Quick viewing(Text Mode)

Food Webs in River Networks M

Food Webs in River Networks M

Blackwell Science, LtdOxford, UK ERE Ecological Research 0912-38142002 Ecological Society of Japan 174July 2002 503 webs in networks M. E. Power and W. E. Dietrich 10.1046/j.0912-3814.2002.00503.x Review Article451471BEES SGML

Ecological Research (2002) 17, 451–471

Food webs in river networks

MARY ELEANOR POWER1* AND WILLIAM ERIC DIETRICH2 Departments of 1Integrative and 2Earth and Planetary Science, University of California Berkeley, Berkeley, California 94720, USA

Food webs and river drainages are both hierarchical networks and complex adaptive systems. How does living within the second affect the first? Longitudinal gradients in , regimes and structure down have long interested ecologists, but their effects on structure and dynamics are just beginning to be explored. Even less is known about how network structure per se influences river and riparian food webs and their members. We offer some preliminary observations and hypotheses about these interactions, emphasizing observations on upstream–downstream changes in food web structure and controls, and introducing some ideas and predictions about the unexplored question of food web responses to some of the network properties of river drainages.

Key words: length; food webs; landscape heterogeneity; river networks; .

INTRODUCTION has access. Ecologists have long pon- dered the historical, environmental and biological Food webs are well described as complex adaptive controls that determine path and chain lengths systems (as lucidly reviewed by Levin (1999)). Like and the impacts of particular web members other complex adaptive systems, food webs have (Hairston et al. 1960; Pimm 1979, 1982; Paine diverse components, linked by flows and (often 1980, 1988; Power 1992a; Power et al. 1996; Post non-linear) interactions, ‘which determine and are et al. 2000). This debate has been somewhat con- reinforced by hierarchical organization of these fused when flow paths and population con- components (Levin 1999; p. 12)’. Paine (1980) trol chains were not distinguished. Our present pointed out that two distinct flows create different understanding of these relationships is particularly in the same food web. Energy flows limited by our rudimentary appreciation of the from more basal resources up to consumers at spatial and temporal contexts and scales of food higher trophic positions, while ‘top-down’ chains webs. The impacts of web members on each other, of population control link consumers to the and the degree to which energy flow predicts inter- populations they regulate or limit, if these action strength, depend largely on which spatial consumers are not suppressed by their own preda- sources of energy and nutrients sustain particular tors. Energy flow paths and population control web members, and how resident versus transient chains are related, but not identical. these members are in their communities. Synthetic should have greater impact in food webs if they studies that link food web dynamics to spatial have access to better, more productive, or more fluxes of energy, matter, organisms and informa- widely distributed energy sources. Conversely, tion across heterogeneous landscapes (Polis et al. interactions and impacts of other may 1997; Nakano & Murakami 2001) will contribute determine the energy source to which a particular much to our understanding of these issues. In studies of spatial food webs, as of any com- plex system, trade-offs exist between realism and *Author to whom correspondence should be mechanistic understanding on the one hand, and addressed. Email: [email protected] scope and generality on the other. To explore the Received 26 September 2001. effects of landscape heterogeneity on multitrophic Accepted 12 November 2001. level dynamics, ecologists sensibly began with

452 M. E. Power and W. E. Dietrich simplified conceptualizations. These depict fluxes of , , organic matter or organisms of organisms, energy or materials across habitat from into mainstems. Longitudinal boundaries (Holt 1984; Polis et al. 1997), over 2- gradients and their impacts on species distribu- D lattices (Oksanen 1990; de Roos et al. 1991), tions have long interested stream ecologists, but or among islands or patches set in an uninhab- upstream–downstream changes in food web inter- itable matrix (chapters in Tilman & Kareiva actions are just beginning to be investigated. Even 1997). Important recent progress towards land- less is known about the network effects per se on scape realism has been made through more organisms and food webs, but these influences may explicit, sometimes experimental, studies of the help explain and predict ‘cumulative watershed ecological effects of specific landscape features, effects’, which are of great concern in such as boundary permeability (Cadennaso & Pick- watershed management (Li et al. 1994; Dunne ett 2000; Laurance et al. 2001; Cadennaso et al. in et al. 2001). press) or geometry (Fagan et al. 1999; Anderson & In this paper, we discuss classical and more Polis 1999), or seasonal shifts in the relative pro- recent ideas for controls on food chain length. We ductivities of coupled by trophic exchange then explore how such controls may vary at differ- (Nakano & Murakami 2001). However, one gen- ent positions in drainage networks where the eral feature of landscapes has so far received very energy sources, habitat structure and disturbance little attention with respect to its influence on regimes differ in channels and adjacent water- spatial food webs. River drainage networks sculpt sheds. We draw mainly on our own work with all terrestrial landscapes, defining their relief, dis- students and colleagues from rivers in north section and many other aspects of their heteroge- coastal California. A more comprehensive review neity. and characteristics influence of the literature relevant to this topic would repay ecologically significant conditions and are partially effort, but is beyond the scope of this paper. predictable from landscape position (Montgomery & Buffington 1993; Sklar & Dietrich 1998). Here, we explore how conditions arising from longitudi- CONTROLS ON LENGTHS OF ENERGY nal and network structure might affect energy flow FLOW PATHS AND FUNCTIONAL and species performances in food webs. FOOD CHAINS IN FOOD WEBS

There are many interesting questions to ask about A TALE OF TWO NETWORKS the properties of food webs (e.g. Cohen 1977; Paine 1980, 1988; Schoener 1989). Questions Food web networks are hierarchical. Energy flows about energy path or chain length are particularly up from lower to higher network positions, and informative for guiding investigations of large- consumer control is exerted down some of these scale variation in the distribution of trophic-level paths, but not others. River drainage networks are (Oksanen et al. 1981; Mittelbach et al. also hierarchical, with gravity driving water, sed- 1988; Power 1992a), as well as issues of practical iment, solutes and organic matter from ridges interest. If we want to conserve , dividing watershed down into channels, and from sustainably harvest a resource, or suppress biologi- headwaters down mainstems to lowland floodplain cal pests, we need to know the energy flow paths rivers and . With the exception of winds that support the groups of interest and the top- and local back eddies in turbulent water, only bio- down controls on their . logical fluxes (migrations and other movements of Classical hypotheses for environmental controls individuals) drive materials upstream or upslope. on the length of food chains (which unfortunately Conditions, resource fluxes and biotic interactions do not distinguish energy flow paths from top- experienced by aquatic, riparian and terrestrial down chains) discuss two environmental variables organisms in watersheds vary down longitudinal and one evolutionary factor. gradients from headwaters to lowland habitats (Vannote et al. 1980; Montgomery & Buffington 1 Productivity/efficiency. Chains should lengthen as 1993, 1997). Conditions also vary abruptly, for fluxes of limiting resources or energy to food example, where network confluences inject pulses webs increase, or as consumers increase their

Food webs in river networks 453

efficiency of resource capture or conversion cient appendages for dismembering prey (Vogel (Fretwell 1977; Pimm 1979; Oksanen et al. 1981; Power 1987; Power et al. 1997). The rela- 1981; Fretwell 1987; Pimm & Kitching 1987). tionship of food chain length to habitat size may 2 Disturbance/stability. Chains should be shorter be less clear, however, in habitats such as rivers in more frequently disturbed environments where cross-habitat exchanges have strong effects (Pimm & Lawton 1977; Pimm & Kitching (Polis et al. 1997; Nakano et al. 1999). Analyses of 1987). river or adjacent terrestrial food webs may over- 3 Design constraints. Pimm (1979) argues that it look crucial interactions if the boundary of obser- may be impossible for to build a vation or experimentation is drawn at the river Pterodactyl predator, for example, because an surface (Wallace et al. 1997; Nakano et al. 1999; large enough to subdue one could not Power & Rainey 2000; Nakano & Murakami fly to catch it. 2001; Sabo & Power in press a, b; Power et al. in press). Despite many food web surveys, particularly Post (unpubl. data, 2001) has recently called for across productivity gradients (McQueen et al. an expanded discourse that acknowledges the 1989; Carpenter et al. 1991; Power 1992b; Pers- influence of many factors on food chain length. In son et al. 1996), and some experimental studies general, lengths of food chains are dynamic, (Carpenter et al. 1987; Jenkins et al. 1992; Woot- responding to a number of non-linear, interacting ton & Power 1993; Wootton et al. 1996; Marks factors (e.g. Levin 1999; Carpenter 1988). As the et al. 2000), little consistent support can be found abundance of particular web members changes for predictions of either the classical productivity over time, so will the lengths of the chains in the or the disturbance hypothesis (Hastings & Conrad food webs in which they are embedded. For exam- 1979). However, recent work on this topic has ple, Hutchinson (1959) argued that if a secondary uncovered one factor that does appear to exert a top predator depleted its primary prey strong, consistent effect: habitat size. Some of the over time, it would, of necessity, drop from the most convincing data ever assembled on this ques- fourth to the third . If predators tion reveal that in temperate lakes, energy flow deplete more edible prey, they enrich primary con- paths lengthen with habitat size (volume), but are sumer guilds with inedible taxa, shortening func- unrelated to productivity (Post et al. 2000). The tional food chains that control green biomass in relationship between food chain length and habitat their habitats (e.g. Power et al. 1996). Multiple size may be quite general, and partially driven by controls on food chain lengths are also likely design constraints (hypothesis 3). Large habitats to interact. Schoener (1989) has proposed a support larger taxa, and many food webs are productivity × area hypothesis for food chain strongly size structured, with larger organisms at length. Disturbance and productivity may also higher trophic levels (Menge & Sutherland 1976.; interact to affect food chain length over timescales Kerfoot & DeMott 1984; Kerfoot & Sih 1987; that encompass species succession (Power et al. Power et al. 1997). Size usually matters more than 1996). species identity in predator–prey relationships: who eats whom depends on the size ( history stage) of the individuals that encounter each other (Polis et al. 1989). Density, however, can some- DEMOGRAPHIC AND times reverse a size-based predator–prey relation- METAPHYSIOLOGICAL MODELS FOR ship as in pack army ants, dogs or PREDICTING THE PERSISTENCE AND piranhas, or whelks grazing on (Barkai & IMPACTS OF FOOD WEB MEMBERS AT McQuaid 1988). Pimm’s (1979) pterodactyl pop- DIFFERENT NETWORK POSITIONS ulations might similarly have been suppressed by parasites or egg predators. The positive relation- To think about how complex dynamic controls on ship between body size and trophic position is food chains might interact over heterogeneous particularly strong in aquatic food webs where landscapes, we revisit three basic conditions that most predators are gape-limited because a viscous must be met if increasing the input of a limiting medium selects against hydrodynamically ineffi- trophic resource to lower trophic levels is to

454 M. E. Power and W. E. Dietrich lengthen food chains or energy-flow paths in a where N(t) is the number of individuals or biomass local web. First, a new top ‘consumer’ must be [i] in the population at time t; N(0) is the number biogeographically available. Consumer is used of individuals or residual biomass (survivors or here sensu latu: this could be a producer or a recolonists) present soon after disturbance resets , if only non-living resources are avail- the environment and t is the time since distur- able and we are asking whether the local chain will bance. The severity of a local disturbance, as well increase from zero to one link. Propagules of the as spatial heterogeneity in the form of refuges from new top consumer must either be able to arrive at disturbance, both affect N(0). the site of the newly enhanced resource or be ‘wait- The influence of spatial heterogeneity also ing in the wings’ as surviving residual biomass, unfolds as we consider R, the function describing often in resistant life history stages. Environmen- the population growth rate. R [individuals pro- tal heterogeneity in the form of refuges, dispersal duced per individual per time, (i i−1 t−1)] here can barriers or conduits plays an obvious role here. be any suitable function describing positive pop- Second, if the flux enters the web several levels ulation growth (e.g. exponential, linear or sigmoi- below the potential new top consumer, lower dal). To investigate the role of environmental trophic levels already present must be able to cap- conditions (as well as natural history traits), we ture enough of the enhanced flux to augment their can unpack the parameter R following Schoener own somatic growth or reproduction. Environ- (1973): mental structure and consumer biomass both Rfcpm=¥[]()- (2) affect this contingency. For example, if more dis- solved nitrogen were added to a smooth channel where p is the proportion time (t) spent with fast laminar flow and a thin film of periphy- [t t−1], c is per capita rate of energy (e) harvest ton, little of this nitrogen would be assimilated [e i−1 t−1], f is conversion efficiency [i e−1] and m locally, even if the were nitrogen lim- is the per capita energetic cost of maintenance ited. Third, lower trophic levels already present [e i−1 t−1]. We can use equations 1 and 2 to ask must use the augmented resource to produce new whether a key population in a functional group can tissue that the new potential top consumer can re-establish after disturbance, so that it could per- ingest and assimilate. Enhancing pro- sist under a given environmental regime, assum- ductivity will not increase lengths of energy flow ing no losses to biotic enemies. In other words, paths or top-down food chains if it is sequestered this analysis would address the question of in defended tissues or individuals. Only if these whether conditions and local resources satisfy three conditions are met can enhanced productiv- requirements that determine a species’ fundamen- ity lengthen food chains (Fig. 26.3 in Persson et al. tal niche (Hutchinson 1959). For example, is there 1996). sufficient productivity area (Schoener 1989) to Timescales are implicit in the verbal arguments support such a population? Habitat productivity above. Before discussing spatial heterogeneity, let [energy area−1 time−1] × foraging area [area] are pos- us scale these arguments to what are perhaps the itively related to c, per capita rate of energy har- most fundamental temporal scales affecting food vest, but increasing foraging area may also increase webs: (i) time since disturbance (defined sensu m, energetic costs, if travel is expensive. Equation Sousa (1984): a discrete event that kills or removes 2 must be parameterized to reflect a particular local biota, freeing space and other resources); and natural history because the costs of travel relative (ii) recovery times following resets for key popu- to exhaustive local grazing on depleted resource lations. To maintain itself in a periodically dis- patches vary greatly among taxa and habitats. turbed habitat, a population, such as a new apex This demographic analysis establishes criteria member of a food web path or chain, needs a for the persistence of a single population under a positive per capita realized rate of natural increase, given environmental regime at a particular drain- R, high enough so that: age network position. While these criteria are nec- essary, they would not necessarily be sufficient for dNt() d t= R()() Nt,0 N t> when Nt()= N() 0persistence, as they do not take interactions among (1) web members into account. For such a food web Food webs in river networks 455 analysis, we turn to the metaphysiological models values of parameters in equations 1 and 2) would of Getz (1991, 1993, 1994), which portray troph- themselves be determined by evolved traits and ically stacked populations with elegant concise- the environmental context. Drawing from our ness. Getz calls his models metaphysiological river and watershed research, we explore how these because they describe the biomass dynamics of parameters might vary among different functional populations in terms of collective intake and groups in food webs and how the values for a given , instead of the birth and rates of functional group may change down drainage net- individuals (although equivalent terms for intrin- works. We next consider aggregated functional sic rate of increase, r, and , K, in groups in a food web fragment that seems suf- traditional logistic equations can be derived from ficient to explain much of the spatio-temporal his formulations [Getz 1993; p. 290: eqs (7) and variation in trophic level biomass, particularly of (8)]. The per biomass growth rate for a population , in the rivers we have studied. First, we will is: compare the constraints and of differ- ent groups that might affect their parameter values dx dt = xf() g- x g (3) iiiiii++11 in the different environmental contexts that occur where f(g) describes a population’s per biomass at different landscape positions. Then we will growth rate for a given collective rate of resource summarize geomorphic studies of longitudinal uptake g (xi, xi+1). Equation 3 holds for all popu- changes in channel environments and offer some lations or functional groups in a food chain with ideas for the influences of network properties of n levels, with gn+1 = 0 for the top predator, n. For rivers that are largely unstudied to date. Finally, the lowest level, the resource flux for x0 might also we will explore, in a preliminary fashion, how be modified to describe a constant or environmen- environmentally mediated shifts in metaphysio- tally determined input rate of photons or dissolved logical or behavioral and demographic parameters nutrients. in a population might influence its impact on food A simple formation for per biomass growth used webs at different channel network positions. by Getz (1993) to produce the hyperbolic relation- ship commonly expected between resource uptake and population growth rates is: FUNCTIONAL GROUPS IN A FOOD WEB FRAGMENT fgii()=-rk i()1 g i (4) where ρ is the maximum growth rate that can be Consider simplified food web fragments (sensu sustained on δ > 0, the maximum per biomass rate Holt 2002) such as those presented in Fig. 1, with of resource extraction achievable at high resource highly aggregated functional groups. Two basal fluxes, and κ is the minimum resource uptake rate energy sources fuel most river food webs: terres- needed for positive population growth (a compen- trial (leaves, fruits, flowers, stems, sation point). or soil carbon and associated fungi and ) The rate of resource uptake for level i is mod- and attached algae. This energy flows either to elled as a functional response with consumer satu- vulnerable primary consumers that support pred- ration and interference (DeAngelis et al. 1975): ators or to invulnerable primary consumers that sequester energy without passing it up the food gx(), x= ()db x()++ x g x (5) ii---111 i ii i i ii web. Vulnerable primary consumers in rivers β where i is the half saturation resource level for an include thin soft fish and naked, mobile inverte- isolated biomass unit that experiences no con- brates (e.g. mayflies and free-living midges). In γ specific influence on its intake, and i is a dimen- contrast, other aquatic primary consumers are pro- γ sionless self interaction scaling parameter ( i > 0 tected with heavy or sessile life styles and γ for interference, i < 0 for social facilitation of are rarely eaten by gape-limited aquatic predators feeding). (Hershey 1987; Power 1987, 1992b). Armored The dynamics of each population in a trophic loricariid catfish are the top consumers in two- stack are governed by the five parameters ρ, κ, β, level food chains in Panamanian . They γ and δ. Values of these parameters (as well as effectively suppress algae across a wide range of 456 M. E. Power and W. E. Dietrich

Fig. 1. Food web fragments depicting aggregated func- Terr. Terr. tional groups that influence pred. arthro. the South Fork food web in headwater (right) and main- Vert pred. stem (left) channels. Solid lines in web diagrams depict Vert pred. linkages that have been dem- onstrated to limit resource Invert pred. populations under certain Invert pred. environmental conditions. For Inedible the mainstem food web, the grazer Inedible longer chains dominate after Edible grazer Edible flood disturbance, and the grazer grazer short, two level chain domi- nates during drought years Algae and during prolonged absence Algae Terr. of scouring floods. Dotted detr. lines for the headwater web depict energy flow linkages for Mainstem web Headwater web which the top-down impacts are still unknown. primary productivities and seasonal conditions is in primary consumers, however, that these dif- (Power 1983, 1984b, a). In temperate streams, ferences have affected the river food webs that we stone cased caddisflies, sessile midges and aquatic have studied in temperate and tropical rivers lepidopterans that live under silk cases attached to (Power 1983, 1984b; Power 1987; Power et al. rocks are similarly invulnerable to most predators 1996; Wootton et al. 1996; Marks et al. 2000). (Hershey 1987) and also suppress algae, unim- There is increasing general recognition that peded by predators (Feminella et al. 1989; Power attributes of intermediate consumers ( et al. 1996). For example, large, armored caddis- and ) may affect energy flow and pop- flies (Dicosmoecus gilvipes) are invulnerable to most ulation regulation in food webs as much as top fish in the upper portions of river networks in the predators or basal resources (Sinclair & Arcese Pacific Northwestern USA (Johansson 1991; Tait 1995; Duffy & Hay 1991; Schlapfer & Schmid et al. 1994; Rader 1997). Defended and vulnerable 1999). taxa may be functionally distinct at other trophic Because of their importance to channel food positions [e.g. edible and inedible algae (Leibold web structure and dynamics, we focus on grazers, 1989; Carpenter et al. 1993; Romo et al. 1996)]. It both edible (e.g. naked or mobile) and inedible Food webs in river networks 457

(armored or sessile), as we hypothesize about how the key metaphysiological parameters in the Getz Edible (1994) models (Fig. 2) might influence landscape δ patterns in food webs. We would predict that e δ i inedible grazers would be more adept at extracting Inedible scant resources (e.g. a sessile or armored scraper would probably leave less residue after grazing a small site than a mobile ‘skimmer’) and so would Uptake rate (g) have lower half saturation constants, β. The max- β i β e imum rate of resource uptake at high resource densities, δ, would probably be larger for mobile, Resource density (x i-1 (t)) edible grazers, by virtue of their better resource ρ tracking [faster conversion of resource pulses to e production instead of defense (Fig. 2a)]. The same Edible κ constraints would make the compensation point ρi (uptake rates needed for positive growth) higher Inedible for inedible grazers that must allocate to defense 0 in addition to growth and offspring (Fig. 2b). Mobile grazers would have higher population growth at a given uptake rate because of more efficient conversion of intake to production, and Population growth rate (f) κ κ δ δ also greater absolute potential uptake rates e i i e (Fig. 2a). Uptake rate (g) Aquatic eat vulnerable aquatic pri- mary consumers, each other and also terrestrial Fig. 2. Possible contrasts in Getzian relationships that fall into streams. By feeding on between resource density, population uptake rate (a) vulnerable grazers, aquatic predators may indi- and population growth (b) for edible, ‘e’ (e.g. naked, rectly release algae from herbivory (in a chain with mobile) grazers and inedible, ‘i’ (e.g. sessile or armored) grazers. Less mobile, inedible grazers often scrub local three functional trophic levels), or they may prey substrates more thoroughly, leaving less residue, so may on key primary aquatic carnivores, releasing cer- have lower half saturation constants, β, than mobile tain herbivores, which then suppress algae, [creat- edible grazers that may ‘skim the cream’ when harvest- ing a four level top-down chain (Power 1990b)]. ing resources. Because of better resource tracking, edi- Whether predators are at the third or fourth ble grazers may have higher maximum rates of resource trophic level positions in the chain that supports uptake, δ. Because of allocations to defense, the resource uptake necessary for inedible grazers to achieve positive them depends on the abundance and activity of Κ primary consumers with predator-specific defenses population growth, , may be higher (but this would depend on how the costs of defense and costs of loco- effective against secondary, but not primary, carni- motion compare in particular systems). Production effi- vores (Power 1990a). Completely invulnerable ciency δ/Κ (Getz 1993) should, therefore, be higher for grazers, if abundant, reduce the access of vulnera- edible grazers, which is frequently observed. The max- ble prey and, therefore, predators to algal carbon. imum rate of population growth at high resource intake Such invulnerable grazers are trophic cul de sacs should, therefore, be higher for edible grazers. These in food webs, reducing energy flow to fishes and trait-based performances may explain, among other other aquatic predators (Tait et al. 1994; Parker & things, why armored grazers are (as observed by Power 1997), and diminishing the top-down con- McNeely, unpubl. data, 2001) abundant in head- water streams of extremely low primary productivity trol that predators exert in food chains (Power (a potential consequence of their putative low half- 1995; Wootton et al. 1996). saturation constants, β) (modified from Getz 1993). After aquatic emerge as flying adults, however, they become potentially vulnerable to terrestrial predators, such as spiders, adult odo- nates, lizards, birds and bats. Rates of 458 M. E. Power and W. E. Dietrich and lateral diffusion into the watershed of aquatic cross the threshold for channel initiation to move insects, as well as the ability of various terrestrial up or down slope (Dietrich & Dunne 1978; Mont- predators to track this prey flux, are strongly influ- gomery & Dietrich 1988). Downstream from the enced by the position within drainage networks. channel head, the upstream limit for production Network position affects not only channel habitats, of aquatic organisms begins where is but also geomorphic and vegetative structure and retained long enough for individuals to complete environmental conditions in the riparian zones and their aquatic life stages. Clearly, this boundary will upland valley slopes adjacent to the channel (Power also change with the precipitation regime and the & Rainey 2000; Power et al. in press). Down drain- permeability of the bed and the surrounding age networks, disturbance regimes, habitat pro- watershed (Hynes 1975). As channels collect dis- ductivity, habitat structure and size, and edge to charge with downstream increases in their drain- area ratios all vary in ways that are partially pre- age areas, their slopes decrease. These channels dictable from general geomorphic relationships widen, deepen and flow faster according to empir- (Vannote et al. 1980; Montgomery & Buffington ical rules of hydraulic geometry that relate channel 1997; Montgomery 1999). Next we present a very width, depth and velocity to that scales brief overview of what is known about downstream with drainage area (Leopold et al. 1964). As one changes in channel environments. moves from headwaters through upstream tribu- taries to mainstems downstream, wider channels and increasing setback of the bordering forest from CHANNEL AND WATERSHED active channel margins increase sunlight to the ENVIRONMENTS DOWN DRAINAGE streambed. As a result, both stream temperature NETWORKS and algal primary productivity increase. Carbon inputs to rivers shift from allochthonous terrestrial Geomorphologists have made considerable recent detritus to algal production along this gradient progress in explaining and predicting systematic (Vannote et al. 1980; Davis-Colley & Quinn downstream changes in river and watershed envi- 1998). As width : depth ratios increase and slopes ronments. Some site-specific features can be pre- decrease, channel habitats change from cascades dicted at real landscape positions from physical and stepped pools constrained by coarse bedrock first principles; others must be empirically pre- and boulder substrates, to plane bed glides and dicted, and still others simply mapped. Large-scale meandering pools and riffles in middle reach main- mapping is more feasible now than it was in the stems, to broad channels with floodplains and past because of newly available high resolution off-channel water bodies in the lowlands (Mont- digital elevation data and the computational gomery & Buffington 1997). Corresponding to power necessary to process these data (Dietrich these slope and drainage area-driven changes, bed et al. 1993). Discharge, slope, channel hydraulic materials change from boulders and bedrock in geometry (Leopold et al. 1964), sediment size and headwaters, to cobbles, pebbles and gravel in transport processes all change systematically mid-reaches, to sand and silt near river mouths downstream. The environmental conditions and (Leopold et al. 1964; Montgomery & Buffington habitat structures that they establish depend, as a 1997). Increasingly downstream, the path and first approximation, on local channel slope (e.g. form of a river becomes less constrained by resis- Montgomery & Buffington 1997) or on drainage tant rock or vegetation, and more ‘the author of area and local channel slope (Sklar & Dietrich its own geometry’ (Leopold et al. 1964), except 1998). where it is engineered (dammed, diked or Stream drainage networks begin some distance diverted) by humans. down from the watershed divide, where the In headwaters, disturbances related to sediment downslope flux of water and sediment first cuts a transport are rare (few or none per millenium), distinct channel head into the hillside. The chan- but catastrophic, deriving from debris flow into nel head tends to persist at a given location unless channels. In upstream tributaries (drainage area environmental changes in runoff or vegetative A < 10 km2), rare superfloods (recurrence intervals cover cause the hillslope position where conditions >30 years) may transport the boulders making up Food webs in river networks 459 most of the bed materials. Transport of finer (sand, volume. and other large fully aquatic taxa do gravel, pebble-sized) through these not usually occur in the steepest headwater reaches reaches is much more frequent, but because these with slopes >10%, where habitats are small fine sediments make up a relatively small propor- (A < 1km2) and sometimes ephemeral. However, tion of the river bed in unimpaired headwaters, food webs in these small headwaters are not pred- fines may not have strong impacts on local biota. ator-free. are highly carnivorous In middle reach mainstems, most of the bed mate- (Parker 1991, 1993) and some cannot withstand rials are smaller cobbles, pebbles and gravels, and fish ; thus, they are restricted to fishless floods typically move them several times per year. headwaters (Petranka 1983). In general, because Here, fine sediments are retained longer and have edge : area ratios increase upstream, small headwa- more devastating effects as disturbance agents ter channels are believed to be more influenced by (Power & Stewart 1987; 1995; K. B. Suttle terrestrial ecosystems. This influence has been con- et al. unpubl. data, 2001). In lowland floodplain sidered largely in terms of allocthonous energy rivers, transport of the fine bed materials (sand and inputs into streams from terrestrial detritus (Van- silt) is chronic (Dietrich & Dunne 1978). note 1980) or insects (Nakano et al. 1999), but may also pertain to the influence of terrestrial predators on stream prey (Jackson & Fisher 1986; Environmental controls on food webs down Power & Rainey 2000; Nakano & Murakami drainage networks 2001; Power et al. in press). How do downstream changes in habitat size and Habitat characteristics of watersheds surround- structure, disturbance, productivity and tempera- ing steeper tributaries may facilitate some terres- ture affect functional groups and food webs? trial predators and impede others. Smyth (cited in Superimposing a food web network over a drainage Power et al. in press) observed that a dominant network and predicting its response is a challeng- riparian spider along the South Fork Eel (S. Fk Eel) ing task. As a first step, we will explore three and its tributaries, Tetragnatha versicolor, was much downstream gradients that influence particular denser along the productive mainstem S. Fk Eel. trophic linkages in webs: (i) changes in habitat However, spiders had longer diel foraging periods geometry that affect aquatic-terrestrial exchange; in upstream, darker tributaries [i.e. upstream (ii) changes in productivity that affect terrestrial increases in its parameter p (the proportion of time and algal carbon sources to different consumers; spent foraging in equation 2)]. Spiders in Elder and (iii) differences in disturbance regimes that Creek (A 17 km2) foraged 24 h day−1, while along mediate the relative of vulnerable and the S. Fk Eel (A 80–130 km2), they foraged only invulnerable primary consumers and, therefore, after dark. Smyth advanced two non-exclusive web connections of predators. hypotheses for this shift across habitats during the We use field observations and a limited litera- foraging period. Tetragnathids are particularly ture review to suggest how selected parameters in vulnerable to desiccation and may have been pre- equations 1 and 2 may change for key functional cluded from foraging along mainstems by day groups at different landscape positions, and how because of wind or heat. Alternatively, they may these changes affect their influence in food webs. have been obliged to forage longer to meet their This is a preliminary exploration intended to energy requirements in the less productive introduce hypotheses that might be studied more upstream tributaries. In either case, the behavioral thoroughly and systematically in the future. change increases the probability that energy cap- tured by these spiders (which feed almost exclu- sively on emergent aquatic insects) is transferred Habitat size and river-to-forest export up food chains to day-foraging terrestrial birds, Obviously, channel habitat size (volume or area) or parasitic flies (A. Smyth in Power et al. increases downstream. Wide, sunny mainstem in press). channels (A > 100 km2) contain more aquatic pred- Other predators are constrained by upstream ators (e.g. large fish, , wading and diving conditions. Sceloporine lizards, which are impor- birds and snakes) due, in part, to larger habitat tant predators on spiders and other terrestrial, as 460 M. E. Power and W. E. Dietrich well as aquatic, arthropod prey in sunny down- summer when the terrestrial productivity is start- stream habitats (Sabo 2000), drop out ing to decline due to drying. W. E. Rainey (pers. of headwater tributaries when solar radiation comm., 2001) has hypothesized that certain becomes insufficient to support the long activity Northern California bats, whose October to March periods necessary to maintain their growth. hibernation is fuelled by aquatic prey cap- Parameter p in equation 2 for these lizards tured during the previous summer, may exert decreases upstream. Declining arthropod abun- time-lagged subsidized predation on terrestrial dance and activity upstream (by lowering param- insect prey when bats emerge in the , before eter c) may also affect the position of this river insect populations recover from the scouring, distribution threshold in drainage networks. Some turbid winter floods. Considerable year-to-year insectivorous bats are deterred from headwater variation in rainfall occurs in California and in hunting because splashing water in these steep other Mediterranean regions (Gasith & Resh reaches interferes with their ultrasonic foraging 1999), however, and multiyear droughts (with no calls. Bat species that forage primarily over quiet scouring winter floods) can alter the hydrologic water, lower in the drainage network, move many regimes experienced by river life and, conse- kilometers downstream for hunting, even if their quently, food web structure. day roosts are located near headwaters (W. E. The importance of terrestrial predation to the Rainey, pers. comm., 2000). dynamics of aquatic populations has not been well Seasonality may influence the strengths of these quantified, but may decrease downstream for a cross-habitat trophic linkages. If insects emerge number of reasons. Increasing setback of terrestrial late in the season they may avoid population con- vegetation from the stream reduces perches, web trol from predation by bats, birds, lizards or spi- sites or cover for terrestrial predators. The aquatic ders that have become seasonally inactive or, in the habitat gains volume and structural complexity, case of some bats and birds have migrated to other offering prey refuges from wading and diving habitats. A similar situation may hold for insects predators. In addition, more aquatic insects should in which development is slowed in cool, unpro- be harvested before they emerge by aquatic pred- ductive portions, high in the drainage networks. ators as these become more abundant, diverse and Life history differences among functional groups larger downstream. An exception to this trend may may interact with landscape impacts on develop- occur when and where channel productivity down- ment rates to influence - and landscape- stream becomes high enough to support floating specific strengths of cross–habitat food web mats of aquatic algae or emergent or floating mac- interactions. Seasonal influences on these cross- rophytes. These interfaces serve as food-rich ther- habitat trophic fluxes may also differ between the mal incubators for certain insect taxa and also as eastern and western of continents. East partial refuges from fish predation (Power 1990b). and central continental temperate watersheds They act as ‘valves’ for river food webs, increasing experience terrestrial productivity peaks during the rate of production and diverting the late spring and summer, while productivity in it from aquatic predators to consumers in terres- small streams in upper watershed positions peaks trial food webs (Power et al. in press). during early spring and autumn, when leaves have fallen from the . Seasonal, reciprocal comple- Productivity and carbon sources mentary stream and watershed production coupled with cross-habitat trophic exchange can support As one moves from headwaters through upstream higher densities of fish and birds than could be tributaries to mainstems downstream, more - maintained by either habitat alone (Nakano & light hits the streambed and stream temperature Murakami 2001). On the western edges of conti- and algal productivity increase. Carbon sources to nents under Mediterranean climates, terrestrial river food webs shift from allochthonous detritus and aquatic productivity are seasonally reversed to algal production along this gradient (Vannote and are less offset than under continental regimes. et al. 1980; Davis-Colley & Quinn 1998), but car- River productivity is minimal during the scour- bon sources to all members of food webs do not ing, turbid, winter floods and peaks during the change in lockstep. Food webs in river networks 461

In a survey of 70 studies of temperate and boreal carbon begins to support both vulnerable primary river food webs, which included his own data from consumers and predators. Alternatively, if fish the South Fork Eel and its tributaries, Finlay found were more dependent on terrestrial invertebrates that algal carbon signatures became more 13C in headwaters than lower in the watershed, remov- enriched downstream, while signatures of terres- ing inedible grazers would have less effect. trial vegetation, detritus and dissolved organic car- McNeely removed all Glossosoma (up to several bon derived from terrestrial sources remained thousand individuals per m2) from replicated constant (Finlay 2001, p. 1057, Figs 1 and 2). whole pool habitats in Creek. Despite the Shredders (primary consumers known to ingest closed forest canopy (<5% open) over this headwa- terrestrial leaves) maintained a terrestrial carbon ter stream, she observed a conspicuous release of signal over the entire range of the drainage areas algal standing crops dominated by the highly edi- (<1–2060 km2). Downstream of small (<10 km2) ble diatoms Melosira and Cymbella following her drainages, all other consumers, scrapers, filter- removal of Glossosoma (F. C. McNeely, unpubl. gatherers, collectors, and and verte- data, 2001). Her experiment suggests that food brate predators, tended to become isotopically web interactions may influence energy sources to heavier downstream, suggesting that they derived web members in these upstream habitats. While much of their carbon from algae. In small headwa- energy sources influence interaction strength in ters with drainage areas <10 km2, all consumers food webs (Elton 1927; Polis et al. 1997), these except scrapers had terrestrial carbon signals. results remind us that the converse is also true. In Scrapers, however, had an algal signal as far up into general, we need more information on how chang- the watershed as they were collected (Finlay 2001, ing conditions across landscapes differentially p. 1058, Fig. 4 and pers. comm., 2000). For exam- influence functionally significant groups in food ple, in the S. Fk Eel watershed, Sugar, McKinley webs (e.g. vulnerable vs invulnerable grazers). and Fox Creeks (drainage areas of 0.6, 1.0 and Some important differential responses involve the 2.5 km2, respectively), armored scrapers (Glossos- resistance or resilience of such taxa following oma penitum) had an algal carbon signal that was disturbance. lighter than terrestrial carbon signatures at these headwater sites (Finlay 2001; Finlay et al. in press). Disturbance regimes Predators (steelhead or Oncorhynchus mykiss or Pacific giant salamanders Dicamptodon Disturbance often acts most severely on basal taxa ensatus) had terrestrial carbon signatures, as did in food webs (e.g. rock bound algae and small their prey (vulnerable mayflies, midges, stoneflies, invertebrates in rivers) because they are usually less and for the fish, terrestrial invertebrates). These mobile than predators. Organisms at lower trophic results led McNeely and Finlay (pers. comm., levels, while generally more susceptible to distur- 2000) to hypothesize that depletion of sparse algae bance, vary in resilience. Early successional pri- by invulnerable Glossosoma at these sites reduced mary consumers are commonly soft-bodied, small access to algal carbon for vulnerable grazers (e.g. and vulnerable to predators, with short generation heptageniid mayflies, chironomids) and aquatic times and rapid growth rates when the environ- predators. In Getz’s models, lower half saturation ment is favorable. In rivers, they may be more coefficients, β, of the Glossosoma could allow them likely to survive scour than invulnerable prey or to persist at lower resource productivity and producer taxa because they are small enough to deplete algal standing crops below the levels seek refuge in pore spaces below or lateral to mobi- needed to sustain populations of edible grazers. lized layers of the river bed. Small prey species may Although inedible grazers are postulated (Fig. 2) also survive dewatering in pockets of residual to need higher uptake rates to maintain positive lateral or subsurface water. Such refuges would population growth, their severe grazing may increase N(0) in equation 1. In addition to this deplete resources below levels that are mechani- resistance to disturbance, early successional taxa cally harvestable by edible grazers. Experimental may exhibit greater resilience (greater R in equa- removal of invulnerable grazers might shift tion 1). Insects with short generation times may upstream the productivity threshold at which algal be flying above rivers during these disturbances in 462 M. E. Power and W. E. Dietrich

‘air force reserves’ of winged adults (Gray & Fisher Creek (A 17 km2) offer preliminary support for this 1981), able to recolonize rapidly following the prediction. Following floods that scoured both disturbance. In their aquatic larval stages, how- channels, invulnerable grazers replaced vulnerable ever, primary consumers with heavy armor or taxa as dominants in the primary consumer sessile life styles are less mobile and, hence, less several weeks later in Elder Creek than in the able to escape disturbance [lower N(0)]. They are South Fork Eel. This interaction of productivity also slower to recover afterwards (lower R) because and disturbance reverses the predictions of the of their allocation to defense rather than to growth classical single factor hypotheses for productivity or progeny. As time since disturbance passes, how- and disturbance effects on food chain length, and ever, these invulnerable grazers may increasingly may be general when food webs reassemble over dominate late successional primary consumer timescales that permit species succession (Power guilds because they do not suffer losses to preda- et al. 1996). tors and they may also be superior competitors for Frequency of bed mobilizations typically space and (McAuliffe 1983, increases down drainage networks. In larger trib- 1984; Li & Gregory 1989). For some taxa, traits utaries and middle reach mainstems, bed scouring influencing roles in successional food webs may floods may move cobbles and gravel substrates sev- differ in different life stages. For example, lim- eral times per year, although flood frequencies nephilid caddisflies (which include Dicosmoecus) are drop during natural droughts or because of human strong fliers as adults (F.C. McNeely, pers. comm., regulation of stream flow (Stanford & Ward 1989; 2001), but as larvae they grow and develop rela- Power 1992b). Headwaters (A < 10 km2), if not tively slowly. In various mainstems of the Eel dewatered, provide the most stable environments River (drainage areas from 114 km2 to 1929 km2), available in river drainages. Sands and gravels are we have found that invulnerable invertebrate transported through these boulder and bedrock grazers are less resilient to flood scour than vulner- dominated reaches, but affect a relatively small able grazers. During drought years that lack scour- area of the bed. In the Eel River drainage, ing winter floods, or in regulated channels where McNeely has found very high densities (>1000 flood scour is artificially eliminated, densities of small individuals per m2) of the invulnerable armored or sessile grazers increase by up to two armored caddisfly Glossosoma in dark headwater orders of magnitude (Power 1992b; Wootton et al. channels. This inedible grazer occurs even in the 1996). At these densities, invulnerable grazers smallest headwaters sampled (e.g. Sugar Creek, sequester algal productivity (in Fig. 2a, this would drainage area <1km2 (Finlay 2000; F.C. McNeely, mean that they lower resource density to the point unpubl. data, 2001). Whether Glossosoma popu- where uptake rates by edible taxa are less than lations are maintained in headwaters by adult those by inedible taxa). As a consequence, top- dispersal, or because local populations are not down and bottom-up food chains shorten to two frequently obliterated by disturbance, remains to trophic levels (Power 1992a, 1995; Power et al. be determined. 1996; Wootton et al. 1996; Parker & Power 1997). Secondary productivity (co-influenced down river networks by both primary productivity and RIVER DISCONTINUUA AND temperature) may interact with disturbance NETWORK EFFECTS regimes to strike different balances between vul- nerable and invulnerable primary consumers at For 20 years, the (Van- different network positions. For example, if distur- note et al. 1980) has been our dominant concep- bance regimes are similar in unproductive tribu- tual framework for structure. taries and more productive mainstems, recovery of This model emphasizes downstream gradients in late successional, inedible taxa should be more energy inputs and other variables. Many key hab- rapid in productive mainstems where these taxa itat features in drainage networks, however, are can garner energy more rapidly. Surveys of food patchy and abruptly discontinuous, on large (Stan- web recovery in the sunny South Fork Eel (A 100– ford et al. 1988; Stanford & Ward 1989; Mont- 130 km2) and its less productive , Elder gomery 1999) or small (Townsend 1989; Pringle Food webs in river networks 463 et al. 1988) scales. In some cases, structures that Two hypotheses were proposed by Kupferberg retain fine sediments or water translocate down- (1996) for this pattern. First, adult frogs that stream conditions to upstream positions. Log jams overwinter in the tributaries may simply move to retain pockets of fine gravels high in tributaries the closest mainstem reach to breed. Alternatively, that would otherwise be too steep to retain them the small deltas that form at these sites from cob- and, consequently, extend upstream spawning ble and boulder accumulations are sought out by habitat for salmonids (Abbe & Montgomery frogs because they provide areas where changes in 1996). Woody debris also provide low flow mainstem discharge impose less risk of detach- refuges during high flows for young salmonids and ment or desiccation to frog egg masses. Deposits other weakly swimming organisms, allowing them of cobbles and boulders from tributaries also accu- to persist in steeper regions of the network during mulate along mainstem shorelines to form cobble- spates. Pools scoured downstream from large logs bar habitats providing important cover and ther- can reach the water table after more aggraded mal environments for riparian lizards (Sabo 2000). channel reaches have dried up. Log jams are less Rana boylii tadpoles and caddis larvae are impor- abundant and more ephemeral downstream after tant primary consumers in the river food web. mainstem channels become wider than the tallest Lizards prey on and are subsidized by aquatic trees and can no longer retain them. Dams, engi- insect emergence (Sabo & Power in press a, b). neered by humans (Stanford et al. 1988) or Food web linkages of larval lampreys are still (Naiman et al. 1988), or imposed by landslides poorly known, but possibly important because also transplant downstream conditions (finer sedi- of the high abundance of these interstitial ments, shallower channel slopes, deep lentic detritivores. habitats) upstream in drainage networks. The A second effect of network structure is to repeat- importance of natural and human-engineered edly juxtapose very different habitats, where small ‘serial discontinuities’ has been pointed out by tributaries enter mainstems. At these confluences, several authors (Stanford et al. 1988; Stanford & the adjacency of habitats with contrasting environ- Ward 1989; Montgomery 1999), but their general mental conditions makes it possible for mobile effects on food webs remain poorly known. organisms to exploit widely varying environments. Network effects are another source of longitudi- Juxtaposition of productive mainstems and unpro- nal discontinuities. We describe three effects that ductive tributary habitats in river networks sets the hierarchical network structure of river drain- the stage for ‘spillover predation’ (Holt 1985; ages impose on organisms and food webs, which Oksanen 1990), in which predators from the more to our knowledge have received little or no atten- productive habitat opportunistically take prey tion from stream ecologists. First, the network from adjacent habitats that would be too unpro- structure of river drainages creates distinct nodes ductive, if isolated, to maintain predator popula- where tributaries join mainstems and inject pulses tions (Fig. 3). Spillover predation at confluences of extra water, sediment, organisms or organic might be intensified by the extra enrichment of matter, ranging from fine particulate matter to resources described above. Therefore, being near large logs. These local loads may create particu- confluences might lengthen and strengthen top- larly rich environments near the nodal points in down food chains in the darker tributaries, with mainstems before flows disperse these materials severe adverse impacts on tributary prey that downstream. In the South Fork Eel, there are would not be resilient to losses to subsidized main- conspicuous loadings of caddisfly larvae at tri- stem predators. This prediction remains, to our butary junctions, where they probably accumulate knowledge, untested. after drifting down tributaries. Lamprey larvae Similar effects can arise when predators (or other (Lampetra tridentata) aggregate at these junctions, web members) need two habitats that provide dif- probably tracking enrichments of fine particulate ferent essential requirements. For example, where organic matter in the sediments. The foothills yel- small, dark, steep, narrow tributaries join low gra- low legged frog, Rana boylii, which is in decline dient, wide, sunny, productive mainstems, foragers throughout much of its range, maintains tradi- such as riparian spiders or salamanders may have tional oviposition sites at tributary confluences. access within their foraging ranges to productive 464 M. E. Power and W. E. Dietrich

Headwater web

Predator

Terrestrial Edible Inedible spider grazer grazer

Detritus Attached algae Fig. 3. Simulation models Physical advection and for interactions of, and fluxes biotic dispersal between, webs and organ- isms at network nodes. These models can be parameterized Biotic dispersal based on trait-based and site- based performances of organ- Mainstem web isms and on disturbance/ recovery regimes representa- Predator tive of the local environ- ments (e.g. unproductive, Predator but less frequently disturbed headwaters, linked to pro- ductive, more frequently

Edible Inedible scoured mainstems). Solid grazer grazer arrows show the direction of energy transfer, dotted arrows depict movements Algae between headwater and mainstem habitats. habitats where food abounds and to cool unpro- propagules are diluted among many alternative ductive habitats where they can escape desiccating channels if they move upstream, while they heat and wind. In terms of Getz’s metaphysiolog- become concentrated if they move downstream. ical parameters, such a population could lower its For organisms, this means that within-channel compensation point κ as individuals take refuge species and genetic diversity may decline upstream from harsh conditions while resting in headwaters, (Horwitz 1978; Hughes et al. 1995; Schmidt et al. while raising its resource uptake rate g as individ- 1995), while beta-diversity (variation among uals foray into more productive mainstem habi- channels) increases. These effects probably exert tats, possibly during times of the day or night more influence on the population genetics and when that habitat is more benign. These adjacent micro-evolution of fully aquatic organisms such as conditions may allow a number of organisms to fish (Turner & Grosse 1980; Turner et al. 1984) persist and influence food webs in habitats that than on or aquatic insects in which the could not, if isolated, support their populations. adults sometimes disperse over drainage divides Third, the hierarchical structure of river net- (Jackson & Resh 1989; Jackson & Resh 1991). works sets up a system of spatially separated, What might this hierarchical structure mean for repeating environments that are rarely directly food web interactions? Headwater channels that linked. To a mobile organism, the channel network subsample mainstem species assemblages should presents a decision . Upstream movement will have lower than mainstems, offset be dispersive, while downstream transport and to some degree if habitat specialists occur there migration will be concentrative. Individuals or and not at lower network positions. Headwater Food webs in river networks 465 webs may include populations that are released the standing crop of organic matter in the channel from competitors, predators or from the parasites was terrestrial detritus, a common local caddisfly that limit them at lower network positions. This was inferred, on the basis of gut contents and release might occur either stochastically because of relative assimilation rates, to be built primarily of the upstream dispersive effects of networks or algae (Mayer & Likens 1987). Isotopic patterns deterministically because the enemy in question suggest that scrapers are built of algae as far up in cannot tolerate conditions imposed upstream by the drainage networks as they can be collected, and the longitudinal environmental gradient. These even where algal accrual on rocks is undetectable alternatives could be distinguished by transplant (Finlay 2000, 2001). Clearly, standing crops of this experiments, as well as by extensive sampling to rapidly growing, rapidly depleted resource do not compare the composition of assemblages at equiv- reveal its importance as an energy base in some alent headwater positions in the networks. If head- headwater streams. water channel food webs were self-contained We speculate, however, that the relationship compartments, we might expect shorter food between biomass and food web impact may be chains with stronger top-down effects of taxa at stronger in headwaters than in lower channels for lower trophic levels. As discussed, however, these , particularly for mobile taxa. This may be impacts might be offset by more intimate links true for at least four reasons: with terrestrial ecosystems in headwaters. If terres- trial consumers limit aquatic prey or if terrestrial 1 Space. First and simplest, habitat volumes are detritus subsidizes aquatic predators so that they smaller, so the same biomass of organisms rep- can persist despite inadequate aquatic productiv- resents a higher density in headwaters than in ity, food chains would be less likely to shorten lower reaches. At the same absolute biomass, upstream. interactions among mobile individuals in head- waters should be more frequent on a per capita basis, unless cooler temperatures or different CONCLUSIONS habitat structure greatly reduce their activity or encounters. We introduced this review by emphasizing that 2 Time. Residence times of individuals in a given population control chains and energy flow paths reach are likely to be longer in headwaters, are distinct in food webs. We close by revisiting where upstream dispersal is curtailed and down- these distinctions in headwaters and lower main- stream washout during bed scour is rare. If stem channels. Disparities between energy flow organisms in lower mainstems are more tran- paths and population control paths are revealed sient, they would be more likely to derive their when the experimental removal of organisms sustenance and translocate nutrients outside the releases ‘hidden trophic levels’ (Paine 1980). local habitat, weakening their local impacts. In Manipulations in the South Fork Eel watershed contrast, nutrient translocation or subsidized have revealed hidden trophic levels in both predation impacts could be locally strong where mainstems and headwaters. In lower mainstems, topography (deep pools, etc.) creates removal of top predators (larger fish) released temporary holding areas or barriers for migrant guilds of smaller predators (e.g. odonate larvae, fishes or other dispersing organisms. fish fry) that had previously been rare and incon- 3 Energy. Because headwater habitats are less pro- spicuous (Power 1990a). In the headwaters, ductive, food extraction and nutrient excretion removal of glossosomatid caddisflies from a reach by individual organisms should have greater per with <5% open canopy released previously incon- biomass effects on local energy flow and mate- spicuous colonial , which grew to substan- rial cycling. These effects would be enhanced in tial biomass (F.C. McNeely, unpubl. data, 2001). headwater channels with vegetative or geomor- Careful observational studies can also detect cases phic structures that increased retention of mate- in which biomass does not correspond to potential rials or organisms (Meyer et al. 1988; Palmer functional importance in food webs. In a New et al. 1996). In productive downstream habi- Hampshire headwater stream, where up to 99% of tats, where biomass turnover could be faster, 466 M. E. Power and W. E. Dietrich

organisms with low biomass may play stronger tive studies to reveal the general effects of spatial dynamic roles in food webs than in less produc- heterogeneity on food web interactions and the tive headwaters. resource flow paths affecting them. Mensurative 4 Chance assembly and shorter food chains. Because and manipulative experiments that examine spe- upstream movements disperse and separate cies performances and interactions at different organisms, headwaters may contain subsets of drainage network positions, and relate these to mainstem species assemblages in which certain habitat structure, disturbance, productivity or populations are released from competitors, other systematically varying conditions, should predators or parasites that limit them at lower provide general insights into the relationships network positions. These releases would be between biomass and energy flow and interaction more likely for taxa at low or intermediate strengths. What are the potential intrinsic growth trophic levels (e.g. primary consumers) because rates of organisms if they are not resource limited predators require their prey for persistence at various landscape positions? Will in situ food (Holt 1996), and in size structured aquatic food webs change when basal resources are experimen- webs they are less likely to be physically accom- tally augmented? Where will consumer manipu- modated as habitat volumes shrink upstream. lations have strong effects? How are species Therefore, relatively high primary consumer interactions influenced by resource fluxes and, con- densities in headwaters probably indicate two- versely, where, when and how do species interac- level, rather than four-level food chains, with tions affect the flow paths of resources and their these grazers exerting top-down control on their capture by particular web members? Certain preferred resources. mobile organisms with wide environmental toler- ances (or life history stages that use different hab- These observations together suggest that energy itats) may experience river food webs as network flow, biomass and interaction strength might be based. These organisms could sample the range of more directly related in headwater food webs than resources and conditions offered throughout much in lower watershed positions, at least for consumer of the system, choosing the best available at any populations. This prediction might be tested ini- given time. By behavioral ‘ideal free’ responses tially by staging population or guild removal (Fretwell & Lucas 1970; Power 1984b; Oksanen experiments at upstream and downstream water- et al. 1995), they could track and damp out local shed positions, to compare per capita or per biom- resource pulses as these arose. Other less mobile or ass impacts on other web members, or other tolerant organisms could be thought of as net- ecosystem properties of interest. To explore the work-controlled. Their distributions or the source topic more deeply, we would have to learn more areas of their resources might be locally restricted than we currently know about the dispersal and because of drainage network features that restrict foraging ranges of organisms in rivers and water- their percolation through the system (e.g. water- sheds, as well as the spatial and temporal scales of falls that block fish movements or predator or fluxes of their resources. Tracers (stable isotopes, warm water barriers in mainstems that preclude trace elements, exotic contaminants or genetic the dispersal of headwater species from one tribu- markers) coupled with experiments will increas- tary to another). More tracking and tracer studies ingly reveal the spatial scales of food webs that are necessary to ascertain where along this contin- influence energy pathways, population limiting uum particular taxa fall. We suggest that synthe- chains and their relationship in food webs (e.g. sizing LaGrangian tracer studies with Eulerian Finlay et al. 1999; Power & Rainey 2000; Finlay manipulative experiments in comparative studies et al. in press). of food webs at different drainage network posi- Habitat structure and environmental conditions tions and at similar network positions across in drainage networks are partially predictable from landscapes will add useful realism to our under- controls on geomorphic processes by local gradient standing of the links between energy flow and and drainage areas, as well as aspect, climate, land interactions in food webs and landscape hetero- use and geologic parent material. These general geneity. Progress along these lines should reveal landscape relationships could be used in compara- new principles linking food web energetics and Food webs in river networks 467 dynamics to the nested scales of heterogeneity in SHEARER J., SPRULES W. G., VANNI M. J. & landscapes that constrain both. ZIMMERMAN A. P. (1991) Patterns of primary production and herbivory in 25 North American Lake ecosystems. In: Comparative Analyses of Eco- ACKNOWLEDGEMENTS systems (eds J. Cole, G. Lovett. & S. Findlay) pp. 67–96. Springer-Verlag, Berlin. CARPENTER S. R., KITCHELL J. F., HODGSON J. We thank Blake Suttle, Camille McNeely and R., COCHRAN P. A., ELSER J. J., ELSER M. M., Wayne Getz for comments that greatly improved LODGE D. M., KRETCHMER D., HE X. & VON the manuscript, the National Science Foundation ENDE C. N. (1987) Regulation of lake primary and the Center for Water and Wildland Resources productivity in food web structure. Ecology 68: of the University of California for financial sup- 1863–1876. port, the University of California Natural Reserve CARPENTER S. R., LATHROP R. C. & MUNOZ- System and the Angelo and Steel families for pro- DEL-RIO A. (1993) Comparison of dynamic viding the Angelo Coast Range Reserve as a site models for edible . Canadian Jour- protected for field research, and the scientists and nal of and Aquatic Sciences 50: 1757–1767. students at the Center for Ecological Research at COHEN J. E. (1977) Ratio of prey to predators in the University of Kyoto and Hokkaido University food webs. 270: 165–167. for their hospitality and intellectual stimulation. DAVIS-COLLEY R. J. & QUINN J. M. (1998) Stream lighting in five regions of North Island, We draw life long inspiration from the and New Zealand: control by channel size and riparian contributions to ecology of Gary Polis, Shigeru vegetation. New Zealand Journal of Marine and Nakano and Masahiko Higashi. Freshwater Research 32: 591–605. DE ROOS A. M., MCCAULEY E. & WILSON W. G. (1991) Mobility versus density-limited REFERENCES predator-prey dynamics on different spatial scales. Proceedings of the Royal Society of London B 246: ABBE T. B. & MONTGOMERY D. R. (1996) Large 117–122. woody debris jams, channel hydraulics and habitat DEANGELIS D. L., GOLDSTEIN R. A. & O’NEILL formation in large rivers. Regulated Rivers 12: 201– R. V. (1975) A model for trophic interaction. 221. Ecology 56: 881–892. ANDERSON W. B. & POLIS G. A. (1999) Nutrient DIETRICH W. E. & DUNNE T. (1978) Sediment fluxes from water to land: affect plant budget for a small catchment in mountainous ter- nutrient status on Gulf of California islands. Oeco- rain. Zeitschrif Für Geomorphologie 29: 191–206. logia (Berlin) 118: 324–332. DIETRICH W. E., WILSON C. K., MONTGOMERY BARKAI A. & MCQUAID C. (1988) Predator-prey D. R. & MCKEAN J. (1993) Analysis of role reversal in a marine benthic ecosystem. Science Thresholds, Channel Networks, and Landscape 242: 62–64. Morphology Using a Digital Terrain Model. The CADENNASO M. L. & PICKETT S. T. A. (2000) Journal of 101: 259–278. Linking forest edge structure to edge function: DUFFY J. E. & HAY M. E. (1991) Amphipods are Mediation of damage. J. Ecology 88: 31– not all created equal: a reply to Bell. Ecology 72: 44. 354–358. CADENNASO M. L., PICKETT S. T. A. & WEATH- DUNNE T., AGEE J., BEISSINGER S., DIETRICH ERS K. C. Effect of landscape boundaries on the W. E., GRAY D., POWER M. E., RESH V. H., flux of nutrients, detritus, and organisms. In: Food RODRIGUES K. (2001) A scientific basis for the Webs and Landscapes (eds G. A. Polis, M. E. Power. prediction of cumulative watershed effects. & G. Huxel) University of Chicago Press, Chicago http://nature.berkeley.edu/forestry.Universityof (in press). CaliforniaWildlandResourceCenterReportno. CARPENTER S. R. (1988) Transmission of variance 46.103 through lake food webs. In. Complex Interactions in ELTON C. (1927) Ecology. Sidgwick & Lake Communities (ed. S. R. Carpenter) pp. 119– Jackson, London. 135. Springer Verlag, New York. FAGAN W. F., CANTRELL R. S. & COSNER C. CARPENTER S. R., FROST T. M., KITCHELL J. F., (1999) How habitat edges change species interac- KITCHELL KRATZ T. K., SCHINDLER D. W., tions. American Naturalist 153: 165–182. 468 M. E. Power and W. E. Dietrich

FEMINELLA J. W., POWER M. E. & RESH V. H. interactions, and the coexistence of prey species. (1989) Periphyton responses to invertebrate graz- American Naturalist 124: 377–406. ing and riparian canopy in three northern Califor- HOLT R. D. (1985) of two- nia coastal streams. 22: 445– patch environments: some anomalous conse- 457. quences of an optimal habitat distribution. Theo- FINLAY J. C. (2001) Stable carbon isotope ratios of retical Population Biology 28: 181–208. river biota: Implications for energy flow in lotic HOLT R. D. (1996) Food webs in space: An island food webs. Ecology 82: 1052–1064. biogeographic perspective. In: Food Webs (eds G. FINLAY J. C., KHANDWALA S. & POWER M. E. W. Polis. & K. O. Winemiller) pp. 313–323. Spatial scales of carbon flow in a river food web. Chapman & Hall, New York. Ecology (in press). HOLT R. D. (2002) Food webs in space: On the FINLAY J. C., POWER M. E. & CABANA G. (1999) interplay of dynamic instability and spatial pro- Effects of water velocity on algal carbon isotope cesses. Ecological Research 17: 261–273. ratios: Implications for river food web studies. HORWITZ R. J. (1978) Temporal variability pat- and 44: 1198–1203. terns and the distributional patterns of stream FRETWELL S. D. (1977) The regulation of plant fishes. Ecological Monographs 48: 307–321. communities by food chains exploiting them. Per- HUGHES J. M., BUNN S. E., KINGSTON D. M. & spectives in Biology and Medicine 20: 169–185. HURWOOD D. A. (1995) Genetic differentiation FRETWELL S. D. (1987) Food chain dynamics: the and dispersal among populations of Paratya aus- central theory of ecology? Oikos 50: 291–301. traliensis (Atyidae) in rainforest streams in south- FRETWELL S. D. & LUCAS H. L. (1970) On terri- east Queensland, Australia. Journal of the North torial behavior and other factors influencing hab- American Benthological Society 14: 158–173. itat distribution in birds. I. Theoretical HUTCHINSON G. E. (1959) Homage to Santa Rosa- development. Acta Biotheoretica 19: 16–36. lia or why are there so many kinds of animals? GASITH A. & RESH V. H. (1999) Streams in med- American Naturalist 93: 145–159. iterranean climate regions: Abiotic influences and HYNES H. B. N. (1975) The stream and its valley. biotic responses to predictable seasonal events. Internationale Vereinigung Fur Theoretische und Annual Review of Ecology and Systematics 30: 51–81. Angewandte Limnologie, Verhandlungen 19: 1–15. GETZ W. M. (1991) A unified approach to multi- JACKSON J. K. & RESH V. H. (1989) Distribution species modeling. Natural Resource Modeling 5: and abundance of adult aquatic insects in the for- 393–412. est adjacent to a northern California stream. Envi- GETZ W. M. (1993) Metaphysiological and evolu- ronmental Entomology 18: 278–283. tionary dynamics of populations exploiting con- JACKSON J. K. & RESH V. H. (1991) Periodicity stant and interactive resources: r – K selection in mate attraction and flight activity of three spe- revisited. 7: 287–305. cies of caddisflies (Trichoptera). Journal of the North GETZ W. M. (1994) Beyond Lotka–Volterra– American Benthological Society 10: 198–209. Gause–Holling and MacArthur–Tilman models. JACKSON J. K. & FISHER S. G. (1986) Secondary In: Lecture Notes in Biomathematics. (ed. S. A. Levin) production, emergence, and export of aquatic pp. 411–442. .Springer-Verlag, New York. insects of a Sonoran Desert stream. Ecology 67: GRAY L. J. & FISHER S. G. (1981) Postflood recol- 629–638. onization pathways of macroinvertebrates in a JENKINS B., KITCHING R. L. & PIMM S. L. (1992) lowland Sonoran Desert stream. American Midland Productivity, disturbance and food web structure Naturalist 106: 249–257. at a local spatial scale in experimental container HAIRSTON N. G., SMITH F. E. & SLOBODKIN L. habitats. Oikos 65: 249–255. B. (1960) Community structure, population con- JOHANSSON A. (1991) Caddis larvae cases trol, and . American Naturalist 94: (Trichoptera, Limnephilidae) as anti-predatory 421–425. devices against brown trout and sculpin. Hydrobi- HASTINGS H. M. & CONRAD M. (1979) Length ologia 211: 185–194. and evolutionary stability of food chains. Nature KERFOOT W. C. & DEMOTT W. R. (1984) Food 282: 838–839. web dynamics: Dependent chains and vaulting. HERSHEY A. E. (1987) Tubes and foraging behavior In: Trophic Interactions Within Aquatic Ecosystems in larval Chironomidae: implications for predator (eds D. G. Meyers & J. R. Strickler) pp. 347–382. avoidance. Oecologia 73: 236–241. Westview Press, Washington. HOLT R. D. (1984) Spatial heterogeneity, indirect KERFOOT W. C. & SIH A. (1987) Predation: Direct Food webs in river networks 469

and Indirect Impacts on Aquatic Communities. Uni- MEYER J. L., MCDOWELL W. H. & BOTT T. L., versity Press of New England, Hanover. ELWOOD J. W., ISHIZAKI C., MELACK J. M., KUPFERBERG S. J. (1996) Hydrologic and geomor- PECKARSKY B. L., PETERSON B. J. & RUBLEE phic factors affecting conservation of a river- P. A. (1988) Elemental dynamics in streams. Jour- breeding frog (Rana boylii). Ecological Applications nal of the North American Benthological Society 7: 6: 1332–1344. 410–432. LAURANCE W. F., DIDHAM R. K. & POWER M. MITTELBACH G. G., OSENBERG C. W. & E. (2001) Ecological boundaries: a search for syn- LEIBOLD M. A. (1988) Trophic relations and thesis. Trends in Ecology and Evolution 16: 70–71. ontogenetic niche shifts in aquatic ecosystems. In: LEIBOLD M. A. (1989) Resource edibility and the Size-Structured Populations (eds B. Ebenman. & L. effects of predators and productivity on the out- Persson) pp. 219–235. Springer-Verlag, Berlin. come of trophic interactions. American Naturalist MONTGOMERY D. R. (1999) Process domains and 134: 922–949. the river continuum. Journal of the American Water LEOPOLD L. B., WOLMAN M. G. & MILLER J. P. Resources Association 35: 397–410. (1964) in . Freeman, MONTGOMERY D. R. & BUFFINGTON J. M. San Francisco. (1993) Channel Classification, Prediction of Channel LEVIN S. A. (1999) Fragile Dominion. Helix Books, Response, and Assessment of Channel Condition. Wash- Reading. ington Department of Natural Resources Report LI J. L. & GREGORY S. V. (1989) Behavioral TFW-SH10-93–002, Olympia, WA. changes in the herbivorous caddisfly Dicosmoecus MONTGOMERY D. R. & BUFFINGTON J. S. (1997) gilvipes (Limnephilidae). Journal of the North Channel-reach morphology in mountain drainage American Benthological Society 8: 250–259. basins. Geological Society of America Bulletin 109: LI H., LAMBERTI G. A., PEARSONS T. N., TAIT 596–611. C. K., LI J. & BUCKHOUSE J. (1994) Cumulative MONTGOMERY D. R. & DIETRICH W. E. (1988) effects of riparian disturbances along high desert Where do channels begin? Nature 336: 232–234. trout streams of the John Day Basin, Oregon. NAIMAN R. J., JOHNSTON C. A. & KELLEY J. C. Transactions of the American Fisheries Society 123: (1988) Alteration of North American Streams by 627–640. . Bioscience 38: 753–762. MARKS J. C., POWER M. E. & PARKER M. S. NAKANO S., MIYASAKA H. & KUHARA N. (1999) (2000) disturbance, algal productivity, and Terrestrial-aquatic linkages: Riparian arthropod interannual variation in food chain length. Oikos inputs alter trophic cascades in a stream food web. 90: 20–27. Ecology 80: 2435–2441. MAYER M. S. & LIKENS G. E. (1987) The impor- NAKANO S. & MURAKAMI M. (2001) Reciprocal tance of algae in a shaded headwater stream as subsidies: Dynamic interdependence between ter- food for an abundant caddisfly (Trichoptera). Jour- restrial and aquatic food webs. Proceedings of the nal of the North American Benthological Society 6: National Academy of Sciences 98: 166–170. 262–269. OKSANEN T. (1990) Exploitation ecosystems in MCAULIFFE J. R. (1983) Competition, colonization heterogeneous habitat complexes. Evolutionary patterns, and disturbance in stream benthic com- Ecology 4: 220–234. munities. In: Stream Ecology. (eds J. R. Barnes. & OKSANEN L., FRETWELL S. D., ARRUDA J. & G. W. Minshall) pp. 137–156. Plenum, New NIEMELA P. (1981) Exploitation ecosystems in York. gradients of primary productivity. American Nat- MCAULIFFE J. R. (1984) Resource depression by a uralist 118: 240–261. stream herbivore: effects on distributions and OKSANEN T., POWER M. E. & OKSANEN L. abundances of other grazers. Oikos 42: 327–333. (1995) Ideal free habitat selection and consumer- MCQUEEN D. J., JOHANNES M. R. S., POST J. R., resource dynamics. American Naturalist 146: 565– STEWART T. J. & LEAN D. R. S. (1989) Bottom- 585. up and top-down impacts on freshwater pelagic PAINE R. T. (1980) Food webs: linkage, interaction community structure. Ecological Monographs 59: strength and community infrastructure. Journal of 289–309. Animal Ecology 49: 667–685. MENGE B. A. & SUTHERLAND J. P. (1976) Species PAINE R. T. (1988) Food webs: road maps of inter- diversity gradients: synthesis of the roles of pre- actions or grist for theoretical development? Ecol- dation, competition, and temporal heterogeneity. ogy 69: 1648–1654. American Naturalist 110: 351–369. PALMER M. A., ARENSBURGER P., MARTIN A. P. 470 M. E. Power and W. E. Dietrich

& DENMAN D. W. (1996) Disturbance and POWER M. E. (1984b) Habitat quality and the dis- patch-specific responses: The interactive effects of tribution of algae-grazing catfish in a Panama- woody debris and floods on lotic invertebrates. nian stream. Journal of Animal Ecology 53: 357– Oecologia 105: 247–257. 374. PARKER M. S. (1991) Relationship between cover POWER M. E. (1987) Predator avoidance by grazing availability and larval Pacific giant fishes in temperate and tropical streams: impor- density. Journal of Herpetology 25: 355–357. tance of stream depth and prey size. In: Predation: PARKER M. S. (1993) Size-selective predation on Direct and Indirect Impacts on Aquatic Communities benthic macroinvertebrates by stream-dwelling (eds W. C. Kerfoot & A. Sih) pp. 333–351. Uni- salamander larvae. Archivs Fur Hydrobiologia 128: versity Press of New England, Hanover. 385–400. POWER M. E. (1990a) Effects of fish in river food PARKER M. S. & POWER M. E. (1997) Effect of webs. Science 250: 411–415. stream flow regulation and absence of scouring POWER M. E. (1990b) Benthic turfs versus floating floods on trophic transfer of biomass to fish in mats of algae in river food webs. Oikos 58: 67– Northern California rivers. Technical Completion 79. Report. In: University Of California Water POWER M. E. (1992a) Top down and bottom up Resources Center, UCAL-WRC-W-825, Davis. forces in food webs: do have primacy? Ecol- PERSSON L., BENGTSSON J., MENGE B. A. & ogy 73: 733–746. POWER M. E. (1996) Productivity and consumer POWER M. E. (1992b) Hydrologic and trophic con- regulation: Concepts, patterns, and mechanisms. trols of seasonal algal blooms in northern Califor- In: Food Webs: Integration of Patterns and Dynamics nia rivers. Archiv Fur Hydrobiologie 125: 385–410. (eds G. A. Polis & K. O. Winemiller) pp. 396– POWER M. E. (1995) , food chains, and eco- 434. Chapman & Hall, Inc., New York. system processes in rivers. In: Linking Species and PETRANKA J. W. (1983) Fish predation: a factor Ecosystems (eds C. G. Jones. & J. H. Lawton) pp. affecting the spatial distribution of a stream- 52–60. Chapman & Hall, New York. breeding salamander. Copeia 1983: 624–628. POWER M. E., KUPFERGERG S. J., MINSHALL PIMM S. L. (1979) The structure of food webs. The- G. W., MOLLES M. C. & PARKER M. S. (1997) oretical Population Biology 16: 144–158. Sustaining Western Aquatic Food Webs. In: PIMM S. L. (1982) Food Webs. Chapman & Hall, Aquatic Ecosystems Symposium. Report to the Western London. Water Policy Review, a Presidential. Advisory PIMM S. L. & KITCHING R. L. (1987) The deter- Commission (ed. W. C. Minckley) pp. 45–61. minants of food chain length. Oikos 50: 302–307. Department of Commerce, National Technical PIMM S. L. & LAWTON J. H. (1977) The numbers Information Service, Springfield. of trophic levels in ecological communities. POWER M. E., PARKER M. S. & WOOTTON J. T. Nature 268: 329–331. (1996) Disturbance and food chain length in POLIS G. A., ANDERSON W. B. & HOLT R. D. rivers. In: Food Webs: Integration of Patterns and (1997) Toward an integration of landscape and Dynamics (eds G. A. Polis & K. O. Winemiller) food web ecology: The dynamics of spatially sub- pp. 286–297. Chapman & Hall, New York. sidized food webs. Annual Review of Ecology and POWER M. E. & RAINEY W. E. (2000) Food webs Systematics 28: 289–316. and resource sheds: Towards spatially delimiting POLIS G. A., MYERS C. A. & HOLT R. D. (1989) trophic interactions. In: Ecological Consequences of The evolution and ecology of intraguild preda- Habitat Heterogeneity (eds M. J. Hutchings, E. A. tion: competitors that eat each other. Annual John & A. J. A. Stewart) pp. 291–314. Blackwell Review of Ecology and Systematics. 20: 297–330. Scientific, Oxford. POST D. M., PACE M. L. & HAIRSTON N. G. POWER M. E., RAINEY W. E., PARKER M. S., (2000) Ecosystem size determines food-chain SABO J. L., SMYTH A., KHANDWALA S., length in lakes. Nature 405: 1047–1049. FINLAY J. C., MCNEELY F. C., MARSEE K. & POWER M. E. (1983) Grazing responses of tropical ANDERSON C. River to watershed subsidies in freshwater fishes to different scales of variation in an old-growth conifer forest. In: Food Webs and their food. Environmental Biology of 9: 103– Landscapes (eds G. A. Polis, M. E. Power & G. 115. Huxel) University Chicago Press, Chicago. (in POWER M. E. (1984a) Depth distributions of press). armored catfish: predator-induced resource avoid- POWER M. E. & STEWART A. J. (1987) Distur- ance? Ecology 65: 523–528. bance and recovery of an algal assemblage follow- Food webs in river networks 471

ing flooding in an Oklahoma [USA] stream. stream power and the influence of sediment sup- American Midland Naturalist 117: 333–345. ply. Geophysical Monograph 107: 237–260. POWER M. E., STEWART A. J. & MATTHEWS W. SOUSA W. P. (1984) The role of disturbance in J. (1988) Grazer control of algae in an Ozark natural communities. Annual Review of Ecology and Mountain stream [USA]: Effects of short-term Systematics. 15: 353–391. exclusion. Ecology (Tempe) 69: 1894–1898. STANFORD J. A., HAUER F. R. & WARD J. V. PRINGLE C. M., NAIMAN R. J., BRETSCHKO G., (1988) Serial discontinuity in a large river system. KARR J. R., OSWOOD M. W., WEBSTER J. R., Verhandlugen Internationale Vereingen Theoretische WELCOMME R. L. & WINTERBOURN M. J. Angewa. Limnologie 23: 1114–1118. (1988) in lotic systems the stream STANFORD J. A. & WARD J. V. (1989) Serial dis- as a mosaic. Journal of the North American Bentho- continuities in a Rocky Mountain river. I. Distri- logical Society 7: 503–524. bution and abundance of Plecoptera. Regulated RADER R. B. (1997) A functional classification of Rivers 3: 169–175. drift: Traits that influence invertebrate availability TAIT C. K., LI J. L., LAMBERTI G. A., PEARSONS to salmonids. Canadian Journal of Aquatic and T. N. & LI H. W. (1994) Relationships between Fisheries Sciences 54: 1211–1234. riparian cover and the community structure of ROMO S., VAN DONK E., GLYSTRA R. & GULATI high desert streams. Journal of the North American R. (1996) A multivariate analysis of phytoplank- Benthological Society 13: 45–56. ton and food web changes in a shallow biomanip- TILMAN D. & KAREIVA P. (1997) . ulated lake. Freshwater Biology 36: 683–696. Princeton University Press, Princeton. SABO J. L. (2000) River-watershed exchange: Effects TOWNSEND C. R. (1989) The patch dynamics con- of rivers on the population and community cept of stream community ecology. Journal of the dynamics of riparian lizards (Sceloporus occiden- North American Benthological Society 8: 36–50. talis). PhD thesis. University California, Berkeley, TURNER B. J. & GROSSE D. J. (1980) Trophic Berkeley (unpubl.). differentitiation in Ilyodon, a genus of stream- SABO J. L. & POWER M. E. Aggregation of lizards dwelling goodeid fishes: versus ecolog- along rivers: subsidy input tracking and short- ical polymorphism. Evolution 34: 259–270. term indirect effects on in situ prey. Ecology (in TURNER B. J., GRUDZIEN T. A., ADKISSON K. press a). P. & WHITE M. M. (1984) Evolutionary genetics SABO J. L. & POWER M. E. River watershed of trophic differentiation in goodeid fishes of the exchange: Effects of riverine subsidies on riparian genus Ilyodon. In: Evolutionary Ecology of Neotropi- lizards and their terrestrial prey. Ecology (in press cal Freshwater Fishes (ed. T. M. Zaret) pp. 81–94 b). Junk, The Hague. SCHLAPFER F. & SCHMID B. (1999) Ecosystem VANNOTE R. L., MINSHALL G. W., CUMMINS K. effects of : a classification of hypothe- W., SEDELL J. R. & CUSHING C. E. (1980) The ses and exploration of empirical results. Ecological river continuum concept. Canadian Journal of Applications 9: 893–912. Fisheries and Aquatic Sciences 37: 130–137. SCHMIDT S. K., HUGHES J. M. & BUNN S. E. VOGEL S. (1981) Life in Moving Fluids. Princeton (1995) Gene flow among conspecific populations University Press, Princeton. of Baetis sp. (Ephemeroptera): adult flight and WALLACE J. B., EGGERT S. L., MEYER J. L. & larval drift. Journal of the North American Bentho- WEBSTER J. R. (1997) Multiple trophic levels of logical Society 14: 147–157. a forest stream linked to terrestrial litter inputs. SCHOENER T. W. (1973) Population growth regu- Science 277: 102–104. lated by intraspecific competition for energy or WATERS T. F. (1995) Sediment in Streams: Sources, time: some simple representations. Theoretical Pop- Biological Effects and Control. American Fisheries ulation Biology 4: 56–84. Society Monograph, Bethesda. SCHOENER T. W. (1989) Food webs from the small WOOTTON J. T., PARKER M. S. & POWER M. E. to the large. Ecology 70: 1559–1589. (1996) Effects of disturbance on river food webs. SINCLAIR A. R. E. & ARCESE P. (1995) Serengeti II: Science 273: 1558–1560. Dynamics, Management and Conservation of an Eco- WOOTTON J. T. & POWER M. E. (1993) Produc- system. University Chicago Press, Chicago. tivity, consumers, and the structure of a river food SKLAR L. & DIETRICH W. E. (1998) River longi- chain. Proceedings of the National Academy of Sciences. tudinal profiles and bedrock incision models: 90: 1384–1387.

Top View