Abstract

Modified minnow traps are the most widely used gear for collecting tertiary burrowing crayfishes. The throats of modified minnow traps are often widened more than 60% to accommodate the capture of larger crayfish. However, widening this crucial chokepoint into the trap may facilitate easier escape of captured individuals, especially as density in the trap increases. Increased escapement rates may reduce catch rates and corresponding estimates of relative abundance and lower detection probability. Incorporating a design feature, that is, throat restriction, that allows entry of all sizes of crayfish while reducing escapement would be an improvement over current designs. Here, I present the results of a paired field and laboratory experiment comparing the effectiveness of modified minnow traps with a throat restriction (restricted) and without such a feature (unrestricted) under varying crayfish densities. I chose rusty crayfish Faxonius rusticus as a study organism because it is widespread and abundant in the Laurentian Great Lakes region and commonly the focus of research and removal efforts. Rusty crayfish capture and escapement were strongly influenced by throat design and crayfish densities. The field component demonstrated that both traps performed similarly under low-to-moderate densities; however, at high densities catch in unrestricted traps plateaued at approximately 50 crayfish/trap, while restricted traps kept accumulating catch up to 155 crayfish/trap. Laboratory trials demonstrated that escapement for both trap types was negligible at low density and slightly higher at medium density: 0.8% for restricted and 11.5% for unrestricted. However, at high density, escapement from restricted traps was 8.8 vs. 45.3% for unrestricted traps. Our findings suggest that inclusion of a throat restriction may increase catch of rusty crayfish by reducing escapement and may be of particular use in removal projects or when sampling in high-density populations.

Introduction

Field sampling for crayfishes, especially tertiary burrowing species (i.e., those that spend most of their lives in flowing or standing waters), is typically performed using a variety of gears including modified minnow traps (MMTs; Gee-style), specially designed crayfish traps, hand collection, and colonization traps (Larson and Olden 2016). However, the most common method used for capturing crayfishes is deploying baited MMTs because of their low cost, ease of use, and effectiveness (Larson and Olden 2016). For crayfish sampling, most conventional Gee-style minnow traps have to be modified by widening the opening to the trap, often referred to as the throat or funnel opening. This modification is necessary to accommodate entry of larger individuals of many species, including the invasive rusty crayfish Faxonius rusticus (Lodge et al. 1986; Somers and Stetchey 1986; Stuecheli 1999).

For sampling crayfishes and other decapods, the trade-off for widening the throat on passive entrapment gears is easier escapement, especially for smaller individuals (Nulk 1978; Miller 1990; Stuecheli 1999). All entrapment gears used to collect crayfishes are highly size selective (Larson and Olden 2016); typically, they are biased towards large individuals, which tend to be males (Lodge et al. 1986; Stuecheli 1999; Ogle and Kret 2008). Some traps for crayfishes and similar decapods (i.e., crabs, lobsters) limit escapement of captured individuals by using spring-operated trap doors that close behind individuals after entering the trap (Miller 1990; Burrell 2017), or by placing the funnel opening on the top of the trap, instead of on the sides (Ulikowski et al. 2017). These modifications are unfeasible for modified minnow traps because of their shape, and how they are deployed. Nonetheless, a modification that reduces escapement of captured crayfishes would be an improvement to current designs. Several potential design solutions already exist for fish and crayfish sampling gears, but to date they have not been rigorously assessed.

Fisheries sampling often involves the use of passive entrapment gears, similar to, and including, minnow traps. Over time, researchers and managers noticed that some species are adept at escaping from entrapment gears (e.g., hoop nets, modified fyke nets) and found simple ways to reduce escapement (Porath et al. 2011; Smith et al. 2016). These design modifications typically involve a restriction around the entrance to the trap (throat) that make it possible for fish to enter the trap but more difficult to escape. Ulikowski et al. (2017) described a similar design for a crayfish trap; their trap design, referred to as Evo, is a collapsible crayfish trap with a series of strings connecting the opposing throat openings, which may interfere with crayfish escapement. A slightly different design from those previously developed would be required for modified minnow traps. For this reason, I designed a device that attaches to the throat of an MMT, hereafter referred to as a throat restriction, and assessed its performance against an MMT lacking this device under field and laboratory conditions.

The study organism I chose for this investigation was the rusty crayfish. This species is among the most widespread and abundant crayfishes in North America and considered invasive across much of its current range (Figure 1, top left; Durland Donahou et al. 2018). Its propensity for colonizing streams and lakes, displacing native crayfishes, and altering aquatic food webs has been well documented (Houghton et al. 1998; Wilson et al. 2004; Olden et al. 2006). The rusty crayfish makes an ideal study organism because it is predominantly surveyed using MMTs, there are many systems where it can be found, and it occurs at a wide range of densities in many habitats. These attributes were ideal for testing differences between MMT designs.

My objectives were to 1) compare catch rate, size structure, and sex ratio between MMTs with and without throat restrictions under field conditions and 2) compare escapement rate, size structure, and sex ratio between MMTs with and without throat restrictions under controlled laboratory conditions. Differences in catch rates were investigated in the field to encompass a wide range of habitats and site-specific abundances of rusty crayfish. I investigated escapement under laboratory conditions to control for density of rusty crayfish in the trap and exclude the possibility of new rusty crayfish entering the trap. I hypothesized that in the field rusty crayfish would enter the two types of traps at similar rates, but that catch rates would be higher in traps with restricted throats due to reduced escapement. Reduced escapement, particularly of small individuals, would likely result in smaller size structure and a higher proportion of females in restricted traps. For laboratory trials, I hypothesized that a larger proportion of rusty crayfish would escape from unrestricted traps, especially smaller individuals and females, and that escapement rates would increase at higher densities for both types of traps.

Methods

Study site

We concentrated field sampling in the lower and middle portions of Baird Creek from June 11 to June 28, 2018, a period when water temperatures exceeded 20°C and rusty crayfish activity was high. Baird Creek is a small stream in Brown County, Wisconsin, with a 40.8-km2 watershed and mean annual discharge of 0.3 m3/s (Figure 1, top right). Baird Creek discharges into the East River in the city of Green Bay (population, ≈105,000) that then empties into the Fox River near its mouth on Green Bay, the largest embayment of Lake Michigan. At base flow, the creek is less than 10 m wide and less than 1 m deep. Ideal habitats for rusty crayfish in this system are pools, logjams, and undercut banks, and ideal substrates are broken bedrock and cobble.

Gear design

Wire mesh MMTs used for this study were 420 mm long and 230 mm wide. We widened the throats of all traps from 22 to 40 mm to accommodate large individuals (Lodge et al. 1986; Somers and Stechey 1986) by removing the ring at the terminal end of the throat and making four evenly spaced cuts perpendicular to the throat. We then forced a 40-mm-wide section of polyvinyl chloride (PVC) pipe through the throat to widen it to the correct diameter. Unrestricted MMTs had only this modification. We constructed throat restrictions using 12-gauge wire and zip ties with clipped ends. We formed two sections of wire into rings, one larger (56-mm-diameter ring) and one smaller (48-mm-diameter ring), placed onto a conical mold, and then we secured eight zip ties with clipped ends (90 mm long × 3 mm wide) to the outside of the wire rings at evenly spaced intervals by using super glue. After setting for 1–2 h, we removed the throat restriction from the mold and secured to the throat of an MMT by using two evenly spaced zip ties with the conical-shaped device facing inwards (Figure 1, bottom left).

Field sampling

The study design for field sampling focused on areas that were likely to hold rusty crayfish (i.e., pools, logjams, undercut banks, rocky areas). Two weeks before field sampling commenced, I identified 45 sites (≥50 m apart) where water depths were sufficient to cover traps completely and had sufficient cover for rusty crayfish. Sites were successively sampled in an upstream direction and only visited once. At each site, a restricted MMT and an unrestricted MMT were set in tandem (i.e., paired samples) and baited with approximately 90 g of locally caught Gizzard Shad Dorosoma cepedianum (Figure 1, bottom right). Water depth (m), habitat type (i.e., run, riffle pool), substrate type (e.g., clay, sand, gravel, cobble), water temperature (°C), and dissolved oxygen (mg/L) were recorded at each site following descriptions and procedures described by Simonson et al. (1993). Water temperature and dissolved oxygen were measured using a handheld meter (model 85, YSI). Rusty crayfish are most active at night; therefore, traps were set in the afternoon, left to fish overnight, and then retrieved the following morning (at least 16-h sets). All captured rusty crayfish were removed from traps, identified to species, measured for carapace length (CL; 0.1-mm resolution), and sexed. We transported approximately 600 rusty crayfish from the field back to the Green Bay Fish and Wildlife Conservation Office (New Franken, Wisconsin) for use in the laboratory portion of the project.

Laboratory experiment

Upon arrival at Green Bay Fish and Wildlife Conservation Office, we placed all rusty crayfish in an aerated 400-gallon (1,514-L) recirculating system with water at ambient temperature, hereafter referred to as the holding tank. We provided sufficient cover to allow crayfish to find shelter and rest, reducing the likelihood of mortality. The laboratory space used for the experiment had multiple windows that allowed crayfish to stay on the same day–night schedule as in the wild. Acclimation of crayfish occurred for 48 h before starting laboratory trials. This was the population from which individuals were drawn for escapement trials.

We performed escapement trials at three stocking densities based on field capture data. Low density (25th percentile) was 2 crayfish/trap, medium density (50th percentile) was 13 crayfish/trap, and high density (75th percentile) was 57 crayfish/trap. We performed escapement trials in eight large plastic tubs (length × width × depth; 80 × 46 × 32 cm; 118 L) for low- and medium-density treatments and in four cattle tubs (length × width × depth; 80 × 46 × 32 cm; 570 L) for high-density treatments. Each tub received only one trap per trial. We aerated all tubs, and they received 50–100% water changes daily to ensure suitable water quality and we filled them sufficiently so that traps were fully submerged. Using both types of tubs, we performed 12 escapement trials each day. In total, we performed 60 escapement trials, 10 replicates for each combination of stocking density (i.e., low, medium, high) and trap type (i.e., restricted, unrestricted). We randomly assigned treatment designation for each tub for each trial period. Based on the density of crayfish to be stocked, each tub had a corresponding density of PVC pipe sections (length × diameter, 100 × 50 mm) to provide cover outside the trap for escaped crayfish (e.g., 50 stocked crayfish had 50 PVC pipe sections). Equivalent volumes of identical bait (20 g of Gizzard Shad) were placed outside the trap in each tub to incentivize escapement.

At the beginning of each trial period, we drew crayfish from the stock tank, measured them for CL (mm), sexed them, randomly assigned them to a treatment group, and placed them in the randomly assigned trap type and tub. We blocked trap throats shut until the experiment began, at which point we unblocked the throats and crayfish could escape. All trials began at 1700 hours and ended at 0800 hours the following morning. At the conclusion of a trial period, we blocked the throats of all traps, thereby preventing further escapement, and ending the experiment. We measured and sexed all crayfish, with crayfish location inside or outside of traps noted. Following each trial period, we returned crayfish to the holding tank to recover for 8 h before potentially being used in the next trial. Using this experimental design, we used approximately half of the crayfish in the stock tank each day, and we likely used each crayfish in multiple trials during the experiment.

Analysis and interpretation

Field sampling.

We assessed the catch relationship between paired traps from the field by plotting catch of restricted traps (x axis) as a function of catch in unrestricted traps (y axis) by using each site where rusty crayfish were captured as replicates. After visualizing the data, nine independent regression models were fit to the data using the easynls package in R including linear, quadratic, linear plateau, quadratic plateau, exponential, logistic, von Bertalanffy, Brody, and Gompertz models (Arnhold 2017). We compared the fit of these models using Akaike's Information Criterion (AIC) and associated model weights (Burnham and Anderson 2003). We used the stats4 package to compute AIC scores and the MuMIn package to calculate model weights (R Core Team 2017). We considered models with ΔAIC scores within 2 units off the best-fit model functionally identical; if multiple models were closely ranked, we chose the model with the highest weight. We plotted the top ranked model along with 95% confidence intervals to aid in interpretation.

We compared size structure and sex ratio between crayfish trap designs. The CLs from all rusty crayfish captured in each gear type were pooled for analysis and compared using the Kolmogorov–Smirnov (K-S) test. Male-to-female sex ratios were calculated across all sites and compared qualitatively between trap designs.

Laboratory experiment.

For laboratory trials, we used a 2 × 3 factorial analysis of variance (ANOVA; two traps [main effect] × three densities [main effect]) to identify differences in escapement rates and potential interactions between trap design and density (interactive effect). If we detected differences during ANOVAs, we used Tukey's honestly significant difference test to identify differences between treatments. If we detected significant interactive effects, we used an interaction plot to identify whether the interaction was ordinal or disordinal. If the interactive effect was ordinal, we interpreted main effects with caution and used post hoc procedures (i.e., Tukey's honestly significant difference). If interactive effects were disordinal, we could not interpret main effects, we did not use post hoc procedures, and we pursued interpretation of just the interactive effect.

Most crayfish were likely used during more than one trial day and had the highest probability of being used in high-density treatments. It was possible that crayfish exposed to similar situations within the same week may have improved ability to escape from traps. To investigate this possibility, escapement rates were plotted for each trap type and density by day and then qualitatively compared across the 5-d experimental period. Sample sizes (i.e., number of days) were too small for tests of correlation.

We compared size structure and sex ratio among stocked, retained, and escaped populations. We compared size structure (i.e., mean CL) between stocked, retained, and escaped crayfish separately for each combination of trap design and stocking density by using 1-way ANOVA with post hoc Tukey's test. We calculated sex ratio of the stocked, retained, and escaped populations for each combination of trap type and density and compared them qualitatively.

Results

Field sampling

During field sampling in Baird Creek, we collected 2,055 rusty crayfishes between restricted (N = 1,111) and unrestricted (N = 944) traps (Data S1, Supplemental Material). Two virile crayfishes Faxonius virilis were also captured, but they were excluded from analyses because they represented less than 0.001% of total catch. We captured crayfish at 38 of 45 sites—sites without crayfish were excluded from statistical analysis.

Habitat conditions in Baird Creek varied widely between sites for dominant substrate (i.e., most abundant substrate type at a site), but they were similar for habitat type (i.e., run, riffle pool), depth, temperature, and dissolved oxygen. Dominant substrates across all sites were bedrock (N = 8 sites), boulder (N = 9 sites), cobble (N = 4 sites), gravel (N = 5 sites), sand (N = 9 sites), silt (N = 5 sites), silt-sand (N = 1 site), sand-gravel (N = 1 site), sand-cobble (N = 1 site), clay-silt (N = 1 site), and detritus (N = 1 site). Of the 45 sites, we classified 39 sites as pools, 6 as runs; no sites were established in riffles because they were too shallow. Depths ranged from 0.21 to 0.92 m (0.38 ± 0.03; mean ± standard error [SE]). Temperature ranged from 16.6 to 20.8°C (19.9 ± 0.19; mean ± SE). Dissolved oxygen ranged from 7.2 to 11.1 mg/L (9.0 ± 0.1; mean ± SE).

Of the nine models fit to paired field capture data of restricted and unrestricted traps, the quadratic model was best supported (model weight [w] = 0.302; Table 1). Three additional models—linear plateau, logistic, and quadratic plateau—had ΔAIC scores within 2 of the best-supported model, quadratic. Because of the similarity among the four top-ranked models, we did not choose just one model to represent the relationship between catch for the two trap types (Figure 2). Instead, these four top-ranked models were treated collectively. Compared with the 1:1 line indicating identical catch rates between paired traps, the fits of all four top-ranked models (i.e., quadratic, linear plateau, logistic, and quadratic plateau) suggested a nonlinear relationship in catch between trap types. Catches between trap types were similar, if not higher, for unrestricted traps for all four top-ranked models, until catches approached approximately 40 crayfish/trap; beyond that threshold up to the maximum value observed (155 crayfish/trap), restricted traps outperformed their unrestricted counterparts. The primary difference between the fits of the top four models was for the highest catch rates for restricted traps. The data point for the highest catch in restricted traps caused the quadratic model to curve downwards. The other three models (i.e., linear plateau, logistic, and quadratic plateau) plateaued instead.

Trap design influenced size structure, but it did not influence sex ratio of captured populations. Unrestricted traps collected larger individuals than restricted traps (K-S test: D = 0.154, P < 0.001; Figure 3). Crayfish captured in unrestricted traps averaged 31.6 ± 0.14 mm (mean CL ± SE) compared with 30.0 ± 0.13 mm in restricted traps. Overall, males averaged 31.4 ± 0.12 mm and females averaged 28.9 ± 0.15 mm. Sex-related biases between traps were minor. In restricted traps, catch was 28.4% females and 71.6% males, whereas catch in unrestricted traps was nearly identical, with 27.4% females and 72.6% males.

Laboratory experiment

Significant differences in escapement rates were observed between trap designs (F1,54 = 25.5, P < 0.001; and stocking densities (F2,54 = 22.7, P < 0.001), and the interaction between these two factors was also significant (F2,54 = 13.0, P < 0.001; Data S2, Supplemental Material). Interpretation of the interaction plot for trap design × stocking density suggested that the interactive effect was ordinal. Therefore, we performed interpretation of main effects cautiously, with the understanding that the interaction (trap design × stocking density) was likely more important to understand than either main effect. Overall, escapement rates were higher in unrestricted traps and at higher densities, but the interaction between these two factors was the most noticeable difference detected by the post hoc Tukey's test (Figure 4). At low densities, trap design did not influence escapement; only one individual escaped from each trap type at low density (5% mean escapement; Table 2). At medium density, only 1 individual escaped from restricted traps (0.8%) compared with 15 for unrestricted traps (11.5%). The significance of the interaction between trap design and stocking density was most apparent at high density; 50 individuals escaped from restricted traps (8.8%) compared with 258 individuals from unrestricted traps (45.3%).

Across the 5-d experiment, escapement rates were similar for low- and medium-density treatments for both trap types. However, in high-density treatments there was a noticeable increase in escapement rates between experiment days 2 and 3 for both trap types (Figure 5). For restricted traps, the escapement rate from high-density treatments increased from less than 1% during days 1 and 2 up to 12–18% for the following 3 d. We observed a similar trend for unrestricted traps. Escapement rates from high-density treatments in unrestricted traps during days 1 and 2 were 29–36% and then increased to 51–55% for the following 3 d.

Trap design and stocking density influenced size structure of retained and escaped crayfish, but they had less influence on sex-related bias. Comparisons of size structure between stocked, retained, and escaped crayfish could only be performed for high density in restricted traps and medium and high density in unrestricted traps. Low sample sizes of escaped individuals at low and medium densities for restricted traps and low density for unrestricted traps prevented valid comparisons (Figure 6). For restricted traps at high density, mean CL was similar between stocked, retained, and escaped individuals (F2,25 = 3.085; P = 0.064). For unrestricted traps at medium density, mean CL differed between stocked, retained, and escaped crayfish (F2,25 = 4.396; P = 0.023). Mean CL of retained and escaped crayfish was similar to that of the stocked population, but mean CL of retained crayfish was 2.8 mm greater than that of escaped crayfish (P = 0.025). Differences were also found for high-density treatments in unrestricted traps (F2,27 = 13.46; P < 0.001). Mean CL of stocked and escaped crayfish was similar, but crayfish retained in the trap were 1.74 mm larger than the stocked population (P < 0.001) and 3.02 mm longer than the escaped population (P = 0.016). Sex-related biases in escapement were unnoticeable for restricted traps regardless of density and for unrestricted traps at low and medium density; however, at high densities, 51.4% of females escaped from unrestricted traps compared with 42.4% for males.

Mortality of stocked crayfish was low; only four individuals died in each trap type during the experiment—all in high-density treatments. Most of the individuals that died had molted in the trap overnight, making them highly vulnerable. For this reason, we did not consider mortality an important factor.

Discussion

Both the field and laboratory portions of this study suggested that capture and escapement of rusty crayfishes in MMTs was density dependent and influenced by throat design, especially at high densities. The influence of both factors was minimal at low and medium densities; thus, use of unrestricted traps may be acceptable at low-to-medium densities. As hypothesized, size structure biases were evident as density increased: smaller individuals disproportionately escaped from the trap, leaving behind larger individuals. Surprisingly, these size biases did not correlate with sex-related biases reported previously (Stuecheli 1991; Houghton et al. 1998; Larson and Olden 2016).

Previous investigations of crayfish sampling have compared different active sampling methods (Williams et al. 2014), different trapping methods (Ulikowski et al. 2017; Barnett and Adams 2018), baits (Somers and Stechey 1986), and biases related to behavior (Collins et al. 1983; Ogle and Kret 2008). The present study builds most directly on the experimental work of Ogle and Kret (2008), who demonstrated how captured crayfish, especially larger individuals, may exclude additional crayfish from entering the trap. Ogle and Kret (2008) used an experimental tub with a stocking density of 30 crayfish per trial (approximately medium-to-high density in the present study) and compared crayfish capture under two scenarios. This study concluded that when rusty crayfish were consistently removed from the trap (via the trap door), the total number of rusty crayfish captured in a fixed time was higher. However, the traps used by Ogle and Kret (2008) lacked throat restrictions. Taken together, the findings of Ogle and Kret (2008) and the present study suggest that not only are crayfish outside the trap deterred from entering the trap but also many crayfish already in the trap tend to escape as density in the trap increases. For conventional MMTs, such as those used in most studies and monitoring programs, catch rate peaked at approximately half the rate of traps with throat restrictions. This suggests the saturation rate for unrestricted traps in Baird Creek was approximately 75 rusty crayfish/trap night, but adding a throat restriction doubled the saturation rate. This also means that MMTs lacking throat restrictions likely severely underestimate relative abundance in high-density populations.

Escapement of crustaceans from baited traps has been most intensively studied in marine systems for explicit fishery management purposes (Nulk 1978; Miller 1980, 1990; Idhe et al. 2006). In many instances, researchers identified similar size and sex biases found in crayfish trapping. For example, targeting of large, predominantly male individuals and exclusion of smaller individuals by larger individuals in the trap (Idhe et al. 2006). In commercial fisheries, traps are often designed specifically to allow escapement of small, sublegal individuals. Maximizing and retaining catch of target species and size classes usually involve proper orientation of the trap, having an appropriate number of openings of the right size, and using devices around trap openings that make it difficult for captured individuals to escape (Nulk 1978; Miller 1980, 1990). These same principles have been adapted to crayfish capture with varying degrees of success, but most researchers still use basic MMTs with a widened throat (Larson and Olden 2016). The present study sought to improve on this model by applying design lessons from marine crustacean traps, specifically, making it more difficult for captured individuals to escape, including smaller individuals.

Differences in catch rates and escapement identified during this study were relatively novel, but the inferences that can be drawn come with limitations. Baird Creek, the only system sampled during this study, has a broad range of habitat supporting crayfish catches of 0–155 crayfish/trap. This range is likely broader than that of most other systems because the sections of Baird Creek sampled included long sections (hundreds of meters) of cobble and broken bedrock streambed that is shallow enough to exclude nearly all fish predators, except during floods. It was in these sections that we observed the highest catch rates. Collins et al. (1983) found that crayfish movement, and therefore catch, were strongly influenced by the presence of fish predators including Rock Bass Ambloplites rupestris and Smallmouth Bass Micropterus dolomieu, but both potential predators were likely absent from the middle and upper reaches of Baird Creek where sampling occurred. In the best crayfish habitat available, many crayfish were observed entering traps in daylight hours within 1 min of deployment. Adding additional study systems would have been ideal but was not logistically feasible because of time and resource constraints. In addition, the results apply most directly to rusty crayfish. It is unclear whether other species would behave similarly, especially for high densities where the largest differences were observed. Sampling in multiple systems would have likely shifted the ranges for low, medium, and high densities used in lab trials lower than they were for Baird Creek.

The observation of higher field catches in unrestricted traps at low-to-medium density was unexpected. Because the two types of traps were set in tandem, a crayfish had a choice of which trap to enter. Unrestricted traps had a more obvious opening, likely allowing easier entry. At low and medium densities, crayfish in the trap usually did not deplete the bait, regardless of trap type. At high densities though, bait in both traps was regularly depleted (i.e., Gizzard Shad were stripped to the bone). In these cases, it is likely that many crayfish entered unrestricted traps, depleted the bait, and then escaped the trap while crayfish continued to accumulate in the adjacent restricted trap, which likely had bait available for a longer duration. This hypothesis of high escapement was well supported by subsequent escapement trials in the laboratory.

One unexpected result from the escapement trials was the increase in escapement rate between days 2 and 3 for high-density treatments of both trap types. Approximately half of the crayfish in the holding tank were used each day for escapement trials. By day 3, nearly all crayfish would have been used at least once, perhaps twice already, most of them in high-density treatments because those required the largest sample sizes. Escapement trial results indicated that rusty crayfish exposed to high-density escapement trials were more likely to escape, up to a point. Escapement rate did not keep increasing after day 3 for either gear type. Because crayfish were not individually marked, it is unclear whether a small portion of the population became more adept at escapement or whether a random portion of the population had improved ability to escape during each trial period from day 3 onwards. The implication of this bias is that escapement rates in high-density treatments were slightly elevated. Using new crayfish collected via a method other than trapping could perhaps have eliminated the increase in escapement rates across trial periods. However, this would have required collecting, housing, and caring for at least 1,400 crayfish—far beyond the capacity of our system to handle, and even less reasonable if required to collect them without traps. Despite these limitations, we obtained valuable insight regarding trap efficiency that can be built on by future studies.

The trade-off for using conventional MMTs is high escapement rates, a problem that has received limited attention. The most immediate implications of these findings apply to situations in which managers are trying to maximize catch, especially removal programs in systems where rusty crayfish (Hein et al. 2006, 2007) or other invasive species (Mouser et al. 2019) are a problem. In these situations, limiting escapement from MMTs would improve efficiency of removals by reducing the amount of effort required to control the species, saving time and money. Estimates of relative abundance are also likely influenced by presence or absence of throat restrictions. Catches from conventional MMTs considered to be high are likely underestimating relative abundance because the traps do not efficiently retain crayfish. However, if the objective is detection, then inclusion of a throat restriction would likely make no difference, because across several sites a series of baited MMTs would at least detect the species, if present (Reid 2015). Beyond the restricted vs. unrestricted comparison, this project provides some of the first estimates of crayfish escapement rates from MMTs, the most widely used crayfish sampling gear (Larson and Olden 2016). This study also provided the first experimental evidence that crayfish already in the trap, particularly small individuals, increasingly try to escape as density in the trap increases. Managers using MMTs for routine monitoring or removal projects can use this study to understand potential biases imposed by throat configuration and decide whether including a throat restriction would be beneficial.

Supplemental Material

Please note: The Journal of Fish and Wildlife Management is not responsible for the content or functionality of any supplemental material. Queries should be directed to the corresponding author for the article.

Data S1. Restricted and unrestricted modified minnow traps were set in tandem at 45 sites in Baird Creek (Brown County, Wisconsin) during 2018, but rusty crayfish Faxonius rusticus were collected at only 38 sites. Sites that lacked rusty crayfish were excluded from further analysis and are not reported here. We measured carapace length (CL) and sexed (male [M]/female [F]) every rusty crayfish captured. We collected habitat data for each sample site and included depth (cm), temperature (°C), dissolved oxygen (DO; mg/L), and substrate.

Found at DOI: https://doi.org/10.3996/032019-JFWM-015.S1 (202 KB XLSX).

Data S2. The laboratory portion of the project required stocking rusty crayfish Faxonius rusticus of known lengths and sex at specific densities (low, medium, high) into restricted or unrestricted traps overnight and then recording how many rusty crayfish were retained in the trap vs. escaped overnight. We conducted 10 trials for each combination of trap type and density (60 trials), and assigned each trial unique ID. We also recorded starting and ending temperatures (°C) and dissolved oxygen (DO; mg/L). We performed the laboratory portion of this project at the Green Bay Fish and Wildlife Conservation Office (New Franken, Wisconsin) during June 2018.

Found at DOI: https://doi.org/10.3996/032019-JFWM-015.S2 (198 KB XLSX).

Reference S1. Simonson TD, Lyons J, Kanehl PD. 1993. Guidelines for evaluating fish habitat in Wisconsin streams. General Technical Report NC-164. St. Paul, Minnesota: U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station.

Found at DOI: https://doi.org/10.3996/032019-JFWM-015.S3 (6.34 MB PDF).

Acknowledgments

This project would not have been possible without the help of numerous staff from the Aquatic Invasive Species Program at the Green Bay Fish and Wildlife Conservation Office including Tyler Harris, Brandon Harris, Joel Wils, Matt Petasek, Dalton Hendricks, Cari-Ann Hayer, and Anthony Reith. Stefan Tucker (University of Wisconsin–Green Bay) helped perform fieldwork and provided equipment and support for the laboratory portion of the project. The Wisconsin Department of Natural Resources provided permits for the transport of rusty crayfish (NR-40 permit) used for the experiment. The Baird Creek Preservation Foundation graciously provided access to Baird Creek through its system of trails. This article was greatly improved by the reviews of the Associate Editor and two reviewers.

Any use of trade, product, website, or firm names in this publication is for descriptive purposes only and does not imply endorsement by the U.S. Government.

References

References
Arnhold
E.
2017
.
easynls: easy nonlinear model. R package version 5.0
.
Barnett
ZC,
Adams
SB.
2018
.
Comparison of two crayfish trapping methods in coastal plain seasonal wetlands
.
North American Journal of Fisheries Management
38
:
911
921
.
Barton
K.
2018
.
MuMIn: multi-model inference. R package version 1.42.1
.
Available: https://CRAN.R-project.org/package=MuMIn (January 2020).
Burnham
KP,
Anderson
DR.
2003
.
Model selection and multimodel inference: a practical information-theoretic approach. 2nd edition
.
New York
:
Springer Science and Business Media
.
Burrell
MA,
inventor; Patent holder, assignee. October 10,
2017
.
Crayfish trap. United States Patent US 9,781,910B.
Collins
NC,
Harvey
HH,
Tierney
AJ,
Dunham
DW.
1983
.
Influence of predator fish density on trapability of crayfish in Ontario lakes
.
Canadian Journal of Fisheries and Aquatic Sciences
40
:
1820
1828
.
Durland Donahou
A,
Conard
W,
Dettloff
K,
Fusaro
A,
Sturtevant
R.
2018
.
Faxonius rusticus (Girard, 1852): U.S. Geological Survey, Nonindigenous Aquatic Species Database
.
U.S. Geological Survey
.
Hein
CL,
Roth
BM,
Ives
AR,
Vander Zanden
MJ.
2006
.
Fish predation and trapping for rusty crayfish (Orconectes rusticus) control: a whole-lake experiment
.
Canadian Journal of Fisheries and Aquatic Sciences
63
:
383
393
.
Hein
CL,
Vander Zanden
JM,
Magnuson
JJ.
2007
.
Intensive trapping and increased fish predation cause massive population decline of an invasive crayfish
.
Freshwater Biology
52
:
1134
1146
.
Houghton
DC,
Dimick
JJ,
Frie
RV.
1998
.
Probable displacement of riffle-dwelling invertebrates by the introduced rusty crayfish, Orconectes Rusticus (Decapoda: Cambaridae) in a north-central Wisconsin stream
.
Great Lakes Entomologist
31
:
13
24
.
Ihde
TF,
Frusher
SD,
Hoenig
JM.
2006
.
Do large rock lobsters inhibit smaller ones from entering traps?
Marine and Freshwater Research
57
:
665
674
.
Larson
ER,
Olden
JD.
2016
.
Field sampling techniques for crayfish
.
Pages
287
323
in
Longshaw
M
and
Stebbing
P,
editors.
Biology and ecology of crayfish
.
Boca Raton, Florida
:
CRC Press
.
Lodge
DM,
Kratz
TK,
Capelli
GM.
1986
.
Long-term dynamics of three crayfish species in Trout Lake, Wisconsin
.
Canadian Journal of Fisheries and Aquatic Sciences
43
:
993
998
.
Miller
RJ.
1980
.
Design criteria for crab traps
.
ICES Journal of Marine Science
39
:
140
147
.
Miller
RJ.
1990
.
Effectiveness of crab and lobster traps
.
Canadian Journal of Fisheries and Aquatic Sciences
47
:
1228
1251
.
Mouser
JB,
Ashley
DC,
Aley
T,
Brewer
SK.
2019
.
Subterranean invasion by gapped ringed crayfish: a removal effort and barrier installation
.
Diversity
11
:
3
.
Nulk
VE.
1978
.
The effects of different escape vents on the selectivity of lobster traps
.
Marine Fisheries Review
40
:
50
58
.
Ogle
DH,
Kret
L.
2008
.
Experimental evidence that captured rusty crayfish (Orconectes rusticus) exclude uncaptured rusty crayfish from entering traps
.
Journal of Freshwater Ecology
23
:
123
129
.
Olden
JD,
McCarthy
JM,
Maxted
JT,
Fetzer
WW,
Vander Zanden
MJ.
2006
.
The rapid spread of rusty crayfish (Orconectes rusticus) with observations on native crayfish declines in Wisconsin (USA) over the past 130 years
.
Biological Invasions
8
:
1621
1628
.
Porath
MT,
Pape
LD,
Richters
LK,
Pope
KL,
Pegg
MA.
2011
.
Influence of throat configuration and fish density on escapement of channel catfish from hoop nets
.
American Fisheries Society Symposium
77
:
563
571
.
R Core Team
.
2017
.
R: a language and environment for statistical computing
.
R Foundation for Statistical Computing
,
Vienna, Austria
.
Available: https://www.R-project.org/ (December 2019).
Reid
SM.
2015
.
Optimizing sampling effort to detect rusty crayfish (Orconectes rusticus) in southern Ontario rivers
.
Management of biological invasions
6
:
303
306
.
Simonson
TD,
Lyons
J,
Kanehl
PD.
1993
.
Guidelines for evaluating fish habitat in Wisconsin streams. General Technical Report NC-164
.
St. Paul, Minnesota
:
U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station (see Supplemental Material, Reference S1)
.
Smith
BJ,
Blackwell
BG,
Wuellner
MR,
Graeb
BD,
Willis
DW.
2016
.
Escapement of fishes from modified fyke nets with differing throat configurations
.
North American Journal of Fisheries Management
36
:
96
103
.
Somers
KM,
Stechey
DP.
1986
.
Variable trappability of crayfish associated with bait type, water temperature and lunar phase
.
American Midland Naturalist
116
:
36
44
.
Stuecheli
K.
1999
.
Trapping bias in sampling crayfish with baited funnel traps
.
North American Journal of Fisheries Management
11
:
236
239
.
Ulikowski
D,
Chybowski
Ł,
Traczuk
P,
Ulikowska
E.
2017
.
A new design of crayfish traps reduces escaping and improves opportunities for long-term catching
.
Turkish Journal of Fisheries and Aquatic Sciences
17
:
363
369
.
Williams
K,
Brewer
SK,
Ellersieck
MR.
2014
.
A comparison of two gears for quantifying abundance of lotic-dwelling crayfish
.
Journal of Crustacean Biology
34
:
54
60
.
Wilson
KA,
Magnuson
JJ,
Lodge
DM,
Hill
AM,
Kratz
TK,
Perry
WL,
Willis
TV.
2004
.
A long-term rusty crayfish (Orconectes rusticus) invasion: dispersal patterns and community change in a north temperate lake
.
Canadian Journal of Fisheries and Aquatic Sciences
61
:
2255
2266
.

Author notes

Citation: Smith BJ. 2020. Density-dependent escapement of rusty crayfish from modified minnow traps with varying throat configurations. Journal of Fish and Wildlife Management 11(1):22–32; e1944-687X. https://doi.org/10.3996/032019-JFWM-015

Competing Interests

The findings and conclusions in this article are those of the author(s) and do not necessarily represent the views of the U.S. Fish and Wildlife Service.

Supplementary data