Monofilament gill nets have increased in popularity in recent decades and are often considered to be superior to multifilament gill nets; however, this claim is still contested by many researchers. Target species, habitat characteristics, and mesh characteristics can all affect the efficacy and selectivity of these nets. Increased angling interest and declining home ranges have led to increased management efforts for gars (Lepisosteidae) among state and federal agencies. However, the sampling of gar species is notoriously difficult and has hindered subsequent management actions in some cases. This study aimed to compare probability and incidence rate of gar capture between net constructions (multi- vs. monofilament), net length (75 vs. 46 m length), sampling habitat (run vs. bend), and deployment orientation (parallel vs. perpendicular vs. diagonal in relation to river channel). We also assessed mode of capture investigate the effects of small mesh sizes (50.8 mm) on capture and potential retention of gars captured in gill nets. Results showed that the use of multifilament gill nets were three times more likely to capture at least one gar (1.3–6.3, 95% CI) than monofilament nets. Multifilament gill nets also averaged approximately twice as many gars per net than monofilament gill nets. Increasing water temperature also significantly increased the probability of gar capture, particularly above 15°C. In addition to mesh type and water temperature, habitat and gill net orientation also had significant effects on the incidence rate of gar capture. We primarily captured gars captured in the gill nets by entanglement (63%), which may have allowed for the wide range of length frequencies sampled in this study. Our study highlights the importance of considering multiple gill net constructions, deployments, and habitats when designing a research project or management plan for gar species. The use of multifilament gill nets can help resource managers to sample gar populations more effectively, resulting in better management strategies and practices.

Fish management requires reliable sampling techniques to assess populations. Sampling gear can be active, such as electrofishing and trawls, or passive, such as gill nets and trap nets. Active gear such as electrofishing has increased in popularity due to its ability to shorten time spent in the field, relatively small personnel requirements (McClelland et al. 2013), and adaptability for large and small habitats (Cooke et al. 1998). However, in many situations researchers and others still rely on passive capture methods to assess fish populations (Njoku 1991; Argent and Kimmel 2005; Clavero et al. 2006). Gill nets are a popular passive capture technique for researchers, managers (Trent and Hassler 1968, Santos et al. 2003), and commercial fishermen (Collins 1979, Reis and Pawson 1992, Barlow and Cameron 2003). Gill nets can be deployed in a range of habitats, fished at any depth, and usually hauled without any special equipment, making them extremely useful in many different scenarios (Hamley 1975; Munprasit et al. 1986; Njoku 1991; Hovgård 1996; Faife and Einarsson 2003; Olin and Malinen 2003; Albert and Einarsson 2004; Argent and Kimmel 2005; Buckmeier et al. 2013; Binion et al. 2015). Relatively low cost and minimal maintenance also make them popular in situations with limited budgets and personnel.

Gill nets consist of a wall of mesh, constructed using linen, nylon, or other materials and can be purchased with monofilament or multifilament strands. Multifilament nets usually consist of twisted linen threads, as opposed to the more rigid nylon strands often used to create the single-stranded monofilament mesh. The mesh is secured to a float line at the top, to provide buoyancy, and a weighted line at the bottom, to help the mesh remain vertical in the water column. Fishers can allow the net to float freely or anchor it to the shore and/or bed at one or both ends (see Hamley 1975 for more details). Several studies (Khan et al. 1975; Collins 1979; Henderson and Nepszy 1992; Faife and Einarsson 2003; Ayaz et al. 2006) have claimed that monofilament gill nets are more efficient than those constructed using multifilament mesh (e.g., Collins 1979; Faife and Einarsson 2003); however, some researchers still dispute this topic based on the results of other studies (e.g. Njoku 1991; Balik 1999). Steinberg (1963; cited in Khan et al. 1975) cited the lower visibility of monofilament gill nets as a major advantage over multifilament mesh, but waters with increased turbidity may reduce the need for this attribute. Njoku (1991) suggested that multifilament mesh entangled fish more easily than monofilament mesh due to the many individual strands in the twine. Many researchers have postulated that wedging and gilling of fishes in gill nets are the most important modes of capture, and that entanglement can be largely ignored (Hamley 1975). However, other studies have found entanglement to be an important mode of capture for certain species such as Rainbow Smelt Osmerus mordax (Collins 1979). Gar possess many anatomical features that may lend them to increased entanglement in gill nets, and entanglement is suggested to be less size-selective than other modes of capture (Hamley 1975; Schlechte et al. 2016). Multifilament gill nets may also translate into higher capture rates due to increased strength (Stewart 1987) and a lack of memory. The construction of multifilament mesh from multiple strands of twine may possess superior tensile strength, compared to monofilament mesh, reducing the potential for larger fish to break through the net. Net memory refers to a net's tendency to retain its shape. Multifilament gill nets tend to lack memory, which means they remain limp and flexible after disturbance in the water or storage. This limpness may increase the likelihood of fish becoming entangled as the net flexes in the water column. Conflicting reports from previous studies highlight the importance of further research on capture rates of gill nets with different constructions, particularly in species that lack a traditional fusiform body.

Gars (Family Lepisosteidae) are notoriously difficult to sample using active capture methods, such as electrofishing, due to their tendency to sink when immobilized. This family of fishes is denser than the surrounding water, which means that they will sink when stunned by electrofishing equipment (Burr 1931), rather than float, as most fish species do. Thus, researchers often employ passive capture methods to sample and monitor gar populations (Ferrara and Irwin 2001; Brinkman 2008; Clay 2009; Buckmeier et al. 2013; Binion et al. 2015; Daugherty et al. 2017). Reintroduction efforts for Alligator Gar Atractosteus spatula and interest in ecosystem services and predator effects, as well as growing angler and bow-fishing interest, has raised important management concerns for all four species of gar and created a surge of research into population monitoring methods in recent decades (Scarnecchia 1992; Ross and Brenneman 2001; Binion et al. 2015; Schlechte et al. 2016). For example, Schlechte et al. (2016) suggested a multifilament experimental gill net design for the capture of Alligator Gar that could provide a more representative sample of the resident population. However, it is important to further investigate aspects of gill net design to fill gaps in the literature. For instance, investigating smaller mesh sizes than those used by Binion et al. (2015) or sampling a wider community of gar species to supplement the findings of Argent and Kimmel (2005) for Longnose Gar Lepisosteus osseus can help to inform researchers and managers about more targeted sampling protocols.

There are other potential aspects that are still essentially missing from the literature, including orientation of net deployment. Hamley (1975) explains, at length, a variety of ways that gill nets can be selective, including the method of fishing. However, this discussion does not address the issue of net orientation. The lack of literature on this topic may be, in part, due to the fact that historically gill nets were favored in marine and lacustrine systems (Argent and Kimmel 2005). But is it is important to investigate the effects of net orientation as this may affect encounter probability, visibility and detectability, and/or “leading” as described by Hamley (1975) for gang sets in which smaller mesh may “lead” fish along the net until they encounter larger net sizes and become captured.

The aim of this paper was to compare the capture rates of two gill net mesh constructions (monofilament and multifilament), two net lengths (75 and 46 m), and three deployment orientations (perpendicular, parallel, and oblique to the stream channel). We also included the mode of capture (gilled, tangled, or wedged) for each gar for analysis as this has been an important topic in previous discussions about gill net use (Hamley 1975; Faife and Einarsson 2003; Binion et al. 2015; Schlechte et al. 2016). Based on previous studies, we expected monofilament gill nets to display a higher capture rate than multifilament nets; however, we expected multifilament nets to show higher capture rates due to the individual net strands and the many protruding structures found on gar. We also hypothesized that longer nets would capture more gar than shorter nets but that capture probabilities would be similar between the two. We hypothesized that perpendicular deployments would be the most efficient orientation for capture of gar species. Finally, we hypothesized that gars would be captured via entanglement significantly more often, compared to gilling and wedging, than would be expected if captured mode occurred at random (equal probabilities of each capture mode).

Study site

We conducted this study on the Clarks River, a fifth-order stream located in western Kentucky (Figure 1). The main-stem Clarks River begins approximately 21 km south of the mouth, at the confluence of the east and west forks. The river then flows northward to mouth, located near Paducah, Kentucky (37°02′47″N, 88°32′36″W; Figure 1) where it flows into the Tennessee River. The Clarks River watershed is dominated by agriculture, forest, and wetlands; relatively little developed area is located within the watershed (Murray State University, Hancock Biological Station and Center for Reservoir Research 2011). A small portion in the upper reaches of the Clarks River is situated within the Clarks River National Wildlife Refuge and is managed by the U.S. Fish and Wildlife Service. The river displays a low gradient, slow flow velocity, and complex substrates consisting of clay, silt, and sand. At base flow conditions, water depth of the Clarks River ranges from approximately 5.5 m at the mouth and within deeper pools/bends to less than 1 m in certain runs. The Ohio and Tennessee rivers greatly influence hydrology, with much of the influence stemming from water-level management practices on Kentucky Lake and the Lower Tennessee River. Seasonal flooding of the surrounding areas is common throughout the lower reaches of the Clarks River, particularly in late winter and early spring.

Figure 1

Map of the Clarks River in Kentucky. Points denote all sampling locations (N = 148 nets) throughout the study period (September 2013–October 2014) and are designated by net length and mesh type. Open triangles = monofilament gill nets; filled dots = multifilament gill nets; red points = long (68.6-m) gill nets; black points = short (34.8-m) gill nets.

Figure 1

Map of the Clarks River in Kentucky. Points denote all sampling locations (N = 148 nets) throughout the study period (September 2013–October 2014) and are designated by net length and mesh type. Open triangles = monofilament gill nets; filled dots = multifilament gill nets; red points = long (68.6-m) gill nets; black points = short (34.8-m) gill nets.

Close modal

Study design

Nets were similar in basic design: 50.8-mm bar mesh, 2.4 m deep, 12.7-mm foamcore float line, and 13.6-kg leadcore lead line. The differences in nets included mesh material (multi- vs. monofilament mesh) and net length (75 vs. 46 m, referred to as “long” and “short,” respectively, hereafter). Multifilament nets were constructed with twisted #139 white twine. Monofilament nets were constructed using clear #8 monofilament webbing. We chose the small mesh size based on findings from Argent and Kimmel (2005), who showed that Longnose Gar were only captured in gill nets with graded mesh panels or gill nets with only smaller mesh sizes (< 102-mm bar). We collected the data in this paper as part of a diet study on all four resident gar species, that did not examine for gill net selectivity. Thus, we have avoided any attempt to calculate selectivity using the direct or indirect approaches due to concerns of violating assumptions of the methods.

The project began with an unequal number of net combinations (two long, multifilament; two short, multifilament; two long, monofilament; one short, monofilament). This limitation, combined with repairs required for net damage from coarse woody debris, limited the number of sets available for each net type, which led to an unbalanced design. Characteristics of the Clarks River also negated our ability to set nets in several of the bends, due to deadheads, coarse woody debris, or other structures. The length of the long nets also did not allow us to deploy them perpendicular to the channel without placing excess fishable net on the shore. This meant that a portion of these nets would not be actively fishing and would thereby negate the comparison of a “long” net to a shorter net construction. We determined netting locations by a combination of random site selection and evidence of gar activity (surface activity, visual presence below the surface, etc.; Figure 1).

Netting occurred during all months from September 2013 to October 2014 (Tables S1 and S2, Supplemental Material), except December 2013 and February 2014 as the river was inaccessible during these months due to high water, which flooded the access point to the river, or ice covering the boat ramp. Net sets in this paper ranged between 0.48 and 25.13 h. We deployed nets in the early morning and pulled them the following morning. We designated the mode of capture for each gar as “gilled,” “tangled,” or “wedged,” based on descriptions by Hamley (1975). We identified and enumerated captured fish. We released all non-gar species immediately, and retained lepisosteids for use in a separate study.

Statistical analyses

We analyzed gar capture probability using logistic regression. We coded individual nets as a “success” if we captured one or more gar, or a “failure” if no gar were captured. We constructed generalized linear models (GLMs; stats::glm) with a binomial (link = logit) distribution using sampling season (Season), net type (Type), net length (Size), deployment orientation (Orientation), habitat type (Habitat), and water temperature (Temp.Mouth) as predictor variables. We originally included Soak.Time as a factor variable representing short net sets (< 11 h) and long net sets (≥ 11 h). This distinction fit the bimodal distribution of our sampling times (Figure 2). However, this distribution was confounded by the fact that it also fit the diurnal nature of sampling. Short sets (< 11 h) were all daylight sampling events, whereas long net sets were overnight sets. Due to our inability to distinguish between the effects of soak time and diurnal fish activity, we excluded Soak.Time from all analyses. Water temperature was the temperature at the mouth of the Clarks River immediately before net retrieval began. We dropped one predictor variable from each nested model and compared to the full model using the dropterm function (MASS package) with a χ2 test. We analyzed goodness-of-fit for the final, reduced model using a Hosmer–Lemeshow Test.

Figure 2

Histogram by net soak-time frequencies in the Clarks River, Kentucky, between September 2013 and October 2014. Two distinct netting patterns are apparent with short sets (< 11 h) being daylight deployments and long sets (≥ 11 h) being overnight deployments. Because we could not separate the effects of soak time and diurnal fish activity, based on these data, we excluded soak time from all analyses.

Figure 2

Histogram by net soak-time frequencies in the Clarks River, Kentucky, between September 2013 and October 2014. Two distinct netting patterns are apparent with short sets (< 11 h) being daylight deployments and long sets (≥ 11 h) being overnight deployments. Because we could not separate the effects of soak time and diurnal fish activity, based on these data, we excluded soak time from all analyses.

Close modal

We used untransformed count data were used to construct a GLM with a negative binomial distribution (MASS::glm.nb) for total gar captures and evaluated against the full suite of predictor variables as the previous analysis. The data in this analysis were conditional on the success of the net to capture at least one gar; thus, only those nets designated as successful in the previous analysis were included. We chose the negative binomial distribution for two reasons: 1) this distribution is useful for overdispersed data and generalizes to a Poisson distribution without overdispersion, and 2) it is better equipped to compare models with different amounts of data points from missing values, as we had in this dataset. We performed a nonparametric dispersion analysis (DHARMa::testDispersion) that compares simulated and observed data to determine if the Poisson model was overdispersed. We used a χ2 test (stats::pchisq) to compare the model fits of between the Poisson and negative binomial distributions. We conducted model comparisons for significant predictors in the same fashion as the previous analysis. We determined significant predictor variables using the dropterm function (MASS package) and compared each nested model to the full model using a χ2 test. We retained all significant predictors and implemented them in the final reduced model. We also calculated catch per unit effort (CPUE) for the suite of predictor variables and reported it as fish per hour.

We analyzed the mode of capture using χ2 tests. We compared observed counts of each capture mode (Table S2, Supplemental Material) to the statistically expected values produced by the assumption that mode of capture occurs at random. We performed all analyses using the statistical software R, version 3.4.1 “Single Candle” (R Core Team 2017). We used significance of α = 0.05 for all analyses.

We deployed a total of 148 nets throughout the duration of this study (Table S1, Supplemental Material). We removed 12 nets with extremely short soak times (< 2 h) that we pulled shortly after deployment as weather reports indicated unsafe boating conditions and inclement weather. This left 136 nets for analysis. We deployed multifilament nets (n = 80) more often than monofilament nets (n = 50). We set long nets (n = 80) nearly twice as often as short nets (n = 43), and deployed nets in runs (n = 102) more often than in bends (n = 29). Net orientation was dictated by net length and river characteristics resulting in nets being deployed parallel to the river channel most often (n = 67), followed by perpendicular sets (n = 29) and, finally, diagonal sets (n = 11).

Length-frequency histograms by species (Figure 3) show that Longnose Gar display a bimodal distribution of sizes with peaks at 800–1,000-mm and 1,200–1,300-mm size classes (Figure 3B). Shortnose Gar Lepisosteus platostomus showed a nearly normal length distribution with a peak at the 700–750-mm size class (Figure 3C), and Spotted Gar Lepisosteus oculatus showed a right-skewed distribution with a peak at the 600–650-mm size class; however, these data, particularly for Spotted Gar (Figure 3D), should be interpreted conservatively due to the small sample size. We show length-frequency histograms for monofilament and multifilament gill nets in Figures 4A and 4B, respectively. The histograms show that the length distributions of the two mesh types agree fairly well, though multifilament gill nets captured more gar of nearly all size classes and show a possible bimodal distribution with a noticeable peak at the 750–800-mm size class.

Figure 3

Length-frequency distributions for gar captures in the Clarks River, Kentucky, between September 2013 and October 2014. Plots show distributions for (A) all gar captures (N = 190), (B) Longnose Gar Lepisosteus osseus (n = 127), (C) Shortnose Gar Lepisosteus platostomus (n = 46), and (D) Spotted Gar Lepisosteus oculatus (n = 14). (A) Longnose Gar show a bimodal distribution of lengths suggesting that either 1) median size classes were not as vulnerable to our fishing gear, or 2) this size class may be present in lower proportions within the actual population. (B) Shortnose Gar show a nearly normal distribution of vulnerable size classes. (C) Spotted Gar show a right-skewed length frequency; however, interpretation of this data requires caution with such a small sample size.

Figure 3

Length-frequency distributions for gar captures in the Clarks River, Kentucky, between September 2013 and October 2014. Plots show distributions for (A) all gar captures (N = 190), (B) Longnose Gar Lepisosteus osseus (n = 127), (C) Shortnose Gar Lepisosteus platostomus (n = 46), and (D) Spotted Gar Lepisosteus oculatus (n = 14). (A) Longnose Gar show a bimodal distribution of lengths suggesting that either 1) median size classes were not as vulnerable to our fishing gear, or 2) this size class may be present in lower proportions within the actual population. (B) Shortnose Gar show a nearly normal distribution of vulnerable size classes. (C) Spotted Gar show a right-skewed length frequency; however, interpretation of this data requires caution with such a small sample size.

Close modal
Figure 4

Length-frequency distributions for gar captured in (A) monofilament and (B) multifilament gill nets in the Clarks River, Kentucky, between September 2013 and October 2014. Both gill nets show similar vulnerable size ranges, capturing gars with total lengths ranging from 550 to 1,400 mm. Multifilament gill nets also captured gar larger than 1,400 mm total length. Captures were fairly consistent across the bulk of the size range for monofilament gill nets; however, multifilament gill nets showed a small peak at the 750–800-mm size range, which may suggest increased vulnerability of this size class to this mesh size (50.8 mm) and mesh type.

Figure 4

Length-frequency distributions for gar captured in (A) monofilament and (B) multifilament gill nets in the Clarks River, Kentucky, between September 2013 and October 2014. Both gill nets show similar vulnerable size ranges, capturing gars with total lengths ranging from 550 to 1,400 mm. Multifilament gill nets also captured gar larger than 1,400 mm total length. Captures were fairly consistent across the bulk of the size range for monofilament gill nets; however, multifilament gill nets showed a small peak at the 750–800-mm size range, which may suggest increased vulnerability of this size class to this mesh size (50.8 mm) and mesh type.

Close modal

Gar capture probability

Approximately half (51.5%) of all net sets captured at least one gar and were considered successful for the purposes of the logistic model. The logistic model included capture success (1 = success, 0 = failure) as the response variable and season (Season), mesh type (Type), net size (Size), habitat type (Habitat), deployment orientation (Orientation), and temperature at the mouth of the Clarks River (Temp.Mouth) as predictor variables. No interactions were significant (all P > 0.05), so the model was restricted to main effects only. The χ2 comparisons showed that Type (P = 0.003) and Temp.Mouth (P = 0.039) were significant predictors of gar capture success (Table 1). Results of the Hosmer–Lemeshow test showed that the reduced model containing these four variables was a suitable fit for the data (χ2 = 8.88, df = 8, P = 0.35).

Table 1

Parameter estimates for logistic model of gar capture probability in the Clarks River, Kentucky, between September 2013 and October 2014. A net was considered a success if at least one gar was captured during that set. Type, Size, and Habitat are factors with two levels. Season and Orientation are factors with three levels. Temp.Mouth is the temperature (°C) at the mouth of the Clarks River at the time we retrieved the net. LR stat is the likelihood ratio statistic. Probability values in bold indicate statistical significance at α = 0.05.

Parameter estimates for logistic model of gar capture probability in the Clarks River, Kentucky, between September 2013 and October 2014. A net was considered a success if at least one gar was captured during that set. Type, Size, and Habitat are factors with two levels. Season and Orientation are factors with three levels. Temp.Mouth is the temperature (°C) at the mouth of the Clarks River at the time we retrieved the net. LR stat is the likelihood ratio statistic. Probability values in bold indicate statistical significance at α = 0.05.
Parameter estimates for logistic model of gar capture probability in the Clarks River, Kentucky, between September 2013 and October 2014. A net was considered a success if at least one gar was captured during that set. Type, Size, and Habitat are factors with two levels. Season and Orientation are factors with three levels. Temp.Mouth is the temperature (°C) at the mouth of the Clarks River at the time we retrieved the net. LR stat is the likelihood ratio statistic. Probability values in bold indicate statistical significance at α = 0.05.

Results show the odds of successfully capturing a gar in a multifilament gill net are 2.9 (1.3–6.3, 95% CI) times the odds of capturing a gar in a monofilament gill net. The odds of capturing a gar in a gill net increase by 1.2 (1.1–1.2) for each 1°C water temperature increase. Gar capture probabilities were always higher for multifilament gill nets than for monofilament gill nets, and both probabilities increased with water temperature (Figure 5).

Figure 5

Plot of gar capture probabilities based on water temperature and net type in the Clarks River, Kentucky, between September 2013 and October 2014. Probability of capturing gar in multifilament gill nets (green dashed line) is significantly higher than in monofilament gill nets (yellow solid line) at all temperatures greater than 15°C. Grey bands represent 95% confidence bands for each net type. Both variables were significant predictors of gar capture probability (P < 0.05).

Figure 5

Plot of gar capture probabilities based on water temperature and net type in the Clarks River, Kentucky, between September 2013 and October 2014. Probability of capturing gar in multifilament gill nets (green dashed line) is significantly higher than in monofilament gill nets (yellow solid line) at all temperatures greater than 15°C. Grey bands represent 95% confidence bands for each net type. Both variables were significant predictors of gar capture probability (P < 0.05).

Close modal

Capture efficiency

To assess gar-capture efficiency, we used only nets designated as a success in the previous analysis (n = 73). We captured 181 gars in the successful nets of this study and all four species that inhabit western Kentucky were represented. Longnose Gar was the most common species captured (n = 122), followed by Shortnose Gar (n = 42) and Spotted Gar (n = 14). Alligator Gar was represented by only three individuals and we excluded it from further analysis, leaving 178 gars for further analysis.

The null negative binomial GLM was created using season (Season), mesh type (Type), net size (Size), habitat type (Habitat), deployment orientation (Orientation), and temperature at the mouth of the Clarks River (Temp.Mouth) as predictor variables. There were no significant interactions (all P > 0.05), so we analyzed only main effects. The nonparametric dispersion analysis revealed that overdispersion was an issue in the Poisson model (Obs:Sim = 1.405, P < 0.001), and the negative binomial model was shown to be a significantly better fit for the data by a χ2 comparison of the negative log-likelihood analysis (P = 0.018). Results from pair-wise χ2 model comparisons showed that Orientation, Habitat, Type, and Temp.Mouth were all significant predictors for the number of gars captured per net (all P < 0.05; Table 2). The Hosmer–Lemeshow test showed no evidence of the reduced model being a bad fit for the data (P > 0.05).

Table 2

Parameter estimates for negative binomial model of gar capture incidence rate in the Clarks River, Kentucky, between September 2013 and October 2014. A net was considered a success if at least one gar was captured during that set. Type, Size, and Habitat are factors with two levels. Season and Orientation are factors with three levels. Temp.Mouth is the temperature (°C) at the mouth of the Clarks River at the time we retrieved the net. LR stat is the likelihood ratio statistic. Probability values in bold indicate statistical significance at α = 0.05.

Parameter estimates for negative binomial model of gar capture incidence rate in the Clarks River, Kentucky, between September 2013 and October 2014. A net was considered a success if at least one gar was captured during that set. Type, Size, and Habitat are factors with two levels. Season and Orientation are factors with three levels. Temp.Mouth is the temperature (°C) at the mouth of the Clarks River at the time we retrieved the net. LR stat is the likelihood ratio statistic. Probability values in bold indicate statistical significance at α = 0.05.
Parameter estimates for negative binomial model of gar capture incidence rate in the Clarks River, Kentucky, between September 2013 and October 2014. A net was considered a success if at least one gar was captured during that set. Type, Size, and Habitat are factors with two levels. Season and Orientation are factors with three levels. Temp.Mouth is the temperature (°C) at the mouth of the Clarks River at the time we retrieved the net. LR stat is the likelihood ratio statistic. Probability values in bold indicate statistical significance at α = 0.05.

The incidence rate of nets deployed parallel to the channel was 0.8 (0.4–1.7, 95% CI) that of nets deployed diagonally; that is nets deployed diagonally across the channel (“reference”; set to 1.0) captured approximately 20% more gar than those deployed parallel to the channel. Nets deployed perpendicular to the channel followed suit, with an incident rate 1.1 (0.5–2.2) that of diagonally deployed nets. So, to clarify, perpendicular nets have approximately equal rates of incidence of capture to diagonal deployments (“reference”; set to 1.0), but parallel nets have significantly lower incidence compared to both of the former orientations.

Deploying gill nets in run habitats produced an incident rate 1.7 (1.0–3.1) that of channel bends; thus, run habitats increase the mean number of captures by nearly 75%. Multifilament gill nets doubled the incidence rate (1.2–3.3; 95% CI) of monofilament gill nets, indicating that multifilament gill nets averaged approximately twice as many individuals as their monofilament counterparts. Lastly, increases in temperature at the mouth of the Clarks River (Temp.Mouth) showed an essentially equal incidence rate with increasing temperature. Figure 6 shows that there is a potential increasing trend in total number of captures per net with increasing temperature. The significance of this variable in the model was potentially driven by the extremely high number of captures in one gill net (17 individuals) and increased variation in total captures as temperature increased. However, consistently low numbers of captures (mean = 2.5 individuals) likely account for the minimal difference in the change of incidence with temperature. We show the CPUE for each predictor variable in Table 3.

Figure 6

Plot of total gar captures per net at specific water temperatures in the Clarks River, Kentucky, from September 2013 to October 2014. Each point represents a single net. Captures are separated by mesh type (triangles = monofilament, circles = multifilament) and net length (yellow = short nets [46 m], green = long nets [75 m]). There is a potential trend of increased captures with increased water temperature; however, this trend is nonsignificant with the given dataset (P > 0.05).

Figure 6

Plot of total gar captures per net at specific water temperatures in the Clarks River, Kentucky, from September 2013 to October 2014. Each point represents a single net. Captures are separated by mesh type (triangles = monofilament, circles = multifilament) and net length (yellow = short nets [46 m], green = long nets [75 m]). There is a potential trend of increased captures with increased water temperature; however, this trend is nonsignificant with the given dataset (P > 0.05).

Close modal
Table 3

Catch per unit effort (CPUE) of total gar captures for each factor predictor variable. Captures is the total combined captures of Longnose Gar Lepisosteus osseus, Shortnose Gar Lepisosteus platostomus, and Spotted Gar Lepisosteus oculatus from the Clarks River, Kentucky, between September 2013 and October 2014. CPUE reported as fish per hour. Data only include complete cases (103 of 148 nets). Total net soak time = 1208.7 h. Total captures = 150 individuals.

Catch per unit effort (CPUE) of total gar captures for each factor predictor variable. Captures is the total combined captures of Longnose Gar Lepisosteus osseus, Shortnose Gar Lepisosteus platostomus, and Spotted Gar Lepisosteus oculatus from the Clarks River, Kentucky, between September 2013 and October 2014. CPUE reported as fish per hour. Data only include complete cases (103 of 148 nets). Total net soak time = 1208.7 h. Total captures = 150 individuals.
Catch per unit effort (CPUE) of total gar captures for each factor predictor variable. Captures is the total combined captures of Longnose Gar Lepisosteus osseus, Shortnose Gar Lepisosteus platostomus, and Spotted Gar Lepisosteus oculatus from the Clarks River, Kentucky, between September 2013 and October 2014. CPUE reported as fish per hour. Data only include complete cases (103 of 148 nets). Total net soak time = 1208.7 h. Total captures = 150 individuals.

Capture mode

Of the 179 gars retained for analysis in this study, we captured 142 in nets that soaked for at least 2 h and our analysis contained information on mode of capture which was designated as “gilled,” “tangled,” or “wedged” as described by Hamley (1975). The 2-h soak time requirement caused us to exclude two fish that we captured in nets that soaked for less than this required time. For completeness, we note here that one of these fish was captured by entanglement and the other by gilling (rows 119 and 130, respectively, in Table S2, Supplemental Material). We captured 90 individuals via entanglement (63%); 31were wedged (22%) and the remaining 22 were gilled (15%). These frequencies deviated significantly from the expected uniform frequencies (χ2 = 58.63, df = 2, P < 0.001). Gar were nearly twice as likely to be captured by entanglement, whereas capture via gilling occurred less than half as often as would be expected (Table 4). There was also a significant difference in modes of capture between gill net mesh types (χ2 = 7.097, df = 2, P = 0.029). In monofilament gill nets, gilling and wedging occurred in equal frequencies; however, wedging occurred approximately 50% more often than gilling in multifilament gill nets. Capture via entanglement was higher than all other capture modes for all three gar species and both net types. Deployment orientation had no significant effect on mode of capture (χ2 = 3.285, df = 4, P = 0.511).

Table 4

Observed and expected frequencies for the three modes of capture. We collected Longnose Gar Lepisosteus osseus, Shortnose Gar Lepisosteus platostomus, and Spotted Gar Lepisosteus oculatus from the Clarks River, Kentucky between September 2013 and October 2014 using gill nets. We based expected frequencies on random capture mode (e.g., equal probabilities for each mode of capture). O/E is the observed frequency of occurrence for each capture mode divided by the expected frequency of occurrence. A χ2 analysis showed that the observed frequencies deviated significantly from the expected values (χ2 = 58.625, df = 2, P < 0.001).

Observed and expected frequencies for the three modes of capture. We collected Longnose Gar Lepisosteus osseus, Shortnose Gar Lepisosteus platostomus, and Spotted Gar Lepisosteus oculatus from the Clarks River, Kentucky between September 2013 and October 2014 using gill nets. We based expected frequencies on random capture mode (e.g., equal probabilities for each mode of capture). O/E is the observed frequency of occurrence for each capture mode divided by the expected frequency of occurrence. A χ2 analysis showed that the observed frequencies deviated significantly from the expected values (χ2 = 58.625, df = 2, P < 0.001).
Observed and expected frequencies for the three modes of capture. We collected Longnose Gar Lepisosteus osseus, Shortnose Gar Lepisosteus platostomus, and Spotted Gar Lepisosteus oculatus from the Clarks River, Kentucky between September 2013 and October 2014 using gill nets. We based expected frequencies on random capture mode (e.g., equal probabilities for each mode of capture). O/E is the observed frequency of occurrence for each capture mode divided by the expected frequency of occurrence. A χ2 analysis showed that the observed frequencies deviated significantly from the expected values (χ2 = 58.625, df = 2, P < 0.001).

The results of the present study suggest that untarred multifilament gill nets display a higher capture probability and incidence rate of gars in the Clarks River, located in western Kentucky, compared to monofilament gill nets. The odds of capturing a gar in a multifilament gill net were nearly triple the odds of gar capture in monofilament gill nets. Also, in successful net sets (i.e., nets that captured at least one individual) multifilament gill nets averaged approximately two times more gar than did monofilament gill nets. These results contrast with several previous studies that compared efficiencies of different gill net meshes on other species and found that monofilament gill nets were superior to their multifilament counterparts (Larkins 1964; Pristas and Trent 1977; Collins 1979; Henderson and Nepszy 1992; Faife and Einarsson 2003). However, most of these studies affirm that differences in capture efficiency and net superiority are species-specific.

Collins (1979) found monofilament gill nets to have superior capture efficiency for Lake Whitefish Coregonus clupeaformis in Lake Huron, the major target fishery in that system, but that multifilament nets were equal or more efficient for the capture of nearly all other species collected in the study. Also, Njoku (1991) found that multifilament gill nets captured a greater weight of fish than monofilament gill nets and better selected for higher quality (more in-demand) species and size classes in Oguta Lake, Nigeria. The increased incidence of gar capture in multifilament gill nets of our study may be related to the increased number of captures via entanglement. Increased water turbidity may have also reduced visual avoidance by individual fish. Multiple studies have commented that, as one might expect, nets and traps are easier for fish to detect and avoid in good light conditions (Hamley 1975; Buijse et al. 1992; Olin and Malinen 2003). A combination of overnight sampling and turbid waters may have negated the expected avoidance of the multifilament gill nets in this study, which researchers have often suggested are easier for fish to see and avoid than monofilament nets (Hovgård 1996; Faife and Einarsson 2003).

We captured gars significantly more often via entanglement than would be expected given capture mode occurred at random. In fact, 91 fish (47%) captured in this study were via entanglement. The increased incidence of entanglement is important for the interpretation of the size selectivity of our gill nets. A high incidence of entanglement suggests that our nets were targeting a size class of gars smaller than the majority of the fish sampled. Entanglement is usually the result of a fish's body (circumference) being too large to successfully penetrate the mesh (Schlechte et al. 2016). Gars collected in this study ranged from 558 to 1,491 mm total length (TL), which may suggest that the largest Longnose Gar were not vulnerable to our mesh size via gilling but were vulnerable to entanglement. We did collect a wide size breadth of all three major gar species (Shortnose: 558–955 mm; Spotted: 590–884 mm; Longnose: 612–1,491 mm TL), which gives promise to targeting a wide range of sizes with a simple net construction. We may also, in part, attribute the increased incidence to the head morphology of the gar species, with its exaggerated snout and many large teeth. Body morphology may also provide some insight into the relatively higher rate of wedging compared to gilling (though both were significantly less than would be expected by chance). Since the maximum head girth of a gar is, in many cases, fairly similar the body girth, it makes sense that a struggling gar may be able to stretch the mesh over a portion of the anterior body (Hamley 1975). The range of size classes sampled by monofilament and multifilament gill nets were identical, except for multifilament gill nets capturing a few fish larger (1,400–1,450 mm TL) than those collected by monofilament gill nets. It is possible that we captured fish of this size in monofilament gill nets but they were able to escape via breaking the mesh or struggling free. Throughout the study, we did see an increased incidence of net damage, primarily to monofilament gill nets. However, we attributed this, more or less exclusively, to coarse woody debris in the Clarks River. It was commonplace to pull gill nets off of woody debris in this river system and many holes were relatively large. Some of this damage to nets could have been caused by large fish; large gars, Freshwater Drum Aplodinotus grunniens, Common Carp Cyprinus carpio, Asian carps Hypophthalmichthys spp., and American Paddlefish Polyodon spathula are all found in the Clarks River (data not shown) and grow to large sizes. Monofilament mesh is more rigid than multifilament mesh (Stewart 1987) and more prone to break when placed under stress. So, it is possible that we captured larger gars in monofilament gill nets but they were able to exert enough force to break the webbing and escape. We were not able to confirm, or quantify, this phenomenon but it is an important quality that researchers and managers should consider when choosing a mesh for their gill nets, particularly if they are sampling larger species such as Alligator Gar.

Our entanglement results agree with a study by Njoku (1991) which also found that multifilament mesh performed better than monofilament mesh on a wide breadth of fish species; the difference in performance was partially attributed to the twisted twines of multifilament mesh that can more easily entangle fish. Collins (1979) explained a similar finding where Rainbow Smelt Osmerus mordax often became entangled by their teeth and, in turn, were more vulnerable to capture via multifilament nets than monofilament nets. The exaggerated snout and large number of teeth possessed by gar may increase the likelihood of entanglement as the teeth become snagged on the many strands of the multifilament mesh; this entanglement effect may be less likely with the single strand in monofilament gill nets. We also captured Channel Catfish Ictalurus punctatus and Yellow Bass Morone mississippiensis in multifilament mesh significantly more often than in monofilament (data not presented). Both species possess structures on the anterior portion of their body (Channel Catfish: pectoral spines, long barbels, dorsal spines; Yellow Bass: dorsal spines, opercle spines) which could increase the potential of snagging and tangling, lending more evidence toward the hypothesis of morphology playing an important role in gill net capture in some species. However, other studies (e.g. Khan et al. 1975; Shon 1975; Hylen and Jakobsen 1979; Ayaz et al. 2006) have found monofilament gill nets to be more efficient than those with multifilament mesh. The variation in results of these studies highlights the importance of selecting the correct sampling gear.

Gill nets are designed to capture fish by the head or body (Hubert et al. 2012); thus, fish that get their entire head and part of their body through the net before attempting to escape should have an increased chance of being captured. Fish swimming directly into the gill net should be more likely to pass a portion of their body through the net before attempting to escape, as opposed to those that meet the net at an angle and are able to change directions and avoid capture. In support of this idea, our results showed that nets deployed perpendicular to the river channel showed the highest incidence rate of gar capture, followed by diagonal net sets and, finally, parallel sets. This finding raises many interesting questions including visibility and/or detectability of different net mesh types by fish and nets deployed in different orientations.

Researchers have described monofilament gill nets as being virtually invisible in the proper water clarity and light conditions (Blaxter et al. 1964). However, the angle at which a fish approaches a net likely plays a part in the visibility and detectability, as well. The water in the Clarks River is quite turbid, which likely made for similar visibility between monofilament and multifilament gill nets. This was probably even more likely after we had deployed multifilament gill nets a few times and they became stained to a tan color. Previous literature has discussed effects of mesh color (Hamley 1975). Also, we deployed many of our net sets at night, which may have further reduced and/or negated the difference in visual detectability between the two net types. However, our study design did not allow us to differentiate between the effects of net soak time and diurnal fish activity. A variety of studies, including Olin and Malinen (2003), have cited diurnal activity as an important factor. So, we cannot make a strong argument for detectability reduction at night from our results, although it has been suggested by other studies (e.g. Albert and Einarsson 2004).

It is also important to consider encounter areas for each deployment orientation. As noted by Hamley (1975), in order for a fish to be captured by a passive gear, the two must present in the same space and time. Deployment orientation can affect whether the sampling gear and a given fish encounter one another in space. A perpendicular deployment (the most efficient in the present study) has a large lateral encounter area with an ability to sample nearly all vulnerable fish traveling up- or downstream and acts similarly to a block net; however, parallel deployments (the least efficient in our study) have a large longitudinal encounter area where fish moving laterally across the river channel have an increased likelihood of capture but fish moving up- or downstream may not encounter the net at all. Diagonal deployments appear to fill a middle ground where some fish are able to pass without encountering the net (opposed to the perpendicular set), but the encounters are not limited solely to those moving laterally between banks (as in the parallel sets). These deployment orientations may have little effect in many marine or lacustrine systems but could be important for researchers and managers working in riverine systems. This may be a further consideration for those that may want to sample migrating fish without completely impeding progress of the entire population. Information on capture rates of different deployment orientations should prove useful for managers and researchers as setting nets in different orientations can help to avoid much of the coarse woody debris along the margins of midorder rivers. The results of our study provide multiple avenues of future research on the topic of visibility and orientation of gill nets.

The differences in mean captures of lepisosteids found in our study have important management implications for reintroduction and monitoring efforts. Our data suggest that multifilament gill nets may provide increased sample sizes for monitoring gar populations, as well as other important recreational and commercial species. However, further research is needed on capture mode, selectivity, and timing of sampling, both seasonal and diurnal timing. New advances in boat electrofishing have shown promise for sampling gar species more effectively (U.S. Army Corps of Engineers at Waterways, personal communication) but additional research needs remain. Tarred gill nets are a popular choice for gar sampling in recent years and should also receive additional scientific attention for their efficacy and selectivity for community sampling, not only for gar but other important management species, as well. The differences in diurnal activity of gar may have contributed, partially, to the sampling composition of this study. Netting occurred during both day and night throughout this study, but for different soak times, which made it impossible for us to determine which variable(s) may have contributed to increased captures between the two periods. The gill nets used in this study were constructed using only 50.8-mm multifilament mesh. Work by Schlechte et al. (2016) found that a weighted-effort gill net with specific ratios of differing mesh sizes could provide a more representative sample of the Alligator Gar community structure for fishes between 1,200 and 2,100 mm TL. Further research on weighted-effort gill nets using smaller mesh sizes could help to increase efficiency and better sample smaller fish sizes, which was noted as a limitation in the net design (Schlechte et al. 2016). Gill nets, as well as other passive capture methods, remain important components of many research and management projects and methods should continue to be tested and improved upon to increase efficiency and reliability. Researchers have also suggested that entanglement is less size-selective than either gilling or wedging, which gives promise to gill nets helping managers and researchers to sample a large proportion of the gar population with relatively simple gear and constructions. The results of this study provide important information on the ability of small mesh sizes to capture wide breadths of sizes for multiple gar species, as well as the prominent modes of capture for these fishes. Further research is needed to assess the viability of these mesh sizes for Alligator Gar as only three were captured in this study.

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.

Table S1. Total gar captures organized by date and net in the Clarks River, Kentucky, between September 2013 and October 2014. All capture values are counts. Date is the date we retrieved the net. Soak.Time is the total soak time of the net. Season is the calendar season. Type is the mesh type of the net. Size is the net length in meters. Orientation is relative to river channel. Habitat is the river habitat in which we deployed the net. Temp.Mouth represents the surface water temperature of the Clarks River at the time we retrieved the net. Total is the total number of gars captured per net.

Found at DOI: https://doi.org/10.3996/112018-JFWM-101.S1 (35 KB DOCX).

Table S2. Individual gar captures by net in the Clarks River, Kentucky between September 2013 and October 2014. Table includes information on net mesh, length, and orientation. Date is the date we retrieved the net. Net is the net number for that day. Species is the gar species captured (Long = Longnose Gar Lepisosteus osseus, Short = Shortnose Gar Lepisosteus platostomus, Spot = Spotted Gar Lepisosteus oculatus). Type is the mesh type of the net. Size is the net length, in meters. Orientation is the deployment orientation code (L = parallel, D = diagonal, R = perpendicular), all delineations are in reference to the stream channel. Capture is the capture mode code (T = tangled, G = gilled, W = wedged). TL is the total length, in millimeters, of the captured individual.

Found at DOI: https://doi.org/10.3996/112018-JFWM-101.S2 (37 KB DOCX).

We would like to thank the Journal reviewers and Associate Editor for their important commentary and edits on previous versions of this manuscript. We would like to thank Dr. Howard Whiteman and the Watershed Studies Institute of Murray State University for their support of this project. We would also like to thank Murray State University's Hancock Biological Station for providing equipment and lab space. Additional thanks go out to Christy Soldo, Benjamin Tumolo, and Whitney Wallett for their many comments and revisions in the development of this manuscript. This project was approved by Murray State University's Institution for Animal Care and Use Committee (Protocol: 2014-007).

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.

Albert
A,
Einarsson
HA.
2004
.
Selectivity of gillnet series in sampling of Perch (Perca fluviatilis L.) and Roach (Rutilus rutilus) in the Coastal Sea of Estonia
.
Fisheries Training Program
,
United Nations University, Reykjavik, Iceland
. .
Argent
DG,
Kimmel
WG.
2005
.
Efficiency and selectivity of gill nets for assessing fish community composition of large rivers
.
North American Journal of Fisheries Management
25
:
1315
1320
.
Ayaz
A,
Acarli
D,
Altinagac
U,
Ozekinci
U,
Kara
A,
Ozen
O.
2006
.
Ghost fishing by monofilament and multifilament gillnets in Izmir Bay, Turkey
.
Fisheries Research
79
:
267
271
.
Balik
I.
1999
.
Investigation of the selectivity of multifilament and monofilament gill nets of Pike Perch (Stizostedion lucioperca (L., 1758)) fishing in Lake Beysehir
.
Turkish Journal of Zoology
23
:
179
183
.
Barlow
J,
Cameron
GA.
2003
.
Field experiments show that acoustic pingers reduce marine mammal bycatch in the California drift gill net fishery
.
Marine Mammal Science
19
:
265
283
.
Binion
GR,
Daugherty
DJ,
Bodine
KA.
2015
.
Population dynamics of Alligator Gar in Choke Canyon Reservoir, Texas: implications for management
.
Journal of the Southeastern Association of Fish and Wildlife Agencies
2
:
57
63
.
Blaxter
JHS,
Parrish
BB,
Dickson
W.
1964
.
The importance of vision in the reaction of fish to driftnets and trawls
.
Pages
529
536
in
Technical Staffs of Fishing News International and Fishing New
.
Modern fishing gear of the world 2
.
London
:
Fishing News Books Ltd.
Brinkman
EL.
2008
.
Contributions to the life history of Alligator Gar, Atractosteus spatula (Lacepede), in Oklahoma. Master's thesis
.
Stillwater
:
Oklahoma State University
. .
Buckmeier
DL,
Smith
NG,
Daugherty
DJ.
2013
.
Alligator Gar movement and macrohabitat use in the lower Trinity River, Texas
.
Transactions of the American Fisheries Society
142
:
1025
1035
.
Buijse
AD,
Houthuijzen
RP.
1992
.
Piscivory, growth, and size-selective mortality of age 0 Pike Perch (Stizostedion lucioperca)
.
Canadian Journal of Fisheries and Aquatic Sciences
49
:
894
902
.
Burr
JG.
1931
.
Electricity as a means of garfish and carp control
.
Transactions of the American Fisheries Society
61
:
174
182
.
Clavero
M,
Blanco-Garrido
F,
Prenda
J.
2006
.
Monitoring small fish populations in streams: a comparison of four passive methods
.
Fisheries Research
78
:
243
251
.
Clay
TA.
2009
.
Growth, survival, and cannibalism rates of Alligator Gar Atractosteus spatula in recirculating aquaculture systems. Master's thesis
.
Thibodaux, Louisiana
:
Nicholls State University
. .
Collins
JJ.
1979
.
Relative efficiency of multifilament and monofilament nylon gill net towards Lake Whitefish (Coregonus clupeaformis) in Lake Huron
.
Journal of the Fisheries Board of Canada
36
:
1180
1185
.
Cooke
SJ,
Bunt
CM,
McKinley
RS.
1998
.
Injury and short-term mortality of benthic stream fishes—a comparison of collection techniques
.
Hydrobiologia
379
:
207
211
.
Daugherty
DJ,
Pangle
KL,
Karel
W,
Baker
F,
Robertson
CR,
Buckmeier
DL,
Smith
NG,
Boyd
N.
2017
.
Population structure of Alligator Gar in a Gulf Coast river: insights from otolith microchemistry and genetic analyses
.
North American Journal of Fisheries Management
37
:
337
348
.
Faife
JR,
Einarsson
HA.
2003
.
Effect of mesh size and twine type on gillnet selectivity of Cod (Gadus morhua) in Icelandic coastal waters. Final Report
.
Reykjavik, Iceland
:
Institute for the Development of Small-Scale Fisheries (IDPPE)
.
Ferrara
AM,
Irwin
ER.
2001
.
A standardized procedure for internal sex identification in Lepisosteidae
.
North American Journal of Fisheries Management
21
:
956
961
.
Hamley
JM.
1975
.
Review of gillnet selectivity
.
Journal of the Fisheries Board of Canada
32
:
1943
1969
.
Henderson
BA,
Nepszy
SJ.
1992
.
Comparison of catches in mono- and multifilament gill nets in Lake Erie
.
North American Journal of Fisheries Management
12
:
618
624
.
Hovgård
H.
1996
.
A two-step approach to estimating selectivity and fishing power of research gill nets used in Greenland waters
.
Canadian Journal of Fisheries and Aquatic Sciences
53
:
1007
1013
.
Hubert
WA,
Pope
KL,
Dettmers
JM.
2012
.
Passive capture techniques
.
Lincoln
:
Nebraska Cooperative Fish & Wildlife
Research Unit–Staff Publications. Paper 111. Available: https://digitalcommons.unl.edu/cgi/viewcontent.cgi?article=1110&context=ncfwrustaff (October 2019).
Hylen
A,
Jakobsen
T.
1979
.
A fishing experiment with multifilament, monofilament, and monotwine gill nets in Lofoten during the spawning season of Arcto-Norwegian Cod in 1974
.
Fiskeridirektoratets Skrifter Serie Havundersøkelser
16
:
531
550
.
Khan
AA,
George
NA,
Pandey
OP.
1975
.
On the fishing power of monofilament and multifilament gill nets
.
Fisheries Technology
12
(
1
):
64
69
.
Larkins
HA.
1964
.
Some epipelagic fishes of the North Pacific Ocean, Bering Sea, and Gulf of Alaska
.
Transactions of the American Fisheries Society
93
:
286
290
.
McClelland
MA,
Irons
KS,
Sass
GG,
O'hara
TM,
Cook
TR.
2013
.
A comparison of two electrofishing programmes used to monitor fish on the Illinois River, Illinois, USA
.
River Research and Applications
29
:
125
133
.
Munprasit
A,
Chanrachkij
I,
Prajakjitt
P,
Manajit
N,
Yasook
N,
and
Petchkham
R.
1986
.
Fishing gear and methods in southeast Asia: Thailand (revisional edition)
.
Samut Prakan, Thailand
:
Training Department, Southeast Asian Fisheries Development Center
.
Murray State University, Hancock Biological Station and Center for Reservoir Research
.
2011
.
Total maximum daily load for Escherichia coli 40 stream segments within the Clarks River watershed Calloway, Graves, Marshall, and McCracken counties, Kentucky. Prepared for the Kentucky Department for Environmental Protection
.
Submitted to the United States Environmental Protection Agency. Available: https://eec.ky.gov/Environmental-Protection/Water/Protection/TMDL/Approved%20TMDLs/TMDL-ClarksRiverEcoli.pdf (October 2019)
.
Njoku
DC.
1991
.
Comparative efficiency and techno-economics of multifilament and monofilament gillnets on the Oguta Lake, Nigeria
.
Fisheries Research
12
:
23
30
.
Olin
M,
Malinen
T.
2003
.
Comparison of gillnet and trawl in diurnal fish community sampling
.
Hydrobiologia
506
:
443
449
.
Pristas
PJ,
Trent
L.
1977
.
Comparisons of catches of fishes in gill nets in relation to webbing material, time of day, and water depth in St. Andrew Bay, Florida
.
Fishery Bulletin
75
:
103
108
.
R Core Team
2017
.
R: A language and environment for statistical computing
.
R Foundation for Statistical Computing
,
Vienna, Austria
.
Available: http://www.R-project.org/ (October 2019).
Reis
EG,
Pawson
MG.
1992
.
Determination of gill-net selectivity for bass (Dicentrarchus labrax L.) using commercial catch data
.
Fisheries Research
13
:
173
187
.
Ross
ST,
Brenneman
WM.
2001
.
The inland fishes of Mississippi
.
Jackson, Mississippi
:
University Press of Mississippi
.
Santos
MN,
Saldanha
HJ,
Gaspar
MB,
Monteiro
CC.
2003
.
Hake (Merluccius merluccius L., 1758) ghost fishing by gill nets off the Algarve (southern Portugal)
.
Fisheries Research
64
:
119
128
.
Scarnecchia
DL.
1992
.
A reappraisal of gars and bowfins in fishery management
.
Fisheries
17
(
5
):
6
12
.
Schlechte
JW,
Bodine
KA,
Daugherty
DJ,
Binion
GR.
2016
.
Size selectivity of multifilament gill nets for sampling Alligator Gar: modeling the effects on population metrics
.
North American Journal of Fisheries Management
36
:
630
638
.
Shon
TJ.
1975
.
On the catch of gill net in the Jeju Island—comparison of Mackerel catch in monofilament and multifilament gill nets
.
Bulletin of the Korean Fisheries Society
8
(
1
):
7
10
.
Stewart
PA.
1987
.
The selectivity of slackly hung Cod gillnets constructed from three different types of twine
.
ICES Journal of Marine Science
43
:
189
193
.
Trent
L,
Hassler
WW.
1968
.
Gill net selection, migration, size and age composition, sex ratio, harvest efficiency, and management of striped bass in the Roanoke River, North Carolina
.
Chesapeake Science
9
:
217
232
.

Author notes

Citation: Richardson BM, Flinn MB. 2019. Increased probability and efficiency of gar capture using multi- vs. monofilament gill nets. Journal of Fish and Wildlife Management 10(2):551–562; e1944-687X. https://doi.org/10.3996/112018-JFWM-101

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.

Supplemental Material