Injury and Mortality Induced by Four Hook Types on Bluegill and Pumpkinseed

Home , Catch and release, Circle hook

North American Journal of Fisheries Management 23:883±893, 2003 ᭧ Copyright by the American Fisheries Society 2003 Injury and Mortality Induced by Four Hook Types on Bluegill and Pumpkinseed

STEVEN J. COOKE*1 Department of Natural Resources and Environmental Sciences, University of Illinois, and Center for Aquatic Ecology, Illinois Natural History Survey, 607 East Peabody Drive, Champaign, Illinois 61820, USA

CORY D. SUSKI Department of Biology, Queen's University, Kingston, Ontario K7L 3N6, Canada

BRANDON L. BARTHEL Department of Natural Resources and Environmental Sciences, University of Illinois, and Center for Aquatic Ecology, Illinois Natural History Survey, 607 East Peabody Drive, Champaign, Illinois 61820, USA

KENNETH G. OSTRAND Sam Parr Biological Station, Center for Aquatic Ecology, Illinois Natural History Survey, Kinmundy, Illinois 62854, USA

BRUCE L. TUFTS Department of Biology, Queen's University, Kingston, Ontario K7L 3N6, Canada

DAVID P. P HILIPP Department of Natural Resources and Environmental Sciences, University of Illinois, and Center for Aquatic Ecology, Illinois Natural History Survey, 607 East Peabody Drive, Champaign, Illinois 61820, USA

Abstract.ÐTackle manufacturers have responded to concerns regarding hooking injury and mortality by attempting to design and market hooks that are less damaging to ®sh (e.g., circle hooks). To date, studies investigating circle hooks have been primarily restricted to large marine species. We compared the injury and short-term (72-h) mortality of bluegills Lepomis macrochirus and pumpkinseeds L. gibbosus angled using number-6 circle hooks and three other conventional hook types (aberdeen, wide-gap, and baitholder) across three water temperatures (18, 22, and 26ЊC). Unlike other hook types, circle hooks were never lodged in the gullet, but they were frequently lodged in the eye. Some ®sh captured on conventional hooks were hooked deeply in the gullet, necessitating line cutting for release. Incidences of bleeding were low using all hook types, and when not lodged in the gullet, all hooks were generally easy to remove. Anatomical hooking locations differed among small (Ͻ145-mm) and large (Ͼ145-mm) bluegills for all hook types but not among pumpkinseeds. Hooking depth differed between small and large ®sh of both species captured on circle hooks; smaller ®sh were hooked more deeply. Mortality in both species was negligible at all water temperatures except for bluegills at 26ЊC (3% mortality). Bluegills that died were smaller than those that survived. Our results con®rm the supposition that circle hooks are less susceptible to deeply hooking ®sh in the gullet. However, circle hooks permanently impaired vision of up to 22% of the ®sh, much more than for other hooks types. Although ef®cient at minimizing injury and mortality in marine ®sh, our study suggests that circle hooks perform similarly to more conventional hook types in ®sheries for small sun®sh.

The most common type of catch-and-release tality (Muoneke and Childress 1994). These stud- studies are assessments of hooking injury and mor- ies generally attempt to describe and quantify the factors that contribute to injury and mortality and provide direction to minimize negative effects. * Corresponding author: [email protected] Based upon the ®ndings of this body of research 1 Present address: Centre for Applied Conservation Re- and their subsequent dissemination to stakehold- search, Department of Forest Sciences, University of Brit- ers, hooking mortality rates across species have ish Columbia, Forest Sciences Center, 3022-2424 Main generally been lowered to 10% or less (Muoneke Mall, Vancouver, British Columbia V6T 1Z4, Canada. and Childress 1994). Still, managers and anglers Received June 3, 2002; accepted January 8, 2003 alike devote efforts and resources to further min-

883 884 COOKE ET AL. imize this mortality. To date, studies have exam- and Crossman 1973) pan®sh targeted by anglers ined the role that water temperature (Wilkie et al. for both catch-and-release ®shing and for harvest. 1997; Wilde 1998), angling duration (Schreer et Anglers harvesting ®sh often return smaller indi- al. 2001), air exposure duration (Ferguson and viduals to the water (Coble 1988). For this reason, Tufts 1992; Cooke et al. 2001), and handling play several studies have investigated the hooking mor- in sublethal physiological effects (Cooke et al. tality of this species, focusing on bait type and 2002) and mortality (Muoneke and Childress water temperature (Burdick and Wydoski 1989; 1994) of many species. In addition, the type of Siewert and Cave 1990; Muoneke 1992). The con- bait (lure, ¯y, organic; Clapp and Clark 1989; generic pumpkinseed L. gibbosus, which ®ght Cooke et al. 2001), presence of a barb (Titus and strongly and erratically (Scott and Crossman Vanicek 1988; Cooke et al. 2001), angler expertise 1973), are also commonly captured and released (Dunmall et al. 2001), hook size (Burdick and Wy- by anglers. Pumpkinseeds, however, have not re- doski 1989; Weidlein 1989), and hook types ceived the same level of study, with regards to the (McEachron et al. 1985; Titus and Vanicek 1988) effects of catch-and-release, as have bluegill. In have also been examined with regard to both de- fact, we were unable to locate any studies that gree of injury and hooking mortality (Muoneke focused on pumpkinseeds. and Childress 1994). The purpose of this study was to compare the Tackle manufacturers have been quick to re- hooking injury and mortality of bluegills and spond to opportunities to further reduce mortality. pumpkinseeds angled with circle hooks and three A recent effort consistent with this goal is the de- other conventional hook types (aberdeen, bait- velopment and vigorous marketing of circle hooks holder, and wide-gap). By comparing a species that to anglers. Circle hooks are more circular in shape is well studied with regards to catch-and-release than the typical J hook (straight shank), with the (bluegill) with a species that has never been ex- point facing perpendicular to the hook shank (Sou- amined (pumpkinseed), we hoped to draw on pre- cie 1994). Although an ancient design (Soucie vious research while extending our understanding 1994), the circle hook has recently been used ex- of hooking injury and mortality to a new species. tensively in longline and trotline ®sheries in ma- Because hooking mortality can vary with water rine (McEachron et al. 1985; Woll et al. 2001) and temperature (Muoneke and Childress 1994), we freshwater (Ott and Storey 1993). More recently, replicated the experiment at three water temper- marine recreational anglers have begun to use atures (18, 22, and 26ЊC). these hooks (Montrey 1999), and use is quickly expanding into the freshwater market (Stange Methods 1999). Studies have clearly documented reduced Field site.ÐAll angling experiments and reten- hooking mortality associated with this design in tion of ®sh were conducted in Lake Opinicon, On- both longline (McEachron et al. 1985) and marine tario. Lake Opinicon is a mesotrophic natural lake angling (Grover et al. 2002; Prince et al. 2002; in eastern Ontario with abundant populations of Skomal et al. 2002) ®sheries. Fish were frequently rock bass Ambloplites rupestris, largemouth bass hooked in the upper maxilla, and circle hooks rare- Micropterus salmoides, smallmouth bass Microp- ly penetrated vital tissues or organs (i.e., gills, gul- terus dolomieu, bluegills, and pumpkinseeds. An- let, eye). Circle hooks do not require aggressive gling was conducted from a 4-m barge or from hook sets, experience fewer snags, make it more docks that extended into the lake to a depth of at dif®cult to hook oneself, and ®sh rarely shake free least 1 m. Experiments were conducted on May once hooked. To date, all circle hook studies have 18, June 6, and June 15, 2001, when the respective focused on large marine ®shes. Anecdotal reports water temperatures were 18, 22, and 26ЊC. Due to emanating from tackle manufacturers and media natural variation in water temperatures (both tem- outlets (Montrey 1999; Stange 1999) indicate that porally and spatially), we conducted experiments circle hooks are also effective for freshwater ®sh- during periods when middepth water temperatures es, although a quantitative rigorous assessment is in the enclosure used for the hooking mortality lacking. study were within 1ЊC of the desired water tem- The hooking injury of small centrarchid ®shes perature. These slight variations in water temper- has been well studied due to their abundance in ature are minimal relative to natural conditions that many freshwater systems and because of their rec- would be faced by free-swimming ®sh in other reational value. The bluegill Lepomis macrochirus parts of the lake. is a popular hard-biting and hard-®ghting (Scott Experimental protocol.ÐWe used four commer- BLUEGILL AND PUMPKINSEED HOOKING MORTALITY 885

Anglers were instructed to cast the bait and let the weight and Crappie Nibble settle until the ¯oat was upright. The anglers were then told to wait for the ¯oat to go under before collecting slack line and gently pulling up on the rod. Instructions on circle hook packaging indicated that strong hook sets were not required. This is atypical for other hook types so we opted for a light hook-set to satisfy all hook types. If upon setting the hook, a ®sh was on the line the angler reeled in the ®sh. We standardized the duration of time that the ®sh was reeled in to 15 s. This interval was suf®cient to ensure that all ®sh could be easily landed without creating undue exhaustion. We operated a two-person sampling crew that was responsible for the hook removal, handling, and enumeration of all ®sh that were captured. The same two individuals recorded the response variables for all ®sh to ensure consistency. Anglers immediately brought angled ®sh to the sampling crew. If the crew was occupied with another ®sh, the angler held the ®sh in a 10-L cooler for up to 1 min before enumeration. Lepomid hybrids and other nontarget species were returned to the lake without enumeration. The sampling crew ®rst determined the anatomical location of the hook (upper jaw, lower jaw, FIGURE 1.ÐDrawings of the four different hook types roof, eye, or gullet). The location of hook pene- used in our study of hooking injury and mortality of bluegills and pumpkinseeds. tration was measured from the anterior aspect of the lower lip to the most posterior point of hook penetration (Dunmall et al. 2001). Hook type and cially available hook types that are frequently used species identity were also recorded. Ease of hook by anglers to target bluegills and pumpkinseeds removal was categorized using slight modi®ca- (Figure 1): baitholder hook (size 6, bronze, offset, tions to the criteria proposed by Cooke et al. down eye; Model 3181UK, Eagle Claw Inc.), aber- (2001). If hemostats were required to remove the deen hook (size 6, gold, light wire, ringed eye; hook, removal was categorized as ``dif®cult.'' If Model 3202UK, Eagle Claw Inc.), circle hook removal of the hook was not possible using he- (size 6, black and chrome, curved in point, mutu mostats, without causing substantial injury, the light; Model 5114±051, Owner Inc.), and wide line was cut and removal was categorized as ``not bend (size 6, bronze, plain shank, offset, ringed possible.'' If hooks could be removed by hand with eye; Model L042, Eagle Claw Inc.). All anglers little effort, removal was categorized as ``easy.'' used the same organic bait (Crappie Nibbles, silver After categorizing ease of hook removal, the crew glitter, Berkley Inc.) that measured 5 ϫ 5mmand looked for the presence of blood and recorded re- weighed about 0.2 g each. Only one Crappie Nib- sponses as either ``none'' or ``some.'' Fish were ble was used on an individual hook at any one also measured for total length (mm). We corrected time. All anglers ®shed with spring ¯oats (3/8'' the depth of hook penetration to total length to [10-mm] Pencil, 6'' [152-mm] stem) placed about permit a comparison of hook penetration depth 0.75 m above the hook. A single 3/0 lead split shot among ®sh of different sizes (Dunmall et al. 2001). was placed 0.15 m above the hook. Anglers used Anglers were queried as to the number of casts both spinning and spin cast reels, but all anglers required before a ®sh was successfully hooked and used medium action rods and 6-lb test line. An- landed. We termed this the relative hooking dif- glers rotated the rods at 10-min intervals to ensure ®culty index. Lower relative hooking dif®culty that all participating anglers used different hook values indicate that a particular hook is performing types. better. The index was reset to zero when the angler 886 COOKE ET AL. successfully landed a ®sh. Fish were suf®ciently small modes. The cutoff that we chose was also abundant and easy to catch that patchiness or an- similar to the median total length values for each gler ability were not considered to be important. species. We compared hooking depths of small and Pumpkinseed and bluegill were provided with a large ®sh for each species and hook type using t- binary spine clip combination so that we could tests. We also compared, using contingency table identify individuals. Including removal of hook analyses, the distributions of anatomical hooking and enumeration, air exposure durations were less location of ®sh in the two size-classes for each than 60 s. Fish were then placed into a 300-L hold- species and hook type. To increase power of these ing tank supplied with ambient lake water for no analyses, all water temperatures were pooled. more than 2 h, at which time they were transferred, Mann±Whitney U-tests were used for exploring using a knotless nylon dip net, to a 16-m3 enclosure differences in hook penetration depths and sizes in the lake. The bottom of the enclosure encom- (continuous variables) of ®sh that died versus passed part of the littoral zone with its accompa- those that survived the 72-h retention period. Al- nying food sources. Floating macrophytes and though data were normal and homogeneous, sam- scattered rock provided cover for the ®sh. The en- ple sizes were substantially different, necessitating closure was monitored for 72 h for the presence the nonparametric approach. Chi square analysis of dead or severely injured ®sh. Moribund indi- was used to test for differences in the distribution viduals were removed from the enclosure and in- of hooking locations, incidences of bleeding and spected for binary clip pattern to determine iden- ease of hook removal (categorical variables) be- tity. tween ®sh that died versus those that survived the Analysis.ÐContinuous data were not trans- 72-h retention period. All analyses were conducted formed because they were determined to be nor- using JMP 4.0 (SAS Institute, Inc.) and Systat 7.0 mally distributed with homogeneous variances. (SAS Institute, Inc.). All values reported are means Homogeneity of variance assumptions were as- Ϯ SE. All tests were assessed for signi®cance at sessed using Levene's test, and normality was ver- ␣ϭ0.1 to minimize type II errors. Data are re- i®ed using normal probability plots and the ported separately for bluegills and pumpkinseeds, Shapiro±Wilk test statistic (W). Differences in to- and formal comparisons between the two species tal length, depth of hook penetration, and relative were not conducted. hooking dif®culty index values among different water temperatures and hook types were assessed Results for each species using two-way analysis of vari- Bluegills ance (ANOVA). When differences were observed, we used Fisher's least signi®cant difference to We angled 685 bluegills across the three water compare means. temperatures (Table 1). Total lengths of the ®sh Categorical data were analyzed with logistic re- were similar across all temperatures and hook ϭ ϭ gressions and contingency tables. For each cate- types (ANOVA: F11, 673 0.864, P 0.550). Rel- gorical (dependent) variable of bleeding, anatom- ative hooking dif®culty differed signi®cantly ical hooking location, and ease of hook removal, among hook type and water temperatures (F ϭ the factors included as independent variables in 1.983, P ϭ 0.072). At 18ЊC, aberdeen hooks had the analysis were the effect of water temperature a higher relative hooking dif®culty index than cir- and hook type. Logistic regressions were restricted cle hooks (Tukey's test, P ϭ 0.064) and wide-gap to main-effects comparisons because of the spar- (P ϭ 0.002) hooks, and at 26ЊC, circle hooks had sity of data in some categories when the combined a higher hooking dif®culty index than aberdeen effects of all factors in the analysis were included. hooks (P ϭ 0.053; Figure 2). Although this may be a limitation, this approach Hooking location was in¯uenced by water tem- has been used elsewhere (Dunmall et al. 2001) in perature (␹2 ϭ 26.93, df ϭ 8, P Ͻ 0.001) (Figure situations where main-effects comparisons were 3) and hook type (␹2 ϭ 36.30, df ϭ 12, P Ͻ 0.001; the focus of the study. Figure 4). In general, ®sh were hooked most fre- We explored size-speci®c differences in hooking quently in the upper jaw (low of 34.29% for circle depth and anatomical hooking location between hooks to high of 47.40% for baitholder hooks) and two size-classes of bluegills and pumpkinseeds: roof of the mouth (low of 32.37% for baitholder small (Յ145 mm) and large (Ͼ145 mm). We gen- hooks to high of 42.78 for aberdeen hooks; Figure erated length frequency distributions that were bi- 4). Circle hooks had the highest incidence of eye- modal and crudely separated ®sh into large and hooking (22.8% or about twice as high as other BLUEGILL AND PUMPKINSEED HOOKING MORTALITY 887

TABLE 1.ÐNumber, mean total length (TL), and standard error (SE) of bluegills and pumpkinseeds angled at three different water temperatures (18, 22, and 26ЊC) using four different hook types (aberdeen, baitholder, circle, and wide- gap).

Bluegill Pumpkinseed Water temperature Mean Mean (ЊC) Hook type N TL (mm) SE N TL (mm) SE 18 Aberdeen 50 141 4.3 8 171 12.6 Baitholder 45 151 3.7 10 166 10.9 Circle 36 138 3.6 11 183 7.9 Wide-gap 36 143 3.9 7 173 13.6 22 Aberdeen 63 140 3.6 12 160 8.8 Baitholder 59 139 3.8 11 165 13.1 Circle 57 147 3.7 28 171 6.4 Wide-gap 53 135 3.3 4 150 15.4 26 Aberdeen 74 140 4.2 18 157 8.4 Baitholder 69 143 4.9 18 166 5.3 Circle 82 143 2.9 18 155 6.4 Wide-gap 61 137 3.9 30 149 4.6

hooks) and the lowest incidence of gullet hooking (0%). Gullet hooking rates for other hook types were about 4%. The depth of hook penetration did not vary with temperature and hook type (Figure ϭ ϭ 5; ANOVA: F11, 673 0.688, P 0.661). Inci- dences of bleeding were uniformly low (from 2.7% for wide-gap hooks to 5.1% for circle hooks) and were not in¯uenced by water temperature (␹2 ϭ 0.01, df ϭ 2, P ϭ 0.939) or hook type (␹2 ϭ 1.50, df ϭ 3, P ϭ 0.683). Both hook type (␹2 ϭ 17.99, df ϭ 6, P ϭ 0.006) and water temperature (␹2 ϭ 27.25, df ϭ 4, P Ͻ 0.001) in¯uenced the ease of hook removal; however, most hooks (about 80%) were usually easy to remove. Circle hooks were the only hook type that was never assigned to the removal category ``not possible.'' For small (Ͻ145-mm) and large (Ͼ145-mm) bluegills, the distribution of anatomical hooking locations differed signi®cantly for all hook types: aberdeen (␹2 ϭ 14.04, df ϭ 4, P ϭ 0.007), baitholder (␹2 ϭ 13.39, df ϭ 4, P ϭ 0.010), circle (␹2 ϭ 6.94, df ϭ 4, P ϭ 0.074), and wide-gap (␹2 ϭ 21.23, df ϭ 4, P Ͻ 0.001). Smaller ®sh were consistently hooked in the eye more frequently than larger ®sh for all hook types: aberdeen (18.0% small, 5.4% large), baitholder (18.9% small, 4.8% large), circle (27.6% small, 16.9% large), and wide-gap (13.6% small, 1.6% large). Conversely, larger ®sh were consistently hooked in the gullet FIGURE 2.ÐRelative hooking dif®culty of bluegills more frequently than smaller ®sh for all hook types (top panel) and pumpkinseeds (bottom panel) for four except circle hooks, which hooked no ®sh of any different hook types (aberdeen, baitholder, circle, and size in the gullet hooked: aberdeen (1.8% small, wide-gap) and three water temperatures (18, 22, and 6.8% large), baitholder (2.2% small, 7.2% large), 26ЊC). Relative hooking dif®culty is the number of casts required to hook a ®sh with a speci®c hook type. Bars and wide-gap (0.0% small, 9.7% large). The depth represent mean values; error bars represent the positive of hook penetration did not differ among the two standard error. size-classes for aberdeen (t ϭϪ0.15, df ϭ 183, 888 COOKE ET AL.

FIGURE 4.ÐAnatomical hooking location of bluegills (top panel) and pumpkinseeds (bottom panel) captured FIGURE 3.ÐAnatomical hooking location of bluegills (top panel) and pumpkinseeds (bottom panel) captured with four different hook types (aberdeen, baitholder, cir- at three water temperatures (18, 22, and 26ЊC). Data for cle, and wide-gap). Data for all water temperatures (18, Њ all hook types (aberdeen, baitholder, circle, and wide- 22, and 26 C) were pooled for each hook type. gap) were pooled for each temperature. the 72-h retention period (U ϭϪ2.03, df ϭ 284, P ϭ 0.042). Depth of hooking was similar among ϭ ϭ ϭ P 0.880, N 111 small, N 74 large), bait- ®sh that died (0.064 Ϯ 0.011, N ϭ 9) and lived ϭϪ ϭ ϭ ϭ holder (t 1.08, df 171, P 0.283, N 90 (0.054 Ϯ 0.002, N ϭ 277; U ϭ 0.78, df ϭ 284, ϭ ϭ small, N 83 large), or wide-gap hooks (t 0.33, P ϭ 0.434). Similarly, ease of hook removal ϭ ϭ ϭ ϭ df 148, P 0.741, N 88 small, N 62 large; (␹2 ϭ 2.81, df ϭ 2, P ϭ 0.245) and hooking lo- Figure 6). However, the depths of hook penetration cation (␹2 ϭ 5.96, df ϭ 4, P ϭ 0.202) did not for circle hooks were deeper in small ®sh (0.058, differ signi®cantly between ®sh that died or sur- ϭ ϭ ϭ N 98) than in large ®sh (0.049, N 77; t vived. Bleeding, however, was more prevalent in Ϫ ϭ ϭ 2.30, df 173, P 0.022). ®sh that died (22.2%, N ϭ 9) than in those that No ®sh died immediately following angling or survived (5.24%, N ϭ 277; ␹2 ϭ 2.87, df ϭ 1, before being introduced into the enclosure. Nine P ϭ 0.091). bluegills diedÐall captured at 26ЊCÐand most of those died within 12 h of retention (Table 2). At Pumpkinseeds 26ЊC, bluegill hooking mortality was 3.2%, and We angled 175 pumpkinseeds across the three over all temperatures this translates to a total of water temperatures (Table 1). Total lengths of the 1.3% hooking mortality for all of the bluegill that ®sh were similar among hook types (ANOVA: ϭ ϭ we angled during our study. All bluegills that sur- F3, 171 1.033, P 0.37); however, total length vived appeared healthy after the 72-h retention pe- decreased signi®cantly with increasing water tem- ϭ ϭ riod and did not exhibit fungal lesions or ®n ero- perature (F2, 172 6.866, P 0.010). Fish captured sion. at 26ЊC (Tukey' test, P Ͻ 0.001) and 22ЊC(P Ͻ Bluegill that died at 26ЊC were signi®cantly 0.013) were smaller than those captured at 18ЊC. smaller (mean TL Ϯ SE ϭ 123 Ϯ 10 mm, N ϭ 9) The relative hooking dif®culty index differed Ϯ ϭ ϭ than those that survived (141 2 mm, N 277) among water temperatures (ANOVA: F2 3.115, BLUEGILL AND PUMPKINSEED HOOKING MORTALITY 889

FIGURE 5.ÐRelative depth of hook penetration for all sizes of bluegills (top panel) and pumpkinseeds (bottom FIGURE 6.ÐRelative depth of hook penetration in panel) for four different hook types (aberdeen, bait- small (Ͻ145-mm) and large (Ͼ145-mm) bluegills (top holder, circle, and wide-gap) and three water tempera- panel) and pumpkinseeds (bottom panel) by four dif- tures (18, 22, and 26ЊC). Relative hooking depth is the ferent hook types (aberdeen, baitholder, circle, and wide- depth of hook ingestion divided by the total length of gap) that are pooled over three water temperatures (18, the ®sh. Bars represent mean values; error bars represent 22, and 26ЊC). Relative hooking depth is the depth of the standard error. hook ingestion divided by the total length of the ®sh. Error bars represent the standard error; bars represent mean values. ϭ ϭ ϭ P 0.047) and hook type (F3 5.207, P 0.002), and the interaction term suggests that individual ϭ ϭ Њ Њ ϭ factors were independent (F6, 174 1.198, P values were higher at 18 C than at 26 C(P 0.310; Figure 2). Overall, circle hooks had sig- 0.022) (Figure 2). ni®cantly higher relative hooking dif®culty index Hooking location was in¯uenced by water tem- values than aberdeen (Tukey's test, P Ͻ 0.001) and perature (␹2 ϭ 8.30, df ϭ 8, P ϭ 0.081; Figure 3) wide-gap (P Ͻ 0.001) hooks, and baitholder hooks and hook type (␹2 ϭ 18.59, df ϭ 12, P ϭ 0.050; had signi®cantly higher index values than aber- Figure 4). In general, most hooking locations were deen hooks (P ϭ 0.043). Hooking dif®culty index in the upper jaw (from 41.03% for baitholder hooks

TABLE 2.ÐFish and hooking data for nine bluegills that died during a 72-h posthooking retention period at 26ЊC; bluegills were held in a 16-m3 enclosure in the lake from which they had been caught.

Total length Posthooking death Hook type (mm) Hook location Bleeding Ease of removal Hooking depth at (in hours) Baitholder 108 Upper jaw None Easy 0.027 12 Wide-gap 109 Upper jaw None Easy 0.046 12 Aberdeen 109 Eye Some Easy 0.055 12 Aberdeen 110 Upper jaw None Easy 0.073 12 Circle 126 Upper jaw None Easy 0.048 12 Aberdeen 173 Gullet None Not possible 0.179 12 Baitholder 116 Roof of mouth None Easy 0.036 12 Aberdeen 120 Upper jaw None Easy 0.050 18 Circle 135 Eye Some Easy 0.066 56 890 COOKE ET AL. to 63.41% for wide-gap hooks) and roof of the jury, and ultimately ®tness. For these reasons, we mouth (26.83% for wide-gap hooks to 46.15% for measured several different response variables in baitholder hooks; Figure 4). Aberdeen hooks were the way that ®sh were hooked in an effort to as- never located in the lower jaw, baitholder hooks certain the effects of hook type on angled bluegills never in the eye, and circle hooks never in the and pumpkinseeds. gullet. The depth of hook penetration did not vary The depth of hook penetration, which can delay ϭ signi®cantly with hook type (ANOVA: F3, 171 or prohibit hook removal and may increase the 0.560, P ϭ 0.646) but did vary signi®cantly with potential for damaging vital organs, has been ϭ ϭ water temperature (F2, 172 7.995, P 0.005; shown to vary with bait type in smallmouth bass Figure 5). Hooks penetrated more deeply in ®sh (Clapp and Clark 1989), rock bass (Cooke et al. angled at 26ЊC than at 22ЊC (Tukey' test, P ϭ 2001), and bluegills (Siewert and Cave 1990). 0.037) and at 18ЊC(P ϭ 0.010). Incidences of Overall, we found that bluegills (about 0.052) were bleeding were generally low (from 1.8% for circle hooked more deeply than pumpkinseeds (about hooks to 10.3% for baitholder hooks) and were not 0.041) for all hook types, and that the depth of in¯uenced by water temperature (␹2 ϭ 0.06, df ϭ hook penetration did not vary in bluegills with 2, P ϭ 0.815) or hook type (␹2 ϭ 3.43, df ϭ 3, hook type or temperature. In pumpkinseeds, how- P ϭ 0.330). Ease of hook removal was in¯uenced ever, hooks penetrated more deeply at 26ЊC than by both hook type (␹2 ϭ 11.18, df ϭ 6, P ϭ 0.083) the other two temperatures. These hooking depth and water temperature (␹2 ϭ 6.11, df ϭ 4, P ϭ values were lower than rock bass captured on 0.047). In general, most ®sh were easily removed worms at 16ЊC (0.086; Cooke et al. 2001) but sim- from the hook (from 76.92% for baitholder to ilar to rock bass captured on jigs (0.047) at 16ЊC 95.12% for wide-gap). Baitholder and circle hooks (Cooke et al. 2001) and smallmouth bass captured were considered dif®cult to remove more frequent- on a variety of different baits and plastics on jig ly than the other two hook types. Some ®sh were heads (about 0.040±0.055) at 15ЊC (Dunmall et al. hooked too deeply for removal, resulting in cutting 2001). In our study, ®sh size in¯uenced the depth the line; however, circle hooks were never hooked of hook penetration in bluegills and pumpkin- that deeply. seeds; however, these differences were only pre- The distribution of anatomical hooking loca- sent for circle hooks. Circle hooks were located tions for all hook types did not differ signi®cantly deeper in small ®sh. All other hook type and ®sh between small and large pumpkinseeds (aberdeen: size-groups had hook penetration depths that were ␹2 ϭ 3.19, df ϭ 4, P ϭ 0.362; baitholder: ␹2 ϭ similar to the depth of the larger circle hook ®sh. 6.04, df ϭ 4, P ϭ 0.110; circle: ␹2 ϭ 2.51, df ϭ Bluegills, therefore, appear to be more susceptible 4, P ϭ 0.473; wide-gap: ␹2 ϭ 7.42, df ϭ 4, P ϭ to deep hooking than pumpkinseeds, especially if 0.115). The depth of hook penetration did not dif- anglers are targeting bluegills with circle hooks. fer between the two size-classes for aberdeen (t ϭ The anatomical hooking location provides in- Ϫ0.05, df ϭ 36, P ϭ 0.99; N ϭ 11 small, N ϭ 27 formation on potential or actual vital tissue dam- large), baitholder (t ϭϪ0.38, df ϭ 37, P ϭ 0.71; age. In our study, ®sh were frequently hooked in N ϭ 10 small, N ϭ 29 large), or wide-gap hooks the eye region, which may directly increase mor- (t ϭϪ0.18, df ϭ 39, P ϭ 0.86; N ϭ 20 small, tality rate (Siewert and Cave 1990) and may also N ϭ 21 large; Figure 6). However, hook penetra- affect ability to locate food, avoid predators, and tion by circle hooks was deeper in small ®sh compete with other ®sh, all of which would have (0.054, N ϭ 15) than in large ®sh (0.040, N ϭ 42; clear impacts on ®tness. Few ®shÐand none cap- t ϭϪ2.06, df ϭ 55, P ϭ 0.045; Figure 6). tured on circle hooksÐwere hooked in the gullet. None of the pumpkinseeds that we captured at In our study, bluegill size in¯uenced the anatom- any temperature died immediately following an- ical hooking location for all hook types. Smaller gling or during the 72-h retention period. All ®sh ®sh were hooked more frequently in the eye, and were in good condition, exhibiting no fungal le- bigger ®sh were hooked more frequently in the sions or ®n erosion at the conclusion of the reten- gullet. These trends, however, were not observed tion period. in pumpkinseed. One of the factors that probably contributed to this interspeci®c variation in hook- Discussion ing location and depth are other differences in The type of hook used can in¯uence several dif- mouth morphology. Pumpkinseeds possess molar- ferent aspects of how ®sh are hooked, each of iform teeth for grinding, whereas bluegill have which can independently in¯uence survival, in- pharyngeal pads covered with ®ne needle-like BLUEGILL AND PUMPKINSEED HOOKING MORTALITY 891 teeth (Keast and Webb 1966). Thus, the roof of closure within the lake they were captured that the bluegill mouth may be more susceptible to incorporated natural substrate, cover, and food. hook penetration than that of the pumpkinseed. Holding bluegills in an arti®cial system may have The molariform teeth also may provide protection heightened mortality due to crowding, con®ne- to the eye in pumpkinseeds, resulting in the less ment stress (Chiszar et al. 1972), or other factors. eye-hooking rates than in bluegills. When ®sh are Although anatomical hooking location is generally not hooked in the gills or heart, incidences of associated with mortality in angled ®sh, we ob- bleeding are typically low (e.g., Cooke et al. 2001; served very little mortality in our study, even in Dunmall et al. 2001). For this reason, incidences ®sh hooked in the gullet and eye. Indeed, there of bleeding were low in both bluegills and pump- was no difference in the anatomical hooking lo- kinseeds. Differences in buccal anatomy, there- cation between the bluegills that died during the fore, suggest that bluegills may be more suscep- 72-h retention period and those that survived. One tible to hook-induced injuries in vital tissues than reason why overall mortality in our study was low are pumpkinseeds. may be due to the fact that deeply ingested hooks Ease of hook removal will strongly in¯uence air were not removed. Other studies have indicated exposure duration and mechanical injury potential that mortality is reduced when deeply ingested and may create additional dif®culties for anglers hooks are not forcibly removed from vital tissue inexperienced with handling ®sh (Cooke et al. or organs (Warner and Johnson 1978; Warner 2001). In this study, baitholder hooks were con- 1979; Mason and Hunt 1967; Weidlein 1989). sidered to be the most dif®cult to remove, and all If managers recommend different hook types to hook types except circle hooks had instances anglers based upon scienti®c ®ndings, anglers will where they were too deep to remove. Circle hooks have questions regarding the capture ef®ciency of could always be removed from ®sh probably be- different gears. Our hooking dif®culty index in- cause they never hooked any ®sh in the gullet, dicated that both water temperature and hook type suggesting an advantage to the use of circle hooks in¯uenced our capture ef®ciency for both bluegills over the other hook types examined. Because the and pumpkinseeds. However, we feel that these point of the circle hook points towards the shank results are not suf®ciently disparate that anglers of the hook, gut hooking with this hook is rare for bluegills or pumpkinseeds would have reser- among a variety of ®sh (e.g., Prince et al. 2002; vations with using any of the hook types we tested. Skomal et al. 2002). Preliminary assessments in marine environments The hooking mortality levels that we observed are showing substantial promise for circle hooks, were low compared with those for many other ®sh which have reduced injury and mortality in species species (Muoneke and Childress 1994) and other ranging from bill®sh (Prince et al. 2002) to chi- studies conducted on bluegills (bluegill mortality nook salmon Oncorhynchus tshawytscha (Grover rates have ranged from 0% to 88%; Burdick and et al. 2002). Our freshwater assessment indicates Wydoski 1989; Siewert and Cave 1990; Muoneke that the bene®t of circle hooks may not be ubiq- 1992). Bluegills experienced mortality only at the uitous across all ®sh species and sizes, and in some highest water temperature (26ЊC); however, the cases may provide little conservation bene®t. Al- mortality levels were extremely low (ϳ3%). In though we found no ®sh were hooked in the gut contrast, pumpkinseeds did not exhibit any mor- using circle hooks, many were hooked in the eye, tality across the range of water temperatures that resulting in an obvious ®tness impairment. Vari- we examined (18±26ЊC). Siewert and Cave (1990) ation in feeding mode and mouth morphology may reported that bluegills angled at 22 Ϯ 2ЊC and held affect the performance of different ®sh hook de- for 10 d experienced substantial mortality. Even signs suggesting that ®ndings from hooking injury control ®sh captured with seines had mortality of and mortality studies may not be applicable to oth- about 12%. We did not have control ®sh in our er organisms, even if they are closely related spe- study, but even our most extreme treatments (an- cies. gling at 26ЊC) had extremely low mortality. We suggest that several factors may account for these Acknowledgments differences. First, Siewert and Cave (1990) held This study bene®ted greatly from the angling their ®sh for 10 d, a period substantially longer prowess of Deanna Barthel, Jana Svec, Julie Claus- than the period that we used. Second, Siewert and sen, Jacko Degner, Mickey Philipp, Madison Phi- Cave (1990) apparently kept ®sh in some form of lipp, Emily Grant, Kate Deters, Brian Jackson, arti®cial tank or raceway. We held ®sh in an en- Ryan Hewitt, Mike Morrissey, and Maggie Bart- 892 COOKE ET AL. miewski. We also thank the staff of the Queen's Keast, A., and D. Webb. 1966. Mouth and body form University Biology Station (Frank Phelan and relative to feeding ecology in the ®sh fauna of a Floyd Connor) for providing logistical support. small lake, Lake Opinicon, Ontario. Journal of the Fisheries Research Board of Canada 23:1845±1874. Karen Metz-Weatherhead skillfully produced the Lucy, J. A., and A. L. Studholme. 2002. Catch and hook illustrations. Earlier versions of this manu- release in marine recreational ®sheries. American script bene®ted from comments provided by Den- Fisheries Society, Symposium 30, Bethesda, Mary- nis Schupp and Steve Quinn. The Illinois Natural land. History Survey, Illinois Department of Natural Re- Mason, J. W., and R. L. Hunt. 1967. Mortality of deeply sources, and the University of Illinois provided hooked rainbow trout. Progressive Fish-Culturist ®nancial assistance for this study. We (SJC and 29:87±91. CDS) also acknowledge the support of the Natural McEachron, L. W., A. W. Green, G. C. Matlock, and G. E. Saul. 1985. A comparison of trotline catches on Sciences and Engineering Research Council for two hook types in the Laguna Madre. Texas Parks scholarship support. and Wildlife Department, Management Data Series 86, References Montrey, N. 1999. Circle hooks ready to boomÐdesign Burdick, B., and R. Wydoski. 1989. Effects of hooking pierces ®sh through jaw, promotes conservation. mortality on a bluegill ®shery in a western reservoir. American Sport®shing 2(1):6±7. Pages 187±196 in R. A. Barnhart, and T. D. Roelofs, Muoneke, M. I. 1992. Seasonal hooking mortality of editors. Catch-and-release ®shing: a decade of ex- bluegills caught on natural baits. North American perience. California Cooperative Fisheries Research Journal of Fisheries Management 12:645±649. Unit, Humbolt State University, Arcata. Muoneke, M. I., and W. M. Childress. 1994. Hooking Chiszar, D., M. Moody, and J. T. Windell. 1972. Failure mortality: a review for recreational ®sheries. Re- of bluegill sun®sh, Lepomis macrochirus, to habit- views in Fisheries Science 2:123±156. uate to handling. Journal of the Fisheries Research Ott, R. A., and K. W. Storey. 1993. Hooking mortality Board of Canada 29:576±578. of channel cat®sh caught by trotline. Proceedings Clapp, D. F., and R. D. Clark, Jr. 1989. Hooking mor- of the Annual Conference Southeastern Association tality of smallmouth bass caught on live minnows of Fish and Wildlife Agencies 45(1991):399±406. and arti®cial spinners. North American Journal of Prince, E. D., M. Ortiz, and A. Venizelos. 2002. A com- Fisheries Management 9:81±85. parison of circle hook and ``J'' hook performance Coble, D. W. 1988. Effects of angling on bluegill pop- in recreational catch-and-release ®sheries for bill- ulations: management implications. North Ameri- ®sh. Pages 66±79 in J. A. Lucy and A. L. Stud- can Journal of Fisheries Management 8:277±283. holme, editors. Catch and release in marine recre- Cooke, S. J., D. P. Philipp, K. M. Dunmall, and J. F. ational ®sheries. American Fisheries Society, Sym- Schreer. 2001. The in¯uence of terminal tackle on posium 30, Bethesda, Maryland. injury, handling time, and cardiac disturbance of Schreer, J. F., S. J. Cooke, and R. S. McKinley. 2001. rock bass. North American Journal of Fisheries Cardiac response to variable forced exercise at dif- Management 21:333±342. ferent temperatures: an angling simulation for Cooke, S. J., J. F. Schreer, K. M. Dunmall, and D. P. smallmouth bass. Transactions of the American Philipp. 2002. Strategies for quantifying sub-lethal Fisheries Society 130:783±795. effects of marine catch-and-release angling: insights Scott, W. B., and E. J. Crossman. 1973. Freshwater ®sh- from novel freshwater applications. Pages 121±134 es of Canada. Fisheries Research Board of Canada in J. A. Lucy and A. L. Studholme (2002). Bulletin 184. Dunmall, K. M., S. J. Cooke, J. F. Schreer, and R. S. Siewert, H. F., and J. B. Cave. 1990. Survival of released McKinley. 2001. The effect of scented lures on the bluegill, Lepomis macrochirus, caught on arti®cial hooking injury and mortality of smallmouth bass ¯ies, worms, and spinner lures. Journal of Fresh- caught by novice and experienced anglers. North American Journal of Fisheries Management 12: water Ecology 5:407±411. 242±248. Skomal, G. B., B. C. Chase, and E. D. Prince. 2002. A Ferguson, R. A., and B. L. Tufts. 1992. Physiological comparison of circle hook and straight hook per- effects of brief air exposure in exhaustively exer- formance in recreational ®sheries for juvenile At- cised rainbow trout (Oncorhynchus mykiss): impli- lantic blue®n tuna. Pages 57±65 in J. A. Lucy and cations for ``catch and release'' ®sheries. Canadian A. L. Studholme (2002). Journal of Fisheries and Aquatic Sciences 49:1157± Soucie, G. 1994. Hook, line and sinker: the complete 1162. angler's guide to terminal tackle. Holt, Rinehart and Grover, A. M., M. S. Mohr, and M. L. Palmer-Zwahlen. Winston, New York. 2002. Hook-and-release mortality of chinook salm- Stange, D. 1999. Inside anglers: curious, marvelous, on from drift mooching with circle hooks: manage- amazingÐcircle hooks for freshwater. In- ment implications for California's ocean sport ®sh- Fisherman 24(2):10±14. ery. Pages 39±56 in J. A. Lucy and A. L. Studholme Titus, R. G., and C. D. Vanicek. 1988. Comparative (2002). hooking mortality of lure-caught Lahontan cutthroat BLUEGILL AND PUMPKINSEED HOOKING MORTALITY 893

trout at Heenan Lake, California. California Fish posium. Humboldt State University, Arcata, Cali- and Game 74:218±225. fornia. Warner, K. 1979. Mortality of landlocked Atlantic salm- Wilde, G. R. 1998. Tournament-associated mortality in on hooked on four types of ®shing gear at the hatch- black bass. Fisheries 23(10):12±22. ery. Progressive Fish-Culturist 41:99±102. Wilkie, M. P., M. A. Brobbel, K. Davidson, L. Forsyth, Warner, K., and P. R. Johnson. 1978. Mortality of land- and B. L. Tufts. 1997. In¯uences of temperature locked Atlantic salmon (Salmo salar) hooked on upon the post-exercise physiology of Atlantic salm- ¯ies and worms in a river nursery area. Transactions on (Salmo salar). Canadian Journal of Fisheries and of the American Fisheries Society 107:772±775. Aquatic Sciences 54:503±511. Weidlein, W. D. 1989. Mortality of released sublegal- Woll, A. K., J. Boje, R. Holst, and A. C. Gundersen. sized smallmouth bass, catch-and-release implica- 2001. Catch rates and hook and bait selectivity in tions. Pages 217±228 in R. A. Barnhart and T. D. longline ®shery for Greenland halibut (Reinhardius Roelofs, editors. Catch-and-release ®shing: a de- hippoglossoides, Walbaum) at east Greenland. Fish- cade of experience. A national sport®shing sym- eries Research 51:237±246.