Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Oecologia (1992) 91:220-228 Oecologia 9 Springer-Verlag 1992 Community organization in streams: the importance of species interactions, physical factors, and chance David D. Hart Academyof Natural Sciences, 1900 Ben Franklin Parkway, Philadelphia,PA 19103, USA Received February3, 1992 / Accepted in revisedform March 30, 1992 Summary. Experimental studies were used to examine the mechanisms governing the distribution and abundance of two major patch types in unshaded reaches of Augusta Creek, Michigan (USA). One patch type is dominated by Cladophora glornerata, a macroalga potentially able to monopolize space, whereas the other type is comprised of a low-growing, epilithic microalgal lawn inhabited by several species of sessile grazers (especially the caddisflies Leucotrichia pictipes and Psychomyia flavida). Cladophora patches are absent from mid-channel sites characterized by current velocities _< ca. 50 cm s- 1; caging experiments indicate that their absence is due to grazing by crayfish (Orconectespropinquus). Cladophora's presence in sites with velocities > 50 cm s-1 apparently results in part because crayfish foraging activity is impaired in high flow regimes. The presence of Cladophora strongly affects various other invertebrates due to its alteration of abiotic and biotic characteristics of the microhabitat. For example, the abundance of sessile grazers (e.g. Leucotrichia and Psychomyia) that inhabit microalgal patches is negatively correlated to the abundance of Cladophora, whereas the abundance of several other invertebrates (e.g. Stenonema mayflies and Taeniopteryx stoneflies) is positively correlated to Cladophora's abundance. Therefore, in some portions of this system, crayfish act as keystone predators because of their ability to regulate the abundance of Cladophora, which in turn has strong positive and negative effects on other components of the community. Cladophora does not always monopolize space at high velocities in the absence of crayfish, however. If sessile grazers arrive at such sites before Cladophora, they can prevent its establishment. Thus, where crayfish are absent, the likelihood that a site will be dominated by either Cladophora patches or sessile grazer microalgal lawn patches depends on two sets of stochastic processes: (1) those that create bare space (e.g. disturbance and grazer emergence); and (2) those controlling the timing of recruitment by Cladophora or grazers at these bare sites. These priority effects (i.e. the ability of grazers and Cladophora to inhibit each other's establishment) contribute to the marked spatial heterogeneity of these two patch types. Collectively, these results demonstrate how interactions between competition, predation, and physical factors can generate a complex mixture of community patterns. Key words: Indirect effects Keystone predators- Plantherbivore interactions - Predator refuges - Priority effects Ecologists continue to seek a unified body of theory that can account for observed variations in the organization of natural communities. In the past, debates often centered on the explanatory power of opposing unitary hypotheses, such as the degree to which communities were regulated by abiotic vs. biotic factors, and the importance of competition vs. predation. There appears to be a subtle but substantive shift occurring in the kinds of questions being asked about ecological communities. For example, more recent approaches explicitly focus on interactions between a variety of ecological factors that can regulate patterns of distribution and abundance within and among communities (e.g. Dunson and Travis 1991). Attempts to ordinate communities with respect to various ecological axes (Schoener 1986; Giller and Gee 1987; Menge and Sutherland 1987) represent a particularly ambitious version of this trend. Thus, there is a growing consensus that pluralistic models are needed to explain how the relative importance of diverse ecological factors governing community organization varies at different spatial and temporal scales. Recently, stream ecologists have also begun to develop and test multifactorial models that can account for patterns of distribution and abundance (e.g. Peckarsky 1983; McAuliffe 1984; Hart 1985a; Power et al. 1985; Schlosser and Ebel 1989; Peckarsky et al. 1990; Power 1990a; Hansen et al. 1991). My study examines how an interplay between physical factors and species interactions controls the organization of a benthic stream community. Specifically, I investigate the complex set of in- 221 teractions controlling the distribution and abundance of two important patch types (one dominated by an arborescent macroalga, the other by a low-growing microalgal lawn) by addressing the following questions: (1) How does macroalgal abundance vary along a gradient in current velocity? (2) Can grazing by crayfish account for the absence of this macroalga from sites with low velocities? (3) Does crayfish grazing, via its effect on macroalgal abundance, have either positive or negative indirect effects on other invertebrates? (4) In the absence of crayfish, what role do sessile grazers play in preventing the macroalga from monopolizing space at high velocities? Taken together, these studies reveal an intricate web of strong interactions that caution against simplistic views of community organization. Methods The study system These studies were carried out in a third order section of Augusta Creek (Michigan, USA), a hard water stream whose flow regime exhibits low variability (Poff and Ward 1989; Hart and Robinson 1990). Previous field experiments in this stream have demonstrated that interactions between epilithic microalgae and grazing insects are tightly coupled, resulting in strong competition among grazers for limited algal resources (Hart 1985b, 1987; Hart and Robinson 1990). Two visually conspicuous and distinctly different patch types often predominate in Augusta Creek's unshaded riffles and runs from spring to fall (Fig. 1). The first is dominated by arborescent growths (several-to-many mm thick) of the filamentous chlorophyte Cladophoraglomerata, along with associated epiphytes and a variety of invertebrates. The second patch type is a low growing ( < < 1 mm thick) epilithic microalgal lawn composed primarily of diatoms and cyanobacteria (Hart 1985 a), along with various sessile and mobile grazers that harvest this microbial resource. Initial observations suggested that the pattern of occurrence of these two patch types was unpredictable, because they were interspersed in a complex mosaic on the stream bottom. Subsequent study, however, indicated that Cladophora patches were absent from midchannel sites that are characterized by low current velocities (see below). I used a combination of experimental and descriptive field studies to examine some of the factors governing the relative abundance of these two patch types during the summer and fall of 1981 and 1982. Stream discharge during the study ranged between 0.9-2.4 m ~ s -1 in 1981 and between 0.8-1.2 m 3 s -1 in 1982. Relationship between Cladophora abundance and current velocity I made visual assessments of Cladophoracover in November, 1981 to determine how its abundance varied with current velocity. Measurements of cover have been shown to be a reliable means of quantifying the abundance of a wide array of marine and freshwater macroalgae (e.g., Lubchenco and Menge 1978; Dudley et al. 1986). Cladophoracover was estimated in a 50 m long stream reach that had an average width and depth of about 6 m and 0.4 m, respectively. Cover was quantified using a 161 cmz hardware cloth frame that was subdivided into twenty-five 6.5 cm 2 grids. The frame was haphazardly tossed from the stream bank into the center of the channel, and an underwater census of cover was made at each of 31 sites where the frame came to rest. Cover was defined as the percentage of grids in which Cladophorafilaments overlaid > 2 5 % of the substrate area enclosed by a grid. This cover criterion, although less standard than techniques based on point estimates, was applied uniformly throughout the study and yielded a high level of precision. The current velocity associated with each cover sample was measured using an Issacs-Kidd flowmeter (propeller diameter = 5 cm) placed immediately above the substrate. Effect of crayfish on Cladophora abundance I assessed the ability of the crayfish Orconectespropinquusto reduce the abundance of Cladophorain a field experiment. In July, 1981, unglazed quarry tiles (upper surface area=2.25 dm 2) were placed on wooden platforms where the current velocity measured over a five minute interval averaged 68 cm s-1. These platforms were elevated about 10-15 cm above the stream bottom to reduce the accessibility of the tiles to crayfish. The tiles were covered with substantial growths of Cladophoraby the time the experiment began two months later. Tiles were randomly assigned to one of three caging treatments (i.e. crayfish enclosures, crayfish exclosures, or open-sided cages), with five replicates per treatment. The cages were 15 cm on a side by 8 cm in height, and were constructed of 0.6 cm mesh hardware cloth. This mesh size made the cages semipermeable (sensu Munger and Brown 1981), since it was fine enough to prevent the passage of all but the smallest crayfish, but large enough to permit passage of all the other common macroinvertebrates in Augusta Creek. Fig. 1. Tile substrates colonized by the two patch types investigated in this study: the filamentous macroalga Cladophora (left); and the microalgal lawn, encrusted with the cases of Psychomyia and Leucotrichia, two sessile, grazing caddisflies (right). Pocket knife shown for scale 222 Enclosure and exclosure cages were constructed with four side walls. In contrast, open-sided cages had side walls on the upstream and downstream faces of the cage, but lacked side walls on cage's two lateral faces, thus permitting crayfish to move in and out of the cages. It is unlikely that crayfish would be differentially attracted to these open-sided cages solely as a source of cover, since natural crayfish shelters are abundant on the cobble bed of Augusta Creek. Cladophora cover on each tile was censused along three crosstile strip transects immediately after the tiles were placed within the cages. Each transect consisted of twenty-four 0.4 cm 2 grids, within which Cladophora was scored as present when its filaments overlaid > 25% of the grid area. Each transect was censused twice, and repeated counts of the number of grids in which Cladophora was present never differed by more than one. The same person measured cover on all tiles to increase the precision of these estimates. Five large crayfish (probably 1 or 2 years old; 2 + 1 S.E. carapace length=28.3 +0.6 mm) were collected from a site about 150 m upstream; one crayfish was added to each of the enclosure cages. One of these enclosure cages was deleted from subsequent analyses because the crayfish within it died midway through the experiment. The cages were haphazardly arranged in an unshaded run where crayfish densities averaged about 5 individuals m -2 (Hart, unpublished). The stream bed in this region was composed primarily of gravel-cobble substrates, and the stream depth was about 30 cm. Current velocity within 5 cm of the stream bed was measured midway through the experiment with an Issacs-Kidd flowmeter immediately upstream from five of the cages; the velocity (2_+1 S.E.) was 53.0 __.2.8 cm s-1. After the experiment was completed, I used a bucket wheel-type Pygmy flowmeter (Buchanan and Somers 1969) to compare velocities inside cages with velocities measured immediately upstream from cages. The current velocity inside a cage was determined by inserting the flowmeter through a hole cut in the cage's roof and measuring the velocity within 2 - 3 cm of the tile surface. This internal velocity (2_+1 S.E.), expressed as a percentage of the external velocity, was 84.0 + 1.0 %, suggesting that the velocity within the experimental cages was about 45 cm s-1. Each day during the experiment, any leaves or other material that accumulated on the cages was removed to reduce alterations of flow and light. The experiment was ended after seven days, at which time Cladophora cover was remeasured. Effect of sessile grazers on Cladophora establishment The purpose of these field experiments was to determine whether sessile grazers could inhibit the establishment of Cladophora. Two of the most abundant sessile grazers occurring in Augusta Creek's unshaded riffles and runs are the larval stages of the caddisflies Leucotrichia pictipes (Hydroptilidae) and Psychomyiaflavida (Psychomyiidae). The abundance of these two species can be readily quantified in situ with minimal disturbance because their larvae usually live in fixed silken shelters (see Hart 1985b; Hart and Robinson 1990). Effect of Leucotrichia. In August, 1982, I manipulated the abundance of Leucotrichia on a series of bricks that had been placed in Augusta Creek about three months earlier. These substrates were initially placed in a shallow riffle where the average current velocity measured with a Pygmy flowmeter exceeded 50 cm s- 1 (Hart, unpublished). At the outset of the experiment, the top of each brick (surface area = 0.97 dm 2) had been colonized by high densities of Leucotrichia and Psychomyia larvae, whereas no Cladophora filaments were visible. I censused the initial number of these two sessile grazers on sixteen bricks. Half of these bricks were then randomly assigned to a treatment in which all Leucotrichia larvae were removed from the upper surface of the brick. The remaining bricks served as controls, in which Leucotrichia larvae were left on the upper surface of the brick. The abundance of Psychomyia larvae was not manipulated in either treatment. Immediately following the establishment of these treatments, the bricks were systematically interspersed on wooden platforms elevated about 10 cm above the stream bottom and exposed to velocities between 5 0 - 6 0 cm s-1 (Hart, unpublished). Previous observations indicated that substrates placed on these platforms were inaccessible to crayfish. About every 10-12 days during the course of the experiment, fine-tipped forceps were used to remove any Leueotriehia larvae that had colonized the bricks assigned to the removal treatment, whereas Leucotriehia were not removed from bricks assigned to the control group. The experiment was terminated after seven weeks, at which time a final census of Leucotrichia and Psychomyia densities and Cladophora cover was made. Cladophora cover on each brick was estimated along seven cross-brick transects according to the method described previously. In addition, five algal biomass samples per brick were collected from haphazardly selected locations on the upper surface of five removal and five control bricks using a 1-cm 3 periphyton syrine sampler (modified from Loeb 1981) that enclosed an area of about 17 mm 2. The algae in each sample were collected onto a precombusted Whatman G F / F filter at a vacuum of < 50 kPa, acid fumed for 15 minutes to remove inorganic carbon (Wetzel 1965), dried at 60~ and stored in a desiccator. Within 2 months the samples were cornbusted in a Carlo-Erba C H N analyzer (model number 1106, Carlo Erba Strumentazione, Milan, Italy) to determine the particulate carbon content of each sample (Stainton et al. 1977). Effect of Psyehomyia. In August, 1982, I selected a set of ten unglazed quarry tiles (upper surface area = 2.25 dm 2) that had been placed in a shallow riffle in Augusta Creek about two months earlier. Psychomyia was the numerically dominant grazer on these tiles, in contrast to the previous experiment using brick substrates, on which Leucotriehia was the most abundant grazer. After performing an initial census of Psychomyia and Leueotrichia densities, all insects were removed from one half of each tile while the other half served as a control. In addition, forceps were used to remove Cladophora filaments from the surface of each tile, after which I scraped off the basal tissue. Each tile was then placed into a crayfish exclosure cage (see above). The current velocity (i_+ 1 S.E.) in the vicinity of these cages, as measured with a Pygmy flowmeter, was 39.0+1.6cm s -1. Cages were cleaned daily, and the grazer removal treatment was repeated at 1-2 week intervals. The experiment was terminated after about two months, at which point I used the previously described methods to estimate Cladophora cover on both halves of each tile, as well as the number of Psychomyia, Leucotrichia, and tube-dwelling chironomids. Indirect effects of crayfish on other invertebrates This field experiment was undertaken to quantify the potential indirect effects of crayfish on other invertebrates due to alterations in Cladophora cover caused by crayfish grazing. Specifically, I examined how variations in Cladophora cover affected the colonization of a diverse assemblage of invertebrates in the absence of crayfish. Varying amounts of Cladophora were present on a group of nine unglazed quarry tiles that had been incubated in August Creek for about two months. In August, 1982, all macroinvertebrates were carefully removed from these tiles using forceps, whereas Cladophora filaments were disturbed as little as possible. Each tile was then placed in an exclosure cage to prohibit access by crayfish. These cages were haphazardly arranged in an unshaded riffle, and the current velocity (2+ 1 S.E.) in the vicinity of these cages, as measured with a Pygmy flowmeter, was 46.2+2.7 em s-i. Cladophora cover was censused visually at about two week intervals during the course of the experiment, using the method described previously. After two months, each cage was removed from the stream bottom while a dip net was positioned immediately downstream to minimize the loss of mobile animals during removal. In the laboratory, all macroinvertebrates that were visible under 7x magnification were removed from the tile and stored in 75% ethanol, after which they were enumerated and identified. 223 100 Statistical analyses I The relationship between Cladophoraabundance and current velocity was measured using Spearman's rank correlation coefficient. One-way ANOVA was used to test for significant treatment effects in the following experiments: (1) the effect of crayfish on Cladophora abundance; (2) the effect of Leueotrichia on Cladophoraestablishment; (3) the effect of Psychomyia on Cladophoraestablishment. Cladophoracover was arcsine-transformed prior to analysis (Sokal and Rohlf 1981). When multiple measures of Cladophora cover or biomass were available for a single experimental unit, ANOVA was performed on the average of these measures to avoid pseudoreplication. Pairwise comparisons among group means were made using the Tukey Studentized Range test. When the assumption of variance homogeneity was rejected using Levene's test, I assessed the treatment effect with Welch's test, in which variances are not assumed to be equal (Dixon 1988). Potential indirect effects of crayfish on other invertebrates were assessed by determining the sign and magnitude of Spearman's rank correlation coefficient between the abundance of various invertebrate taxa and the cover of Cladophorawithin the crayfish exclosure cages. Rare taxa (i.e., those whose average abundance per tile was < one individual) were not included in these analyses. Results Relationship between Cladophora abundance and current velocity There was a significant positive correlation between the abundance of Cladophora and current velocity (rs = 0.83, N = 3 1 , P < 0 . 0 0 1 ) (Fig. 2). Cladophora cover was zero for samples with velocities less than 50 cm s-2. Even when the analysis was restricted to those samples with velocities greater than 50 cm s - 1 Cladophora cover increased significantly with velocity (rs = 0.69, N = 23, P < 0.001), although this relationship exhibited considerable scatter. For example, at a velocity of about 80 cm s-1, there was a three-fold range in Cladophora cover. Effect of crayfish on Cladophora abundance After seven days, Cladophora was virtually eliminated from the enclosure cages, and cover was reduced by V V 40 V V V 30 V > o V (D o o J~ o_ o q~ _o (~ v 20 v v V 10 v V i_ 10 _ r 20 I I ,.30 50 40 Velocity (cm 60 70 80 Initial Final 80 Z_ 60 2_ 40 20 Exclosure Open Enclosure Caging Treatment Fig. 3. Initial and final Cladophoracover on tites randomly assigned to three different caging treatments. Error bat's represent + one standarderror nearly 70% in the open-sided cages (Fig. 3). In contrast, average cover increased by more than 50% in the exclosure cages. Changes in Cladophora cover between the beginning and end of the experiment differed significantly a m o n g the three treatments (F 2.12 = 9 I. 2, P < 0.000 l). For example, the decline in Cladophora cover within enclosure cages was significantly different f r o m the change in cover within the other two caging treatments ( P < 0.01, Tukey test). The high effective density of crayfish in enclosure cages (ca. 44 ind m -z) m a y overestimate the natural effect of crayfish on Cladophora, however, making it especially important to compare changes in Cladophora cover between the exclosure and open-sided treatments. Cladophora cover in open-sided cages also declined significantly c o m p a r e d to the exclosure cages ( P < 0 . 0 1 , Tukey test). Because the major difference between these two treatments is accessibility by crayfish (see Discussion), this strongly suggests that crayfish at natural densities reduce Cladophora's abundance. Effect of sessile grazers on Cladophora establishment Effect of Leucotrichia. Before the experiment began, the density (s _+ 1 S.E. larvae d m - 2 ) of Leucotrichia on control bricks (63.9+_4.5) and Leucotrichia removal bricks { 50 I 90 s-') Fig. 2. Relationship between Cladophoracover and current velocity (56.7 +_3.2) did not differ significantly (F1,14= 1.70,P= 0.21). There was also no significant difference between Psychomyia's density (2 4-1 S.E. larvae din-2) on control bricks (44.4_+3.7) and Leucotrichia removal bricks (41.1+3.5) (F2n~=0.43, P = 0 . 5 2 ) . By the end of the experiment, the average density of Leucotrichia larvae was significantly lower on removal bricks than on control bricks (Ft,:4 =439.1, P < 0.0001) (Fig. 4), thus confirming the success of the density manipulation. In contrast, there was no significant difference between the average density of Psychomyia larvae on removal and control bricks (FI,t4 = 0.03, P = 0.87). The average percent cover of Cladophora on removal bricks at the end of this experiment was more than 20fold greater than on Leucotrichia control bricks (F1,14 = 224 Table 1. Relationship between the abundance of common invertebrate taxa (i.e., those averaging > 1 individual per tile) and Cladophora cover on tiles within exclosure cages. For each taxon, the value of Spearman's rank correlation coefficient is given 48 7 I00 40 32 Taxon Correlation with Cladophora cover Leucotrichia Psychomyia Antocha -0.93"*** 8 40 24 s 16 o 2O Oligochaetes Hydropsyche Baetis Larval Chironomidae 0 Psychomyio Leucotrichia C[adophora Fig. 4. Final insect density and Cladophora cover on control bricks (hatched bars) and on Leucotrichia removal bricks (open bars). Error bars represent + one standard error 7 Pupa| Chironomidae * 0.05<P<0.1, ** P<0.05, *** P<0.01, **** P<0.005, NS=not significant 1~f 5_ / / 60 45 o 40 g a0 ~z s t~ o g o c Cheumatopsyche Isoperla Stenonema Taeniopteryx -0.86**** - 0.66'* - 0 . 3 8 NS - 0 . 1 5 NS 0.18 NS 0.42 NS 0.50* 0.58** 0.68** 0.80*** 0.86**** 20 15 Leucotrlchi~ Psychomyio Chironomids of the grazing effect on Cladophora (see below) can be attributed to Psychomyia. In contrast to both Psychomyia and Leucotrichia, chironomid densities were significantly greater on tile halves from which grazers (including midges) were removed than on control halves (F1,18 =25.47, P=0.0001). The percent cover of Cladophora on tile halves from which grazers were removed was nearly 20-fold greater than on control halves (FL18=416.4, P < 0 . 0 0 0 1 ) (Fig. 5). Clado ~hora Fig. 5. Final insect density and Cladophora cover on control halves of tiles (hatched bat's) and on grazer removal halves of tiles (open bars). Error bars represent • one standard error 139.4, P < 0 . 0 0 0 1 ) (Fig. 4). Algal biomass (2___1 S.E. lag C r a m - 2 ) on the upper surface of Leucotrichia removal bricks (19.9_+1.1) was nearly four times greater than on control bricks (5.6 ___0.4) (F1,7 = 132.9, P < 0.0001). Effect of Psychomyia. The initial density (2_+ 1 S.E. larvae dm -z) of Psychomyia on control halves (58.1 _+4.8) and on grazer removal halves (50.7+4.9) of the tiles did not differ significantly (F1.18=1.19, P=0.2887). There was also no significant difference between the initial density of Leucotrichia on control tiles (5.3_+ 1.6) and grazer removal tiles (10.8_+2.8) (Fl,~S=2.55, P = 0.1277). At the end of the experiment, the density of Psychomyia larvae on the control halves was about twice as large as on grazer removal halves (F~,18=67.4, P < 0.0001) (Fig. 5). The density of Leucotrichia larvae on grazer removal and control sections also differed significantly (Fl,18=14.63, P=0.0012), but Leucotrichia's density in this experiment was 1-2 orders of magnitude lower than that of Psychomyia. Because Psychomyia and Leucotrichia larvae are very similar in body mass (Hart and Robinson 1990), this marked difference in population density between the two species suggests that most Indirect effects of crayfish on other invertebrates A total of 22 invertebrate taxa colonized the tiles used in these experiments, but only 12 of these taxa were abundant enough to include in the correlation analyses. Two-thirds of these taxa exhibited statistically significant positive or negative correlations with Cladophora cover (Table 1). For example, the abundances of Psychomyia, Leucotrichia, and the tube-dwelling tipulid Antocha were negatively correlated to Cladophora cover, whereas the abundances of the mayfly Stenonema, the stoneflies Taeniopteryx and Isoperla, and chironomid pupae were positively correlated to Cladophora cover. There was also a marginally significant positive correlation between Cladophora and net-spinning caddisfly Cheumatopsyche. Discussion Cladophora's absence from sites with velocities <50 cm s -1 At least three hypotheses can potentially explain the observed absence of Cladophora from mid-channel microhabitats with current velocities _< ca. 50 cm s - l : (1) failure to disperse to these sites; (2) physiological unsuitability of the sites; (3) biological interactions that either 225 prevent Cladophora from establishing at the sites or that eliminate it after establishment. Because bare substrates protected from crayfish and sessile grazers are readily colonized by Cladophora, however, its absence is not the result of dispersal limitations. Moreover, Cladophora cover on tiles transplanted to the crayfish exclosure cages (in which the internal velocity averaged 45 cm s- 1) increased by more than 50%, demonstrating that these sites are favorable for growth. Thus, although the growth of Cladophora may be strongly limited where flow is very low or absent (e.g. Dodds 1991), the current velocity at which the caging experiment was conducted was more than adequate to support Cladophora's growth. The results from the enclosure cages clearly demonstrate that crayfish are able to control the abundance of Cladophora, but they do not ensure that the observed reduction of Cladophora in the open-sided cages is necessarily caused by crayfish. Two considerations strongly support the conclusion that Cladophora's decline was due to crayfish, however. First, other than crayfish, the only common animal at the study site that was too large to pass through the 0.6 cm mesh cages was the mottled sculpin (Cottus bairdi). Underwater observations in both the day and night confirmed that crayfish entered the open-sided cages and fed upon Cladophora, whereas sculpin were never observed in these cages. Second, a preliminary analysis of the summer diets of these two species indicated that 86% of the crayfish (N=7) contained an abundance of Cladophora fragments in their guts, whereas Cladophora fragments were never found in the guts of the sculpin ( N = 3). These results, although based on small sample sizes, are consistent with other data indicating that Cladophora commonly occurs in the guts of O. propinquus (e.g. Capelli 1980) whereas sculpin rarely ingest plant material (e.g. Koster 1937; Dineen 1951). Flow-mediated refuges The caging experiment demonstrates that crayfish can eliminate Cladophora from sites with velocities < ca. 50 cm s- 1, and the descriptive data indicate that Cladophora is often abundant in sites whose velocities are faster than this. Collectively, these results strongly suggest that the ability of crayfish to control Cladophora is constrained by flow. Maude and Williams (1983) quantified the relationship between current velocity and the activity of eight crayfish species in a laboratory flume. The average velocity at which these crayfish were dislodged from the Plexiglas flume bed ranged between 26 and 50 em s -1, and the 95% confidence interval bracketing this average '~slip speed" for O. propinquus ranged between 31 and 38 cm s-1. Similarly, Howard and Nunny (1983) demonstrated that juvenile lobsters (Homarus gammarus), whose size range bracketed the crayfish sizes used in my study, were dislodged from semi-natural substrates in lab flumes at velocities of about 40-50 cm s-1. They concluded that lobsters in the field are restricted to microhabitats in which velocities are lower than this. My underwater field observations have confirmed that high flows can dislodge O. propinquus from the stream bed. For example, I have observed crayfish easily climb up the downstream (and therefore sheltered), vertical face of a large boulder, only to be "blown off" the boulder once they reach the top, where they encounter the full force of the current. The probable role of flow in constraining the ability of crayfish to control Cladophora is but one example of a growing body of work indicating that flow can mediate the intensity of predator-prey interactions in a variety of freshwater and marine environments (e.g. Menge 1978; Palmer 1988; Peckarsky etal. 1990; Richardson and Brown 1990; DeNicola and McIntire 1991 ; Hansen et al. 1991). The results of Maude and Williams (1983) clearly demonstrate that flow can constrain the foraging activity of crayfish, yet the velocity threshold suggested by their slip speed estimate for O. propinquus is somewhat lower than that implied by the observed relationship between Ctadophora cover and current velocity (Fig. 2). Their lab study might be expected to underestimate this velocity threshold, however, since the ability of crayfish to resist dislodgement is probably lower on the relatively smooth Plexiglas flume bed used by Maude and Williams (1983) than on the topographically complex stream bed where my study was conducted. Recent studies have shown that Cladophora's abundance in Augusta Creek sometimes exhibits a negative correlation with depth as well as a positive correlation with current velocity (Creed 1990). One possible explanation for the negative correlation between Cladophora cover and depth is that crayfish avoid shallow sites where they may experience greater predation risk (see Power, 1987 for a similar pattern involving grazing stream fish). Further experiments are required to assess the relative importance of constraints involving depthrelated predation risk vs. flow in controlling the effects of crayfish on Cladophora, especially because stream depth and current velocity often exhibit a strong negative correlation. Cladophora monopolies, sessile grazers, priority effects, and chance If crayfish were always excluded from sites with velocities > 50 cm s - 1 and if crayfish herbivory was the only factor limiting Cladophora's abundance, then Cladophora should be able to monopolize space in these sites. Cladophora cover in sites with velocities > 50 cm s-1 was highly variable, however, and rarely exceeded 50%. Two factors may prevent Cladophora from monopolizing these sites. First, some sites with velocities > 50 cm s-1 are presumably accessible to crayfish. For example, there is probably no single velocity threshold at which crayfish foraging becomes impossible. Rather, it seems likely that there is a broad velocity zone over which the probability of crayfish grazing varies as a function of numerous factors, including crayfish size and posture (Maude and Williams 1983), bed topography, and near-bed velocity gradients. 226 Second, even in the absence of crayfish, it is clear that both Leucotrichia and Psychornyia can prevent Cladophora from monopolizing sites where these sessile grazers have established prior residence. Similarly, grazing insects have been observed to inhibit the establishment of Cladophora in a California stream (Dudley and D'Antonio 1991). A reciprocal interaction also occurs in Augusta Creek, since Cladophora patches can inhibit the establishment of sessile grazers. Although such priority effects have been observed in a variety of marine and terrestrial settings (see references in Connell and Slatyer 1977; Yodzis 1986), they have rarely been documented in streams (Downes 1990; but see McAuliffe 1984). Further studies are needed to determine whether priority effects influence species interactions in a broader array of benthic stream communities. Priority effects can generate considerable spatial and temporal heterogeneity in natural communities as a result of two sets of stochastic processes influencing the colonization history of sites. One set of processes controls the creation of bare space at a site. In Augusta Creek, space becomes available when stones are overturned during spates, when sessile grazers pupate and emerge following the completion of their larval feeding stage, and when Cladophora sloughs from stone surfaces as a result of either winter senescence or high drag during floods. The second class of processes controlling patterns of patchiness is linked to the factors governing the arrival of propagules at a site. For example, differences between sessile grazers and Cladophora in the phenology of propagule dispersal will necessarily cause their establishment probabilities to differ in space and time. Therefore, interactions between species-specific life history characteristics and factors controlling the availability of bare space can cause microhabitats that are abiotically similar to become dominated by very different species assemblages (see also Sousa 1985). The manifestation of priority effects in ecological communities will often involve a large element of chance, which may contribute directly to the complex spatial mosaic of these two patch types in Augusta Creek. This strong but stochastic interaction also serves as a potent reminder that the observation of patchiness per se provides little information about the degree to which a community is organized by abiotic vs. biotic factors, since either set of factors has the potential to generate marked heterogeneity. Indirect effects of crayfish The direct effect of crayfish grazing on Cladophora is transmitted to other species via Cladophora's modification of the benthic environment. For example, the recruitment of species such as Leucotrichia and Psychomyia is inversely related to the abundance of Cladophora. Several mechanisms probably account for these negative effects. The tendency of Cladophora to form a thick mat (hence its popular name, "blanket-weed"; Whitton 1970) significantly modifies the abiotic environment experienced by organisms living on the stone surface underneath the mat. For example, Dudley et al. (1986) showed that current velocities beneath a Cladophoraglornerata mat were reduced by more than 40% compared to velocities in the absence of the mat. Cladophora mats also intercept light that would otherwise reach the epilithic microalgal lawn, and field experiments have indicated that shading reduces the standing crop of this lawn (Hart, unpublished; see also Feminella and Resh 1991). As a consequence, shading by Cladophora will negatively affect various species that feed upon microalgae, because many of these grazers are food-limited in Augusta Creek (Hart 1987; Hart and Robinson 1990). Although the biomass of the Cladophora mat is usually much greater than the microalgal lawn (Hart et al. 1991), Cladophora is not an acceptable alternative food source for these grazers. Once the Cladophora mat becomes established at a site, these grazers are apparently unable or unwilling to consume it, probably because of various chemical and physical properties that deter herbivores (LaLonde et al. 1979; Patrick et al. 1983). In contrast, Cladophora may lack such defenses during its initial growth phase at a site, which could explain why grazers are able to control it at this time. Similar size-related changes in vulnerability to grazers have been documented in marine plantherbivore interactions (Lubchenco and Gaines 1981). The abundances of several other insect taxa (i.e., Stenonema, Taeniopteryx, Isoperla, chironomid pupae, and Cheumatopsyche) are positively related to Cladophora cover. A variety of mechanisms appear to underlie these patterns, including: (1) enhancement of epiphytic food resources on Cladophora filaments; (2) accumulation of fine particulate organic matter beneath the Cladophora mat; (3) the use of Cladophora filaments as structural support for filtering nets; and (4) the role of Cladophora mats in providing shelter from high current velocities or potential predators. Similar explanations were advanced by Dudley et al. (1986) to account for positive associations between various invertebrate taxa and the abundance of Cladophora. Crayfish can have negative effects on the abundances of invertebrates via at least two paths. First, as described above, taxa whose abundances are positively correlated to Cladophora cover will experience a negative indirect effect as the result of Cladophora consumption by crayfish. Second, crayfish may directly feed upon various invertebrates. Estimating the direct consumption of invertebrates by crayfish was not a primary focus of my study. Nonetheless, preliminary dietary analyses and laboratory experiments suggest that relatively few of the invertebrate taxa described above are heavily preyed upon by O. propinquus in Augusta Creek (Hart, unpublished data; see also Capelli 1980). Because crayfish reduce the abundance of Cladophora, which in turn can have a strong negative effect on the sessile grazer-microalgal lawn patch type, crayfish therefore act as keystone predators (Paine 1966) in this community. The general importance of keystone predation and other strong interactions (sensu MacArthur 1972) in stream communities remains poorly understood (Hildrew 1992), but these interactions are obviously more widespread than previously recognized (McAuliffe 1984; Hart 1985a; Power et al. 1985). Because crayfish 227 a n d / o r Cladophora o c c u r in a w i d e v a r i e t y o f s t r e a m c o m m u n i t i e s , a n d b e c a u s e each species has been s h o w n to i n t e r a c t s t r o n g l y in v a r i o u s settings w h e r e t h e y h a v e been s t u d i e d (e.g. see D u d l e y et al. 1986; P o w e r 1990a, 1990b r e g a r d i n g Cladophora glomerata; see L o d g e a n d L o r m a n 1987; W e b e r a n d L o d g e 1990; C r o w l a n d Schnell 1990 r e g a r d i n g crayfish), s t r o n g i n t e r a c t i o n s m a y be i m p o r t a n t in m a n y s t r e a m c o m m u n i t i e s w h e r e either o r b o t h o f these species occur. M o r e generally, there is a pressing n e e d for m o d e l s t h a t c a n p r e d i c t the k i n d s o f e c o l o g i c a l a t t r i b u t e s a n d e n v i r o n m e n t a l settings t h a t l e a d to s t r o n g i n t e r a c t i o n s a m o n g species (see also F a u t h a n d R e s e t a r i t s 1991). In c o n c l u s i o n , the c o m p l e x set o f i n t e r a c t i o n s d o c u m e n t e d in this s t u d y c a n n o t be easily r e c o n c i l e d with v a r i o u s d i c h o t o m o u s r e p r e s e n t a t i o n s o f c o m m u n i t y org a n i z a t i o n (e.g. c o n t r o l b y a b i o t i c vs. b i o t i c factors, c o m p e t i t i o n vs. p r e d a t i o n , o r s t o c h a s t i c vs. d e t e r m i n i s t i c p r o cesses). I f ecologists wish to a c c o u n t fully for the p r o cesses c o n t r o l l i n g c o m m u n i t y p a t t e r n s , t h e n they will need to d e v e l o p m o r e d e t a i l e d m o d e l s t h a t explicitly ack n o w l e d g e the m u l t i f a c t o r i a l o r g a n i z a t i o n o f c o m m u n i ties, a n d to design r e s e a r c h p r o g r a m s t h a t c a n identify h o w the relative c o n t r i b u t i o n s o f these factors v a r y in space a n d time. Acknowledgements. I am very grateful to Earl Werner and Don Hall for their support and friendship while I worked at the Kellogg Biological Station. The Snyder family kindly provided access to the study site. Steve Latta, Chris Robinson, Jean Marek, and Deb Wohl assisted with the field and lab work, and Deb Wohl helped analyze the data and prepare figures. Bob Wetzel and especially Mike Klug generously provided equipment that permitted the CHN analyses to be performed, and Greg Walker helped run the CHN analyzer. I thank Joaquin Feliciano, Steve Kohler, Mark Leighton, Pete Peterson, Peter Petraitis, LeRoy Poff, Deb Wohl and an anonymous reviewer for their detailed and constructive criticisms of earlier drafts of the manuscript. This work was supported by NSF grants DEB-8111305, BSR-8312629, and BSR8918608. KBS Publication No. 730. References Buchanan TJ, Somers WP (1969) Discharge measurements at gaging stations. US Geol Survey Techniques Water-Resources Inv, book 3, chap A8, 65 p Capelli GM (1980) Seasonal variation in the food habits of the crayfish Oreonectes propinquus (Girard) in Trout Lake, Vilas County, Wisconsin, U.S.A. (Decapoda, Astacidae, Cambaridae). Crustaceana 38:82-86 Connell JH, Slatyer RO (1977) Mechanisms of succession in natural communities and their role in community stability and organization. Am Nat 111 : 1119-1144 Creed RP Jr (1990) The direct and indirect effects of omnivorous crayfish on stream benthic communities. Dissertation, Michigan State University, East Lansing, Michigan Crowl TA, Schnell GD (1990) Factors determining popuiation density and size distribution of a freshwater snail in streams: effects of spatial scale. Oikos 59:359-367 DeNicola DM, McIntire CD (1991) Effects of hydraulic refuge and irradiance on grazer-periphyton interactions in laboratory streams. J N Am Benthol Soc 10:251-262 Dineen CF (1951) A comparative study of the food habits of Cottus bairdii and associated species of Salmonidae. Am Midl Nat 46 : 640-645 Dixon WJ (1988) BMDP statistical software manual. University of California Press, Berkeley, California, USA Dodds WK (1991) Community interactions between the filamentous alga Cladophora glomerata (L.) Kuetzing, its epiphytes, and epiphyte grazers. Oecologia 85: 572-580 Dowries BJ (1990) Patch dynamics and mobility of fauna in streams and other habitats. Oikos 59:411-413 Dudley TL, Cooper SD, Hemphill N (1986) Effects of macroalgae on a stream invertebrate community. J N Am Benthol Soc 5:93-106 Dudley TL, D'Antonio CM (1991) The effects of substrate texture, grazing, and disturbance on macroalgal establishment in streams. Ecology 72:297 309 Dunson WA, Travis J (1991) The role of abiotic factors in community organization. Am Nat 138:1067-1091 Fauth JE, Resetarits WJ Jr (1991) Interactions between the salamander Siren intermedia and the keystone predator Notophthalmus viridescens. Ecology 72:827-838 Feminella JW, Resh VH (1991) Herbivorous caddisflies, macroalgae, and epilithic microalgae: dynamic interactions in a stream grazing system. Oecologia 87:247-256 Giller PS, Gee JHR (1987) The analysis of community organization: the influence of equilibrium, scale and terminology. In: Gee JHR and Giller PS (eds) Organization of Communities: Past and Present. Blackwell, Oxford, pp 519-543 Hansen RA, Hart DD, Merz RA (1991) Flow mediates predatorprey interactions between triclad flatworms and larval black flies. Oikos 60:187-196 Hart DD (1985a) Grazing insects mediate algal interactions in a stream benthic community. Oikos 44:40-46 Hart DD (1985b) Causes and consequences of territoriality in a grazing stream insect. Ecology 66:404~14 Hart DD (1987) Experimental studies of exploitative competition in a grazing stream insect. Oecologia 73:41-47 Hart DD, Robinson CT (1990) Resource limitation in a stream community: phosphorus enrichment effects on periphyton and grazers. Ecology 71 : 1494-1502 Hart DD, Kohler SL, Carlton RG (1991) Harvesting of benthic algae by territorial grazers : the potential for prudent predation. Oikos 60:329-335 Hildrew AG (1992) Food webs and species interactions. In: Petts GE and Calow P (eds) The Rivers Handbook. Blackwell, Oxford (in press) Howard AE, Nunny RS (1983) Effects of near-bed current speeds on the distribution and behaviour of the lobster, Homarus gammarus (L.). J Exp Mar Biol Ecol 71:2%42 Koster WJ (1937) The food of sculpins (Cottidae) in central New York. Trans Am Fish Soc 66:374-382 LaLonde RT, Morris CD, Wong CF, Gardner LC, Eckert D J, King DR, Zimmerman RH (1979) Response ofAedes triseriatus larvae to fatty acids of Cladophora. J Chem Ecol 5:371-381 Lodge DM, Lorman JG (/987) Reductions in submersed macrophyte biomass and species richness by the crayfish Orconectes rusticus. Can J Fish Aquat Sci 44:591 597 Loeb SL (1981) An in situ method for measuring the primary productivity and standing crop of the epilithic periphyton community in lentic systems. Limnol Oceanogr 26 : 394-399 Lubehenco J, Menge BA (1978) Community development and persistence in a low rocky intertidal zone. Ecol Monogr 48 : 67-94 Lubchenco J, Gaines SD (1981) A unified approach to marine plant-herbivore interactions. I. Populations and communities. Ann Rev Ecol Syst 12:405-437 MacArthur RH (1972) Strong, or weak, interactions? Trans Conn Acad Arts Sci 44:177-188 Maude SH, Williams DD (1983) Behavior of crayfish in water currents: hydrodynamics of eight species with reference to their distribution patterns in southern Ontario. Can J Fish Aquat Sci 40: 68-77 McAuliffe JR (1984) Competition for space, disturbance, and the structure of a benthic stream community. Ecology 65 : 894-908 Menge BA (1978) Predation intensity in a rocky intertidal community. Relation between predator foraging activity and environmental harshness. Oecologia 34:1-16 228 Menge BA, Sutherland JP (1987) Community regulation: variation in disturbance, competition, and predation in relation to environmental stress and recruitment. Am Nat 130:730-757 Munger JC, Brown JH (1981) Competition in desert rodents: an experiment with semipermeable exclosures. Science 211:510512 Paine RT (1966) Food web complexity and species diversity. Am Nat 100:65-75 Palmer MA (1988) Epibenthic predators and marine meiofauna: separating predation, disturbance, and hydrodynamic effects. Ecology 69:1251-1259 Patrick R, Rhyne CF, Richardson RW III, Larson RA, Bott TL, Rogenmuser K (1983) The potential for biological controls of Cladophora glomerata. EPA-600/3-83-065. U.S. Environmental Protection Agency, Washington, D.C Peckarsky BL (1983) Biotic interactions or abiotic limitations? A model of lotic community structure. In: Fontaine TD, Bartell SM (eds) Dynamics of Lotic Ecosystems, Ann Arbor Science, Ann Arbor, Michigan pp 303 323 Peckarsky BL, Horns SC, Statzner B (1990) Stonefly predation along a hydraulic gradient: a field test of the harsh-benign hypothesis. Freshwat Biol 24:181 191 Poff NL, Ward JV (1989) Implications of streamflow variability and predictability for lotic community structure: a regional analysis of streamflow patterns. Can J Fish Aquat Sci 46:18051818 Power ME (1987) Predator avoidance by grazing fishes in temperate and tropical streams: importance of stream depth and prey size. In: Kerfoot WC, Sih A (eds) Predation: direct and indirect impacts on aquatic communities. University Press of New England, Hannover, New Hampshire, pp 333-351 Power ME (1990a) Effects of fish in river food webs. Science 250:811 814 Power ME (1990b) Benthic turfs vs floating mats of algae in river food webs. Oikos 58 : 67-79 Power ME, Matthews WJ, Stewart AJ (1985) Grazing minnows, piscivorous bass, and stream algae: dynamics of a strong interaction. Ecology 66:1448-1456 Richardson TD, Brown KM (1990) Wave exposure and prey size selection in an intertidal predator. J Exp Mar Biol Ecol 142:105-120 Schoener TW (1986) Overview: Kinds of ecological communities - ecology becomes pluralistic. In: Diamond J and Case TJ (eds) Community Ecology, Harper and Row, New York, pp 467-479 Schlosser IJ, Ebel KK (1989) Effects of flow regime and cyprinid predation on a headwater stream. Ecol Monogr 59:41-57 Sokal RR, Rohlf FJ (1981) Biometry. Second edition. W.H. Freeman and Company, San Francisco, California, USA Sousa WP (1985) Disturbance and patch dynamics on rocky intertidal shores. In: Pickett STA and White PS (eds) The Ecology of Natural Disturbance and Patch Dynamics. Academic Press, New York, pp 101 124 Stainton MP, Capel M J, Armstrong FAJ (1977) The chemical analysis of fresh water. Second edition. Miscellaneous Special Publication Number 25, Fish and Environment Canada Weber LM, Lodge DM (1990) Periphytic food and predatory crayfish: relative roles in determining snail distribution. Oecologia 82:33 39 Wetzel RG (1965) Necessity for decontamination of filters in 14C measured rates of photosynthesis in fresh waters. Ecology 46: 540-542 Whitton BA (1970) Biology of Cladophora in freshwaters. Water Res 4:457 476 Yodzis P (1986) Competition, mortality, and community structure. In: Diamond J and Case TJ (eds) Community Ecology, Harper and Row, New York, pp 480-491