, Volume 716, Issue 1, pp 103–114 | Cite as

Habitat-specific effects of particle size, current velocity, water depth, and predation risk on size-dependent crayfish distribution

  • Jennifer M. Clark
  • Mark W. Kershner
  • Justin J. Montemarano
Primary Research Paper


This study assessed effects of abiotic (current velocity, water depth, particle size) and biotic (predation risk for crayfish, size distribution and densities of predatory fish) variables on habitat- and size-specific distribution patterns of lotic crayfish (Orconectes obscurus) using field surveys and tethering experiments. Additionally, particle size manipulations were used with predation assays to assess habitat-specific interactions since the average particle size increased from deep pools to shallow pools to riffles. Large crayfish had the highest densities in deep pools and were associated with increased water depth, whereas small and medium crayfish had the highest densities in shallow pools and were strongly associated with increased particle size and decreased water depth. Regardless of size, crayfish in deep pools had significantly lower survival than in shallow pools and riffles. However, only small crayfish showed consistent differences in predation risk by habitat type and were significantly more vulnerable to predation than larger crayfish. Additionally, large rocky refugia resulted in significantly higher survival of small crayfish in the combined particle manipulation/tethering experiment. Overall, predation appears to be a key mechanism structuring habitat-specific distribution patterns for only small O. obscurus. Large substrates may be particularly important in habitats where both small crayfish density and predation risk are high.


Orconectes obscurus Stream Pools Riffles Substrate Refugia 


Lotic systems are extremely dynamic and prone to disturbance (i.e., flooding, seasonal drying), often leading to a dynamic environment for stream biota. Further, flow alters habitat characteristics (e.g., substrate type/distribution), organism dispersal patterns (active and passive), predation risk, competitive interactions, and resource acquisition (Hart & Finelli, 1999). While small headwater stream communities are primarily structured by abiotic factors (e.g., seasonal drying) (Creed, 2006), trophic complexity tends to be driven by both abiotic and biotic factors in larger permanent streams (Creed, 2006).

Crayfish are often associated with permanent streams and when present tend to dominate benthic communities (Griffith et al., 1994). Due to their large size and aggressive behavior (Renai & Gherardi, 2004), crayfish are often important drivers of community structure (Lodge et al., 1994; Nyström et al., 1996), trophic cascades (Lodge et al., 1994; Nyström et al., 1996), organic matter processing (Nyström et al., 1996; Usio & Townsend, 2001), and species replacements (Hill & Lodge, 1999). Further, crayfish have been characterized as both keystone species (Nyström et al., 1996) and ecosystem engineers (Statzner et al., 2003; Creed & Reed, 2004). Given the important roles played by these organisms in stream ecosystems and the fact that many North American crayfish species are in decline (Lodge et al., 2000), understanding their natural habitat requirements and the factors influencing distribution patterns is essential for management and restoration efforts.

Previous studies show that crayfish inhabit a wide variety of aquatic habitats with their distributions influenced by many abiotic and biotic factors including water depth (Englund & Krupa, 2000; Usio & Townsend, 2000; Flinders & Magoulick, 2007a, b), substrate type (Lodge & Hill, 1994; Kershner & Lodge, 1995; Usio & Townsend, 2000; Flinders & Magoulick, 2003, 2005; Clark et al., 2008), current velocity (Usio & Townsend, 2000; Bubb et al., 2004; Parkyn & Collier, 2004; Flinders & Magoulick, 2007b; Clark et al., 2008), and predation risk (Mather & Stein, 1993a, b; Englund & Krupa, 2000; Usio & Townsend, 2000; Light, 2003; Magoulick, 2004). Furthermore, in many streams, habitat-specific crayfish distribution is size dependent (Englund & Krupa, 2000; Usio & Townsend, 2000; Flinders & Magoulick 2003, 2007a, b; Clark et al., 2008) and appears to be driven by the presence of predators (Mather & Stein, 1993a; Englund & Krupa, 2000; Magoulick, 2004), potentially resulting from complex interactions between abiotic variables and predator density.

While a few studies have addressed how multiple factors interact within streams to influence size-specific crayfish distribution patterns (Flinders & Magoulick, 2007a; Englund & Krupa, 2000; Usio & Townsend, 2000; Nyström et al., 2006), fewer studies have addressed interactions between abiotic characteristics and crayfish predation risk, with interactions between substrate type and crayfish predation risk being studied most intensively (Stein & Magnuson, 1976; Kershner & Lodge, 1995; Olsson & Nyström, 2009). However, little is known about how small changes in microhabitat influence predation risk for crayfish.

This study investigated independent and interactive effects of predation risk and abiotic variables on size-specific lotic crayfish (O. obscurus) habitat distributions. Field surveys were used to assess the effects of water depth, particle size, current velocity, and fish predators (size and density) on size-specific crayfish distribution in deep pool, shallow pool, and riffle habitats in a fourth-order, forested stream. Coupled with this survey, we used tethering experiments and particle size manipulations to assess habitat-specific biotic (e.g., predatory fish density) and abiotic effects on size-dependent predation risk for crayfish.

As in other streams, we expected O. obscurus to show size-dependent distribution patterns corresponding with habitat characteristics that minimize predation risk (e.g., substrate size, predator abundance/size distribution) and those that physically limit crayfish presence (e.g., current velocity; Clark et al., 2008). Given these size-specific constraints, smaller crayfish would primarily occupy riffle and shallow pool habitats, while larger crayfish would tend to be associated with shallow pools and deep pools. For tethering experiments, the survival of smaller crayfish was expected to be the lowest in deep pools due to the presence of large fish predators and small substrate sizes (i.e., less refugia), while larger crayfish were expected to have the highest mortality in riffles and shallow pools due to greater difficulty finding refugia in the available substrate and the fact that terrestrial predators have access to these shallow-water habitats. Lastly, for the particle size manipulation study, we hypothesized that small crayfish survival (the only size class tested) would be the highest on larger particle sizes (small cobble) and the lowest on smaller particle sizes (sand), regardless of habitat type.


Field site and habitat descriptions

This study was conducted in the West Branch of the Mahoning River, a fourth-order forested stream near Ravenna, Portage County, Ohio (41°09′41″N 81°11′50″W). The stream reach is dominated by pool–riffle sequences with benthic substrates ranging from clay to boulders. For this study, stream habitat type was characterized into three categories consisting of deep pools (water depth at baseflow ≥61 cm), shallow pools (maximum water depth at baseflow = 60 cm), and riffles (maximum water depth at baseflow = 12 cm with noticeable breaks in the surface water). All channel units used were within a 500-m stretch of the stream, with channel units averaging 40 m in length and 7 m in width.

Habitat-specific crayfish/predatory fish survey and abiotic measurements

We assessed size-specific crayfish (O. obscurus) and predatory fish density and size distribution, in addition to water depth, current velocity, and average particle size in four deep pools, four shallow pools, and four riffles (n = 12 channel units) during summer 2005 to allow for comparison between size-specific crayfish distribution patterns and habitat-specific fish predator and abiotic variables. Crayfish were divided into three size categories consisting of small (carapace length (CL) < 20 mm), medium (CL = 20–30 mm), and large (CL > 30 mm). Size classes were chosen based on sexual development patterns with the small size class consisting of immature juveniles and medium and large size classes being sexually mature (Fielder, 1972). Replicate channel units (ca. ≤10 m long) were enclosed by upstream and downstream block nets (mesh size = 6 mm), and crayfish and fish were collected using a four-pass removal method (Zippin, 1958) with a backpack electroshocker (LR-24 Electrofisher, Smith-RootTM, Vancouver, Washington, USA). During each pass, crayfish and fish were collected by two netters (dip net mesh size = 3 mm) walking upstream in a zig-zag pattern as a third person operated the backpack electroshocker. Electroshocking is considered be an effective technique to sample both crayfish (Rabeni et al., 1997) and fish populations (Poos et al., 2007) and our method met the assumptions required for this method of population estimation (Zippin, 1958). Furthermore, many studies have successfully surveyed crayfish and fish populations using electroshocking in combination with pass removal (Usio & Townsend, 2000; Hicks, 2003; Seiler & Turner, 2004).

All crayfish collected were counted and measured (CL, mm) using Vernier calipers and all fish were identified to species and measured [total length (TL), mm] using a standard fish measuring board. To allow calculation of surface area (m2) for density estimation, lengths (along both stream edges) and widths (at both upstream and downstream block nets) were measured in each channel unit.

Water depth and average current velocity were measured during baseflow conditions using a wading rod and a Marsh-McBirney® flow meter (Model 2000, FLO-MATETM, Loveland, Colorado, USA) in 15 randomly selected sites within each of the 12 channel units sampled for crayfish and predatory fish (n = 180 estimates for each abiotic variable across all channel units). Average particle size (n = 100 particles) was also determined in each habitat type replicate using Wolman pebble counts (Wolman, 1954, for a full description of methods see Clark et al., 2008).

Statistical analyses

To estimate crayfish density, we used a four-pass removal method (Zippin, 1958) where cumulative total catch was plotted against catch per unit effort (CPE) with data points fitted with a regression line and solving for CPE = 0 to determine the population estimate for each channel unit. To calculate size-specific density estimates, total crayfish density (population estimate divided by the area calculated for each channel unit) for a given channel unit was multiplied by the proportion of crayfish captured within each size class for that sampling area. Only adult fish (minimum adult size determined from Trautman, 1957) were used in predatory fish analyses, as this age group is most likely to consume crayfish. Due to limitations in gape size, young-of-year fish rarely consume crayfish (Roell & Orth, 1993; Rabeni, 1992; Dorn & Mittelbach, 1999). Depletion was not reached for every electroshocking effort for predatory fish, so density was calculated by dividing the total catch from the four passes by the enclosed sample area for all channel units. Predatory fish in this system include Cottus bairdi Girard (central mottled sculpin), Semotilus atromaculatus Mitchill (creek chub), Ambloplites rupestris Rafinesque (rock bass), and Lepomis cyanellus Rafinesque (green sunfish) and were determined by examining gut contents for crayfish (unpublished data) and/or were observed preying upon crayfish during predation assays.

A redundancy analysis (RDA) was used to assess associations of size-specific crayfish densities with current velocity, water depth, average particle size (see Clark et al., 2008 for a full description of methods), fish predator density, and average fish predator TL at all 12 sample sites during summer 2005. RDA was used to express covariation among variables in a smaller number of composite variables (axes) to allow assessment of relationships among crayfish size classes and habitat variables (McCune et al., 2002). Statistical significance (P < 0.05) of the RDA model, RDA axes, and habitat variables was obtained via permutation ANOVA with 999 permutations. Crayfish density data were found to be non-normally distributed according to Shapiro–Wilk tests and were transformed [using an “ln(x + 1)” transformation] before RDA was performed. The percent of variance in the distance matrix was measured using a Euclidean distance measure (see Legendre & Legendre, 1998 for further description of RDA). RDA and associated analyses were performed using the “vegan” package (Oksanen et al., 2013) in R 2.15.3 (R Core Team 2013).

Habitat-specific predation risk: tethering experiment

To assess the effects of crayfish size and habitat on predation risk and allow for comparison with habitat-specific crayfish distribution patterns, predation assays were conducted during August 2005, testing three size categories of crayfish (small, medium, and large as described above) in three different habitat types (deep pools, shallow pools, and riffles as described above). All crayfish used for predation assays were collected by electroshocking and hand netting downstream from the survey and experimental sites. Only males were collected and used in the assays in order to control for sex-specific differences in predation risk. Following capture, crayfish were brought back to the lab, measured using Vernier calipers (CL, mm), and a black, brass barrel snap swivel (size 14) was super-glued to the middle of the dorsal portion of their carapace. All crayfish were housed overnight in aerated kiddie pools to insure that all swivels were secure. In the field, 10-meter sections of the stream were randomly selected for the tethering experiment in the same 12 sample habitats (four deep pools, four shallow pools, four riffles) used during the summer 2005 survey. A grid composed of 1 m × 1 m cells was visualized over the entire channel unit, and five crayfish from each size class were tethered in randomly selected cells on that grid in each of the 12 channel units (for a total of n = 180 crayfish tethered across all channel units). Crayfish were tethered by tying a 17-cm piece of a 6-pound test, transparent monofilament fishing line to their snap swivel (on one end) and to a tent stake (on the other end of the line). Each tent stake was then hammered into the streambed. Rebar stakes were used instead of tent stakes in deep pools to increase stake stability due to looser substrate in this habitat. Mortality was checked daily at each tethering plot for 5 days.

At the end of the experiment, water depth and average current velocity were measured at each tethering location using a wading rod and a Marsh-McBirney® flow meter. Additionally, characteristics of the top layer of substrate were measured at each tethering location. To do this, a Hess sampler (34 cm diameter, which corresponded with two times the tether length) was placed down with the stake at its center and percent cover of particles less than or equal to 4 mm was estimated. All particles greater than 4 mm in diameter [b-axis (intermediate axis of the particle)] were measured in the field using Vernier calipers and all particles less than or equal to 4 mm in diameter (b-axis) were brought back to the lab. Once in the lab, organic matter was removed from all samples by soaking them in 3% hydrogen peroxide for 24 h. Each sample was then rinsed with tap water, dried at 70°C, and analyzed using a Camsizer® particle analyzer (Retsch® Technology, Haan, Germany).

In the lab, three individual small, medium, and large crayfish (n = 9) were tethered for 5 days in 189-l aquaria to determine whether or not the glue would fail during that time interval and also to determine if crayfish were able to clip the monofilament line. All crayfish remained tethered for the full 5 days. Given that many studies have successfully tethered crayfish using this technique (DiDonato & Lodge, 1993; Kershner & Lodge, 1995; Englund & Krupa, 2000; Flinders & Magoulick, 2007a), it was assumed that all missing crayfish in the stream-based predation assays were either eaten or removed by predators.

Statistical analyses

The Cox Proportional Hazard Model (Cox, 1972) was used to compare patterns of mortality for crayfish through the use of Cox regressions to test effects of crayfish size class and habitat type separately (using SAS 8.01, SAS Institute Inc., 2000, Cary, North Carolina, USA). As noted in previous studies (e.g., Englund & Krupa, 2000; Clark et al., 2008), this model not only uses actual losses from the population but also uses data from “censored observations.” Censored observations for our study included crayfish that molted or died due to reasons other than predation. Mortality was said to occur when a piece of carapace was left on the snap swivel, the fishing line was cut, or if a crayfish was partially eaten. Further, pairwise Cox regressions were used as post hoc tests to test for differences in survival between habitat types and size classes.

Differences in habitat-specific abiotic variables (water depth, current velocity, and particle size) were analyzed using separate one-way ANOVAs for each abiotic variable (using JMP 7.0.1, SAS Institute Inc., 2007, Cary, North Carolina, USA). All data were log(x + 1)-transformed prior to analysis. Wolman pebble count and Camsizer® particle analyzer data were combined to calculate the average particle size for each tethering plot. Particle size was calculated using a fractional particle size curve just as with RDA particle size data. Separate fractional particle size curves were generated for each tethering plot to produce the data described above.

Habitat-specific predation risk: particle size manipulation experiment

In order to further explore the effects of refugia as a mediator of habitat-specific predation risk, a predation assay where small crayfish (CL < 20 mm) were tethered on manipulated substrate patches was conducted during August 2007. The previously described predation assay methods were used to set up this experiment with a couple of important exceptions. As a result of patterns of size-specific predation risk observed during summer 2005, this experiment focused on only small crayfish and both males and females were used (due to low crayfish densities). Further, particle size was manipulated using three particle size treatments chosen based on the Wentworth scale (Gordon et al., 2004). The particle size treatments were sand (range = 0.38–0.57 mm), small gravel (range = 16–32 mm, b-axis), and small cobbles (range = 64–128 mm, b-axis). All particles were collected downstream of the field site and pre-measured with the exception of sand. To measure the particle size of sand, ten samples were brought back to the lab and analyzed with a Camsizer® particle analyzer (Retsch Technology®). Tethering plots in four deep pools, four shallow pools, and four riffles were selected using the methods outlined above. Each tethering plot was randomly assigned one of the three particle size treatments such that all three particle treatments were tested in each habitat type. For the experimental setup, a Hess sampler (34 cm diameter) was placed down at each tethering plot and a complete monolayer of particles associated with the appropriate treatment was placed within the area enclosed by the Hess sampler. A crayfish was then tethered on each of these manipulated substrate patches of a single particle size treatment using previously described methods. Five crayfish per particle size treatment were tethered in each replicate of each channel unit type in this experiment (i.e., five tethered crayfish per particle size treatment in each of four replicates each of three channel unit types for a total of n = 180 tethering plots).

Statistical analyses

The Cox Proportional Hazard Model (Cox, 1972) was used to compare mortality profiles for small crayfish associated with each habitat type and particle size treatment testing the effects of habitat type and particle size treatment separately (SAS 8.01, SAS Institute Inc., 2000). The same censored variables and protocol for assessment of mortality were used as in the predation assays (described above). Survival of male and female crayfish was not significantly different, so data were pooled across sex. Pairwise Cox regressions were used as post hoc tests to test for differences in survival between particle treatments and habitat types (SAS 8.01, SAS Institute Inc., 2000).


Habitat-specific crayfish/predatory fish survey and abiotic measurements

Of the three RDA axes produced performing RDA (with predator density, predator size, particle size, depth, and current velocity as constraints), RDA Axes 1 and 2 were statistically significant (P < 0.01) and explained 62.4 and 14.0% of the variance in the crayfish density data, respectively (Fig. 1; Table 1). Additionally, permutation ANOVA of habitat variables included in the RDA indicated that particle size and depth were statistically significant factors in structuring size-specific crayfish density (Table 1). In contrast, predator density, predator size, and current velocity were not statistically significant (Table 1). Large crayfish density was related more strongly to Axis 2 compared to small and medium crayfish, exhibiting strong positive associations with deep pool habitats and increased depth (Fig. 1). In contrast, large crayfish were negatively associated with riffle habitats (Fig. 1). Small and medium crayfish densities were more strongly related to Axis 1, exhibiting strong positive associations with shallow pool habitats, increased average particle size, and decreased depth (Fig. 1). Means (±1 SE) and ranges for all abiotic and biotic parameters measured can be found in Table 2.
Fig. 1

RDA plot displaying Axes 1 and 2 with small, medium, and large crayfish densities constrained with current velocity, water depth, average particle size, fish predator density, and average fish predator TL in all 12 sample sites during summer 2005. Sample sites are presented in ordination space according to Euclidean distance. Direction of habitat variable and crayfish density vectors indicates increasing values, such that vectors pointing in the same direction are positively correlated and vectors pointing in the opposite direction are negatively correlated. Statistically significant habitat variables are underlined

Table 1

Summary table of RDA results from size-specific crayfish densities of all 12 sampling sites in summer 2005 constrained by predator density, predator size, current velocity, depth, and particle size


RDA Axis 1

RDA Axis 2

RDA Axis 3

P value






% cumulative variance





P value





Habitat variable bi-plot scores

 Predator density





 Predator size





 Current velocity










 Particle size





Crayfish density scores

 Small crayfish




 Medium crayfish




 Large crayfish




Habitat variable and crayfish density scores indicate correlations to the respective RDA axes

P values associated with RDA axes and habitat variables were calculated using permutation ANOVA

Table 2

Summary table of averages (±1 SE) and ranges of abiotic and biotic variables measured during summer 2005 in deep pools, shallow pools, and riffles


Deep pools

Shallow pools


Abiotic variables

 Average water depth (cm)

78 ± 3.0

21 ± 2.0

6.0 ± 0.30

 Water depth range (cm)




 Average current velocity (m/s)

0.00 ± 0.00

0.02 ± 0.00

0.23 ± 0.02

 Current velocity range (m/s)




 Average particle size (mm)

24 ± 5.0

36 ± 4.0

66 ± 4.0

 Particle size range (mm)




Biotic variables

 Average predator density (#/m2)

0.02 ± 0.01

0.11 ± 0.04

0.32 ± 0.04

 Predator density range (#/m2)




 Average fish predator size (cm)

8.3 ± 0.74

4.9 ± 1.6

5.6 ± 0.14

 Fish predator size range (cm)




 Average crayfish size (CL, mm)

29.0 ± 0.577

24.3 ± 0.349

22.9 ± 0.331

 Crayfish size range (CL, mm)




Habitat-specific predation risk: tethering experiment

The main effects of crayfish size class (Cox regression, χ2 = 10.04, P = 0.0015, df = 2) and habitat type (Cox regression, χ2 = 17.53, P < 0.0001; df = 2) significantly affected crayfish survival during summer 2005 (Fig. 2). Regardless of habitat type, small crayfish survival (n = 60) was significantly lower than both medium (Cox regression, χ2 = 12.95, P = 0.0003, n = 60) and large crayfish (Cox regression, χ2 = 7.41, P = 0.0065, n = 60) (Fig. 2). However, medium and large crayfish had similar survival (Cox regression, χ2 = 1.23, P = 0.2680) (Fig. 2). Regardless of crayfish size class, survival was significantly lower in deep pools than in either riffles (Cox regression, χ2 = 14.57, P = 0.0001) or shallow pools (Cox regression, χ2 = 5.96, P = 0.0146) (Fig. 2). Survival of crayfish was similar in shallow pools and riffles (Cox regression, χ2 = 2.11, P = 0.1460) (Fig. 2).
Fig. 2

Average percent survival (mean ± 1 SE) of small, medium, and large crayfish in deep pools, shallow pools, and riffles during summer 2005. Dark circles deep pools, light circles shallow pools, and dark squares riffles

Abiotic variables measured at each tethering location in each habitat type further supported how habitat type was defined in this system and showed significant differences in current velocity, water depth, and average particle size. Current velocity was significantly faster in riffles than in both deep and shallow pools (ANOVA, F(2,177) = 202.27, P < 0.0001; Tukey’s HSD, P < 0.05) with deep and shallow pools having similar velocities (Tukey’s HSD, P > 0.05). Water depth was significantly deeper in deep pools than in both riffles and shallow pools (ANOVA, F(2,177) = 535.58, P < 0.0001; Tukey’s HSD, P < 0.05) with riffles and shallow pools having similar water depth (Tukey’s HSD, P > 0.05). Lastly, average particle size was the largest in riffles followed by shallow pools and then deep pools with the smallest average size (ANOVA, F(2,177) = 39.01, P < 0.0001; Tukey’s HSD, all P < 0.05).

Habitat-specific predation risk: particle size manipulation experiment

Habitat type (Cox regression, χ2 = 57.50, P < 0.0001) and particle size (Cox regression, χ2 = 13.56, P = 0.0002) significantly affected survival of small crayfish (Fig. 3). Regardless of particle size, survival in riffles was significantly higher than in both deep pools (Cox regression, χ2 = 41.62, P < 0.0001) and shallow pools (Cox regression, χ2 = 28.76, P < 0.0001) (Fig. 3). Additionally, survival of small crayfish was significantly higher in shallow pools than deep pools (Cox regression, χ2 = 6.85, P = 0.0089) (Fig. 3).
Fig. 3

Average percent survival (mean ± 1 SE) of small crayfish in deep pools, shallow pools, and riffles on sand, gravel, and cobble plots during summer 2007. Dark squares sand plots, light circles gravel plots, and dark circles cobble plots

Furthermore, survival on cobble plots was significantly higher than on sand (Cox regression, χ2 = 6.66, P = 0.0099) plots, but similar on cobble and gravel plots (Cox regression, χ2 = 0.60, P = 0.4383) (Fig. 3). Small crayfish survival appeared to be higher on gravel than on sand, but was not significant (Cox regression, χ2 = 3.13, P = 0.0768) (Fig. 3).


All size classes of crayfish were vulnerable to predation. However, small crayfish were significantly more vulnerable than larger crayfish regardless of habitat type, with medium and large crayfish having similar survivorship. Although predation risk was similar for crayfish ≥20 mm CL, their spatial distribution differed, with the largest size class strongly correlated with deeper pools and medium crayfish more strongly correlated (along with small crayfish) with shallow pools. These differences in habitat use suggest that interactions other than predation, such as intra- or interspecific competition and/or abiotic variables, structure larger crayfish distribution patterns.

When the spatial distribution of a given size class does not match with predictions based on habitat-specific predation risk, other factors, such as substrate type, may be driving observed habitat use. In this study, small crayfish showed the lowest survival in deep and shallow pools, and considerably higher survival in riffles. Field surveys, however, indicated that small crayfish occurred in the highest densities in shallow pool habitats where risk of mortality was relatively high. This suggests that abiotic variables such as particle size may provide important refugia from predators for these smaller crayfish, allowing them to persist in these high-risk habitats. Tethering experiments in this study confirmed that larger particles can increase survivorship of small-bodied crayfish regardless of habitat type. Similarly, in Wisconsin lakes, tethered O. rusticus (Girard) had higher mortality rates on sand than on cobble substrate (Kershner & Lodge, 1995). Likewise, mortality was lower for three species of Orconectes crayfish on dense cobble than smaller-diameter substrates in mesocosm experiments (Garvey et al., 1994). In the field, larger substrates may be preferred by smaller crayfish since interstitial spaces between larger substrates provide better refugia against predation than smaller substrates (Stein & Magnuson, 1976). In fact, crayfish abundance is often positively associated with increasing substrate size (Lodge & Hill, 1994; Streissl & Hödl, 2002; Parkyn & Collier, 2004; Flinders & Magoulick, 2005; Pockl & Streissl, 2005) with crayfish density generally negatively associated with sand substrate (this study, Kershner & Lodge, 1995). These general patterns are further borne out by the tight associations between small juvenile crayfish abundance and habitats with large particle sizes (e.g., cobbles, boulders), and large adult crayfish occupation of patches with small particle sizes (e.g., sand and clay) (Usio & Townsend, 2000; Clark et al., 2008). Thus, while predation is an important source of mortality in the early life stages of organisms and can restrict juveniles to refugia, adults can likely range more freely (Stein & Magnuson, 1976; Werner et al., 1983a, b; Schlosser, 1987), given their reduced predation risk and corresponding decreased need for refugia.

While substrate type is important, predation risk for crayfish can also be strongly associated with interactions between crayfish size and other abiotic variables (e.g., water depth). For example, in Kentucky streams, body size played a strong role in predation risk with small O. putnami (Faxon) showing higher mortality than larger crayfish in the presence of fish predators and larger crayfish being relatively unaffected by fish predators (Englund & Krupa, 2000). Similarly, in New York streams, small O.propinquus (Girard) and O.rusticus had higher mortality than larger individuals (Kuhlmann et al., 2008). Water depth in these two studies also interacted with body size, with small crayfish having higher mortality rates than larger crayfish in deep water. However, this trend was reversed in shallower water with smaller crayfish having higher survival rates (Englund & Krupa, 2000; Kuhlmann et al., 2008) with the exception of one stream that had a high abundance of terrestrial predators (Englund & Krupa, 2000). This pattern is similar to Midwestern streams, where small, juvenile O. sanborni (Faxon) had much higher mortality in deep pools than in shallow pools and riffles, which shared similarly low juvenile mortality (Mather & Stein, 1993b).

While our study found similar trends in size-specific survival, water depth appeared to have limited effects on crayfish survival regardless of body size, but did play a strong role in size-dependent distribution patterns with smaller crayfish occupying shallower habitats. This suggests that other abiotic parameters (i.e., particle size, current velocity) may interact with predation risk to structure crayfish distribution patterns. For example, particle size in our system increased in size from deep pools to shallow pools to riffles and may be large enough in shallow pools to mitigate the impacts of water depth.

Interestingly, while predation risk was low for small crayfish in riffles and particle size was the largest, field surveys found no preference for this habitat by small crayfish. It is possible that current velocity and/or the density of predaceous Cottus bairdi may limit small crayfish densities in riffles. However, in this stream, typical riffle current velocities were within tolerable limits for this species (Clark et al., 2008) and therefore are unlikely to substantively limit crayfish distribution. While diet data indicate that Cottus bairdi were weak predators, they have relatively high densities in riffles (in this stream) and may play a more significant role as a competitor than as a predator.

Some streams do not follow these size- and habitat-specific patterns. In Ozark streams, O. marchandi (Hobbs) had higher mortality in deeper water than shallow habitats (similar to our study), but did not show size-selective mortality (Flinders & Magoulick, 2007a). Further, in some streams, size-dependent crayfish survival can have complex interactions between water depth and terrestrial predator abundance (Englund & Krupa, 2000). Since size- and habitat-specific mortality is not consistent across streams, predator types (i.e., aquatic vs. terrestrial) and the complexity of the predator assemblage (i.e., diversity and abundance) may also play a major role in affecting size- and habitat-specific crayfish mortality and distribution patterns in some systems.

In our system, fish predators were present in each habitat type, but did not appear to affect distribution patterns. Their impact on crayfish distribution and survival was likely limited by the low abundance of crayfish that were susceptible to predation based on fish gape limitations and crayfish relative size. For example, the abundant central mottled sculpins (abundant in riffles and shallow pools) reach a maximum TL of 10 cm, have gape sizes capable of consuming only very small crayfish, and are weak predators of crayfish (J.M. Clark, unpublished data). While creek chubs were relatively abundant in both shallow and deep pools and were sometimes large (TL > 15 cm), they also tend to be relatively weak predators of crayfish in this system (J.M. Clark, unpublished data). Other predatory fish, such as green sunfish and rock bass, are more common in deep pools than in shallow-water habitats and are efficient crayfish predators (rock bass, Rabeni, 1992; green sunfish, Englund & Krupa, 2000). These two fish species (in low abundance in the stream in this study) were observed preying upon small (but not medium or large) crayfish in deep pools during tethering experiments (as were creek chubs), and this is most likely the reason that medium and large crayfish experienced similar predation risk overall.

Small, juvenile crayfish are generally more vulnerable to predation as many predatory fish are gape limited (Stein & Magnuson, 1976; Stein, 1977; Rabeni 1992; DiDonato & Lodge, 1993; Kershner & Lodge, 1995; Englund & Krupa, 2000, this study). Further, even when not gape limited, predatory fish tend to prey upon smaller crayfish, reducing handling costs and increasing capture rates as a function of smaller crayfish chelae (Stein, 1977). Chela displays are often used by crayfish to deter fish predators (Bovbjerg, 1956; Stein & Magnuson, 1976), but are likely only a threat from larger crayfish, contributing to their lower vulnerability to fish predation (Rabeni, 1992). Ultimately, size-selective feeding exhibited by predaceous fish (Rabeni, 1992; Mather & Stein, 1993a) may be an important driver of crayfish population dynamics.

Other aquatic predators such as spiny softshell turtles (Apalone spinifera L.), painted turtles (Chrysemys picta Schneider), and common snapping turtles (Chelydra serpentina L.); semi-aquatic predators such as North American river otter; and terrestrial predators such as belted kingfishers (Megaceryle alcyon L.), great blue herons, green herons (Butorides virescens L.), and raccoons (Procyon lotor L.) were common visitors to this system and all can be predators of crayfish. These diving and wading predators are likely culprits (over fish) for predation on medium and large crayfish due to the larger gape sizes of many of these predators. Half-eaten medium and large crayfish occurred in each of the three habitat types and, in some cases, body parts were scattered around tethering locations, suggesting that aquatic, semi-aquatic, and/or terrestrial predators were all preying upon crayfish. Although few studies have examined the role of terrestrial predators on crayfish, mammals and birds have been shown to be important predators of small and large O. putnami (Englund & Krupa, 2000) and Procambarus clarkii (Girard) (Correia, 2001).

While predation and the presence of large substrates appear to drive distribution patterns of small crayfish, predation is likely not the only factor structuring the spatial distribution of larger crayfish given minimal habitat-specific differences in survival for ≥20 mm CL. Other factors such as territoriality, current velocity, and competition likely play a stronger role for larger crayfish. Future studies should assess habitat- and size-specific trade-offs made between predation risk and resource availability and how combined effects of abiotic parameters can impact these trade-offs and recruitment. Given the potential importance of crayfish in aquatic systems [as keystone species (Nyström et al., 1996; Creed & Reed, 2004) and ecosystem engineers (Statzner et al., 2003)], their distribution patterns may have substantial effects not only on stream communities but also on organic matter processing and nutrient cycling. For example, increased crayfish densities in low flow and high depositional habitats may greatly increase the amount of fine particulate organic matter generated through their foraging activities on leaves accumulating in these habitats.



We would like to thank Matt Begley, Doug Kapusinski, Emily Faulkner, Joe Faulkner, Eric Floro, Emma Kennedy, Jackie Johnston, Joel Mulder, Jamie Stamberger, Matt Walker, Raja Vukanti, Maureen Drinkard, Adam Leff, Ben Leff, Laura Leff, Denise Walker, Constance Hausman, Stephanie Hovatter, Alex Arp, and Mark DuFour for assisting with fieldwork. Andrew Moore provided assistance with analysis of particle populations. Special thanks to two anonymous reviewers for providing helpful comments for an earlier version of this manuscript.


  1. Bovbjerg, R. V., 1956. Some factors affecting aggressive behavior in crayfish. Physiological Zoology 29: 127–136.Google Scholar
  2. Bubb, D. H., T. J. Thom & M. C. Lucas, 2004. Movement and dispersal of the invasive signal crayfish Pacifastacus leniusculus in upland rivers. Freshwater Biology 49: 357–368.CrossRefGoogle Scholar
  3. Clark, J. M., M. W. Kershner & J. R. Holomuzki, 2008. Grain size and sorting effects on size-dependent responses by lotic crayfish to high flows. Hydrobiologia 610: 55–66.CrossRefGoogle Scholar
  4. Correia, A. M., 2001. Seasonal and interspecific evaluation of predation by mammals and birds on the introduced red swamp crayfish Procambarus clarkii (Crustacea, Cambaridae) in a freshwater marsh (Portugal). Journal of Zoology 255: 533–541.CrossRefGoogle Scholar
  5. Cox, D. R., 1972. Regression models and life tables. Journal of the Royal Statistical Society 34: 187–200.Google Scholar
  6. Creed, R. P., 2006. Predator transitions in stream communities: a model and evidence from field studies. Journal of the North American Benthological Society 25: 533–544.CrossRefGoogle Scholar
  7. Creed, R. P. & J. M. Reed, 2004. Ecosystem engineering by crayfish in a headwater stream community. Journal of the North American Benthological Society 23: 224–236.CrossRefGoogle Scholar
  8. DiDonato, G. T. & D. M. Lodge, 1993. Species replacements among Orconectes crayfishes in Wisconsin Lakes-The role of predation by fish. Canadian Journal of Fisheries and Aquatic Sciences 50: 1484–1488.CrossRefGoogle Scholar
  9. Dorn, N. J. & G. G. Mittelbach, 1999. More than predator and prey: a review of interactions between fish and crayfish. Vie et Milieu-Life and Environment 49: 229–327.Google Scholar
  10. Englund, G. & J. J. Krupa, 2000. Habitat use by crayfish in stream pools: influence of predators, depth, and body size. Freshwater Biology 43: 75–83.CrossRefGoogle Scholar
  11. Fielder, D. D., 1972. Some aspects of the life histories of three closely related crayfish species, Orconectes obscurus, O. sanborni, and O. propinquus. The Ohio Journal of Science 72: 129–145.Google Scholar
  12. Flinders, C. A. & D. D. Magoulick, 2003. Effects of stream permanence on crayfish community structure. American Midland Naturalist 149: 134–147.CrossRefGoogle Scholar
  13. Flinders, C. A. & D. D. Magoulick, 2005. Distribution, habitat use and life history of stream- dwelling crayfish in the spring river drainage of Arkansas and Missouri with a focus on the imperiled Mammoth Spring crayfish (Orconectes marchandi). American Midland Naturalist 154: 358–374.CrossRefGoogle Scholar
  14. Flinders, C. A. & D. D. Magoulick, 2007a. Effects of depth and crayfish size on predation risk and foraging profitability of a lotic crayfish. Journal of the North American Benthological Society 26: 767–778.CrossRefGoogle Scholar
  15. Flinders, C. A. & D. D. Magoulick, 2007b. Habitat use and selection within Ozark lotic crayfish assemblages: spatial and temporal variation. Journal of Crustacean Biology 27: 242–254.CrossRefGoogle Scholar
  16. Garvey, J. E., R. A. Stein & H. M. Thomas, 1994. Assessing how fish predation and interspecific prey competition influence a crayfish assemblage. Ecology 75: 532–547.CrossRefGoogle Scholar
  17. Gordon, N. D., T. A. McMahon, B. L. Finlayson, C. J. Gippel & R. J. Nathan, 2004. Stream Hydrology: An Introduction for Ecologists, 2nd ed. Wiley, West Sussex, England.Google Scholar
  18. Griffith, M. B., S. A. Perry & W. B. Perry, 1994. Secondary production of macroinvertebrate shredders in headwater streams with different baseflow alkalinity. Journal of the North American Benthological Society 13: 345–356.CrossRefGoogle Scholar
  19. Hart, D. D. & C. M. Finelli, 1999. Physical-biological coupling in streams: the pervasive effects of flow on benthic organisms. Annual Review of Ecology and Systematics 30: 363–395.CrossRefGoogle Scholar
  20. Hicks, B. J., 2003. Distribution of fish and crayfish in a Waikato stream in relation to basin area. New Zealand Journal of Zoology 30: 149–160.CrossRefGoogle Scholar
  21. Hill, A. M. & D. M. Lodge, 1999. Replacement of resident crayfishes by an exotic crayfish: the roles of competition and predation. Ecological Applications 9: 678–690.CrossRefGoogle Scholar
  22. Kershner, M. W. & D. M. Lodge, 1995. Effects of littoral habitat and fish predation on the distribution of an exotic crayfish, Orconectes rusticus. Journal of the North American Benthological Society 14: 414–422.CrossRefGoogle Scholar
  23. Kuhlmann, M. L., S. M. Badylak & E. L. Carvin, 2008. Testing the differential predation hypothesis for the invasion of rusty crayfish in a stream community: laboratory and field experiments. Freshwater Biology 53: 113–128.Google Scholar
  24. Legendre, P. & L. Legendre, 1998. Numerical Ecology. Elsevier Science, Amsterdam, The Netherlands.Google Scholar
  25. Light, T., 2003. Success and failure in a lotic crayfish invasion: the roles of hydrologic variability and habitat alteration. Freshwater Biology 48: 1886–1897.CrossRefGoogle Scholar
  26. Lodge, D. M. & A. M. Hill, 1994. Factors governing species composition, population size, and productivity of cool-water crayfishes. Nordic Journal of Freshwater Research 69: 111–136.Google Scholar
  27. Lodge, D. M., M. W. Kershner, J. E. Aloi & A. P. Covich, 1994. Effects of omnivorous crayfish (Orconectes rusticus) on a freshwater food web. Ecology. 75: 265–1281.CrossRefGoogle Scholar
  28. Lodge, D. M., C. A. Taylor, D. M. Holdich & J. Skurdal, 2000. Nonindigenous crayfishes threaten North American freshwater biodiversity: lessons from Europe. Fisheries 25: 7–20.CrossRefGoogle Scholar
  29. Magoulick, D. D., 2004. Effects of predation risk on habitat selection by water column fish, benthic fish, and crayfish in stream pools. Hydrobiologia 527: 209–221.CrossRefGoogle Scholar
  30. Mather, M. E. & R. A. Stein, 1993a. Direct and indirect effects of fish predation on the replacement of a native crayfish by an invading congener. Canadian Journal of Fisheries and Aquatic Sciences 50: 1279–1288.CrossRefGoogle Scholar
  31. Mather, M. E. & R. A. Stein, 1993b. Using growth-mortality tradeoffs to explore a crayfish species replacement in stream riffles and pools. Canadian Journal of Fisheries and Aquatic Sciences 50: 88–96.CrossRefGoogle Scholar
  32. McCune, B., J. B. Grace & D. L. Urban, 2002. Analysis of Ecological Communities. MjM Software Design, Gleneden Beach, OR.Google Scholar
  33. Nyström, P., C. Brönmark & W. Granéli, 1996. Patterns in benthic food webs: a role for omnivorous crayfish? Freshwater Biology 36: 631–646.CrossRefGoogle Scholar
  34. Oksanen, J., F. G. Blanchet, R Kindt, P. Legendre, P. R. Minchin, R. B. O’Hara, G. L. Simpson, P. Solymos, M. Henry, H. Stevens & H. Wagner, 2013. vegan: Community Ecology Package. R package version 2.0-6. http://CRAN.R-project.org/package=vegan/.
  35. Olsson, K. & P. Nyström, 2009. Non-interactive effects of habitat complexity and adult crayfish on survival and growth of juvenile crayfish (Pacifastacus leniusculus). Freshwater Biology 54: 35–46.CrossRefGoogle Scholar
  36. Parkyn, S. M. & K. J. Collier, 2004. Interaction of press and pulse disturbance on crayfish populations: flood impacts in pasture and forest streams. Hydrobiologia 527: 114–124.CrossRefGoogle Scholar
  37. Pockl, M. & F. Streissl, 2005. Austropotamobius torrentium as an indicator for habitat quality in running waters? Bulletin Francais De La Peche Et De La Pisciculture 376–377: 743–758.CrossRefGoogle Scholar
  38. Poos, M. S., N. E. Mandrak & R. L. McLaughlin, 2007. The effectiveness of two common sampling methods for assessing imperiled freshwater fishes. Journal of Fish Biology 70: 691–708.CrossRefGoogle Scholar
  39. R Core Team, 2013. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. http://www.R-project.org/.
  40. Rabeni, C. F., 1992. Trophic linkage between stream Centrarchids and their crayfish prey. Canadian Journal of Fisheries and Aquatic Sciences 49: 1714–1721.CrossRefGoogle Scholar
  41. Rabeni, C. F., K. J. Collier, S. M. Parkyn & B. J. Hicks, 1997. Evaluating techniques for sampling stream crayfish (Paranephrops planifrons). New Zealand Journal of Marine and Freshwater Research 31: 693–700.CrossRefGoogle Scholar
  42. Renai, B. & F. Gherardi, 2004. Predatory efficiency of crayfish: comparison between indigenous and non-indigenous species. Biological Invasions 6: 89–99.CrossRefGoogle Scholar
  43. Roell, M. J. & D. J. Orth, 1993. Trophic basis of stream-dwelling smallmouth bass, rock bass, and flathead catfish in relation to invertebrate bait harvest. Transactions of the American Fisheries Society 122: 46–62.CrossRefGoogle Scholar
  44. SAS Institute, 2000. The SAS System for Windows, version 8.01. SAS Institute, Inc., Cary, North Carolina.Google Scholar
  45. SAS Institute, 2007. JMP, version 7.0.1. SAS Institute, Inc., Cary, North Carolina.Google Scholar
  46. Schlosser, I. J., 1987. The role of predation in age- and size-related habitat use by stream fishes. Ecology 68: 651–659.CrossRefGoogle Scholar
  47. Seiler, S. M. & A. M. Turner, 2004. Growth and population size of crayfish in headwater streams: individual- and higher-level consequences of acidification. Freshwater Biology 49: 870–881.CrossRefGoogle Scholar
  48. Statzner, B., O. Peltret & S. Tomanova, 2003. Crayfish a geomorphic agents and ecosystem Engineers: effect of a biomass gradient on baseflow and flood-induced transport of gravel and sand in experimental streams. Freshwater Biology 48: 147–163.CrossRefGoogle Scholar
  49. Stein, R. A., 1977. Selective predation, optimal foraging, and the predator-prey interaction between fish and crayfish. Ecology 58: 1237–1253.CrossRefGoogle Scholar
  50. Stein, R. A. & J. J. Magnuson, 1976. Behavioral response of a crayfish to a fish predator. Ecology 57: 751–761.CrossRefGoogle Scholar
  51. Streissl, F. & W. Hödl, 2002. Habitat and shelter requirements of the stone crayfish, Austropotamobius torrentium Schrank. Hydrobiologia 477: 195–199.CrossRefGoogle Scholar
  52. Trautman, M. B., 1957. Fishes of Ohio. The Ohio State University Press, Ohio.Google Scholar
  53. Usio, N. & C. R. Townsend, 2000. Distribution of the New Zealand crayfish Paranephrops zealandicus in relation to stream physico-chemistry, predatory fish, and invertebrate prey. New Zealand Journal of Marine and Freshwater Research 34: 557–567.CrossRefGoogle Scholar
  54. Usio, N. & C. R. Townsend, 2001. The significance of the crayfish Paranephrops zealandicus as shredders in a New Zealand headwater stream. Journal of Crustacean Biology 21: 354–359.CrossRefGoogle Scholar
  55. Werner, E. E., J. F. Gilliam, D. J. Hall & G. G. Mittelbach, 1983a. Experimental tests of optimal habitat use in fish: the role of relative habitat profitability. Ecology 64: 1525–1539.CrossRefGoogle Scholar
  56. Werner, E. E., J. F. Gilliam, D. J. Hall & G. G. Mittelbach, 1983b. An experimental test of the effects of predation risk on habitat use in fish. Ecology 64: 1540–1548.CrossRefGoogle Scholar
  57. Wolman, M. G., 1954. A method of sampling coarse river-bed material. Transactions of the American Geophysical Union 15: 951–956.Google Scholar
  58. Zippin, C., 1958. The removal method of population estimation. Journal of Wildlife Management 22: 82–90.CrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media Dordrecht 2013

Authors and Affiliations

  • Jennifer M. Clark
    • 1
  • Mark W. Kershner
    • 2
  • Justin J. Montemarano
    • 2
    • 3
  1. 1.Biology DepartmentHiram CollegeHiramUSA
  2. 2.Department of Biological SciencesKent State UniversityKentUSA
  3. 3.Department of BiologyArmstrong Atlantic State UniversitySavannahUSA

Personalised recommendations