North American Journal of Fisheries Management 26:921–925, 2006 [Management Brief] Ó Copyright by the American Fisheries Society 2006 DOI: 10.1577/M05-084.1

Nonlethal Sampling of and Slimy Sculpin for Stable Isotope Analysis: How Scale and Fin Tissue Compare with Muscle Tissue

MAUREEN H. KELLY* AND WILLIAM G. HAGAR Department of Biology, University of Massachusetts, Boston, Massachusetts 02125, USA

TIMOTHY D. JARDINE AND RICHARD A. CUNJAK Canadian Rivers Institute, Department of Biology, University of New Brunswick, Fredericton, New Brunswick E3B 5A3, Canada

Abstract.—We found that the sampling of tissues that do not 1984; Vander Zanden et al. 1997; Post 2002). Certain result in the death of the fish, such as scale and fin tissue, may biochemical reactions in an organism discriminate be substituted for muscle tissue in stable isotope analysis (SIA) against the heavier isotopes, causing a fractionation of fishes. Comparisons were made between the values of d13C 13 15 in the amounts of isotopes in tissues. For example, C and d N found in muscle tissue with the corresponding scale is discriminated against in lipogenesis, causing lipid- tissue of three sunfish species (bluegill Lepomis macrochirus, containing tissues to be lower in 13C values than other pumpkinseed L. gibbosus, and redbreast sunfish L. auritus) and with caudal fin tissue of slimy sculpin Cottus cognatus. tissues (DeNiro and Epstein 1977). Urea and ammonia 15 The fish showed strong linear correlation in d13C values have been shown to contain less N than the tissues of between their nonlethally sampled scale or fin tissue and their the organism that excreted them (Macko et al. 1982). muscle tissue (combined sunfish: r ¼ 0.97; slimy sculpin: r ¼ Muscle has been the tissue of choice for SIA studies 0.84). Sunfish d13C values were higher in scale tissue than in of fish (Pinnegar and Polunin 1999), but other tissues, muscle tissue and required a correction factor for converting such as liver, heart, brain, and gonadal tissues, have also the scale values to the muscle values (regression equation: y ¼ 13 been analyzed. The sampling of these tissues almost 1.1673x þ 1.0531). Slimy sculpin d C fin and muscle values always requires killing the fish. Nonlethally sampled were similar and did not require a correction factor. The tissues, such as scale or fin tissue, are potential correlation of d15N values between the tissues was also strong in both sunfish (r ¼ 0.94) and slimy sculpin (r ¼ 0.90). A substitutes for muscle and other lethally sampled correction factor was needed to convert d15N values from scale tissues. The stable isotope values of scale and fin tissue to muscle in the three sunfish species (y ¼ 0.8504x þ 2.6698) can be compared with known values in lethally sampled and from fin to muscle in slimy sculpin (y¼1.2658x – 3.3234). tissues (Hobson and Clark 1993; Shannon et al. 2001), Results of this study and other literature support the use of and nonlethally sampled tissues could become the nonlethally sampled tissues for SIA of fish. These methods standard for sampling in the future. Sampling these may be used for investigations of rare and endangered species tissues allows researchers to study populations of fish and also allow for analysis of archived fish scales. without affecting their numbers, without removing their resources from the ecosystem, and without reducing Stable isotope analysis (SIA) has become a reliable their gene pools. As early as 1985, Estep and Vigg method for quantifying trophic relationships and (1985) stated that the consistent difference in d13C dietary habits of animals, including fishes, in the wild. between scale and muscle tissue could be an important The ratios of heavy-to-light stable isotopes found in nondestructive tool for distinguishing wild fish from organisms are used to follow the flow of energy and fish reared in different hatcheries. For conservation 13 nutrients through an ecosystem. The amount of Cin value, the use of nonlethally sampled tissues would an organism increases a negligible amount (0–1%) allow the study of endangered and locally rare species. from its diet and is used to identify the base of the Researchers need to know how scale and fin tissues carbon source for the organism (Peterson and Fry 1987; compare in isotope values with muscle and other 15 Post 2002). The larger increase of N from diet to tissues to be able to substitute them in previous and consumer (3–5%) is useful when quantifying an current food web studies (Pinnegar and Polunin 1999; organism’s trophic position (Minagawa and Wada Perga and Gerdeaux 2003). To better understand the degree of isotope fractionation between different * Corresponding author: [email protected] tissues in individual fish and to guide the choice of 13 Received May 25, 2005; accepted April 28, 2006 tissues for future studies, we compared the d C and Published online November 20, 2006 d15N values of scale and muscle tissue in benthopelagic

921 922 KELLY ET AL. centrarchid fishes from two ponds in Massachusetts then dried and cut to appropriate weights for SIA. The (bluegill Lepomis macrochirus and pumpkinseed L. Ecosystem Center Laboratory at the Marine Biological gibbosus) and two ponds in New Brunswick (pump- Laboratory, Woods Hole, Massachusetts, performed kinseed and redbreast sunfish L. auritus). We also the isotope analysis for the U.S. sunfish and the Stable compared caudal fin tissue and muscle tissue of a Isotopes in Nature Laboratory (SINLAB), University of benthic cottid, slimy sculpin Cottus cognatus, collected New Brunswick, performed the isotope analysis on from two streams in New Brunswick. Finally, we sunfish and slimy sculpin from Canada. compared our results for scale and fin tissue with those For SIA, we chose not to acid wash our scale or fin of previous studies to determine the generality of the tissue. Although some of the carbon in scales is outcome across several species of fish. inorganic carbonate, Bunn et al. (1995) reported that Fish were sacrificed in this study to compare the acid washing may alter the d15N values of natural 13 15 amounts of Cand N in lethally as well as materials and increase the variability of the isotope nonlethally sampled tissues. However, in previous values. The fact that both d13C and d15N values of studies that sampled only scale or fin tissue, the fish scale tissue have been shown to increase with acid were returned to the water without apparent negative treatment (Perga and Gerdeaux 2003) suggests that it is effects (Gjerde and Refstie 1988; Conover and Sheehan not just the inorganic carbon that is removed with acid 1999; Tyus et al. 1999; Pratt and Fox 2002). If scale treatment. and fin tissues provide adequate data for analysis, the Stable isotope values are expressed in delta notation unnecessary sacrifice of sampled organisms may be (d) and are determined by the equation avoided. dsample ¼ðRsample=Rreference 1Þ 3 1; 000; Methods where R is the ratio of the heavier isotope : lighter Pumpkinseeds and bluegills were collected from isotope found in the sample and reference material (R ¼ Maquan Pond (4280303800 N, 7085100800 W) and Furnace 13C/12Cor15N/14N). The reference material for carbon Pond (4280302000 N, 7084904000 W) in southeastern is the Peedee Belemnite Limestone Formation, and the Massachusetts. Five pumpkinseeds were collected in reference material for nitrogen is atmospheric nitrogen. Mills Pond (458530900 N, 6680004900 W) adjacent to Positive delta values show a greater proportion of the Grand Lake and five redbreast sunfish were caught in Yoho Lake (4584603500 N, 6685200500 W); both of these heavier isotope than the reference, while negative delta ponds are located in New Brunswick. All sunfish were values show a lesser proportion of the heavier isotope collected by angling. Scale and muscle tissue were than the reference. The delta value is expressed in parts extracted from individual fish, except for one pooled per thousand (%). Samples were calibrated with sample of five age-2 pumpkinseeds from Maquan International Atomic Energy Agency standards CH6, Pond. Collection of slimy sculpin was by electrofishing CH7, N1, and N2. Standard deviations of replicate in two stream reaches of the Miramichi River, New samples were never greater than 0.1% for carbon and Brunswick. Site 1 was located in Brook 0.2% for nitrogen. (CBK; 4685204700 N, 6680606600 W), a third-order brook The three species and the four sites of sunfish were that drains to the Little Southwest Miramichi River. combined to examine Pearson’s correlations between 13 15 Site2waslocatedinStewartBrook(SBK; muscle and scale tissue for d C and d N. For slimy 4685704700 N, 6583906600 W), a second-order brook that sculpin, data from the two sites were combined to 13 15 drains directly to the main Northwest Miramichi River. examine d C and d N correlations between muscle Sacrificed fish were put on ice within 1 h of capture and fin tissue. Simple linear regression models were and were kept frozen at 208C in the laboratory. When used to determine the relationship between muscle 13 15 processed, the fish were slightly thawed and rinsed with tissue d C and d N values and the corresponding water. Scales were removed from various locations values in scale or fin tissues. Statistical analyses were above and below the lateral line and dried at room done with the Statistical Package for the Social temperature for 48 h. After drying, whole scales were Sciences (SPSS 2003). Throughout the rest of this filed using a slim taper file. To obtain the muscle tissue, paper, ‘‘sunfish’’ denotes all three species of sunfish the flesh was cut on the left side of the fish under the from all four sites and ‘‘slimy sculpin’’ denotes the fish dorsal fin and the skin was removed. A 2.5- 3 1.0-cm from both sites. rectangle of muscle tissue was excised, dried at 608C for 48 h, and ground with a mortar and pestle. Fin tissue Results from the slimy sculpin was obtained by taking a small Both sunfish and slimy sculpin showed strong sample of caudal tissue with scissors; the tissue was correlations in d13C and d15N values between their MANAGEMENT BRIEF 923

13 FIGURE 1.—Scale versus muscle tissue d C values in 15 sunfish (pooled data for bluegills, pumpkinseeds, and FIGURE 2.—Scale versus muscle tissue d N values in d13 sunfish (pooled data for bluegills, pumpkinseeds, and redbreast sunfish) and caudal fin versus muscle tissue C 15 values in slimy sculpin. The line associated with data points redbreast sunfish) and caudal fin versus muscle tissue d N represents the fit of linear regression for the three sunfish values in slimy sculpin. The line associated with data points species (y ¼ 1.1673x þ 1.0531) and for slimy sculpin (y ¼ represents the fit of the linear regression for the three sunfish 0.6969x – 8.2162). species (y ¼ 0.8504x þ 2.6698) and for slimy sculpin (y ¼ 1.2658x – 3.3234). muscle and scale or fin tissue (Figures 1, 2). Linear regression models for all relationships tested (sunfish sculpin could be used interchangeably without a 13 13 correction factor. Researchers will need to sacrifice a d C: F ¼ 530, p , 0.001; slimy sculpin d C: F ¼ 69, p , 0.001; sunfish d15N: F ¼ 258, p , 0.001; slimy few fish to determine the actual conversion factor for 15 their individual species and locations. sculpin d N: F ¼ 122, p , 0.001) were highly 13 significant. Regression slopes were tested for equality The high correlation of the sunfish d C values is with a 1:1 ratio (slope ¼ 1.0) between muscle tissue notable because this data set represents the values of values and the scale or fin tissue values. All slopes three species of sunfish from four ponds, two in the were significantly different from 1.0 (z-test: p , 0.05). USA and two in Canada (Figure 1). Literature values To test the hypothesis that correction factors were for anadromous, saltwater, and freshwater fishes from necessary to convert the values of scale or fin tissues to wide-ranging locations show similar mean increases in 13 the values of muscle tissues, paired t-tests were scale d C values over muscle values (Table 1). The 13 performed on d13C and d15N values of slimy sculpin difference in d C between these two tissues is 13 and sunfish. Both groups of fish showed significant probably a result of the discrimination of Cin differences in the values of scale or fin and muscle lipogenesis because of the substantial lipid content in tissues: sunfish d13C(p , 0.001), sunfish d15N(p , muscle tissue and the lack of lipid in scale tissue 0.001), slimy sculpin d13C(p , 0.005), and slimy (DeNiro and Epstein 1977; Jobling et al. 1998). sculpin d15N(p , 0.001). Since the slopes of the Unlike scales, fin tissue is a mixture of bone, muscle, regression model lines did not equal 1.0 for either and cartilage and would be expected to contain some 13 group of fish or for either isotope measured, a single lipid, but not as much as muscle tissue. The d C correction factor could not be used for each group of values of slimy sculpin fin tissue, however, were very predicted values. The regression equations were used close to those of the muscle tissue. Although there was 13 to predict the isotope values of the muscle tissues from a statistical difference between the d C values of fin the isotope values of the scale and fin tissues (sunfish and muscle tissue in slimy sculpin, the difference was d13C: y ¼ 1.1673x þ 1.0531; slimy sculpin d13C: y ¼ small and the use of a correction factor is probably not 0.6969x þ 8.2162; sunfish d15N: y ¼ 0.8504x þ 2.6698; warranted. This pattern of comparable values between slimy sculpin d15N: y ¼ 1.2658x – 3.3234). fin and muscle tissue has held up over a variety of species (Rounick and Hicks 1985; McCarthy and Discussion Waldron 2000; Jardine et al. 2005; Table 1). The values of d13C and d15N in the scale and muscle Bluegill, pumpkinseed, and redbreast sunfish scale tissues of sunfish and in the fin and muscle tissues of d15N values can be used for SIA with a correction slimy sculpin were strongly correlated, leading us to factor, but other species may fractionate 15N between conclude that the values of these stable isotopes in muscle and scale differently. Our values between scale scale and fin tissues can be substituted for the values in and muscle tissues in d15N were consistent with results muscle tissues in these and other species of fish. for whitefish Coregonus lavaretus (Perga and Ger- Correction factors were needed in three of the four deaux 2003), but significantly differed from published categories we tested; only the d13C values in the slimy values for sockeye salmon Oncorhynchus nerka, chum 924 KELLY ET AL.

13 15 TABLE 1.—Literature results on the differences between d C and d N values in scale or fin and muscle tissues of fish. Values are means 6 SDs except where noted with SE. Values without parentheses are our calculations; values in parentheses denote other researchers’ reported data, when scales were acid washed or muscle tissue was lipid normalized.

Species Location n d13C(%) d15N(%)

Scale–muscle Bluegill, pumpkinseed and redbreast sunfish Lepomis macrochirus, L. gibbosus, and L. auritusa Massachusetts and New Brunswick 38 2.7 6 0.7 1.3 6 0.6 Whitefish Coregonus lavaretusbc European lakes 144 2.7 (4.0) 1.5 (0.2) Sockeye salmon Oncorhynchus nerkade Alaska 89 3.0 (3.7 6 0.5) 0.1 6 0.6 Chum salmon O. ketade Southeast Alaska 25 3.2 (3.9 6 0.7) 0.2 6 0.6 Pink salmon O. gorbuschade Southeast Alaska 23 2.6 (3.2 6 0.6) 0.2 6 0.5 Chinook salmon O. tshawytschade Southeast Alaska 15 3.0 (3.7 6 0.4) 0.1 6 0.6 Coho salmon O. kisutchde Southeast Alaska 12 2.9 (3.6 6 0.3) 0.4 6 0.4 Cui-ui Chasmistes cujusf Nevada 6 2.6 6 0.4 0.2 6 1.4 Haddock Melanogrammus aeglefinusgh Georges Bank 2 3.2 6 0.6 (SE) 0.0 6 0.2 (SE) Redfinned bully Gobiomorphus huttonii Hutt River, New Zealand 2 0.5 Upland bully G. brevicepsi Hutt River, New Zealand 2 0.4 Fin–muscle Slimy sculpin Cottus cognatusa Canadian rivers 30 0.3 6 0.5 0.5 6 0.5 Brown trout Salmo truttaj European rivers 38 0.1 0.4 Brook trout Salvelinus fontinalisk Pointe Wolfe River, Canada 28 0.7 6 0.9 0.2 6 0.5 Atlantic salmon Salmo salark Miramichi River, Canada 24 0.2 6 0.4 0.8 6 0.4 Redfinned bully G. huttoni i Hutt River, New Zealand 2 0.2 Upland bully G. brevicepsi Hutt River, New Zealand 2 0.3 Longfinned eel Anguilla dieffenbachii i Hutt River, New Zealand 2 0.3 Shortfinned eel A. australisi Hutt River, New Zealand 2 0.1 a This study. b Perga and Gerdeaux (2003). c Values reported in the Perga and Gerdeaux paper for the differences between scale and muscle tissues were 4.0% and 0.2% for d13C and d15N, respectively. They acid-washed their scales to remove carbonates, which increased d13C and d15N values by 1.3%. We adjusted their values by 1.3% to compare with our scales, which were not acid washed. d Satterfield and Finney (2002). e Satterfield and Finney (2002) lipid-normalized their muscle values, which increased the d13C by 0.7%. We subtracted 0.7% from their d13C values to compare with our muscle samples, which were not lipid normalized. f Estep and Vigg (1985). g Wainwright et al. (1993). h Composite samples of five fish. i Rounick and Hicks (1985). j McCarthy and Waldron (2000). k Jardine et al. (2005). salmon O. keta, pink salmon O. gorbuscha, Chinook Our results, taken with those from previous studies salmon O. tshawytscha, coho salmon O. kisutch, cui-ui (Estep and Vigg 1985; Rounick and Hicks 1985; Chasmistes cujus, and haddock Melanogrammus Wainwright et al. 1993; McCarthy and Waldron 2000; aeglefinus (Estep and Vigg 1985; Wainwright et al. Satterfield and Finney 2002; Perga and Gerdeaux 1993; Satterfield and Finney 2002; Table 1). Satterfield 2003; Jardine et al. 2005), indicate that slimy sculpin and Finney (2002) used only the outer annulus of the fin tissue and sunfish scale tissue can be used in place scale in their analysis of the salmon species and Estep of muscle tissue for isotope analysis with the correction and Vigg (1985) had a small sample size of six. These factor determined from the linear regression equation factors may have accounted for the less-pronounced stated. Similar scale–muscle and fin–muscle relation- differences between scale and muscle d15N. ships are expected for other fishes but may initially Factors that affect the fractionation of 15N may be require comparison of tissue before extrapolation. location specific. Both sunfish and slimy sculpin had smaller differences between isotope levels in tissues at Given that scale and fin tissues have been sampled in higher d15N values (Figure 2). A more complete a variety of species with relatively little effect on understanding of the processes that cause the fraction- growth and survival (Gjerde and Refstie 1988; Con- ation of 15N in fish would help us explain these over and Sheehan 1999; Tyus et al. 1999; Pratt and Fox differences in locations. We conclude that slimy 2002), we advocate the use of these tissues in dietary sculpin fin values of d15N can be substituted for investigations using SIA for both 13C and 15N. Further muscle values with a correction factor. study is required to understand the factors that affect MANAGEMENT BRIEF 925 the fractionation of 15N in organisms and why that composition. Fish Physiology and Biochemistry fractionation may differ with location. 18:225–240. Macko, S. A., W. Y. Lee, and P. L. Parker. 1982. Nitrogen and Acknowledgments carbon isotope fractionation by two species of marine amphipods: laboratory and field studies. Journal of We thank Shaheen Kanchwala, Jessica Thomas, Experimental Marine Biology and Ecology 63:145–149. Hallie Lee, and Ray Forget for collecting fish in the McCarthy, I. D., and S. Waldron. 2000. Identifying migratory Massachusetts ponds and helping with the processing of Salmo trutta using carbon and nitrogen stable isotope samples. We thank Solange Brault for her consultation ratios. Rapid Communications in Mass Spectrometry 14:1325–1331. on the statistical analyses. Support for this study came Minagawa, M., and E. Wada. 1984. Stepwise enrichment of from National Science Foundation Research Experience d15N along food chains: further evidence and the relation for Undergraduates grant number DBI-0097685 (Jind- between d15N and animal age. Geochimica et Cosmochi- hart, University of Massachusetts, Boston); University mica Acta 48:1135–1140. of Massachusetts, Boston Undergraduate Research Perga, M. E., and D. Gerdeaux. 2003. Using the 13C and 15N Grant 2002 (M.H.K.); Fulbright Fellowship 2003 of whitefish scales for retrospective ecological studies: changes in isotopic signatures during the restoration of (W.G.H.); Contribution Number 83 of the Catamaran Lake Geneva. Journal of Fish Biology 63:1197–1207. Brook Research Project; and the Canada Research Peterson, B. J., and B. Fry. 1987. Stable isotopes in ecosystem Chairs Program (R.A.C.). Aaron Fraser, Alex Ready, studies. Annual Review of Ecology and Systematics and Dave Parsons assisted in slimy sculpin collections. 18:293–320. Anne McGeachy and Mireille Savoie conducted isotope Pinnegar, J. K., and N. V. C. Polunin. 1999. Differential 13 15 analyses at SINLAB. fractionation of d C and d N among fish tissues: implications for the study of trophic interactions. References Functional Ecology 13:225–231. Post, D. M. 2002. Using stable isotopes to estimate trophic Bunn, S. E., N. R. Loneragan, and M. A. Kempster. 1995. position: models, methods, and assumptions. Ecology Effects of acid washing on stable isotope ratios of C and 83:703–718. N in penaeid and seagrass: implications for food- Pratt, T. C., and M. G. Fox. 2002. Effect of fin clipping on web studies using multiple stable isotopes. Limnology overwinter growth and survival of age-0 walleyes. North and Oceanography 40:622–625. American Journal of Fisheries Management 22:1290– Conover, G. A., and R. J. Sheehan. 1999. Survival, growth, 1294. and persistence in juvenile black crappies marked Rounick, J. S., and B. J. Hicks. 1985. The stable carbon with fin clips, freeze bands, or oxytetracycline. North isotope ratios of fish and their invertebrate prey in four American Journal of Fisheries Management 19:824–827. New Zealand rivers. Freshwater Biology 15:207–214. DeNiro, M. J., and S. Epstein. 1977. Mechanism of carbon Satterfield, F. R., and B. P. Finney. 2002. Stable isotope isotope fractionation associated with lipid synthesis. analysis of Pacific salmon: insight into trophic status and Science 197:261–263. oceanographic conditions over the last 30 years. Progress Estep, M. L., and S. Vigg. 1985. Stable carbon and nitrogen in Oceanography 53:231–246. isotope tracers of trophic dynamics in natural populations Shannon, J. P., D. W. Blinn, G. A. Haden, E. P. Benenati, and and fisheries of the Lahontan Lake system, Nevada. K. P. Wilson. 2001. Food web implications of d13C and Canadian Journal of Fisheries and Aquatic Sciences d15N variability over 370 km of the regulated Colorado 42:1712–1719. River USA. Isotopes in Environmental and Health Gjerde, B., and T. Refstie. 1988. The effect of fin-clipping on Studies 37:179–191. growth rate, survival and sexual maturity of rainbow SPSS. 2003. SPSS for Windows, release 6.0. SPSS, Chicago. trout. Aquaculture 73:383–389. Tyus, H. M., W. C. Starnes, C. A. Karp, and J. F. Saunders. Hobson, K. A., and R. G. Clark. 1993. Turnover of 13C in 1999. Effects of invasive tissue collection on rainbow cellular and plasma fractions of blood: implications for trout, razorback sucker, and bonytail chub. North nondestructive sampling in avian dietary studies. Auk American Journal of Fisheries Management 19:848–855. 110:638–641. Vander Zanden, M. J., G. Cabana, and J. B. Rasmussen. 1997. Jardine, T. D., M. A. Gray, S. M. McWilliam, and R. A. Comparing trophic position of freshwater fish calculated Cunjak. 2005. Stable isotope variability in tissues of using stable nitrogen isotope ratios (del 15N) and temperate stream fishes. Transactions of the American literature dietary data. Canadian Journal of Fisheries Fisheries Society 134:1103–1110. and Aquatic Sciences 54:1142–1158. Jobling, M., S. J. S. Johansen, M. Foshaung, I. C. Burkow, Wainwright, S. C., M. J. Fogarty, R. C. Greenfield, and B. and E. H. Jorgensen. 1998. Lipid dynamics in anadro- Fry. 1993. Long-term changes in the Georges Bank food mous Arctic charr, Salvelinus alpinus (L.): seasonal web: trends in stable isotope compositions of fish scales. variations in lipid storage depots and lipid class Marine Biology 115:481–493.