebrate groups listed by Hynes (17) as typical components of stream communities were found, with a few exceptions. No amphipods or isopods were encountered. Gastropods were also absent, probably due to the low concentrations of calcium and other dissolved minerals, typically encountered in blackwater streams (23). Another unique characteristic of the Jaú’s invertebrate community was the dominance of Palaemonidae (shrimps) in the Crustacea Group. Freshwater shrimp are rare in north temperate streams but quite common in small Amazonian tributaries. The taxonomic composition of invertebrates in the Jaú was similar to that reported for the Tarumã River, another blackwater tributary of the Negro River (16, 32).

Table 2. Taxonomic and functional composition of aquatic macroinvertebrates found in the Jaú River system.

The variation encountered in the relative abundances of FFGs with stream size is indicated in Figure 2. The similarity of these distributions to those predicted by the RCC and their significance for ecosystem structure and function are discussed below.

Figure 2. Relative abundance of macroinvertebrate functional feeding groups in tributaries of different size in the Jaú River system.

Shredders - With the exception of the 4th order tributary, the relative abundance of shredders declined progressively with increasing stream size (Fig. 2). This trend, which is predicted by the RCC, has also been encountered in many north temperate river systems (15, 3), and apparently reflects the predominance of large particulate organic matter, the main resource for this group, in smaller order streams. The unusually high proportion of shredders in the 4th order stream may reflect additional inputs of fresh litter from fringing flooded forests, a habitat that was not considered in Vannote et al.’s (25) original RCC model. Adis et al. (1) encountered high rates of litter production in black water flooded forests of the Central Amazon. A large combined input of litter inputs from this and upstream sources, together with the existence of large log dams and slower current speeds, apparently account for the large litter banks commonly encountered in this environment. The greater proportion of shredders in this tributary when compared to smaller tributaries may also reflect the lower frequency of hydraulic disturbance commonly encountered in higher order streams (5). In another study involving a large number of low order streams (< 3rd order) in the Jaú river system, no significant correlation was found between the abundance of shredders and litter and was attributed this to between stream variation in flow disturbance (6). Increased meander length and a lack of natural barriers in larger order streams (> 5th order) apparently reduced the residence time of leaf litter in these channels and prevented litter bank formation.

Summing the results from all tributaries, shredders accounted for 6.1% of all invertebrates collected (Table 3). This is significantly lower than the proportion of shredders commonly encountered in north temperate river systems (~10%, (11)), but is similar to that encountered in New Zealand (35, 36),in French (7) and in the Central Amazonian streams (6). The low relative abundance of shredders encountered in the Jaú river system may reflect the lower food quality and higher levels of secondary compounds commonly found in leaves from tropical forests (18, 13).

Table 3. Absolute abundance of invertebrates encountered in different functional feeding groups in tributaries of different size in the Jaú River system

.

Scrapers: The relative abundance of this group was consistently high in all stream sizes except 7th order (Fig. 2). This trend differed significantly from that predicted by the RCC, where scraper abundance is expected to peak in intermediate sized streams. The distinct pattern, in this case, may reflect the use of a different resource by this group. Scrapers are thought to be feeding predominantly on periphytic algae attached to benthic substrates (10, 22). Algal production is expected to be highest in intermediate sized streams which are wide enough to receive full incident radiation during at least part of the day and shallow enough so that most benthic surfaces are well illuminated. Fungi growing on the surfaces of benthic litter is an alternative resource for scrapers that can be quite important in many river systems. Fungi can account for as much as 9% of total detrital biomass in streams (14) and are exceptionally high in food value (2). Walker (28) has argued that fungi are the primary consumers of terrestrial leaf litter in Amazonian streams and the first dominant link in aquatic food chains of these systems. The production of fungi presumably depends on the availability of fresh benthic substrates which were plentiful in all study streams except for the 7th order Jaú main channel. The use of fungi by scrapers could thus explain the distribution encountered for this group. The high contribution of scrapers to total invertebrate numbers, 15.6% (Table 3), reflects the relative importance of this group and the fungus-scraper link to organic carbon flow in this system.

Collectors: The relative abundance of benthic particle feeders varied considerably showing a slight tendency to increase with stream size (Fig.2). This trend contrasts with the decline in collector biomass predicted by the RCC for larger streams and suggests that fine benthic organic matter may be more abundant in these tributaries than expected. The velocity of the Jaú main channel is exceptionally low (mean = 0.15 m s-1), especially when the mouth of the Jaú is flooded by the rising Negro River (Forsberg et al., unpublished data). Meade et al. (20) found that a significant part of the fine organic sediment load of the Lower Amazon was stored seasonally in the river channel when river surface slope and velocity were reduced. A similar phenomenon may also occur in the Jaú main channel, resulting in a significant accumulation of fine benthic organic matter during at least part of the year. The large numbers of scrapers and shredders encountered in 4th order streams (Table 3) may also provide a significant upstream source of fine benthic organic matter for these larger channels. No significant change was found in the collector density with stream size in a large number of streams in the Jaú river system. However he only examined small to intermediate sized streams (6).

 

 

 

Filterers: The relative abundance of planktonic particle feeders increased significantly in the largest tributaries (Fig. 2) as predicted by the RCC and observed in small North American streams (24). The increase in filterers in larger tributaries was expected due to the higher concentrations of suspended particles encountered in these systems (Forsberg et al., unpublished). The high percentage of filterers in small streams is more difficult to explain. Wallace et al. (34) argued that at the low current speeds common in most small streams, planktonic filtration is extremely inefficient. However, Walker (33) suggested that many filterers can also feed by biting or sucking organic particles. This would allow filterers to survive on both benthic and planktonic particles. This hypothesis is supported by the results of Cargnin-Ferreira (6) who found a strong positive relationship between the densities of shredders and filterers in low order tributaries of the Jaú River. This result suggested that filterers could be feeding directly on the larger particles produced by shredders. Yule (37) attributed the high proportion of filterers (65%) which she encountered in a Papua New Guinea river system to inputs of fine particulate organic matter from riparian forests. Filterers were also the largest functional group in the Jaú river system, accounting for 26.0% of total invertebrate abundance (Table 3). The high representativity of this guild and the dominating presence of riparian forests along low order tributaries of the Jaú, suggest that the forest could be an important source of fine particulate matter in this system, as well.

Predators: In contrast to the homogenous distribution of predators predicted by the RCC and encountered in other systems (15, 12, 22, 3), the relative abundance of this guild varied significantly with tributary size in the Jaú (Fig. 2). When all non-predators were considered as potential prey, absolute predator numbers were positively correlated with prey numbers (Spearman Rank, r = 0.90), suggesting that this variation was primarily due to resource limitation. The exceptionally low prey:predator ratios encountered in all tributaries (Table 4) and the inverse correlation encountered between this ratio and predator densities (Spearman Rank, r = - 0.7) tend to support this hypothesis. Similar results were found in a more extensive analyses of invertebrate communities in low order tributaries of the Jaú (6). Not all predator guilds in Amazonian streams though have been found to be resource limited. Walker (29) encountered a prey:predator ratio of 300 in the blackwater Tarumã River and concluded that prey availability was not a limiting factor for predators in this system. The higher prey:predator ratio encountered in this case may have been due, in part, to the different sampling method used. The author in this study collected invertebrates by thoroughly washing submersed litter, a procedure which allowed her to collect organisms significantly smaller (<1mm) than those collected in the present study. The greater selectivity of our sampling method may have therefore resulted in a systematic underestimate of prey abundance and resource availability for predators. Predators were the second largest functional group, accounting for 19.6% of total invertebrate abundance (Table 3).

Table 4. Prey/Predator and Prey/Omnivores ratios encountered in tributaries of different size in the Jaú River system.

Omnivores: The only clear omnivores encountered in the Jaú river system were freshwater shrimp (Decapoda). Shrimp in Amazonian streams have been found to consume algae and fungi attached to submersed litter and also feed as predators on small fish and invertebrates (19, 28, 30, 31). The distribution of shrimp, specifically, and omnivores, in general, was not predicted by the RCC (25). However, considering the feeding habits of shrimp, tendencies similar to those encountered for predators, collectors, filterers or scrapers might be expected. In general, the relative abundance of shrimp decreased

 

 

progressively with stream size (Fig 2). This pattern was similar to that encountered for shrimp in a larger number of small order tributaries of the Jaú (6) and resembles, in part, those encountered for shredders, scrapers and filterers in the present study. The lack of a positive correlation between the absolute abundances of shrimp and total prey (all invertebrates except shrimp, Spearman Rank, r = -0.10) and the existence of a positive correlation between shrimp numbers and prey: omnivore ratios (Spearman Rank, r = 0.70) suggest that prey-availability is not a limiting factor for shrimp. The greater abundance of shrimp in smaller streams suggests that fungi associated with submersed litter may be a more important resource for this group. The large concentrations of submerged litter commonly found in small streams may also provide an important refuge from large vertebrate predators (fish, alligators). Large fish are rarely encountered in first and second order streams but quite common in larger tributaries and may effectively reduce shrimp densities in these environments. Omnivores were the third most important functional group, numerically, accounting for 18.5% of total invertebrate abundance (Table 3). However, shrimp were the largest macroinvertebrates in the system and probably accounted for the majority of invertebrate biomass.

Ecosystem Structure and Function

As expected, the relative abundances of most functional feeding groups varied significantly with stream size in the Jaú River system. The specific patterns encountered for several FFGs though were quite distinct from those predicted by the RCC and these differences have important implications for the structure and functioning of this fluvial ecosystem.

In the ecosystem envisioned by Vannote et al. (25), particulate allochthonous organic matter enters the river system primarily in small order streams in the form of coarse leaf litter. This material is processed initially by shredders who reduce the material to smaller particles and pass it along to a sequence of guilds further downstream specialized in processing and reducing it even further. The higher percentage of shredders observed here in small order streams and the predominance of collectors and filterers in high order streams (Fig.2) indicate that a similar food chain may exist in the Jaú river system. However the predominance of omnivores and scrapers in small tributaries (Fig. 2) suggests that this is not the dominant path of organic carbon flow.

 

In the RCC, scrapers are assumed to feed predominantly on periphytic algae. In Amazonian streams both scrapers and shrimp have been found to feed on fungi as well (19, 16, 30). Most low order tributaries of the Jaú (< 3rd order) are completely shaded by riparian forest and periphyton growth is exceptionally low (Forsberg, unpublished data). The most likely energy source for scrapers and shrimp in these streams is therefore fungi attached to submerged leaf litter. The dominance of scrapers and shrimp in these environments thus suggests that fungi, and not shredders, are the primary consumers of coarse allochthonous particulate matter in the Jaú River system. Henderson & Walker (16) and Walker (28) arrived at a similar conclusion regarding organic carbon flow in the Tarumã River, another blackwater tributary of the Amazon. Based on meticulous in-vivo and in-vitro observations of macroinvertebrate feeding behavior, she identified fungi, associated with submerged leaf litter, as the first link in the heterotrophic foodweb of this system. The direct consumption of submerged litter by shredders and other herbivores is apparently inhibited by the low nutricional quality and high phenolic content of tropical forest leaves (13). Fungi transform this basically impalatable substrate into a nutritious, protein rich food (i.e. fungal biomass) which is easily assimilated by aquatic invertebrates. Once consumed by scrapers and omnivores, fungi are either metabolized or transformed into feces, providing a nutritious secondary resource for particle collectors and filterers further down the river size continuum.

Our results thus point to the existence of two distinct pathways for the processing of allochthonous particulate organic

 

matter in the Jaú River system: 1) a dominant pathway beginning with the colonization and consumption of leaf liter by fungi and 2) a minor pathway, similar to that proposed by other authors (25), initiating with the consumption of submerged leaves by shredders. An important characteristic of the shredder pathway in the Jaú, that was not predicted by the RCC, was the existence of large shredder populations in fourth order streams (Fig. 2). These populations were apparently sustained by lateral inputs of fresh leaf litter from fringing floodplain forests, a habitat not included in the original continuum model. Future refinements of the RCC will clearly have to consider both fungal food chains and allochthonous carbon inputs from floodplain environments. Food quality also appeared to have an important influence on the pattern of carbon flow in the Jaú and will have to be considered in any model developed to describe the trophic dynamics of tropical black water river systems.

Implications for Management and Conservation

The results described here have important implications for the management and conservation of this and other fluvial ecosystems in the Amazon region. The functional inter-dependence of invertebrate communities inhabiting different parts of the Jaú river continuum make the system especially vulnerable to perturbation. The impacts of anthropogenic or natural disturbances are expected to be especially severe when they occur in headwater streams. A disturbance which affects the integrity of invertebrate communities in these tributaries is likely to have a cascading impact on all interdependent communities downstream, greatly altering the integrated structure and functioning of the fluvial ecosystem. These impacts are likely to effect not just invertebrate communities, but all of the aquatic flora and fauna which depend on them either directly or indirectly. The effects of deforestation in headwater areas are likely to be especially devastating, since litter inputs from headwater forests are the primary source of allochthonous organic matter to the river system. To maintain the structural and functional integrity of the fluvial ecosystem, it is essential that these areas be preserved in their natural state.

 

ACKNOWLEDGMENTS

We thank José Palheta and Gerley Díaz Castro for help with field work and Dr. Ilse Walker for constructive comments throughout the study. This work was supported by the National Science Foundation, Victoria Amazônica Foundation, INPA, NASA, and CNPq (Brazilan National Research Council).

--------------------------------------------------------------------------------

RESUMO

Estrutura Trófica de Comunidades de Macroinvertebrados no Sistema do Rio Jaú (Amazônia Central, Brasil). O Conceito do Rio Contínuo (CRC) provê uma linha teórica para o entendimento da estrutura trófica e funcional do ecosistema fluvial. Os grupos funcionais (GFs) de macroinvertebrados têm uma importante função-chave nestes sistemas, processando de forma semelhante classes específicas de tamanho e tipos de recursos orgânicos. O CRC faz previsões específicas sobre a distribuição dos GFs ao longo do contínuo de tamanho do rio. Nós apresentamos aqui, o primeiro teste dessas previsões em um ecossistema fluvial neotropical, o rio Jaú. As distribuições dos GFs de macroinvertebrados aquáticos, determinadas ao longo de um contínuo de tamanho do rio dentro da bacia do rio Jaú, foram comparadas àquelas previstas pelo CRC.

 

 

Os macroinvertebrados aquáticos foram coletados em tributários do Jaú de 1a, 2a, 3a, 4a e 7a ordens, usando um esquema de amostragem otimizado em quatro diferentes ocasiões entre março e novembro de 1995. Como esperado pelo CRC, as abundâncias relativas da maioria dos GFs variaram significativamente com o tamanho do rio. Os padrões específicos encontrados para muitos GFs entretanto, foram muito distintos daqueles previstos pelo CRC. As comunidades nas cabeceiras dos rios, as quais de acordo com o CRC seriam dominadas por cortadores e coletores de partículas do bentos, foram dominadas por raspadores e onívoros, aparentemente consumindo fungos. A abundância relativa de predadores , a qual é esperada ser constante no CRC, variou consideravelmente entre os tributários no Jaú, devido primeiramente às diferenças na disponibilidade de presas. Duas rotas de processamento da matéria orgânica particulada alóctone foram identificads no sistema do rio Jaú: 1) uma rota dominante, iniciando com a colonização e consumo de serrapilheira por fungos e 2) uma rota menor, iniciando com o consumo de folhas submersas pelos cortadores.

Palavras-chave: rio, macroinvertebrado, grupos funcionais, Amazônia.

--------------------------------------------------------------------------------

REFERENCES

Adis, J.K.; Furch, K; Irmler, U. Litter production of a Central Amazonian inundation forest. Trop. Ecol. 20:236-245, 1979.

 

Bärlocher, F.; Kendrick, B. Dynamics of the fungal population on leaves in a stream. J.Ecol. 72: 761-791, 1974.
Basaguren, A.; Elosegui, A.; Pozo, J. Changes in the trophic structure of benthic macroinvertebrate communities associated with food availability and stream flow variations. Hydrobiol. 81: 79-91, 1996.
Braioni, M.G. Faunistic research in a fluvial ecosystem study: the case of the River Adige. Boll. Zool. 61:415-424, 1994.
Brussock, P.P.; Brown, A.V. Riffle-pool geomorphology disrupts longitudinal patterns of streams benthos. Hydrobiol. 220: 109-117, 1991.
Cargnin-Ferreira, E. Fatores influindo na distribuição de grupos funcionais de macroinvertebrados aquáticos em pequenos tributários do rio Jaú, Amazônia Central. Manaus, 1998, 65 p. (M.S. Thesis, Curso de Ecologia, INPA).
Chauvet, E.; Giani, N.; Gessner, M.O. Breakdown and invertebrate colonization of leaf litter in two contrasting streams: significance of oligochaetes in large river. Can. J. Fish. Aquat. Sci. 50: 488-495, 1993.
Culp, J.M.; Davies, R.W. Analysis of longitudinal zonation and the River Continuum Concept in the Oldman-South Saskatchewan river system. Can. J. Fish. Aquat. Sci. 39: 1258-1266, 1982.
Cummins, K.W. Trophic relations of aquatic insects. Ann. Rev. Entomol. 18: 183-206, 1973.
Cummins, K.W.; Klug, M.J. Feeding ecology of stream invertebrates. Ann. Ver. Ecol. Syst. 10: 147-172, 1979.
Cummins, K.W.; Wilzbach, M.A.; Gates, D.M.; Perry, J.B.; Taliaferro, W.B. Shredders and riparian vegetation. Leaf litter that falls into streams influences communities of stream invertebrates. BioSci. 39: 24-30, 1989.
Dudgeon, D. Longitudinal and temporal changes in functional organization of macroinvertebrate communities in Lam Tsuem River, Hong Kong. Hydrobiol. 111: 207-217, 1984.
Forsberg, B.R.; Devol, A.H.; Richey, J.E.; Martinelli, L.A.; Victoria, R.L.; Bonassi, J.A.. Autotrophic arbon sources for fish of the Central Amazon. Ecol. 74: 643-652, 1993.
Gessner, M.O.; Schoerbel, J. Fungal biomass associated with leaf litter in a stream. Oecol. 87: 602-603, 1991.
Hawkins, C.P.; Sedell, J.R. Longitudinal and seasonal changes in functional organization of macroinvertebrate communities in four Oregon streams. Ecol. 62: 387-397, 1981.

 

Henderson, P.A.; Walker, I. On the leaf litter community of the Amazonian blackwater stream Tarumazinho. J. Trop. Ecol. 2: 1-17, 1986.
Hynes, H.B.N. The Ecology of Running Waters. Liverpool University Press. Liverpool, 1970, 555p.
Janzen, D. H. Tropical blackwell rivers, animals, and mast fruiting by the Dipterocarpaceae. Biot. 6: 69-103, 1974.
Kensley, B.; Walker, I. Palaemonid shrimps from the Amazon basin, Brasil (Crustacea: Decapoda: Natantia). Smithsonian Contributions to Zoology. Whashington, 1982, 362, 28p.
Meade, R.H.; Dunne, T.; Richey, J.E.; Santos, U.M.; Salati, E. Storage and remobilization of suspended sediment in the lower Amazon River of Brazil. Sci. 228: 488-490, 1985.
Minshall, G.W. Autotrophy in stream ecosystems. BioSci. 28: 767-771, 1978.
Riaño, P.; Basaguren, A.; Pozo, J. Variaciones espacio-temporales de las comunidades de macroinvertebrados bentónicos del río Agüera (País Vasco-Cantabria). Limnet. 9: 19-28, 1993.
Schmidt, G.W. Chemical properties of some waters in the tropical rain forest region of Central Amazonia. Amazon. 3: 199-207, 1972.

 

 

Smock, L.A.; Gilinsky, E.; Stoneburner, D.L. Macroinvertebrate production in a southeastern United States blackwater stream. Ecol. 66: 1491-1503, 1985.
Vannote, R.L., Minshall, G.W., Cummins, K.W., Sedell, J.R.; Cushing, C.E. The river continuum concept. Can. J. Fish. Aquat. Sc., 37: 130-137, 1980.
Walker, I. Rede de alimentação de invertebrados das águas pretas do sistema do rio Negro. 1. Observações sobre a predação de uma ameba do tipo Amoeba discoides. Ac. Amazon. 8: 423-438, 1978.
Walker, I. On the structure and ecology of the micro-fauna in the Central Amazonian forest stream ‘Igarapé da Cachoeira’. Hydrobiol. 122: 137-152, 1985.
Walker, I. The biology of streams as part of Amazonian forest ecology. Exp. 43: 279-287, 1987.
Walker, I. Study of benthic micro-faunal colonization of submerged litter leaves in the central Amazonian black-water stream Tarumã-Mirim (Tarumazinho). Ac. Limnol. Brasil. 2: 623-648, 1988.
Walker, I.; Henderson, P.A.; Sterry, P. On the patterns of biomass transfer in the benthic fauna of an Amazonian black-water river, as evidenced by 32P label experiment. Hydrobiol. 215: 153-162, 1991.
Walker, I. Life history traits of shrimps (Decapoda: Palaemonidae) of Amazonian inland waters and their phylogenetic interpretation. Stud. Neotrop. Fauna Environ. 27: 131-143, 1992.
Walker, I. The benthic litter-dwelling macrofauna of the Amazonian forest stream Tarumã-Mirim: patterns of colonization and their implications for community stability. Hydrobiol. 291: 75-92, 1994.
Walker, I. Amazonian streams and small rivers in: Tundisi, J.G.; Bicudo, C.E.M. e Matsumura-Tundisi, T. (eds.). Limnology in Brazil. Rio de Janeiro, 1995, p.167-193.
Wallace, J.B.; Webster, J.R.; Woodal, W.R. The role of filter feeders in flowing waters. Arch. Hydrobiol. 79: 506-532, 1977.
Winterbourn, M.J.; Cowie, B.; Rounick, J.S. Food resources and ingestion patterns of insects along a West Coast, South Island, river system. N. Z. J. Mar. Freshwat. Res. 18: 43-51, 1984.
Winterbourn, M.J.; Rounick, J.S.; Cowie, B. Are New Zealand stream ecosystems really different? N. Z.J. Mar. Freshwat. Res. 15: 321-328, 1981.
Yule, C.M. Trophic relationships and food webs of the benthic invertebrate fauna of two aseasonal tropical streams on Bougainville Island, Papua New Guinea. J. Trop. Ecol. 12: 517-534, 1996.