This article was downloaded by: [Department Of Fisheries] On: 25 November 2013, At: 22:50 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Transactions of the American Fisheries Society Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/utaf20 Fish Assemblages in the Oligohaline Stretch of a Southwest Florida River during Periods of Extreme Freshwater Inflow Variation Philip W. Stevens a , Marin F. D. Greenwood b c & David A. Blewett a a Florida Fish and Wildlife Conservation Commission, Fish and Wildlife Research Institute, Charlotte Harbor Field Laboratory , 585 Prineville Street , Port Charlotte , Florida , 33954 , USA b Florida Fish and Wildlife Conservation Commission, Fish and Wildlife Research Institute , 100 8th Avenue Southeast , St. Petersburg , Florida , 33701 , USA c ICF International , 630 K Street, Sacramento , California , 95814 , USA Published online: 20 Nov 2013.

To cite this article: Philip W. Stevens , Marin F. D. Greenwood & David A. Blewett (2013) Fish Assemblages in the Oligohaline Stretch of a Southwest Florida River during Periods of Extreme Freshwater Inflow Variation, Transactions of the American Fisheries Society, 142:6, 1644-1658, DOI: 10.1080/00028487.2013.824920 To link to this article: http://dx.doi.org/10.1080/00028487.2013.824920

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions Transactions of the American Fisheries Society 142:1644–1658, 2013 C American Fisheries Society 2013 ISSN: 0002-8487 print / 1548-8659 online DOI: 10.1080/00028487.2013.824920

ARTICLE

Fish Assemblages in the Oligohaline Stretch of a Southwest Florida River during Periods of Extreme Freshwater Inflow Variation

Philip W. Stevens* Florida Fish and Wildlife Conservation Commission, Fish and Wildlife Research Institute, Charlotte Harbor Field Laboratory, 585 Prineville Street, Port Charlotte, Florida 33954, USA Marin F. D. Greenwood1 Florida Fish and Wildlife Conservation Commission, Fish and Wildlife Research Institute, 100 8th Avenue Southeast, St. Petersburg, Florida 33701, USA David A. Blewett Florida Fish and Wildlife Conservation Commission, Fish and Wildlife Research Institute, Charlotte Harbor Field Laboratory, 585 Prineville Street, Port Charlotte, Florida 33954, USA

Abstract Maintenance or restoration of the oligohaline stretch (i.e., salinity 0.5–5 psu) of coastal rivers is becoming an increasingly important goal of water managers striving to balance human consumption of water with the ecological integrity of estuaries. The objectives of this study were to compare fish assemblage structure and species abundance of the oligohaline stretch to those of the lower river mouth during periods of varying freshwater inflow (wet and dry periods) in one of southwest Florida’s largest rivers, the Peace River. The abundance of several estuarine residents and estuarine transients captured in 21.3-m seines—Sand Seatrout Cynoscion arenarius, Tidewater Mojarra Eucinostomus harengulus, Red Drum Sciaenops ocellatus, and Spot Leiostomus xanthurus—were similar between river sections, which is consistent with the premise that the oligohaline stretch is an extension of the juvenile habitat known to be important for fish in lower rivers. Estuarine residents known to have affinities for marsh habitat—Eastern Mosquitofish Gambusia holbrooki, Rainwater Killifish Lucania parva, and Sailfin Molly Poecilia latipinna—were at least an order of magnitude more abundant in the oligohaline stretch, probably the result of greater production at low salinity, greater marsh area, or less competition. During a dry period, the oligohaline fish assemblage became more similar to the assemblage of the lower river mouth. Reductions in the abundance of species characteristic of the oligohaline stretch were offset by increases in the abundance of Bay Anchovy Anchoa mitchilli. This study provides

Downloaded by [Department Of Fisheries] at 22:50 25 November 2013 information to managers that can be used in the restoration of oligohaline waters by identifying characteristic fishes in the oligohaline stretch of a large river, providing insight into how this river section functions as fish habitat, and determining the changes in fish assemblages that occur during low freshwater inflow conditions.

The oligohaline zone of an estuary is the region where increase exponentially in this salinity range, resulting in unique freshwater first meets and starts to mix with the saline water physical, chemical, and biological processes (Deaton and of the ocean; it is defined as water having a salinity of 0.5–5.0 Greenberg 1986). For example, freshwater diatoms that have psu (Anonymous 1958). Ion ratios of the water (e.g., Ca:Na) depleted silica in the water, a limiting nutrient for diatom

*Corresponding author: [email protected] 1Present address: ICF International, 630 K Street, Sacramento, California 95814, USA. Received November 15, 2012; accepted July 9, 2013

1644 FISH ASSEMBLAGES OF AN OLIGOHALINE RIVER STRETCH 1645

growth, are lysed as they enter higher-salinity water down- and scientists (Ogden et al. 2005). Restoring the oligohaline stream, a process that accelerates remineralization of silica zone and the fish assemblages that inhabit the areas between in the estuary (Anderson 1986). Watershed-derived inorganic freshwater and coastal mangroves of the Florida Everglades is nitrogen is rapidly processed where a river first meets the a primary goal of the Comprehensive Everglades Restoration estuary, which affects denitrification rates and the availability Plan (Davis et al. 2005). In river systems, the position of low- of nitrogen to estuarine organisms downstream (Holmes salinity water is used as a gauge for meeting the state of Florida’s et al. 2000; Hughes et al. 2000; Merrill and Cornwell 2002). mandates for minimum flows and water levels in several south- Vegetation communities in the lower-salinity reaches of an west Florida river systems, such as the Caloosahatchee, Hills- estuarine river are distinct from those of other sections (Latham borough, Myakka, and Peace rivers (SFWMD 2010a, 2010b). et al. 1994; Visser et al. 1998). Fish diversity is generally low The premise is that maintenance or restoration of the oligoha- at salinities approaching zero, but diversity increases markedly line zone will ensure that ecologically meaningful isohalines as freshwater, estuarine, and marine species overlap (Whitfield are produced downstream, promoting the general health of es- et al. 2012). Similarly, fish assemblage structure changes tuarine habitats such as oyster reefs and seagrass beds. Even rapidly between salinities of 0.1 and 1 psu, in contrast to the though the position of the oligohaline zone is increasingly used gradual change along the remainder of the estuarine salinity as a management target, recent river-specific studies that char- gradient (Greenwood 2007). This rapid change is probably a acterize oligohaline fish assemblages are few (but see Catalano result of changes in river morphology and vegetation (Stevens et al. 2006; Greenwood et al. 2007; Rehage and Loftus 2007; et al. 2010) and of significant changes in ionic ratios, to which Stevens et al. 2010). The purpose of this study was to evaluate many fishes are not well adapted (Deaton and Greenberg 1986). differences between fish assemblages of the oligohaline stretch The oligohaline zone is an extension of habitat for various (the section of river where low salinity water most often occurs) estuarine fish guilds (estuarine-use functional guilds are similar and the lower river mouth during periods of varying freshwa- to those described in Elliott et al. 2007). Several estuarine ter inflow (wet and dry periods) in one of southwest Florida’s residents (those species that can complete their life cycle within largest rivers, the Peace River. By comparing fish assemblages in the estuary) and estuarine transients (those species that spawn these two river sections, hypotheses regarding the structure and offshore and enter estuaries in large numbers, particularly function of oligohaline fish assemblages can be tested. Specifi- as juveniles), such as juvenile Spot Leiostomus xanthurus, cally, based on the literature presented above, we predicted that Atlantic Menhaden Brevoortia tyrannus, and juvenile Weakfish (1) estuarine residents and estuarine transients are as abundant in Cynoscion regalis, are as abundant in the oligohaline portions of the oligohaline stretch as they are in the lower river mouth, and temperate estuaries as they are in higher-salinity regions (Rozas (2) estuarine residents that have strong affinities for marsh habi- and Hackney 1983; Rogers et al. 1984). Estuarine residents tat are more abundant in the oligohaline stretch than in the lower known to have affinities for marsh habitat, such as Eastern river mouth. Mosquitofish Gambusia holbrooki, Rainwater Killifish Lucania parva, and Sailfin Molly Poecilia latipinna, are abundant in oligohaline zones (Brown-Peterson and Peterson 1990; Lorenz STUDY SITE and Serafy 2006; Rehage and Loftus 2007). Water management The Peace River is a 182-km-long river that originates in practices that reduce the size and volume of the oligohaline central Florida and flows southwest into the Charlotte Harbor zone can lead to a reduction in numbers of estuarine transients estuary (Figure 1). Once the river leaves its headwaters at Lake (Serafy et al. 1997) and of estuarine residents that support Hancock, no artificial structures (e.g., weirs, dams) impede its higher-level consumers (Lorenz and Serafy 2006). flow. For this study, the oligohaline stretch of the Peace River is

Downloaded by [Department Of Fisheries] at 22:50 25 November 2013 In recent years, maintaining or restoring the oligohaline zone defined on the basis of changes in river morphology and of veg- has become a primary goal of water managers who must balance etation typically found in low-salinity (<5 psu) water. As part of human consumption of freshwater with the ecological integrity regulatory requirements of water withdrawal in the lower Peace of estuaries. In California’s San Francisco Estuary, for example, River, riparian vegetation was mapped using several decades water managers recognize that the abundance and survival of of aerial photography and it was found that the position of many estuarine species are correlated with the position of the various vegetation types correlated well with long-term salinity oligohaline zone, and specific requirements are placed on estu- conditions (R. Montgomery, Atkins Global, unpublished data; ary inflow to maintain that position seaward of certain locations see also similar results from other river systems: Latham et al. (Kimmerer 2002). The precise mechanisms that make the po- 1994; Visser et al. 1998). The use of river morphology and sition of the oligohaline zone so important are not known; one vegetation as indicators of low-salinity conditions was thought possibility is that habitat size (area or volume) changes as the to provide a better measure of the general placement of low- oligohaline zone moves upstream or downstream (Kimmerer salinity water than the use of situ salinity measurements, which et al. 2009; Peebles and Greenwood 2009). can be variable in the short term. As the river winds around a The oligohaline zone is well recognized in several conceptual large bend in the lower river at latitude 27◦ 04.9 N, the vege- models of Florida estuaries developed by resource managers tation changes abruptly from hardwoods such as bald cypress 1646 STEVENS ET AL.

FIGURE 1. The Peace River, Florida, showing river sections (oligohaline stretch and lower river mouth) used in the analysis.

Taxodium distichum to oligohaline plants such as bulrush 1999 (termed the “wet study period”) and July 2007–April 2010 Scirpus spp., soft rush Juncus effusus, giant leather fern (termed the “dry study period”). The two study periods were Acrostichum danaeifolium, and black needlerush Juncus represented by very different freshwater flow regimes (Figure 2). roemerianus. This abrupt transition was defined as the upriver An El Nino˜ event (1997–98) marked by high rainfall and fresh- boundary of the oligohaline stretch. This part of the river is water inflow characterized the early part of the wet study period. relatively narrow (<0.2 km) and has steep banks along its outer The dry study period began during a severe drought (2007) bends. Downstream from the confluence with Shell Creek, marked by record low water levels in the Peace River and ended a major tributary, the river abruptly widens to >1.5 km, the in 2010 during more typical rainfall and inflow conditions. Downloaded by [Department Of Fisheries] at 22:50 25 November 2013 littoral zone is characterized by expansive, shallow flats with The two studies provided an opportunity to characterize Peace water < 1.5 m deep, and the vegetation becomes dominated River fish assemblages during periods of varying freshwater by red mangrove Rhizophora mangle. This point (latitude, inflow. 26◦ 58.5 N) defined the downstream extent of the oligohaline Monthly stratified-random sampling was conducted in the stretch. Below this point, the river widens toward its mouth, oligohaline stretch by using a small center-bag seine (21.3 m where two cities, Punta Gorda and Port Charlotte, are located; long, 1.8 m deep, 3 mm stretch mesh). Six sampling locations portions of the shoreline have been hardened with seawalls, were chosen randomly each month from numerous possible sites but much of the native vegetation remains. Submerged aquatic that contained shoreline along the main stem of the river (n = vegetation is largely absent from both the oligohaline stretch 147 total samples in the wet study period, 204 total samples in and the lower river mouth. the dry study period). To ensure that the six sampling sites did not occur too closely together in a given month and in an effort METHODS to spread the samples along the river, the sites were selected Field Sampling.—Fishes were sampled in the oligohaline from three subzones of equal length (two sites per subzone). stretch of the Peace River in two studies, during April 1997–May These subzones were used for site selection purposes only and FISH ASSEMBLAGES OF AN OLIGOHALINE RIVER STRETCH 1647

Wet study period Dry study period 700

600 ) -1

s 500 3

400

300 River flow (m River flow 200

100

0 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Year

FIGURE 2. River flow at the U.S. Geological Survey Arcadia staff gauge in the lower Peace River, Florida, during the two study periods, April 1997–May1999 (termed the “wet study period”) and July 2007–April 2010 (termed the “dry study period”).

were not used in the analysis. The seine was deployed from a we included two swimming invertebrates, blue crab Callinectes boat in a shallow arc along the shore in depths suitable for the sapidus and pink shrimp Farfantepenaeus duorarum,intheterm gear (0.3–1.8 m at the center bag). The two ends of the seine fish assemblage. were pulled together while keeping the net along the shore (e.g., Due to frequent hybridization and extreme difficulty in iden- mangrove, leatherfern, seawall), sampling an area of ∼68 m2. tification of smaller individuals, some fishes were identified To compare fish assemblages of the oligohaline stretch with to genus only. These fishes included menhaden species of the those downriver, the periods corresponding to the oligohaline genus Brevoortia (B. patronus, B. smithi, and hybrids), silver- studies (April 1997–May 1999 and July 2007–April 2010) were side species of the genus Menidia (M. peninsulae, M. beryllina, selected from a long-term (1996–2010) database of 21.3-m seine and hybrids), small individuals of the mojarra genus Eucinosto- collections from the lower mouth of the river (Idelberger and mus (<40 mm SL), the goby genus Gobiosoma (<20 mm SL), Greenwood 2005). In the lower river mouth, four sampling loca- and the sunfish genus Lepomis (<20 mm SL). These fishes were tions were chosen randomly each month from numerous possi- included and treated as individual species in analyses because ble sites that contained shoreline along the main stem of the river of their great abundance in the study area, but caution should be (n = 109 total samples in the wet study period, 136 total sam- used in interpreting these results because members of these gen- ples in the dry study period). The sites in the lower mouth were era may have different habitat and physicochemical affinities. stratified by overhanging (e.g., mangrove) and nonoverhanging Further details of site selection, sampling techniques, and tax-

Downloaded by [Department Of Fisheries] at 22:50 25 November 2013 (e.g., salt marsh) shoreline at a ratio of 3:1, which approximated onomic groupings can be found in Idelberger and Greenwood the ratio of available shoreline habitats that could be sampled. (2005). Gear deployment was identical to that for the sampling in the Data Analysis.—Water conditions measured during fish sam- oligohaline stretch. pling were compared between river sections (oligohaline, lower Each time a seine was deployed, water conditions— mouth) and study periods (wet, dry) using ANOVA. The physic- ◦ temperature ( C), salinity (psu), and dissolved oxygen (mg/L)— ochemical variables were loge transformed to improve normal- were profiled with a water quality datasonde (measurements ity and better satisfy ANOVA assumptions. Comparisons were taken at 0.2 m, 1.0 m if the water was deep enough, and the bot- deemed significant at P < 0.05. Significant differences were tom). The profiles were averaged at each site for analysis. Fishes further analyzed using Tukey’s pairwise comparisons. collected in each sample were identified to the lowest practical Spatial patterns in fish assemblage structure were analyzed taxonomic level (nomenclature for fishes follows Page et al. using multivariate techniques. The factors of interest included 2013), measured (SL), enumerated, and released in the field. To river section (oligohaline stretch and lower river mouth) and verify field identification, representative subsamples of organ- study period (wet and dry). Sample abundance indices for each isms were retained for identification in the laboratory. To avoid species (fish/100 m2) were square-root transformed to reduce introducing additional taxonomic terms throughout this paper, the influence of highly abundant species. Multivariate analyses 1648 STEVENS ET AL.

were conducted separately for three seasons (January–April, ± SD) in the oligohaline stretch during the wet study period May–August, and September–December) approximating those (1.7 ± 2.8 psu) was less than that of dry study period (4.6 defined for fish assemblages of the lower river mouth in Idel- ± 6.0). In the lower river mouth, salinity during the wet berger and Greenwood (2005). These known seasonal differ- study period (7.8 ± 7.1 psu) and the dry study period (14.7 ences in the fish assemblage can mask any spatial effects. Thus, ± 8.7 psu) was greater than that in the oligohaline stretch. conducting the analyses separately for each season allowed for For dissolved oxygen, significant differences between river better isolation of the factors of interest. A two-way analysis of sections and study periods were found by using ANOVA, but similarity (ANOSIM; Clarke and Warwick 2001) was used to the differences were relatively small (≤0.3 mg/L). There were test for differences in fish assemblages by river section and study no significant differences in temperature between river sections period. Nonmetric multidimensional scaling (MDS; Clarke and or study periods (ANOVA: P-values > 0.10). Warwick 2001) was used to graphically depict relative differ- Fish assemblages differed by river section and by study ences in fish assemblages by river section and study period. period within each season (Table 1; Figure 3). Values of R Before performing ANOSIM, Bray–Curtis similarity matrices from ANOSIM ranged from 0.20 to 0.40, indicating that fish were calculated for data averaged by sampling event in an effort assemblages were fairly well separated. In the MDS plots, to include an appropriate level of variability in the statistical test. the fish assemblages of the oligohaline stretch and lower river Before performing MDS, Bray–Curtis similarity matrices were mouth were distinct from one another during the wet study calculated for data averaged by river section and study period period during each of the seasons. However, the oligohaline in an effort to allow for better visual interpretation of the fac- fish assemblages were more similar to those of the lower river tors of interest. Similarity percentage analysis (SIMPER; Clarke mouth during the dry study period than during the wet period, and Warwick 2001) was used to identify species representative particularly during May–August. The oligohaline stretch was of similarities between the groups determined from MDS. All distinguished by a high abundance of estuarine residents, multivariate analyses were conducted with PRIMER version 6 namely G. holbrooki, Lucania parva, and Hogchoker Trinectes (PRIMER-E, Plymouth, UK). maculatus (SIMPER). The lower river mouth was distinguished Analysis of fish abundance was conducted using a gener- by estuarine transients, namely Spot Leiostomus xanthurus and alized linear model (PROC GLIMMIX in SAS version 9.1 Striped Mullet Mugil cephalus, during January–March and by [GLM]; SAS Institute, Cary, North Carolina). For each species, estuarine residents, namely Bay Anchovy Anchoa mitchilli and differences in adjusted (least squares) mean abundance index Sand Seatrout Cynoscion arenarius, during the other seasons (fish/100 m2) were tested between river section, study period, (SIMPER). and their interaction. A negative binomial distribution with log- Abundance of common species differed by river section and link function and an offset of 0.68 (to convert count data to by study period (GLM; Figures 4–6). The estuarine resident density estimates) was chosen to best represent the data. Ad- G. holbrooki, Lucania parva, and Poecilia latipinna were at justed mean abundance indices and associated 95% confidence least an order of magnitude more abundant in the oligohaline limits were back-transformed following analysis. Species cho- stretch of the river. One of these species, L. parva,wasmore sen for comparison were the six most common (total number abundant during the wet study period than during the dry pe- collected) estuarine residents and three common estuarine tran- riod. The estuarine resident T. maculatus was most abundant in sients. The estuarine transients chosen were those known to the oligohaline stretch and during the wet study period. Two be more abundant in upper estuaries and rivers than in estuar- estuarine residents (A. mitchilli and juveniles of C. arenarius) ine areas of high salinity (e.g., Red Drum Sciaenops ocellatus and two estuarine transients (Tidewater Mojarra Eucinostomus rather than Pinfish Lagodon rhomboides). When determining harengulus and S. ocellatus) were equally abundant in the oligo-

Downloaded by [Department Of Fisheries] at 22:50 25 November 2013 the most common species in each fish guild, species complexes haline stretch and the lower river mouth. The abundance of these (e.g., Menidia spp. Eucinostomus spp.) were not considered be- species did not differ between the two study periods, except for cause members within each complex have differing habitat and A. mitchilli, which was more abundant during the dry study physicochemical affinities. For species with distinct recruitment period. A species that was more abundant in the lower river periods (e.g., S. ocellatus), only the months in which the species mouth was the estuarine transient M. cephalus. This species occurred in relatively high abundance were included in the was also more abundant during the dry study period than during analysis. the wet period. Significant interaction terms in the models sug- gested shifts in the distributions of dominant species between river sections depending on freshwater inflow regime. During RESULTS the wet study period, C. arenarius, E. harengulus, and S. ocel- Water conditions between the oligohaline stretch and lower latus were more abundant in the lower river mouth, but during river mouth differed, as did conditions during the two study the dry study period these species were more abundant in the periods. Salinity differed between river sections and study oligohaline stretch. Note that although Leiostomus xanthurus periods (ANOVA: P-values < 0.001; all pairwise comparisons was the most abundant estuarine transient overall, it was not were significant, Tukey’test: P-values < 0.001). Salinity (mean abundant enough during the majority of the study period to FISH ASSEMBLAGES OF AN OLIGOHALINE RIVER STRETCH 1649

TABLE 1. Species and number of fish collected in 21.3-m seines used in the oligohaline stretch and lower mouth of the Peace River, Florida, April 1997–May 1999 (termed the “wet study period”) and July 2007–April 2010 (termed the “dry study period”). n = number of seine sets. Taxa are sorted by abundance.

Lower mouth Oligohaline stretch Wet study Dry study Wet study Dry study period period period period Total Species (n = 109) (n = 136) (n = 147) (n = 204) (n = 596) Bay Anchovy Anchoa mitchilli 40,282 60,979 25,319 38,135 164,715 Silversides Menidia spp. 3,092 3,404 3,415 5,505 15,416 Eastern Mosquitofish Gambusia holbrooki 144 93 5,762 9,155 15,154 Mojarras Eucinostomus spp. 829 2,423 810 5,013 9,075 Rainwater Killifish Lucania parva 405 28 4,114 2,444 6,991 Spot Leiostomus xanthurus 10 4,100 2 2,101 6,213 Hogchoker Trinectes maculatus 179 66 3,760 1,187 5,192 Striped Mullet Mugil cephalus 370 1,949 76 1,551 3,946 Tidewater Mojarra Eucinostomus harengulus 656 694 652 1,381 3,383 Sailfin Molly Poecilia latipinna 30 46 797 1,854 2,727 Seminole Killifish Fundulus seminolis 20 1,730 915 2,665 Sand Seatrout Cynoscion arenarius 708 102 171 313 1,294 Rough Silverside Membras martinica 425 107 44 711 1,287 Menhadens Brevoortia spp. 93 187 239 651 1,170 Pinfish Lagodon rhomboides 178 415 114 449 1,156 Gulf Killifish Fundulus grandis 95 373 22 502 992 Striped Anchovy Anchoa hepsetus 155 502 155 152 964 Naked Goby Gobiosoma bosc 104 99 205 451 859 Gobies Gobiosoma spp. 1 60 163 634 858 Clown Goby Microgobius gulosus 47 87 216 394 744 Silver Perch Bairdiella chrysoura 661 36 32 12 741 Longnose Killifish Fundulus similis 591 138 6 1 736 Pink shrimp Farfantepenaeus duorarum 311 197 50 157 715 Blue crab Callinectes sapidus 259 91 227 111 688 Southern Kingfish Menticirrhus americanus 459 127 10 77 673 Diamond Killifish Adinia xenica 35 252 350 637 Striped Mojarra Eugerres plumieri 48 200 86 298 632 Marsh Killifish Fundulus confluentus 2 20 20 573 615 Red Drum Sciaenops ocellatus 64 70 47 259 440 Scaled Sardine Harengula jaguana 22 4 335 361 Sheepshead Minnow Cyprinodon variegatus 13 47 1 286 347 Silver Jenny Eucinostomus gula 26 240 3 6 275 Downloaded by [Department Of Fisheries] at 22:50 25 November 2013 Atlantic Thread Herring Opisthonema oglinum 118 10 3 73 204 Spotted Seatrout Cynoscion nebulosus 95 32 15 31 173 Leatherjack Oligoplites saurus 63 35 29 32 159 Whirligig Mullet Mugil gyrans 86 44 2 132 Coastal Shiner Notropis petersoni 114 3 117 Lined Sole Achirus lineatus 354151284 Sheepshead Archosargus probatocephalus 1 31 2 39 73 Code Goby Gobiosoma robustum 1265472 Mullets Mugil spp. 69 69 Needlefishes Strongylura spp. 17 26 10 53 Threadfin Shad Dorosoma petenense 151 52 Brook Silverside Labidesthes sicculus 39 13 52 Shiners Notropis spp. 50 50 Ladyfish Elops saurus 7 3 15 24 49 Frillfin Goby Bathygobius soporator 10 14 5 13 42 Redear Sunfish Lepomis microlophus 31 8 39 1650 STEVENS ET AL.

TABLE 1. Continued.

Lower mouth Oligohaline stretch Wet study Dry study Wet study Dry study period period period period Total Species (n = 109) (n = 136) (n = 147) (n = 204) (n = 596) Gulf Pipefish Syngnathus scovelli 13 11 5 10 39 Inshore Lizardfish Synodus foetens 7224 5 38 Bluegill Lepomis macrochirus 328536 Common Snook Centropomus undecimalis 2 18 3 12 35 Blackcheek Tonguefish Symphurus plagiusa 2201 9 32 Timucu Strongylura timucu 22 1 2 25 Crested Goby Lophogobius cyprinoides 19 3 22 Skilletfish Gobiesox strumosus 2144 1 21 Gray Snapper Lutjanus griseus 1111 7 20 Largemouth Bass Micropterus salmoides 11920 Bluefin Killifish Lucania goodei 13 4 17 Redfin Needlefish Strongylura notata 411 1 16 Florida Gar Lepisosteus platyrhincus 17715 Atlantic Stingray Dasyatis sabina 11 1 2 14 Southern Puffer Sphoeroides nephelus 211 1 14 White Mullet Mugil curema 54 312 Chain Pipefish Syngnathus louisianae 515112 Least Killifish Heterandria formosa 2911 Atlantic Needlefish Strongylura marina 4610 Crevalle Jack Caranx hippos 34 7 Mayan Cichlid Cichlasoma urophthalmus 66 Bullhead catfishes Ictaluridae 6 6 Sailfin catfish Loricariidae 1 2 3 6 Lane Snapper Lutjanus synagris 66 Anchovies Anchoa spp. 55 Swamp Darter Etheostoma fusiforme 55 Taillight Shiner Notropis maculatus 41 5 Bighead Searobin Prionotus tribulus 325 Longnose Gar Lepisosteus osseus 22 4 Leopard Searobin Prionotus scitulus 22 4 Tilapias Tilapia spp. 4 4 Atlantic Spadefish Chaetodipterus faber 33

Downloaded by [Department Of Fisheries] at 22:50 25 November 2013 Goldspotted Killifish Floridichthys carpio 33 Spotted Sunfish Lepomis punctatus 33 Hardhead Catfish Ariopsis felis 22 Flagfish Jordanella floridae 22 Golden Shiner Notemigonus crysoleucas 11 2 Gulf Flounder Paralichthys albigutta 22 Smalltooth Sawfish Pristis pectinata 22 Bowfin Amia calva 11 Gafftopsail Catfish Bagre marinus 11 Horse-eye Jack Caranx latus 11 Florida Blenny Chasmodes saburrae 11 Bluntnose Jack Hemicaranx amblyrhynchus 11 African Jewelfish Hemichromis letourneuxi 11 Sunfishes Lepomis spp. 11 Gulf Toadfish Opsanus beta 11 Drums Sciaenidae 1 1 Total 50,797 77,545 48,840 76,399 253,581 FISH ASSEMBLAGES OF AN OLIGOHALINE RIVER STRETCH 1651

January–April 2D Stress: 0 Oligohaline zone Stress < 0.01 Lower mouth wet period ANOSIM Results River section: R = 0.33; p = 0.001 Study period: R = 0.23; p = 0.002 dry period

dry period

wet period

May–August 2D Stress: 0 Oligohaline zone Stress < 0.01 Lower mouth wet period ANOSIM Results River section: R = 0.25; p = 0.002 Study period: R = 0.20; p = 0.005

dry period

dry period

wet period

September–December 2D Stress: 0 Oligohaline zone Stress < 0.01 Lower mouth wet period ANOSIM Results River section: R = 0.40; p = 0.001 Study period: R = 0.30; p = 0.001 Downloaded by [Department Of Fisheries] at 22:50 25 November 2013

dry period

dry period

wet period

FIGURE 3. Two-dimensional nonmetric scaling (MDS) ordination of fish assemblages collected in 21.3-m seines in the oligohaline stretch and lower mouth of the Peace River, Florida, during two study periods (wet and dry). Plots are shown separately by season. The results of two-way ANOSIM comparing river section and study period for each season are given. 1652 STEVENS ET AL.

140 River section Study period River section x study period P < 0.0001 P = 0.2903 P < 0.1434 mosquitofish 120

100 Gambusia holbrooki 80

60

40

20

0

100 River section Study period River section x study period P < 0.0001 P < 0.0001 P = 0.0002 rainwater killifish

± 95 % CL) ± 95 % 80 -2

60 Lucania parva

40

20

0 Adjusted mean density (number·100 m 25 River section Study period River section x study period P < 0.0001 P = 0.6864 P = 0.9387 sailfin molly 20

15 Poecilia latipinna Downloaded by [Department Of Fisheries] at 22:50 25 November 2013

10

5

0 1 2 et Wet Dry w

Oligohaline Lower mouth ligohaline x dry Oligohaline x O Lower mouth xLower wet mouth x dry

FIGURE 4. Mean abundance (± 95% confidence limits) of common estuarine residents with known affinities for marsh habitat in the oligohaline stretch and lower mouth of the Peace River, Florida, during two study periods (wet and dry). FISH ASSEMBLAGES OF AN OLIGOHALINE RIVER STRETCH 1653

1800 River section Study period River section x study period P = 0.3148 P = 0.0073 P = 0.2135 1600 bay anchovy

1400

1200

1000 Anchoa mitchilli

800

600

400

200

0

60 River section Study period River section x study period P < 0.0001 P < 0.0001 P = 0.4226 hogchoker 50 ± 95 % CL) -2

40 Trinectes maculatus

30

20

10

0 Adjusted mean density (number·100 m 40

River section Study period River section x study period sand seatrout P = 0.2836 P = 0.1100 P = 0.0041

30 Cynoscion arenarius Downloaded by [Department Of Fisheries] at 22:50 25 November 2013 20

10

0 1 2 Wet Dry wet dry

Oligohaline Lower mouth Oligohaline x wetOligohaline x Lower mouth xLower mouth x dry

FIGURE 5. Mean abundance (± 95% confidence limits) of the remaining common estuarine residents in the oligohaline stretch and lower mouth of the Peace River, Florida, during two study periods (wet and dry). For Cynoscion arenarius, data were limited to their recruitment period of April–November. 1654 STEVENS ET AL.

20 River section Study period River section x study period P = 0.6097 P = 0.0727 P = 0.0065 Eucinostomus harengulus

15 tidewater mojarra

10

5

0

160 River section Study period River section x study period P = 0.0267 P < 0.0001 P = 0.3746

140 striped mullet ± 95 % CL) -2 120

100

80 Mugil cephalus

60

40

20

0 Adjusted mean density (number·100 m 8 River section Study period River section x study period P = 0.6395 P = 0.1708 P = 0.0289 red drum

6 Sciaenops ocellatus Downloaded by [Department Of Fisheries] at 22:50 25 November 2013 4

2

0 1 t 2 t We Dry th x we Oligohaline Lower mouth r mou owe Oligohaline x wetOligohaline x dry L Lower mouth x dry

FIGURE 6. Mean abundance (± 95% confidence limits) of the common estuarine transients in the oligohaline stretch and lower mouth of the Peace River, Florida, during two study periods (wet and dry). For Sciaenops ocellatus, data were limited to their recruitment period of October–March. For Mugil cephalus, data were limited to their recruitment period of January–April. FISH ASSEMBLAGES OF AN OLIGOHALINE RIVER STRETCH 1655

analyze using the generalized linear model. The abundance of in piscivorous fish abundance in the low-salinity portions of this species (mean ± SE) during a strong recruitment year in the estuary compared with areas downstream. Piscivorous fish 2010 did not differ between the oligohaline stretch (1.3 ± 0.3 abundance, however, does not necessarily reflect predation fish/100 m2) and lower river mouth (4.0 ± 1.8 fish/100 m2; t-test pressure; habitat structural complexity, water clarity, and a suite after loge(x + 1) transformation on abundance: P > 0.4). of other environmental variables can affect predator feeding Trends in total fish abundance in the generalized linear model success (Sheaves 2001). Obvious differences between the two were sensitive to the inclusion of A. mitchilli, a species that oc- areas of the river are vegetation type, river morphology, and curs in great abundance in the study area (Figure 7). For exam- shoreline slope. Although the abundances of some fishes were ple, the trends in total fish abundance reflect those of A. mitchilli; similar between the oligohaline stretch and lower river mouth, abundance did not differ by river section, but abundance was comparative studies determining relative survivorship, growth, greater during the dry study period than during the wet period. and contribution to adult populations for these species in To illustrate this point, total fish abundance excluding anchovies different areas of the estuary are needed to elucidate the relative is also shown. The trends, excluding anchovies, reflect those of nursery function of the oligohaline stretch (Beck et al. 2001). common estuarine residents (namely G. holbrooki and P. latip- The greater abundance of the estuarine-resident G. holbrooki, inna); abundance was greater in the oligohaline stretch and did Lucania parva, and P. latipinna in oligohaline water than in not differ by study period. either freshwater or higher-salinity waters has been well doc- umented (e.g., Brown-Peterson and Peterson 1990; Peterson and Meador 1994; Lorenz and Serafy 2006; Rehage and Loftus DISCUSSION 2007; Martin et al. 2009). These species have broad salinity tol- The oligohaline stretch of the Peace River provided an ex- erances, which is a requirement for surviving conditions in the tension of juvenile habitat for estuarine residents and estuarine oligohaline stretch. The great abundance of these three species transients, which is consistent with studies conducted in tem- in the oligohaline stretch may also result from greater production perate estuaries (Rozas and Hackney 1983; Rogers et al. 1984). at low salinity, greater availability of marsh habitat, or less com- For example, A. mitchilli and juvenile C. arenarius, E. haren- petition. For example, Gambusia spp. and P. latipinna collected gulus, L. xanthurus, and S. ocellatus were as abundant in the from marsh types and salinities similar to those in the present oligohaline stretch as they were in the lower river mouth, an area study had greater reproductive output (GSI and brood size) than known to provide juvenile habitat for these species (Idelberger those collected from tidal freshwater with a salinity of zero and Greenwood 2005). An exception was the estuarine transient (Brown-Peterson and Peterson 1990; Martin et al. 2009), sug- M. cephalus, which was more abundant in the lower river than gesting that the great abundance of these species in low-salinity in the oligohaline stretch. Rogers et al. (1984) also noted that habitat results, in part, from greater production. These species M. cephalus were less abundant in low-salinity waters. Sev- when present in oligohaline water may also benefit from less eral other fishes are known to use the lower Peace River as competition with marine and freshwater taxa. Although both juveniles—Silver Perch B. chrysoura, Southern Kingfish Men- freshwater and estuarine fishes contribute to species richness ticirrus americanus, and Hardhead Catfish Ariopsis felis—but in the oligohaline stretch, obligate freshwater species were un- these species are better represented in trawl samples (Idelberger common; for example, Coastal Shiner Notropis petersoni and and Greenwood 2005), and thus whether the oligohaline stretch Bluegill Lepomis macrochirus were each represented by fewer is an extension of juvenile habitat for them could not be deter- than 120 individuals in this 5-year study. Other common estuar- mined in the present study. ine residents such as Anchoa mitchilli occupy different habitats Juveniles of estuarine residents and estuarine transients and have different diets than those of G. holbrooki, Lucania

Downloaded by [Department Of Fisheries] at 22:50 25 November 2013 are thought to benefit from reduced predation in oligohaline parva, and P. latipinna (Pattillo et al. 1997). The low-salinity water (e.g., Rozas and Hackney 1983; McIvor and Odum marshes that dominate in the oligohaline stretch may favor G. 1988). However, comprehensive studies addressing predation holbrooki, L. parva, and P. latipinna. pressure in different regions of estuaries are lacking (Sheaves The fish assemblage of the oligohaline stretch was affected 2001). Comparing data from haul seines (i.e., 61 m long, by differing freshwater inflow regimes. For example, the oligo- 25.4 mm stretch mesh; 183 m long, 38 mm stretch mesh) haline and lower-river-mouth fish assemblages that had been that targeted larger fishes as part of a fisheries-independent clearly separate during the wet study period became more simi- monitoring program and sampled throughout regions of the lar during the dry study period. The differences were attributed Charlotte Harbor estuary (Winner et al. 2010; Call et al. 2011) partly to a reduction in the estuarine resident L. parva and found the abundance of piscivorous fish in the upper estuary T. maculatus. The abundance of these species depicted in the (mean = 11 fish per 100 m shoreline, SE = 2) to be similar to generalized linear models was significantly lower during the dry that in the oligohaline stretch (mean = 10, SE = 1); both of study period. These changes could have occurred for at least two these river sections had approximately 33% fewer piscivores reasons: (1) the species that characterized the oligohaline stretch than was observed in the lower estuary (mean = 29, SE = 3) (i.e., L. parva and T. maculatus) simply moved upriver out of the (authors’ unpublished data). Thus, there is a general reduction sampling area, or (2) production of these species decreased as 1656 STEVENS ET AL.

River section Study period River section x study period 1400 P = 0.9745 P < 0.0001 P = 0.0233

1200

1000 Total fish

800

600

± 95 % CL) CL) % 95 ± 400 2 -

200

0

350 River section Study period River section x study period P < 0.0001 P = 0.9817 P = 0.3453

300 Total fish exclu

250

200 Adjusted mean density (number·100 m (number·100 mean density Adjusted ding anchovies

150

100

50

0 1 2 Wet Dry wet x wet Downloaded by [Department Of Fisheries] at 22:50 25 November 2013

Oligohaline Lower mouth mouth x dry Oligohaline Oligohaline x dry Lower mouth xLower

FIGURE 7. Mean abundance (± 95% confidence limits) of total fish collected in the oligohaline stretch and lower mouth of the Peace River, Florida, during two study periods (wet and dry). The lower panel excludes anchovies in the calculation of total fish to illustrate the effect of these abundant species on total fish abundance.

a result of increased salinities. We acknowledge the limitations deed moved upriver beyond our study area. If the characteristic of the study design in tracking major shifts in fish distribution oligohaline fishes tracked the upriver movement of isohalines, farther upriver from what is typically the oligohaline stretch. this could explain their reduced abundance in our study area. The isohaline of 6 psu moved 20 km upriver from its long-term If this was the case, fishes that have affinities for marsh habitat mean position for several months during the dry study period (R. (i.e., L. parva) would have left the marshes and moved upriver Montgomery, Atkins Global, unpublished data), and approached into cypress. Such a disconnect between dynamic (salinity) and the upriver boundary of our seining. Thus, oligohaline water in- static (vegetation, geomorphology) habitat can also result in FISH ASSEMBLAGES OF AN OLIGOHALINE RIVER STRETCH 1657

decreased production of the species affected (Sklar and Browder River through adoption of a minimum flow and level. As noted 1998). by Estevez (2002), single-species analyses have tended to be The overall effect of drought in the Peace River appears to more common than community- or assemblage-level assess- be a shift in fauna. The losses of L. parva and T. maculatus ments in estuarine inflow studies; the present study integrated in the oligohaline stretch that occurred during the dry study both methods. Changes in habitat extent of the oligohaline zone period were offset by increases in estuarine transients and other and associated biota form an important component in the deter- estuarine residents, especially A. mitchilli. Also, species that mination of appropriate freshwater inflows to estuaries, along had been more abundant in the lower river mouth during the wet with other factors such as secondary production (e.g., abundance study period (i.e., C. arenarius, E. harengulus, and S. ocellatus) and growth of fish; Purtlebaugh and Allen 2010). In studies that were more abundant in the oligohaline stretch during the dry aim to track fish distributions in relation to changing isohalines, study period. Overall, the number of small fishes available as we recommend sampling much farther upriver than mean condi- prey to large piscivores did not differ between the two river tions indicate would be necessary, to capture the effects of events sections. This result may explain the similarity in the number of such as drought or the long-term effects of water withdrawal. piscivores found in the oligohaline stretch to that of the upper estuary in general. It is possible, however, that the prey types ACKNOWLEDGMENTS available in each of the river sections could be important to some piscivores. Species that specialize in the consumption of We thank the staff of the Charlotte Harbor Field Labo- estuarine residents, for example, would benefit by foraging in ratory for their dedicated assistance in the field. This study the oligohaline stretch. The function of the oligohaline stretch was supported by the Southwest Florida Water Manage- in providing food to piscivores depends on the relative value of ment District (Numbers 08POSOW0490, 08POSOW1743, and the characteristic species as prey. Studies comparing predator 10POWOW0239), a Florida State Wildlife grant (T-13-R-1), feeding success and relative mortality of small fishes in the funds collected from the state of Florida saltwater fishing li- oligohaline stretch and areas farther downstream would provide cense sales, and the U.S. Fish and Wildlife Service, Federal Aid greater insight into the function of estuarine habitats (Sheaves for Sport Fish Restoration Grant Number F-43. We appreciate 2001). the comments of several anonymous reviewers and the editorial Oligohaline regions of estuaries throughout the world are at suggestions of Bland Crowder. risk as water managers search for sources of potable water for an ever-growing human population. The oligohaline stretch in REFERENCES many southwest Florida rivers has been constrained or elim- Anderson, G. F. 1986. Silica, diatoms and a freshwater productivity maximum inated with structures (e.g., dams, locks, weirs) that prevent in Atlantic coastal plain estuaries, Chesapeake Bay. Estuarine, Coastal and movement of the low-salinity water upriver during conditions Shelf Science 22:183–197. of low inflow (Catalano et al. 2006; Peebles and Greenwood Anonymous. 1958. The Venice system for the classification of marine waters according to salinity. Limnology and Oceanography 3:346–347. 2009). Water withdrawals from free-flowing rivers may reduce Beck, M. W., K. L. Heck Jr., K. W. Able, D. L. Childers, D. B. Eggleston, the areal expanse of oligohaline water. Although these losses B. M. Gillanders, B. Halpern, C. G. Hays, K. Hoshino, T. J. Minello, R. J. Orth, can be estimated for both constrained and free-flowing systems P. F. Sheridan, and M. P. Weinstein. 2001. The identification, conservation, based on movement of isohalines, the connection between fish and management of estuarine and marine nurseries for fish and invertebrates. population size and physicochemical habitat quantity is not al- BioScience 51:633–641. Brown-Peterson, N., and M. S. Peterson. 1990. Comparative life history of ways apparent (Kimmerer et al. 2009). In the lower Peace River, female mosquitofish, Gambusia affinis, in tidal freshwater and oligohaline water managers have focused on the oligohaline stretch as their habitats. Environmental Biology of Fishes 27:33–41.

Downloaded by [Department Of Fisheries] at 22:50 25 November 2013 management target (SWFWMD 2010). The minimum flow and Call, M. E., P. W. Stevens, D. A. Blewett, D. R. Sechler, S. Canter, and level on the Peace River is set to allow no more than a 15% T. R. Champeau. 2011. Peace River fish community assessment. Report to decrease in the volume of low-salinity water (defined as a salin- Southwest Florida Water Management District, Florida Fish and Wildlife Conservation Commission, St. Petersburg. ity of 2 psu). Maintenance of low-salinity water was considered Catalano, M. J., M. S. Allen, and D. J. Murie. 2006. Effects of variable flows on to be biologically meaningful, and hydrodynamic model runs water chemistry gradients and fish communities at the Hillsborough River, found that maintaining the area of low-salinity water was more Florida. North American Journal of Fisheries Management 26:108–118. conservative than using higher salinities as a metric. The present Clarke, K. R., and R. M. Warwick. 2001. Change in marine communities: an study identified the characteristic fishes of the river’s oligohaline approach to statistical analysis and interpretation, 2nd edition. PRIMER-E, < Plymouth, UK. stretch (mean salinity was 5 psu during both study periods), Davis, S. M., D. L. Childers, J. J. Lorenz, H. R. Wanless, and T. E. Hopkins. provided insight into how this river section functions as fish 2005. A conceptual model of ecological interactions in the mangrove estuaries habitat, and highlighted the changes in fish assemblages that of the Florida Everglades. Wetlands 25:832–842. occur during low freshwater inflow conditions. The appreciable Deaton, L. E., and M. J. Greenberg. 1986. There is no horohalinicum. Estuaries changes in fish assemblage structure that were observed with 9:20–30. Elliott, M., A. K. Whitfield, I. C. Potter, S. J. M. Blaber, D. P. Cyrus, low freshwater inflow in the present study validate the impor- F. G. Nordlie, and T. D. Harrison. 2007. The guild approach to categorizing tance of maintaining salinity gradients, as is done on the Peace estuarine fish assemblages: a global review. Fish and Fisheries 8:241–268. 1658 STEVENS ET AL.

Estevez, E. D. 2002. Review and assessment of biotic variables and analytical Peebles, E. B., and M. F. D. Greenwood. 2009. Spatial abundance quantiles as a methods used in estuarine inflow studies. Estuaries 25:1291–1303. tool for assessing habitat compression in motile estuarine organisms. Florida Greenwood, M. F. D. 2007. Nekton community change along estuarine salinity Scientist 72:277–288. gradients: can salinity zones be defined? Estuaries and Coasts 30:537–542. Peterson, M. S., and M. R. Meador. 1994. Effects of salinity on freshwater Greenwood, M. F. D., R. E. Matheson Jr., R. H. McMichael Jr., and T. C. fishes in coastal plain drainages in the southeastern U.S. Reviews in Fisheries MacDonald. 2007. Community structure of shoreline nekton in the estuarine Science 2:95–121. portion of the Alafia River, Florida: differences along a salinity gradient and Purtlebaugh, C. H., and M. S. Allen. 2010. Relative abundance, growth, and inflow-related changes. Estuarine, Coastal and Shelf Science 74:223–238. mortality of five age-0 estuarine fishes in relation to discharge of the Suwannee Holmes, R. M., B. J. Peterson, L. A. Deegan, J. E. Hughes, and B. Fry. 2000. River, Florida. Transactions of the American Fisheries Society 139:1233– Nitrogen biogeochemistry in the oligohaline zone of a New England estuary. 1246. Ecology 81:416–432. Rehage, J. S., and W. F. Loftus. 2007. Seasonal fish community variation in Hughes, J. E., L. A. Deegan, B. J. Peterson, R. M. Holmes, and B. Fry. 2000. headwater mangrove creeks in the southwestern Everglades: an examination Nitrogen flow through the food web in the oligohaline zone of a New England of their role as dry-down refuges. Bulletin of Marine Science 80:625–645. estuary. Ecology 81:433–452. Rogers, S. G., T. E. Targett, and S. B. Van Sant. 1984. Fish-nursery use in Geor- Idelberger, C. F., and M. F. D. Greenwood. 2005. Seasonal variation in fish gia salt-marsh estuaries: the influence of springtime freshwater conditions. assemblages within the estuarine portions of the Myakka and Peace rivers, Transactions of the American Fisheries Society 113:595–606. southwest Florida. Gulf of Mexico Science 23:224–240. Rozas, L. P., and C. T. Hackney. 1983. The importance of oligohaline estuarine Kimmerer, W. J. 2002. Physical, biological, and management responses to wetland habitats to fisheries resources. Wetlands 3:77–89. variable freshwater flow into the San Francisco Estuary. Estuaries 25:1275– Serafy, J. E., K. C. Lindeman, T. E. Hopkins, and J. S. Ault. 1997. Effects 1290. of freshwater canal discharge on fish assemblages in a subtropical bay: Kimmerer, W. J., E. S. Gross, and M. L. MacWilliams. 2009. Is the response of field and laboratory observations. Marine Ecology Progress Series 160:161– estuarine nekton to freshwater flow in the San Francisco Estuary explained 172. by variation in habitat volume? Estuaries and Coasts 32:375–389. SFWMD (South Florida Water Management District). 2010a. Requirements Latham, P. J., L. G. Pearlstine, and W. M. Kitchens. 1994. Species association for minimum flows and levels for Florida’s water bodies prevent significant changes across a gradient of freshwater, oligohaline, and mesohaline tidal harm by permitted water withdrawals. SFWMD, West Palm Beach. Available: marshes along the lower Savannah River. Wetlands 14:174–183. my.sfwmd.gov/portal/page/portal/xweb%20protecting%20and%20restoring/ Lorenz, J. J., and J. E. Serafy. 2006. Subtropical wetland fish assemblages and minimum%20flows%20and%20levels%20%28everglades%29. (November changing salinity regimes: implications for Everglades restoration. Hydrobi- 2012). ologia 569:401–422. SFWMD (South Florida Water Management District). 2010b. Minimum flows Martin, S. B., A. T. Hitch, K. M. Purcell, P. L. Klerks, and P. L. Leberg. 2009. and levels (environmental flows). SFWMD, West Palm Beach. Available: Life history variation along a salinity gradient in coastal marshes. Aquatic www.swfwmd.state.fl.us/projects/mfl/mfl reports.php. (November 2012). Biology 8:15–28. Sheaves, M. 2001. Are there really few piscivorous fishes in shallow estuarine McIvor, C. C., and W. E. Odum. 1988. Food, predation risk, and microhabitat habitats? Marine Ecology Progress Series 222:279–290. selection in a marsh fish assemblage. Ecology 69:1341–1351. Sklar, F. H., and J. A. Browder. 1998. Coastal environmental impacts brought Merrill, J. Z., and J. C. Cornwell. 2002. The role of oligohaline marshes in about by alterations to freshwater flow in the Gulf of Mexico. Environmental estuarine nutrient cycling. Pages 425–441 in M. P. Weinstein and D. A. Management 22:547–562. Kreeger, editors. Concepts and controversies in tidal marsh ecology. Kluwer Stevens, P. W., M. F. D. Greenwood, C. F. Idelberger, and D. A. Blewett. 2010. Academic Publishers, Dordrecht, The Netherlands. Mainstem and backwater fish assemblages in the tidal Caloosahatchee River: Ogden, J. C., S. M. Davis, K. J. Jacobs, T. Barnes, and H. E. Fling. 2005. The implications for freshwater inflow studies. Estuaries and Coasts 33:1216– use of conceptual ecological models to guide ecosystem restoration in south 1224. Florida. Wetlands 25:795–809. Visser, J. M., C. E. Sasser, R. H. Chabreck, and R. G. Linscombe. 1998. Marsh Page, L. M., H. Espinosa-Perez,´ L. T. Findley, C. R. Gilbert, R. N. Lea, N. vegetation types of the Mississippi River deltaic plain. Estuaries 21:818–828. E. Mandrak, R. L. Mayden, and J. S. Nelson. 2013. Common and scientific Whitfield, A. K., M. Elliott, A. Basset, S. J. M. Blaber, and R. J. West. 2012. names of fishes from the United States, Canada, and Mexico, 7th edition. Paradigms in estuarine ecology: a review of the Remane diagram with a American Fisheries Society, Special Publication 34, Bethesda, Maryland. suggested revised model for estuaries. Estuarine, Coastal and Shelf Science Pattillo, M. E., T. E. Czapla, D. M. Nelson, and M. E. Monaco. 1997. Distribution 97:78–90. and abundance of fishes and invertebrates in Gulf of Mexico estuaries: vol- Winner, B. L., D. A. Blewett, R. H. McMichael Jr., and C. B. Guenther. 2010. Downloaded by [Department Of Fisheries] at 22:50 25 November 2013 ume II—species life history summaries. National Oceanic and Atmospheric Relative abundance and distribution of Common Snook along shoreline habi- Administration, Strategic Environmental Assessments Division, Estuarine tats of Florida estuaries. Transactions of the American Fisheries Society Living Marine Resources Report 11, Rockville, Maryland. 139:62–79. This article was downloaded by: [Department Of Fisheries] On: 25 November 2013, At: 22:51 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Transactions of the American Fisheries Society Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/utaf20 Introduction to a Special Section: Ecology, Culture, and Management of Burbot Martin A. Stapanian a & Charles P. Madenjian b a U.S. Geological Survey , 6100 Columbus Avenue, Sandusky , Ohio , 44870 , USA b U.S. Geological Survey, Great Lakes Science Center , 1451 Green Road, Ann Arbor , Michigan , 48105 , USA Published online: 20 Nov 2013.

To cite this article: Martin A. Stapanian & Charles P. Madenjian (2013) Introduction to a Special Section: Ecology, Culture, and Management of Burbot, Transactions of the American Fisheries Society, 142:6, 1659-1661, DOI: 10.1080/00028487.2013.837097 To link to this article: http://dx.doi.org/10.1080/00028487.2013.837097

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions Transactions of the American Fisheries Society 142:1659–1661, 2013 American Fisheries Society 2013 ISSN: 0002-8487 print / 1548-8659 online DOI: 10.1080/00028487.2013.837097

SPECIAL SECTION: BURBOT

Introduction to a Special Section: Ecology, Culture, and Management of Burbot

Martin A. Stapanian U.S. Geological Survey, 6100 Columbus Avenue, Sandusky, Ohio 44870, USA Charles P. Madenjian U.S. Geological Survey, Great Lakes Science Center, 1451 Green Road, Ann Arbor, Michigan 48105, USA

The Burbot Lota lota is the only truly freshwater member agement, and culture of Burbot and to identify research needs, of the cod family (Gadidae) and one of only two species of particularly those related to efforts to rehabilitate or restore Bur- freshwater fish that have a circumpolar range (McPhail and bot populations of conservation concern. Lindsey 1970; McPhail and Paragamian 2000). Two subspecies The first four papers (Hardy and Paragamian 2013; Stephen- of Lota lota have been documented: Lota lota maculosa, which son et al. 2013; Barron et al. 2013; Ashton et al. 2013) address is found exclusively in North America from south of Great Slave the formulation of an International Conservation Strategy (ICS) Lake in Canada to the southern limit of its range; and Lota lota by the United States and Canada for Burbot rehabilitation in the lota, which is found over the remainder of the species’ Nearctic Kootenai (Canadian spelling: Kootenay) River and Lake system. range and its entire Eurasian range (Hubbs and Schultz 1941; Shortly after the construction of Libby Dam in 1972, the once Van Houdt et al. 2003). However, many recent authorities (e.g., robust Burbot fishery in the Kootenai system collapsed. Hardy Scott and Crossman 1973) do not designate subspecies. Bur- and Paragamian (2013) review the system’s history, efforts to bot occupy the widest range of depths of all fishes found in rehabilitate the population, and additional management actions the Laurentian Great Lakes basin (i.e., from small streams to needed to rehabilitate the Burbot population to a self-sustaining at least 300 m in Lake Superior; Boyer et al. 1989). World- level. Of particular importance in the ICS are (1) changes in dam wide, many Burbot populations are threatened or endangered operation to remain below the winter water temperature and dis- or have been extirpated (reviewed by Stapanian et al. 2010). charge volume thresholds that provide suitable conditions for Due in part to its unpopularity as a sport and commercial fish Burbot spawning and migration and (2) enhanced aquaculture in much of its range, the species is often ignored in fish man- techniques for stocking Burbot. agement and conservation programs (McPhail and Paragamian Stephenson et al. (2013) monitored hatchery-reared Burbot 2000; Stapanian et al. 2008, 2010). Even basic information on age 1–3 years over a 3-year period in Kootenay Lake and the

Downloaded by [Department Of Fisheries] at 22:51 25 November 2013 Burbot ecology, particularly its early life history and spawning Kootenay River in British Columbia using passive integrated habitats and sites, is lacking. This lack of information is partic- transponder (PIT) and ultrasonic tags. Their results provide ularly troubling because Burbot are an indicator of the health valuable new information on habitat use by juveniles and adults of coldwater systems (Stapanian et al. 2010). Efforts to reha- and on postrelease survival. For example, younger Burbot bilitate or restore imperiled populations include culturing early released in tributaries remained in the tributaries for longer pe- life stages. Burbot larvae are difficult to culture for a variety riods, whereas older Burbot dispersed quickly throughout both of reasons, including their delicate body structure, small size at riverine and lacustrine habitats. These results are encouraging hatch, and live-feed requirement for at least 5 weeks following news for rehabilitating the Kootenay Burbot population. alimentary tract development. Barron et al. (2013) advance the knowledge of culturing The following collection of papers represents a subset of the Burbot through a 2-year study of the survival of larval Bur- presentations at the 4th International Symposium on Burbot, bot stocked at different densities in in-ground outdoor ponds. which was held at the 10th International Congress on the Biol- These authors used an exponential decay model to determine ogy of Fish in Madison, Wisconsin, in 2012. The objectives of the stocking density of larval Burbot that maximizes the num- this symposium were to present research on the ecology, man- ber of surviving juveniles. Although Burbot larvae exhibited a

1659 1660 STAPANIAN AND MADENJIAN

preference for copepods, the total zooplankton densities used Lake Huron. Heretofore, all of the breeding strategies of Bur- by the authors provided sufficient prey for Burbot larvae during bot in the Great Lakes basin were thought to include spawning the culture period. in tributaries or in comparatively shallow water nearshore dur- Ashton et al.’s (2013) study tested the effects of seven types ing winter or early spring. The authors contend that the newly of artificial markers (fin clips, freeze brands, visible implant discovered behavior constitutes an adaptation that improves re- elastomers at two different body locations, and PIT tags) on the cruitment or survival in the face of catastrophic events. survival and growth of juvenile Burbot in the laboratory. They Another paper by Cott et al. (2013a) indicates that Burbot evaluated six criteria, including affordability and ease of use, show little variation in life history traits across environmental for marker selection. Knowledge of the best markers will assist gradients in North America. For example, despite wide differ- in monitoring juvenile Burbot released from hatcheries for the ences in the duration of ice cover over a latitudinal gradient, restoration of imperiled wild populations. spawning time varied little. These results support the conclu- Cott et al.’s (2013b) study provides the first description of the sions of early natural history studies, which suggested that the hearing ability of three size-classes of adult and juvenile Burbot life history traits of Burbot are holdovers from the species’ ma- in a laboratory setting. Sound production and reception are po- rine ancestry. The authors propose that the variability of under- tentially important mating cues for Burbot because the species ice conditions is comparatively low and predictable, allowing is not outwardly sexually dimorphic, mates under conditions of for life history stability among Burbot populations. low light (at night, during winter, and often under ice cover), and Gorman and Sitar’s (2013) paper describes the trends in the is known to vocalize during the spawning period (Cott 2013). abundance of Burbot from 60 years of gill-net surveys in Lake A practical implication may be that anthropogenic noise, such Superior. It discusses the patterns in abundance, such as the as that of ice road traffic in northern latitudes during winter, rapid recovery of Burbot following control of Sea Lampreys disturbs Burbot spawning by masking or interfering with their Petromyzon marinus in the 1960s and the decline of Burbot in vocalizations. the mid-1980s following the recovery of Lake Trout Salvelinus Taylor and Arndt (2013) report on a study of the trends in namaycush, the main predator on Burbot in the lake. The Bur- the abundance of age-0, juvenile, and adult Burbot in Columbia bot’s high fecundity, longevity, and early maturation give it the Lake, British Columbia, for the 1991 to 1999 year-classes. These potential to rebound from low population levels. results suggest that cohort strength is highly variable and inde- The papers in this collection will contribute significantly to pendent of spawning stock size and that survival at the various rehabilitating imperiled or restoring extirpated Burbot popu- life stages is primarily density independent. The exceptional lations, managing established populations, and understanding survival of juveniles from one of the year-classes was associ- the basic biology and ecology of Burbot. Moreover, better ated with unusually high water levels and a temporary increase understanding of the basic biology and ecology of Burbot in interstitial habitat. These results will be useful in predictive may have applications to other sciences. To illustrate, consider models of Burbot rehabilitation. the case of the giant nerve axons of certain species of squid, George et al.’s (2013) paper provides much-needed informa- including opalescent inshore squid Loligo opalescens and tion on the diet of larval Burbot in nearshore and offshore waters bigfin reef squid Sepioteuthis lessoniana, which, unlike those of northern Lake Huron in Michigan. Their results suggest that of most other animal species, were sufficiently large to facilitate larval Burbot have a comparatively inflexible diet. Clearly, this the laboratory study of the basic properties of single nerve cells has important implications for larval Burbot culture, as a narrow as early as the 1930s (Allen et al. 1982; Gilbert et al. 1990; diet could pose a problem in some of the Great Lakes, which Lee et al. 1994). The theory of action potential conduction was have experienced zooplankton declines in recent years. developed based on laboratory investigations using squid axons.

Downloaded by [Department Of Fisheries] at 22:51 25 November 2013 Stapanian et al.’s (2013) paper examines the temporal These basic studies eventually led to medical advances not only patterns in the spatial distribution of Burbot from 15 years in neuroscience but also in physiology (cardiac, circulatory, of gill-net surveys in Lake Erie. Their results support a sensory, and muscle), immunology, molecular biochemistry, hypothesis derived from bioenergetics studies (Madenjian nutritional biochemistry, oncology, aging, and ethology. Burbot 2011; Madenjian et al. 2013), namely, that adult males tend to exhibit several unique characteristics. First, as mentioned, be more active than adult females. Although more males than the Burbot is the only truly freshwater member of the family females were caught in the gill nets in 14 of the 15 years of the Gadidae (Scott and Crossman 1973) and is one of only two study, bottom trawl data indicated a 1:1 sex ratio. The authors freshwater fishes with a circumpolar distribution (McPhail and stress the need for telemetry studies to describe the temporal Lindsey 1970). Second, Burbot appear to be one of the relatively movements of Burbot and assess sex ratios so as to verify the few freshwater fishes that communicate via vocalizations (Cott inferences from active and passive capture gear. 2013). Third, compared with most other fishes, the testes of Jude et al.’s (2013) study provides evidence of a heretofore Burbot are very large; the gonadosomatic indices of male undocumented breeding strategy for Burbot, i.e., spawning dur- Burbot can well exceed 10%, whereas those of the males of ing June and July in unconsolidated and rocky substrates in other fishes are typically well under 5% (Cott et al. 2013b). offshore, hypolimnetic waters of southern Lake Michigan and Perhaps basic research on Burbot biology, especially with INTRODUCTION 1661

regard to the Burbot’s unique characteristics, will eventually Gorman, O. T., and S. P. Sitar. 2013. Ups and downs of Burbot and their preda- not only facilitate the rehabilitation and management of Burbot tor Lake Trout in Lake Superior, 1953–2011. Transactions of the American populations but also lead to advances in other fields of science. Fisheries Society 142:1757–1772. Hardy, R., and V. L. Paragamian. 2013. A synthesis of Kootenai River Burbot stock history and future management goals. Transactions of the American ACKNOWLEDGMENTS Fisheries Society 142:1662–1670. This special section is dedicated to the memory of Adrienne Hubbs, C. L., and L. P. Shultz. 1941. Contributions to the ichthyology of Alaska with description of two new fishes. Occasional Papers of the Museum of M. (Petersen) Stapanian. We thank C. Grimes, R. Nicholson, L. Zoology University of Michigan. Hendee, and the other members of the staff at the Transactions Jude, D. J., Y. Wang, S. R. Hensler, and J. Janssen. 2013. Burbot early life of the American Fisheries Society for their patience and coop- history strategies in the Great Lakes. Transactions of the American Fisheries eration. Martin A. Stapanian and Charles P. Madenjian served Society 142:1733–1745. as associate editors for all of the papers in this section. We also Lee, P. G., P. E. Turk, W. T. Yang, and R. T. Hanlon. 1994. Biological char- acteristics and biomedical applications of the squid Sepioteuthis lessoni- thank the contributing authors and the numerous reviewers for ana cultured through multiple generations. Biological Bulletin 186:328– their efforts, which resulted in an improved collection. P.Kocov- 341. sky, P. Seelbach, C. Grimes, and numerous contributing authors Madenjian, C. P. 2011. Sex effect on polychlorinated biphenyl concentrations provided comments on this paper. The International Congress in fish: a synthesis. Fish and Fisheries 12:451–460. on the Biology of Fish and the University of Wisconsin pro- Madenjian, C. P., M. A. Stapanian, R. R. Rediske, and J. P. O’Keefe. 2013. Sex difference in polychlorinated biphenyl concentrations of Burbot vided the symposium facilities. Use of trade, product, or firm Lota lota from Lake Erie. Archives of Environmental Contamination and names does not imply endorsement by the U.S. Government. Toxicology 65:300–308. This article is contribution 1785 of the U.S. Geological Survey, McPhail, J. D., and C. C. Lindsey. 1970. Freshwater fishes of northwestern Great Lakes Science Center. Canada and Alaska. Bulletin of the Fisheries Research Board of Canada 173. McPhail, J. D., and V. L. Paragamian. 2000. Burbot biology and life history. REFERENCES Pages 11–23 in V. L. Paragamian and D. W. Willis, editors. Burbot: biology, Allen, R. D., J. Metuzals, I. Tasaki, S. T. Brady, and S. P. Gilbert. 1982. Fast ecology, and management. American Fisheries Society, Fisheries Manage- axonal transport in squid giant axon. Science 218:1127–1129. ment Section Publication 1, Spokane, Washington. Ashton, N. K., S. C. Ireland, and K. D. Cain. 2013. Artificial marker selection Scott, W. B., and E. J. Crossman. 1973. Freshwater fishes of Canada. Bulletin and subsequent tagging evaluations with juvenile Burbot. Transactions of the of the Fisheries Research Board of Canada 184. American Fisheries Society 142:1688–1698. Stapanian, M. A., C. P. Madenjian, C. R. Bronte, M. P. Ebener, B. F. Lantry, and Barron,J.M.,N.R.Jensen,P.J.Anders,J.P.Egan,S.C.Ireland,andK.D.Cain. J. D. Stockwell. 2008. Status of Burbot populations in the Laurentian Great 2013. Effects of stocking density on survival and yield of North American Lakes. Pages 111–130 in V.L. Paragamian and D. H. Bennett, editors. Burbot: Burbot reared under semi-intensive conditions. Transactions of the American ecology, management, and culture. American Fisheries Society, Symposium Fisheries Society 142:1680–1687. 59, Bethesda, Maryland. Boyer, L. F., R. A. Cooper, D. T. Long, and T. M. Askew. 1989. Burbot (Lota Stapanian, M. A., V. L. Paragamian, C. P. Madenjian, J. R. Jackson, J. Lap- lota) biogenic sedimentary structures in Lake Superior. Journal of Great palainen, M. J. Evanson, and M. D. Neufeld. 2010. Worldwide status of Lakes Research 15:174–185. Burbot and conservation measures. Fish and Fisheries 11:34–56. Cott, P. A. 2013. Life history and reproductive ecology of a mid-winter spawner: Stapanian, M. A., L. D. Witzel, and A. Cook. 2013. Temporal changes and the Burbot (Lota lota). Doctoral dissertation. Laurentian University, Sudbury, sexual differences in spatial distribution of Burbot in Lake Erie. Transactions Ontario. of the American Fisheries Society 142:1724–1732. Cott, P. A., T. A. Johnston, and J. M. Gunn. 2013a. Stability in life history Stephenson, S. M., M. D. Neufeld, S. C. Ireland, S. Young, R. S. Hardy, and P. characteristics among Burbot populations across environmental gradients. Rust. 2013. Survival and dispersal of sonic-tagged, hatchery-reared Burbot Transactions of the American Fisheries Society 142:1746–1756. released into the Kootenay River. Transactions of the American Fisheries Cott, P. A., T. A. Johnston, J. M. Gunn, and D. M. Higgs. 2013b. Hearing Society 142:1671–1679. sensitivity of the Burbot. Transactions of the American Fisheries Society Taylor, J. L., and S. K. A. Arndt. 2013. Variability in Burbot cohort abundance at Downloaded by [Department Of Fisheries] at 22:51 25 November 2013 142:1699–1704. juvenile and adult stages in Columbia Lake, British Columbia. Transactions George, E. M., E. F. Roseman, B. M. Davis, and T. P. O’Brien. 2013. Feeding of the American Fisheries Society 142:1705–1715. ecology of pelagic larval Burbot in northern Lake Huron, Michigan. Trans- Van Houdt, J. K., B. Hellemans, and F. A. M. Volckaert. 2003. Phylogenetic actions of the American Fisheries Society 142:1716–1723. relationships among Palearctic and Nearctic Burbot (Lota lota): Pleistocene Gilbert, D. L., W. J. Adelman Jr., and J. M. Arnold, editors. 1990. Squid as extinctions and recolonization. Molecular Phylogenetics and Evolution 29: experimental animals. Plenum Press, New York. 599–612. This article was downloaded by: [Department Of Fisheries] On: 25 November 2013, At: 22:51 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Transactions of the American Fisheries Society Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/utaf20 A Synthesis of Kootenai River Burbot Stock History and Future Management Goals Ryan Hardy a & Vaughn L. Paragamian a b a Idaho Department of Fish and Game, 2885 West Kathleen Avenue , Coeur d’Alene , Idaho , 83815 , USA b Codfather Fisheries Consultant , 6399 South Tall Pines Road, Coeur d’Alene , Idaho , 83814 , USA Published online: 20 Nov 2013.

To cite this article: Ryan Hardy & Vaughn L. Paragamian (2013) A Synthesis of Kootenai River Burbot Stock History and Future Management Goals, Transactions of the American Fisheries Society, 142:6, 1662-1670, DOI: 10.1080/00028487.2013.790845 To link to this article: http://dx.doi.org/10.1080/00028487.2013.790845

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions Transactions of the American Fisheries Society 142:1662–1670, 2013 C American Fisheries Society 2013 ISSN: 0002-8487 print / 1548-8659 online DOI: 10.1080/00028487.2013.790845

SPECIAL SECTION: BURBOT

A Synthesis of Kootenai River Burbot Stock History and Future Management Goals

Ryan Hardy* and Vaughn L. Paragamian1 Idaho Department of Fish and Game, 2885 West Kathleen Avenue, Coeur d’Alene, Idaho 83815, USA

Abstract In Idaho, Burbot Lota lota are endemic only to the Kootenai River, where they once provided an important winter fishery to the indigenous people and European settlers. This fishery and that of Kootenay Lake in British Columbia may have been the most robust Burbot fisheries in North America. However, the fishery in Idaho rapidly declined after the construction of Libby Dam by the U.S. Army Corps of Engineers in 1972, and it closed in 1992. Concomitant to the collapse in Idaho was the collapse of the Burbot fishery in Kootenay Lake and the Kootenay River. The operation of Libby Dam for hydroelectric power generation and flood control created major changes in the river’s nutrient concentration, temperature, and seasonal discharge, particularly during the winter when Burbot spawn. Libby Dam operations were implicated as the major limiting factor to Burbot recruitment, giving rise to higher winter temperatures and widely fluctuating flows. Because the Burbot in the Kootenai River are at risk of demographic extinction, a conservation strategy was prepared to outline the measures necessary to rehabilitate the Burbot population to a self-sustaining level. The strategy indicated that operational discharge changes at Libby Dam are required during winter to provide suitable temperature and discharge conditions for Burbot migration and spawning. Studies recommend that the discharge at Bonners Ferry average 176 m3/s for a minimum of 90 d (mid- November through mid-February). Furthermore, preferred Burbot water temperatures of about 6◦C are necessary for migration and cooler temperatures of 1–4◦C for spawning. With each passing year, Burbot stock limitations increasingly constrain rehabilitation. Thus, coordination of intensive culture, extensive rearing, and pen rearing among the Kootenai Tribe of Idaho, the British Columbia Ministry of Environment, the University of Idaho’s Aquatic Research Institute, and the Idaho Department of Fish and Game is important for restoration.

In Idaho, Burbot Lota lota are native only to the and remedial measures to sustain this once-thriving population. Kootenai River2 (Simpson and Wallace 1982) a transbound- The objective with this paper was to report and examine these ary river that flows through both Montana and British Columbia case studies and the results of the remedial strategies currently Downloaded by [Department Of Fisheries] at 22:51 25 November 2013 (Figure 1). in place, along with providing additional management direction The Burbot in the Kootenai River historically provided one for rehabilitating the Kootenai River Burbot stock to the point of the most robust Burbot fisheries in North America. However, at which it is again a self-sustaining and fishable population. this population collapsed in the 1970s due to a series of en- Managers tasked with the recovery of similarly declining stocks vironmental changes (Paragamian et al. 2000). A progression will find this information useful as a guide to identifying popu- of investigations soon followed, examining the consequences lation limitations, recovery needs, and approaches to sustaining of these changes as well as population viability, stock indices, Burbot stocks.

*Corresponding author: [email protected] 1Present address: Codfather Fisheries Consultant, 6399 South Tall Pines Road, Coeur d’Alene, Idaho 83814, USA. 2The term “Kootenai” represents the U.S. spelling, the term “Kootenay” the Canadian spelling. In this article the former is used when it refers to locations or entities in the United States, the latter when it refers to locations in Canada. Received September 4, 2012; accepted March 25, 2013

1662 BURBOT STOCK HISTORY AND FUTURE MANAGEMENT GOALS 1663 Downloaded by [Department Of Fisheries] at 22:51 25 November 2013

FIGURE 1. Map showing the locations of the Kootenai/Kootenay River, Kootenay Lake, Lake Koocanusa, Libby Dam, Bonners Ferry, the Columbia River, Boundary Creek, and important points within the system. River distances are measured in river kilometers (rkm) from the northernmost reach of Kootenay Lake. [Figure available in color online.] 1664 HARDY AND PARAGAMIAN

BACKGROUND 2000). Jeppson also reported in 1958 that three commercial fish- erman were estimated to have harvested over 2,000 kg of Burbot Origin of Burbot in the Kootenai River during one winter. Additional studies were not carried out until Ten thousand years ago, the Great Wisconsin Glacier re- the late 1970s (Partridge 1983), and more intense investigations ceded, opening up the landscape throughout the northern re- were performed in the 1990s through the first decade of the 21st gions of North America and exposing northern Idaho and century. Montana (Alt and Hyndman 1991). The glacial activity and The Burbot in the Kootenai River (Figure 1) continued to pro- retreat changed the landscape, causing water courses and lakes vide an important winter and spring fishery into the early 1970s. to reform, including the Kootenai River. This glacial retreat led This fishery and the one in Kootenay Lake (Paragamian et al. to colonization by a wide diversity of flora and fauna, includ- 2000) may have been the most robust ones in North America ing Burbot. Recent genetic analysis revealed that the Kootenai (Paragamian and Hoyle 2005). However, after the construction River, Idaho, contains a mixture of Burbot from the Pacific and of Libby Dam by the U.S. Army Corps of Engineers (USACE) Mississippi refugia representing two different clades (Powell in 1972, the fishery in Idaho rapidly declined, and it closed in et al. 2008). The Burbot above Kootenai Falls, Montana, are of 1992. Concomitant to the collapse in Idaho was the collapse of the Mississippi clade, while those below the falls are an inter- the Burbot fishery in Kootenay Lake and the Kootenay River grade of the Pacific and Mississippi clades (Powell et al. 2008). in British Columbia (Paragamian et al. 2000), which resulted in The geological mechanism that allowed Burbot to enter the those fisheries being closed in 1997. Kootenai River from the Pacific refugia is thought to be a wa- tercourse that flowed south into the Pend Oreille River drainage Ecosystem Changes from the Kootenai River through the Purcell Trench (Smyers and The history of environmental changes to the Kootenai River Breckenridge 2003). The Kootenai River originates in Koote- is well documented (Northcote 1973; Cloern 1976; Dailey et al. nay National Park in southeastern British Columbia (Figure 1). 1981; Partridge 1983; Redwing Naturalist 1996). Logging and It flows southward into northwestern Montana, where it is im- mining were always important economic activities in the Koote- pounded by Libby Dam, forming Lake Koocanusa. From there nai River basin but were not without environmental conse- the river flows westward into the northeastern portion of the quences, such as tributary discharge flashes (Northcote 1973) Idaho Panhandle, northward back into British Columbia to form and the release of heavy metals (Partridge 1983). A fertilizer Kootenay Lake, and finally into the Columbia River at Castlegar, plant on the St. Mary River in British Columbia was responsible British Columbia. The Kootenai River is the second largest of for eutrophication (Northcote 1973), and the river was diked the Columbia River tributaries and the third largest in drainage from the early 1900s to the 1970s to reduce the flooding of 2 (approximately 50,000 km ; Bonde and Bush 1975). Burbot agricultural lands (Northcote 1973; Redwing Naturalist 1996). were present throughout this drainage prior to the construction The construction of Libby Dam has been implicated as of Libby Dam. the primary cause of Burbot population collapse in Idaho and British Columbia (Paragamian 2000; Paragamian et al. 2005; Kootenai River Fishery Paragamian and Wakkinen 2008). The dam was built for flood Burbot once thrived in the Kootenai River, providing an im- control and hydropower generation, so river regulation has been portant seasonal staple for Native Americans and early European an important objective for the USACE and the Bonneville Power settlers as well as more recent residents of the Kootenai Valley Administration. As a result of the intense river regulation during in the 1920s–1960s (KVRI Burbot Committee 2005). Burbot the winter months (December–February), discharge and temper- fishing was unregulated, and abundant spawners were harvested ature have significantly increased during the Burbot spawning

Downloaded by [Department Of Fisheries] at 22:51 25 November 2013 from tributaries with spears and pitchforks. They were also har- period (Partridge 1983). Additional changes may also have af- vested from the river in the winter with set lines placed under fected Burbot (Paragamian et al. 2000; Anders et al. 2002), and the ice. There were several markets in the Bonners Ferry, Idaho overfishing has been implicated as a possible cause of popula- area(Figure 1), that bought and sold Burbot during the early to tion collapse in Kootenay Lake (Ahrens and Korman 2002). mid-1900s. Studies of Burbot were few prior to the 1970s. The first Population Vital Statistics study on record was by a state biologist (Paul Jeppson) in the An examination of the proportional stock density (PSD; late 1950s (Idaho Department of Fish and Game, Panhandle Anderson and Weithman 1978; Gablehouse 1984) of Burbot Region archives). Jeppson caught 199 Burbot with hoop nets from the combined periodic hoop net sampling over 46 years near the mouth of Boundary Creek over a very short period of showed that their mean length increased from 459 mm TL in winter sampling from 1957 to 1958 (Figure 1). The Burbot in the 1957–1958 sample to 615 mm in the 2002–2004 sample his catch ranged from 350 to 500 mm in TL, with apparently (Paragamian et al. 2008a). The PSD did not increase appre- normal distributions of length and age. Jeppson’s records show ciably (only from 92 to 98) between the two samples, but the that the combined annual catch by sport and commercial an- relative stock density increased from 17 to 86, indicating that glers likely exceeded thousands of kilograms (Paragamian et al. the population was increasing in average length with very few BURBOT STOCK HISTORY AND FUTURE MANAGEMENT GOALS 1665

young fish entering the population. Concomitant to the change ing (Woods 1982; Snyder and Minshall 1996). The reduction of in length structure was a significant decline in catch per unit of nutrients and primary production is another factor thought effort (Paragamian et al. 2008a). to have contributed to the decline of Burbot in this system A review of the population characteristics and extinction risk (Paragamian et al. 2000; Ahrens and Korman 2002). Since Bur- for Burbot in the Kootenai River suggested that the population bot larvae may remain in the limnetic zone for 16–27 d (Ghan would be extirpated by 2015 in the absence of remediation and Sprules 1993) feeding on phytoplankton and zooplankton (Pyper et al. 2004; Paragamian et al. 2008b; Paragamian and (McPhail and Paragamian 2000), a shift in food availability Hansen 2009). Mark–recapture models indicated that the Burbot could affect planktivorous larvae such as those of Burbot. population had declined from about 150 fish in the mid-1990s to only 50 fish in the early 2000s, with total annual mortality estimated at 63% (Pyper et al. 2004). Population numbers prior CONSERVATION STRATEGY AND PROGRESS to the construction of Libby Dam were thought to be in the In 2000, the Burbot population in the Kootenai River was thousands (Paragamian and Hansen 2009). petitioned for listing as threatened under the U.S. Endangered Species Act. However, in 2003 the U.S. Fish and Wildlife Ser- vice found that listing was not warranted because the population LIMITING FACTORS did not represent a distinct population segment (USOFR 2003). Discharge and Temperature Because the Burbot in the Kootenai River were culturally and Studies have shown that dam operations affect the migra- recreationally important before their collapse, an international tion patterns of fish (Raymond 1979; Marschall et al. 2011). Burbot conservation strategy was developed by a communi- Early investigations from 1993 to 1998 used radio- and sonic tywide working group to restore the population (Paragamian telemetry to determine the effects of Libby Dam on Burbot et al. 2002; KVRI Burbot Committee 2005; Ireland and Perry movements. In controlled tests comparing river discharge with 2008). The strategy outlined rehabilitation measures, includ- Burbot movement during winter migration, it was determined ing changes to the operation of Libby Dam and development that high winter discharges (especially during power peaking) of conservation aquaculture to supplement the wild stock dur- disrupted the spawning movements of Burbot. In many cases, ing population rehabilitation. Although the simple stocking of as discharge increased Burbot abandoned upstream movements Burbot is not viewed as the sole means of increasing natural and drifted downstream (Paragamian 2000). In performance tri- recruitment, the combination of these remedial measures, along als, Burbot were previously determined to have relatively low with large-scale nutrient restoration (Holderman and Hardy swimming endurance (Jones et al. 1974). Further examination 2004) and habitat improvements in the basin, are viewed as of Burbot movement in the Kootenai River suggested that dis- the way to restore the Burbot population in the river. As pointed charges above 300 m3/s inhibit Burbot migration (Paragamian out, the operation of Libby Dam is thought to play the most et al. 2005). Significantly greater upstream movements were important role in limiting the natural recruitment of Burbot, and found when the discharges from Libby Dam averaged less than therefore any multifaceted effort to restore the Burbot popula- 176 m3/s. tion should include the management of dam outflow to ensure Additional studies examined the consequences of both higher suitable temperature and discharge conditions. winter temperatures and discharge from Libby Dam. Logistic Demographic data (Paragamian et al. 2008b) were used regression using the temperature and discharge of the Kootenai to establish Burbot restoration targets in the Kootenai River River at Bonners Ferry pre– and post–Libby Dam suggested that (Paragamian and Hansen 2009). Because the density of the the Burbot spawning migration pre–Libby Dam was predictable Burbot population in the river prior to the operation of Libby

Downloaded by [Department Of Fisheries] at 22:51 25 November 2013 and similar to that expected in unregulated rivers in Alaska, Dam was unknown, the density of the species in Alaskan rivers where Burbot populations are considered healthy (Paragamian (Evenson 1993) was used as a surrogate target for the restoration and Wakkinen 2008). The movement of Burbot post–Libby Dam of Burbot in the Kootenai River. The interim target included an was unpredictable, with migrations occurring almost a month abundance of 5,500 age-4 and older individuals (45 fish/km; later and the highest probability of movement being on holidays, 3.0 fish/ha) within 25 years (when each adult produced 0.85 re- when the discharge from Libby Dam was significantly reduced. cruits per year) and the ultimate target an abundance of 17,500 Sonic telemetry of spawning adults suggested that the migration individuals (143 fish/km; 9.6 fish/ha) (when each adult pro- and spawning of Burbot in the river is inhibited by the higher duced 1.1 recruits per year). It was believed that these targets winter discharges and warmer water temperatures (Paragamian are achievable because Burbot populations elsewhere have been and Wakkinen 2008). resilient when the factors limiting population recovery were re- duced (Bruesewitz 1990; Taube and Bernard 1995; Stapanian Reduced Primary Production et al. 2006, 2008). While the dam affected the seasonal discharge and tempera- The strategy also identified conservation aquaculture as a re- ture of the Kootenai River, the reservoir behind the dam formed medial measure to help strengthen the depressed Burbot stock. a nutrient trap, reducing primary production and nutrient spiral- Paragamian and Hansen (2011) developed an age-structured 1666 HARDY AND PARAGAMIAN

simulation model to estimate the number of age-0 Burbot (fall Determining Natural Production fingerlings) to stock annually to rebuild the population in the Early in the planning process, there was a need to find a Kootenai River. They found that with an annual survival of way to mass-mark stocked Burbot to determine survival and about 38%, 110,000–900,000 age-0 Burbot per year will need whether these stocked fish would spawn and produce year- to be stocked to rebuild the population in the river in 25 years, classes on their own. The number and size of progeny needing depending on the restoration goal (interim or ultimate). If sur- to be tagged precluded conventional mechanical-type tags, and vival is 61%, the stocking numbers could range from 12,000 to we turned to other recently developed approaches. Originally 35,000 age-0 fingerlings per year. After stocked Burbot in the proposed by Anderson and Garza (2005), parentage-based ge- population reach age 4, the discharge from Libby Dam must netic tagging (PBT) is currently being used in other fisheries be regulated to provide suitable temperatures and flows during as an alternative to mechanical tagging methods (Steele et al. the Burbot prespawning and spawning periods in order to enable 2011). This type of tagging involves sampling and genotyping the population to reproduce and return to self-sustaining status. of hatchery broodstock annually to create a parental genotype database. Progeny from any of these genotyped parents can be assigned to their release group and brood year. To deter- Burbot Culture mine whether the Burbot captured in the Kootenai River origi- The importance of Burbot stock limitations to restoration nated from hatchery stockings, PBT techniques for this species is increasing with each passing season. Because the Kootenai were developed by the Idaho Department of Fish and Game’s River Burbot stock is so limited, the introduction of a donor (IDFG) Eagle Fish Genetics Laboratory in 2012 (Matt Camp- stock was found to be useful in enhancing the population. Re- bell, IDFG, personal communication). Therefore, every Burbot cent analysis of the cytochrome b region of mtDNA indicated captured in the basin will have tissue sampled to assign it to that the Burbot in Columbia and Moyie lakes were of a simi- a particular parental crossing or determine whether it is a re- lar phylogenetic group as Kootenai River Burbot (Powell et al. sult of naturally production. This method will also provide a 2008) and thus suitable as a donor stock. Although the Burbot way to evaluate survival by brood year and release location as from Moyie Lake are of the Pacific clade (Powell et al. 2008), well as survival rates based on time of release and a multi- genetic concerns may be moot because Burbot from that clade tude of other possible factors involved in producing a strong are presently migrating into the Kootenai River from above year-class. Kootenai Falls and Libby Dam (Skarr et al. 1996). Moyie Lake is in the Kootenai River basin, and Burbot from the lake had previously been provided to the University of Idaho’s Aquacul- Recent Stocking Success and Findings ture Research Institute (UI–ARI) by the Kootenai Tribe of Idaho The first stocking of larval Burbot reared at the UI–ARI for spawning and experimental intensive culture (Jensen et al. occurred in 2009 with 209 larval fish. This was scaled up to 2008a). over 400,000 larval and juvenile Burbot stocked by the Koote- Although still evolving, intensive rearing techniques for Bur- nai Tribe of Idaho in 2012. Gill-net sampling by the IDFG bot have proven extremely successful through stockings that in the main river showed that these stockings were successful have increased the catch rates of juvenile Burbot in the main and the stocked fish eventually recruited to hoop net sampling river. Extensive rearing of larvae has also been effective at in- (Figure 2). Catch rates of Burbot are now as high as they were a creasing the growth and survival of Burbot for the purpose of decade ago, and this is without the largest release numbers from restoration when released as fingerlings (Dillen et al. 2008; 2011 and 2012 being fully recruited to sampling gear. The first Vught et al. 2008) and is considered important to an initial of their kind, these results are extremely positive, since efforts

Downloaded by [Department Of Fisheries] at 22:51 25 November 2013 restoration strategy for Kootenai River Burbot (Jensen et al. to restore Burbot through intensive and extensive culture tech- 2008a, 2008b; Vught et al. 2008). Intensive culture and exten- niques failed to produce measureable numbers in other studies sive pond rearing techniques developed by the Kootenai Tribe (Dillen et al. 2008; Vught et al. 2008). of Idaho, the British Columbia Ministry of Environment, the Recent experimental releases of sonically tagged, hatchery- UI–ARI, and Idaho Department of Fish and Game have become reared age-2 and -3 Burbot into the Kootenai system showed an important restoration measure and show promise through in- rapid, widespread dispersal from one stocking site, with some creased catch rates of hatchery juveniles in the river. To rebuild individuals staying in nearby traditional Burbot spawning loca- the Burbot population in the Kootenai River, large-scale meth- tions (Neufeld et al. 2011). This suggests that a limited number ods for intensive culture and rearing of age-0 Burbot will need of stocking locations is sufficient to allow Burbot to access crit- further refinement. If spawning conditions and habitat can be ical rearing habitats within a short period of time following re- improved, hatchery recruits will add to the collapsed reproduc- lease (Neufeld et al. 2011). This also suggests that lake-origin, tive stock and thereby help to restore a Burbot population to the hatchery-reared Burbot are suitable for stocking in a riverine river. In addition, more Burbot in the river will help researchers environment. These fish were also detected in the vicinity of refine the correlation between dam operations and winter move- known spawning locations during the spawning season, which ments, which will help in finalizing a system operation request. may indicate that these hatchery-released fish will contribute to BURBOT STOCK HISTORY AND FUTURE MANAGEMENT GOALS 1667

FIGURE 2. Catch per unit of effort (fish/net-day) and effort (number of net-days) for Burbot caught in hoop nets in the Kootenai River from 1992 to 2012. A net-day is defined as a 24-h set. [Figure available in color online.]

the Kootenai River spawning population in the future (Neufeld Nutrient Restoration et al. 2011). As noted above, the decline in primary production is also thought to have contributed to the decline in Burbot populations Preliminary Results of Tributary Use in the Kootenai River and Kootenay Lake. A large-scale nutri- As noted, it is believed that a combination of limiting factors ent addition program was jointly implemented by the KTOI and led to the dramatic decline of Burbot following the construc- IDFG, with planning and study design beginning in 2000 and tion of Libby Dam. Although the primary limiting factor is actual additions starting in 2005. Nutrient additions have been hypothesized to stem from winter dam operations, efforts to shown to dramatically increase all levels of the food web, includ- reestablish spawning in major historic spawning tributaries are ing fish populations that feed directly on macroinvertebrates and under way. The strong year-class survival associated with these diatoms (Hardy 2008). In addition to this, the Bonneville Power

Downloaded by [Department Of Fisheries] at 22:51 25 November 2013 stocking events may provide valuable information on the winter Administration (through the KTOI) is funding a large-scale nu- dam operations associated with this early success. Several trib- trient addition in the South Arm of Kootenay Lake. Preliminary utary release sites were chosen to test whether age-0 larvae that results there also show a dramatic increase in zooplankton num- are released into these tributaries will return to these locations bers and a concomitant increase in kokanee Oncorhynchus nerka to summer, overwinter, or spawn. The IDFG operates a PIT (lacustrine Sockeye Salmon) (Matt Neufeld, British Columbia tag array (Biomark. Inc.) in one of these tributary release sites Ministry of Environment, personal communication). With the (Deep Creek). The array, which is located 7 rkm from the con- limiting factor of nutrients being mitigated in this fashion, Bur- fluence, showed that 98% of the age-0 juveniles from a recent bot survival at the early life stages at which first feeding occurs stocking of 3,000 PIT-tagged juveniles released 13–26 rkm up- may be significantly increased. stream were not detected and possibly survived in Deep Creek to overwinter their first year. In addition, year-classes from prior MANAGEMENT GOALS stocking events were detected by this array, suggesting that they were using this tributary for summer feeding or overwintering Population Structure as well. These results provide new insight into the importance Since the wild stock of Burbot is very depressed, culture in tributaries in the early life stages of Burbot. and rearing methods must now be developed at a scale that is 1668 HARDY AND PARAGAMIAN

large enough to rebuild the population of Burbot in the Kootenai The efficacy of stocking in key locations and at different River. This is a major task that has been taken on by the KTOI. life stages (larval or juvenile) is still unclear, but it is certain An initial restoration goal of 5,500 age-4 and older Burbot has to reveal itself with the monitoring program that is currently been established as an interim goal, with 17,500 such fish as in place. This will give us an indication of the strategies that the ultimate goal for population rehabilitation. The age and size we should continue and those that may need to be modified or at release, density, release location and timing, and survival of discontinued. the stocked fish will be assessed and is part of the overall IDFG monitoring program to determine whether restoration goals are SUMMARY being reached. This monitoring program must also ensure that Considerable progress in Burbot restoration has been made stocking does not affect the wild component of the population in through the collaboration of state, tribal, university, and provin- terms of growth, survival, and recruitment. With PBT marking cial agencies. The restoration of the Burbot stock in the Kootenai in place, this assessment should be much more efficient than River will further rely on these cooperative efforts. Winter oper- those with conventional tagging methods. ations at Libby Dam that foster the migration and spawning of System Operation Request Burbot, in addition to other remedial measures to improve the survival of both early life stages and facilitate spawning, are nec- As noted above, past research efforts have identified limiting essary and will include the cooperation of multiple stakeholders factors due to higher than normal winter flows and temperatures to be accomplished. It is important to understand that we are con- as well as the loss of nutrients and habitat alterations, the com- strained by the existence of a significantly altered ecosystem, bination of which resulted in the Burbot population decline. which is at best difficult to manipulate entirely for the recovery Recommendations to improve the migration and spawning of of all affected fish species. To date, stocking hatchery-reared wild Burbot—over and above the nutrient restoration and habitat Burbot has been successful in adding many more Burbot to the improvements currently under way—include providing cooler total Kootenai River population. Given the species’ relatively winter water temperatures and lower discharges from December short maturation time of 3–5 years, the results of this effort to through February to reduce the disruption in the spawning mi- boost natural production should soon be evident. gration. Burbot recovery will may ultimately depend on some type of modification to winter dam operations, above all other measures. It is also very clear that this is the most difficult to ACKNOWLEDGMENTS achieve due to social and political constraints. Further, it will be Special thanks to Sue Ireland, Jack Siple, Eric Wagner, Jose important not only to determine the spawning success of intro- Ponce, and Chris Lewandowski of the Kootenai Tribe of Idaho duced Burbot but also the timing of spawning and how it relates (KTOI). Thanks to Matt Neufeld, Sarah Stephenson, and Don to the USACE winter water management schedule. Miller of the British Columbia Ministry of Environment. Thanks to Ken Cain, Nate Jensen, Neil Ashton, and Josh Eagan of the Research Needs University of Idaho Aquaculture Research Institute (UI–ARI). Although previous research from 1993 to 2003 (Paragamian We thank the KTOI and tribal members for their sponsorship of et al. 2005; Paragamian and Wakkinen 2008) identified the ef- Burbot culture studies at UI–ARI. Thanks to Pete Rust, Corie fects of winter operations on upstream migrations of Burbot, it Laude, and Shaun Lacy with the Idaho Department of Fish and was done using active tracking techniques with an inherent level Game for their river sampling efforts. Thanks to Dan Schill for of error. Passive tracking technologies now exist that will better his suggestion of using parentage-based genetic tagging as a define the specific effects of flow and temperature on migration. marking technique to identify the year-class strength of stocked

Downloaded by [Department Of Fisheries] at 22:51 25 November 2013 Tagging prespawn individuals is of high importance to confirm Burbot. We thank the Kootenai Valley Resource Initiative for past evaluations and possibly make more precise requests for Burbot restoration efforts in collaboration with many agencies. Libby Dam winter operations favorable to spawning. The Bonneville Power Administration provided the funding for In addition to tracking evaluations, evaluating stocking strat- this study and all ongoing Burbot restoration in the basin. egy in relation to food availability is important to the recov- ery of natural production in the Kootenai River. It has been hypothesized that larval survival and recruitment are condi- REFERENCES Ahrens, R., and J. Korman. 2002. What happened to the West Arm Burbot tioned by the match of larvae with prey fields in time and space stock in Kootenay Lake? Use of an age-structured population model to de- (the match–mismatch hypothesis; Cushing 1972). Therefore, termine the possible causes of recruitment failure. Report prepared for the Burbot larvae must locate food patches during this critical period British Columbia Ministry of Water, Land, and Air Protection, the Habitat before irreversible starvation or a point of no return is reached. Conservation Trust Fund, and the Bonneville Power Administration, Nelson. Investigation of temporal food availability vis-a-vis` preferences Alt, D. D., and D. W. Hyndman. 1991. Roadside geology of Idaho. Mountain Press Publishing, Missoula, Montana. and how it changes with growth and larval abundance would Anders, P. J., D. L. Richards, and M. S. Powell. 2002. The first endangered be of considerable utility in ensuring that stocked or naturally White Sturgeon population: repercussions in an altered large-river–floodplain produced larvae survive to perpetuate the population. ecosystem. Pages 67–82 in W. Van Winkle, P. J. Anders, D. H. Secor, and D. BURBOT STOCK HISTORY AND FUTURE MANAGEMENT GOALS 1669

A. Dixon, editors. Biology, management, and protection of North American McPhail, J. D., and V. L. Paragamian. 2000. Burbot biology and life history. sturgeon. American Fisheries Society, Symposium 28, Bethesda, Maryland. Pages 11–23 in V. L. Paragamian and D. W. Willis, editors. Burbot: biology, Anderson, E. C., and J. C. Garza. 2005. The power of single-nucleotide poly- ecology, and management. American Fisheries Society, Fisheries Manage- morphisms for large-scale parentage inference. Genetics 172:2567–2582. ment Section Publication 1, Spokane, Washington. Anderson, R. O., and A. S. Weithman. 1978. The concept of balance for cool- Neufeld, M. D., K. D. Cain, N. R. Jensen, S. C. Ireland, and V. L. Paragamian. water fish populations. Pages 371–381 in R. L. Kendall, editor. Selected 2011. Movement of lake origin Burbot reared in a hatchery environment and coolwater fishes of North America. American Fisheries Society, Special Pub- released into a large river. North American Journal of Fisheries Management lication 11, Bethesda, Maryland. 31:56–62. Bonde, T. H., and R. M. Bush. 1975. Kootenai River water quality investiga- Northcote, T. G. 1973. Some impacts of man on Kootenay Lake and its tions, Libby Dam pre-impoundment study 1967–1972. U.S. Army Corps of salmonids. Great Lakes Fishery Commission Technical Report 25. Engineers, Washington, D.C. Paragamian, V. L. 2000. The effects of variable flows on Burbot spawning mi- Bruesewitz, R. E. 1990. Population dynamics and movement of Burbot (Lota grations in the Kootenai River, Idaho, USA, and British Columbia, Canada. lota) in western Lake Michigan and Green Bay. Master’s thesis. University Pages 111–123 in V.L. Paragamian and D. W. Willis, editors. Burbot: biology, of Wisconsin, Stevens Point. ecology, and management. American Fisheries Society, Fisheries Manage- Cloern, J. E. 1976. Recent limnological changes in southern Kootenay Lake, ment Section Publication 1, Spokane, Washington. British Columbia. Canadian Journal of Zoology 54:1571–1578. Paragamian, V. L., and M. J. Hansen. 2009. Rehabilitation needs for Burbot Cushing, D. H. 1972. The production cycle and numbers of marine fish. Sym- in the Kootenai River, Idaho, USA and British Columbia, Canada. North posium of the Zoological Society of London 29:213–232. American Journal of Fisheries Management 29:768–777. Dailey, R. J., E. C. Carmack, C. B. Gray, C. H. Pharo, S. Jasper, and R. C. Paragamian, V. L., and M. J. Hansen. 2011. Stocking goals for rehabilitation Wiegland. 1981. The effects of upstream impoundment on Kootenay Lake, of Burbot in the Kootenai River, Idaho, USA and British Columbia, Canada. B. C. Canada Inland Waters Waters Directorate, Research Institute, Scientific Journal of Applied Ichthyology 27(Supplement 1):22–26. Series, West Vancouver. Paragamian, V. L., R. Hardy, and B. Gunderman. 2005. Effects of regu- Dillen, A., I. Vught, D. De Charleroy, D. Monnier, and J. Coeck. 2008. A lated discharge on Burbot migration. Journal of Fish Biology 66:1199– preliminary evaluation of reintroductions of Burbot in Flanders, Belgium. 1213. Pages 179–183 in V. L. Paragamian and D. H. Bennett, editors. Burbot ecol- Paragamian, V. L., and G. Hoyle. 2005. Kootenai River fisheries investiga- ogy, management, and culture. American Fisheries Society, Symposium 59, tion: Burbot stock status. Annual Progress Report to the Bonneville Power Bethesda, Maryland. Administration, Project 88-65, Portland, Oregon. Evenson, M. J. 1993. A summary of abundance, catch per unit of effort, and Paragamian, V. L., J. McCormick, and C. Laude. 2008a. Changes in population mean length estimates of Burbot sampled in rivers of interior Alaska. Alaska indices of a diminishing Burbot population in the Kootenai River, Idaho, Department of Fish Game, Fisheries Data Series 93-15, Anchorage. USA and British Columbia, Canada. Journal of Freshwater Ecology 23: Gablehouse, D. W. Jr. 1984. A length categorizations system to assess fish 553–563. stocks. North American Journal of Fisheries Management 4:37I–384. Paragamian, V. L., B. J. Pyper, M. J. Daigneault, R. P. Beamesderfer, and S. Ghan, D., and W. G. Sprules. 1993. Diet, prey selection and growth of larval C. Ireland. 2008b. Population dynamics and extinction risk of Burbot in the and juvenile Burbot Lota lota (L.). Journal of Fish Biology 42:47–64. Kootenai River, Idaho, USA and British Columbia, Canada. Pages 213–234 Hardy, R. S. 2008. Kootenai River fisheries recovery investigations: ecosys- in V. L. Paragamian and D. H. Bennett, editors. Burbot: ecology, manage- tem rehabilitation. Progress Report to the Bonneville Power Administration, ment, and culture. American Fisheries Society, Symposium 59, Bethesda, Project 1988-06500, Portland, Oregon. Maryland. Holderman, C., and R. Hardy, editors. 2004. Kootenai River ecosystem project: Paragamian, V.L., and V.D. Wakkinen. 2008. Seasonal movement and the inter- an ecosystem approach to evaluate and rehabilitate a degraded, large riverine action of temperature and discharge on Burbot in the Kootenai River, Idaho, ecosystem. Annual Report to the Bonneville Power Administration, Project USA, and British Columbia, Canada. Pages 55–77 in V. L. Paragamian and 94-49, Portland, Oregon. D. H. Bennett, editors. Burbot: ecology, management, and culture. American Ireland, S., and P. Perry. 2008. Burbot restoration in the Kootenai River Basin: Fisheries Society, Symposium 59, Bethesda, Maryland. using agency, tribal, and community collaboration to develop and implement Paragamian, V. L., V. Whitman, J. Hammond, and H. Andrusak. 2000. Collapse a conservation strategy. Pages 251–256 in V. L. Paragamian and D. H. Ben- of the Burbot fisheries in Kootenay Lake, British Columbia, Canada, and nett, editors. Burbot: ecology, management, and culture. American Fisheries the Kootenai River, Idaho, USA, post-Libby Dam. Pages 155–164 in V. L. Society, Symposium 59, Bethesda, Maryland. Paragamian and D. W. Willis, editors. Burbot: biology, ecology, and manage- Jensen, N. J., S. R. Williams, S. C. Ireland, J. T. Siple, M. D. Neufeld, and K. D. ment. American Fisheries Society, Fisheries Management Section Publication Downloaded by [Department Of Fisheries] at 22:51 25 November 2013 Cain. 2008a. Preliminary captive Burbot spawning observations. Pages 155– 1, Spokane, Washington. 165 in V. L. Paragamian and D. H. Bennett, editors. Burbot: ecology, man- Paragamian, V. L., S. Yundt, J. Hammond, C. Spence, S. Ireland, G. Hoffman, agement, and culture. American Fisheries Society, Symposium 59, Bethesda, and J. Laufle. 2002. An international recovery strategy for Burbot in the Maryland. Kootenai River and Kootenay Lake, Idaho and British Columbia. Intermoun- Jensen, N. J., S. C. Ireland, J. T. Siple, S. R. Williams, and K. D. Cain. 2008b. tain Journal of Science 8:178–190. Evaluation of egg incubation methods and larval feeding regimes for North Partridge, F. 1983. Kootenai River fisheries investigations. Idaho Department of American Burbot. North American Journal of Aquaculture 70:162–170. Fish and Game, Federal Aid in Fish and Wildlife Restoration, Job Completion Jones, D. R., J. W. Kiceniuk, and O. S. Bamford. 1974. Evaluation of swimming Report F-73-R-5, Boise. performance of several species of fish from the Mackenzie River. Journal of Powell, M., V. L. Paragamian, and J. Dunnigan. 2008. Mitochondrial variation the Fisheries Research Board of Canada 31:77–99. in western North American Burbot with special reference to the Kootenai KVRI (Kootenai Valley Resource Initiative) Burbot Committee. 2005. Kootenai River in Idaho and Montana, USA. Pages 3–28 in V. L. Paragamian and River/Kootenay Lake Burbot conservation strategy. Kootenai Tribe of Idaho, D. H. Bennett, editors. Burbot: ecology, management, and culture. American Bonners Ferry. Fisheries Society, Symposium 59, Bethesda, Maryland. Marschall, E. A., M. E. Mather, D. L. Parrish, G. W. Allison, and J. R. McMen- Pyper, B. J., M. J. Daigneault, R. C. P. Beamesderfer, V. L. Paragamian, and emy. 2011. Migration delays caused by anthropogenic barriers: modeling S. C. Ireland. 2004. Status and population dynamics of Burbot in the Kootenai dams, temperature, and success of migrating salmon smolts. Ecological Ap- River, Idaho and British Columbia, Canada. Annual Progress Report to the plications 21:3014–3031 Bonneville Power Administration, Project 88-65, Portland, Oregon. 1670 HARDY AND PARAGAMIAN

Raymond, H. L. 1979. Effects of dams and impoundments on migrations of ecology, management, and culture. American Fisheries Society, Symposium juvenile Chinook Salmon and steelhead from the Snake River, 1966 to 1975. 59, Bethesda, Maryland. Transactions of the American Fisheries Society 108:505–529. Stapanian, M. A., C. P. Madenjian, and L. D. Witzel. 2006. Evidence that Redwing Naturalists. 1996. History of diking on the Kootenay River flood- Sea Lamprey control led to recovery of the Burbot population in Lake Erie. plain in British Columbia. Report prepared for Habitat Enhancement Branch, Transactions of the American Fisheries Society 135:1033–1043. Department of Fisheries and Oceans, Vancouver. Steele, C., M. Ackerman, J. McCane, M. Campbell, M. Hess, and S. Narum. Simpson, J., and R. Wallace. 1982. Fishes of Idaho. University Press of Idaho, 2011. Parentage based tagging of Snake River hatchery steelhead and Chinook Moscow. Salmon. Annual Progress Report to the Bonneville Power Administration, Skarr, D., J. DeShazer, T. Garrow, T. Ostrowski, and B. Thornburg. 1996. Quan- Project 2010-031-00, Portland, Oregon. tification of Libby Reservoir levels needed to maintain or enhance reservoir Taube, T., and D. R. Bernard, 1995. Stock assessment and biological charac- fisheries. Completion Report to the Bonneville Power Administration, Project teristics of Burbot in Lake Louise and Tolsona Lake, Alaska, 1994. Alaska 83-467, Portland, Oregon. Department of Fish and Game, Fisheries Data Series 95-14, Anchorage. Smyers, N. B., and R. M. Breckenridge. 2003. Glacial Lake Missoula, Clark USOFR (U.S. Office of the Federal Register). 2003. Endangered and threatened Fork ice dam, and the floods outburst area: Northern Idaho and west- wildlife and plants; 12-month finding for a petition to list the lower Koote- ern Montana. Pages 1–15 in T. W. Swason, editor. Western cordillera and nai River Burbot (Lota lota) as threatened or endangered. Federal Register adjacent areas. Geological Society of America, Field Guide 4, Boulder, 68:47(11 March 2003):11574–11579. Colorado. Vught, I., A. S. Harzevili, J. Auwerx, and D. De Charleroy. 2008. Aspects of re- Snyder, E. B., and G. W. Minshall. 1996. Ecosystem metabolism and nutri- production and larviculture of Burbot under hatchery conditions. Pages 167– ent dynamics in the Kootenai River in relation to impoundment and flow 178 in V. L. Paragamian and D. H. Bennett, editors. Burbot: ecology, man- enhancement for fisheries management. Idaho State University, Stream Ecol- agement, and culture. American Fisheries Society, Symposium 59, Bethesda, ogy Center, Completion Report, Pocatello. Maryland. Stapanian, M. A., C. P. Madenjian, C. R. Bronte, M. P. Ebener, B. F. Lantry, and Woods, P. F. 1982. Annual nutrient loadings, primary productivity, and trophic J. D. Stockwell. 2008. Status of Burbot populations in the Laurentian Great state of Lake Koocanusa, Montana and British Columbia, 1972–80. U.S. Lakes. Pages 91–107 in V. L. Paragamian and D. H. Bennett, editors. Burbot: Geological Survey Professional Paper 1283. Downloaded by [Department Of Fisheries] at 22:51 25 November 2013 This article was downloaded by: [Department Of Fisheries] On: 25 November 2013, At: 22:52 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Transactions of the American Fisheries Society Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/utaf20 Survival and Dispersal of Sonic-Tagged, Hatchery- Reared Burbot Released into the Kootenay River Sarah M. Stephenson a , Matthew D. Neufeld a , Susan C. Ireland b , Shawn Young b , Ryan S. Hardy c & Pete Rust c a British Columbia Ministry of Forests , Lands and Natural Resource Operations , 401-333 Victoria Street, Nelson , British Columbia , V1L 4K3 , Canada b Kootenai Tribe of Idaho , Post Office Box 1269, Bonner's Ferry , Idaho , 83805 , USA c Idaho Department of Fish and Game , 2750 Kathleen Avenue, Coeur d’Alene , Idaho , 83815 , USA Published online: 20 Nov 2013.

To cite this article: Sarah M. Stephenson , Matthew D. Neufeld , Susan C. Ireland , Shawn Young , Ryan S. Hardy & Pete Rust (2013) Survival and Dispersal of Sonic-Tagged, Hatchery-Reared Burbot Released into the Kootenay River, Transactions of the American Fisheries Society, 142:6, 1671-1679, DOI: 10.1080/00028487.2013.774293 To link to this article: http://dx.doi.org/10.1080/00028487.2013.774293

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions Transactions of the American Fisheries Society 142:1671–1679, 2013 C American Fisheries Society 2013 ISSN: 0002-8487 print / 1548-8659 online DOI: 10.1080/00028487.2013.774293

SPECIAL SECTION: BURBOT

Survival and Dispersal of Sonic-Tagged, Hatchery-Reared Burbot Released into the Kootenay River

Sarah M. Stephenson* and Matthew D. Neufeld British Columbia Ministry of Forests, Lands and Natural Resource Operations, 401-333 Victoria Street, Nelson, British Columbia V1L 4K3, Canada Susan C. Ireland and Shawn Young Kootenai Tribe of Idaho, Post Office Box 1269, Bonner’s Ferry, Idaho 83805, USA Ryan S. Hardy and Pete Rust Idaho Department of Fish and Game, 2750 Kathleen Avenue, Coeur d’Alene, Idaho 83815, USA

Abstract As part of recovery efforts for the Kootenay population of Burbot Lota lota, we monitored 109 sonic-tagged, hatchery-reared Burbot released at 1–3 years of age throughout Kootenay Lake and the Kootenay River over a 3-year period. Our objectives were to evaluate broodstock choice, assess differences in survival and dispersal by release site and age, and evaluate spawning movements. Overall, release survival was high (66%) and there was dispersal throughout the system (up to 235 km), involving both lacustrine and riverine habitat. Spawning movements were extensive (up to 59 km/d upstream) and suggest the use of known spawning locations. However, most age-1 releases had lower survival and remained in the release tributaries for 1 year postrelease, which was longer than expected and which warrants further investigation. Overall, this telemetry study provides a positive outlook on the current aquaculture rehabilitation efforts for Kootenay Burbot and provides direction for further work.

In British Columbia, Idaho, and Montana, the population of population in the lower Kootenai River, the Kootenai Tribe of Burbot Lota lota in Kootenay Lake and the Kootenay River Idaho coordinated a collaborative process through the Kootenai (spelled Kootenai in the United States) is at risk of demographic Valley Resource Initiative’s Burbot Subcommittee to develop

Downloaded by [Department Of Fisheries] at 22:52 25 November 2013 extinction. Prior to 1972 the harvest of fishermen (both sport a conservation strategy to guide Burbot restoration activities and commercial) in the Kootenay River was estimated at tens (KVRI 2005; Ireland and Perry 2008). Although our current of thousands of kilograms, and in Kootenay Lake anglers annu- knowledge of the specific causes of the decline in the system ally harvested over 20,000 Burbot in the late 1960s and early is incomplete, habitat loss and degradation are the most likely, 1970s (Paragamian et al. 2000; Ahrens and Korman 2002). With as has been seen with other riverine populations at risk (KVRI the completion of Libby Dam and other habitat changes, these 2005; Stapanian et al. 2010). popular fisheries neared collapse by the late 1970s and have As a stopgap measure while habitat restoration projects are not recovered since (Paragamian et al. 2000; Neufeld 2005). being completed, hatchery-reared Burbot are now annually Studies on both sides of the Canadian–U.S. border date back released into the Kootenay River (Neufeld et al. 2011a). to 1979, when the Kootenai River Fisheries Investigation was Broodstock for the hatchery program are procured from a wild initiated by the Idaho Department of Fish and Game (Partridge stock at Moyie Lake within the Kootenay River basin. Under 1980). Additionally, in response to the decline in the Burbot the Burbot Conservation Strategy and the first Kootenai River

*Corresponding author: [email protected] Received September 4, 2012; accepted January 30, 2013 1671 1672 STEPHENSON ET AL.

Burbot 5-Year Operational Research Plan (2006–2011; KVRI TABLE 1. Size of hatchery-reared Burbot prior to sonic tagging, for all re- 2005; Neufeld et al. 2009), studies were initiated to evaluate the leases from 2009 to 2011 (n = 121). success of the hatchery efforts as well as the release survival, Weight (g) TL (mm) movements, and habitat use of hatchery-raised Burbot. In addition to in-river recapture efforts, sonic-tagged subadult Age at release Min Max Mean Min Max Mean and adult hatchery releases can be used as a passive method of 1 62 115 83 205 270 240 tracking movement, habitat use, and survival. One variable that 2 65 818 326 215 270 266 is best evaluated with telemetry is the dispersal of the hatchery- 3 409 800 413 350 480 548 reared Burbot from various release sites. Due to the spatial and temporal variability of the natural environment, dispersal to suitable habitat is an important factor influencing survival (Hofmann and Fischer 2002; Slavik et al. 2005) and conse- from spawning adults captured via ice fishing; eggs were fertil- quently the success of a hatchery-supported population. In the ized on site, and the adult Burbot were tagged and released back Kootenay River and Kootenay Lake, Neufeld et al. (2011a) eval- into Moyie Lake (further described in Neufeld et al. 2011b). uated the first ever release and tracking of sonic-tagged age-2 These gametes were then transported to the Aquaculture Re- and age-3 hatchery-reared Burbot in North America. The 5- search Institute (ARI) at the University of Idaho, Moscow. month pilot study documented a high survival rate and rapid Rearing at the ARI occurred in a recirculating system where dispersal for these fish; progeny from lake-origin broodstock photoperiod and water temperature were generally maintained showed successful behavioral plasticity in a natural riverine similar to the conditions in the Kootenay River. The Burbot used habitat (Neufeld et al. 2011a). In that study, 28 out of 30 Burbot in this study were reared to 1–3 years of age prior to tagging and survived the initial transport and release and the mean distance release. All fish were weighed and measured prior to tagging of dispersal was 80 km (range = 9.8–137.5 km). Furthermore, (Table 1), and the mean sizes of the three ages were significantly the sonic-tagged hatchery fish were detected in known spawn- different (Figure 2). ing locations during the winter spawning months, suggesting that some hatchery progeny have contributed to spawning in the Study Period and Releases wild during their first year at large (Neufeld et al. 2011a). This study was initiated with the first release of sonic-tagged In the current study, we expand upon this previous work by Burbot into the Goat River, a Kootenay River tributary, on Oc- evaluating two and half years of continuous telemetry data from tober 21, 2009. Four subsequent releases occurred in 2010 and 121 sonic-tagged Burbot released over 3 years. Specifically, our 2011 (Table 2). We used six release locations stretching across objectives were (1) to determine the differences in survival and the study area that included both tributary and main-stem Koote- dispersal by release location and age at release, (2) to evalu- nay River releases (Figure 1). Sonic receivers tracked Burbot ate the movements of potential spawners during the spawning movement from these release groups until the end of our study season that would indicate spawning locations, and (3) to in- period on March 31, 2012. vestigate the potential for imprinting from release locations for Tagging hatchery-reared Burbot from Moyie Lake broodstock. A total of 121 hatchery-reared Burbot 1–3 years old were tagged with passive integrated transponder tags (Biomark, METHODS Inc., Boise, Idaho) and sonic V9-2 L tags (VEMCO Divi- sion, AMIRIX Systems, Inc., Halifax, Nova Scotia). All sonic Study Area transmitters were 9 × 29 mm in length and weighed 4.7 g in Downloaded by [Department Of Fisheries] at 22:52 25 November 2013 The study area extended along the Kootenay River from river kilometer (RKM; measured from the outlet of Duncan Lake, TABLE 2. Release dates, locations, and numbers for all sonic-tagged, which flows into the north end of Kootenay Lake) 259 in Idaho hatchery-reared Burbot released into the Kootenay River and its tributaries downstream into British Columbia (Figure 1). The portion in from 2009 to 2011, in chronological order. British Columbia includes both riverine habitat (RKM 121– 165) and lacustrine habitat in Kootenay Lake (RKM 18–120). Release date Release location Number released 2 Kootenay Lake has a surface area of 390 km and is a fjord- Oct 21, 2009 Goat River 30 like lake, running north–south in the trench formed between the Aug 10, 2010 Boundary Creek 20 Selkirk and Purcell mountains. The Kootenay River is regulated Aug 10, 2010 Moyie River 20 by Libby Dam in Montana, which was completed in 1972 to Nov 3, 2010 Goat River 15 provide flood control and power generation. Aug 2, 2011 RKM 259 (mainstem) 6 (+ 3 delay-start tags) Aug 2, 2011 RKM 240 (mainstem) 7 (+ 3 delay-start tags) Burbot Culture Aug 2, 2011 RKM 170 (mainstem) 7 (+ 4 delay-start tags) The Burbot reared in the hatchery program originated from Oct, 2011 Goat River 4 (+ 2 delay-start tags) a lacustrine population in Moyie Lake. Gametes were collected SURVIVAL AND DISPERSAL OF HATCHERY-REARED BURBOT 1673 Downloaded by [Department Of Fisheries] at 22:52 25 November 2013

FIGURE 1. Map of study area, which extended from RKM 18 to RKM 259 in Kootenay Lake and the Kootenay River. The river flows from Montana through Idaho into British Columbia. The circles indicate the locations of the sonic receivers and the stars the release locations for the sonic-tagged hatchery Burbot. [Figure available in color online.] 1674 STEPHENSON ET AL.

up as point stations on the river and as many as four receivers together serve as gate stations at specific transects on the lake. The mean distance between stations was 4.4 km (range, 13– 100 m; further described in Neufeld and Rust 2009). All of the tracking locations of fish resulted from detections by this passive monitoring system.

Telemetry Data Analysis Survival and data verification.—Detections were first evalu- ated to eliminate erroneous ones (methods are further described in Neufeld and Rust 2009). Release survival was defined as surviving a minimum of 1 month postrelease, as indicated by detections by receivers. Long-term survival was evaluated by identifying tags that were detected by sonic receivers a mini- mum of 1 year postrelease. To investigate the factors influencing survival, survival was compared across release locations and age FIGURE 2. Mean TLs (whiskers = 95% CIs) of hatchery-reared Burbot, by at release. age at release. The fish were measured prior to being tagged at the Aquaculture Dispersal.—Dispersal from release sites was calculated as a Research Institute at the University of Idaho from 2009 to 2011. linear distance using RKM locations for each fish, subtracting air and 2.9 g in water. In 2009, V9 tags with two different pulse the most downstream detection location from the most upstream rates were used to optimize the trade-off between battery life and one. Dispersal period was defined as the first 6 months postre- pulse frequency. One-third of the tags (n = 10) were set with a lease. Because tag codes were not recorded before the release delay time of 90–240 s between pulses (nominal delay = 165 s; of several fish, the specific release locations for several Burbot battery life = 751 d), and the other two-thirds (n = 20) were set were assigned using the first detection location. Comparisons with a delay time of 60–180 s (nominal delay = 120 s; battery were made across years, release locations, and ages at release. A life = 573 d). All of the tags in 2010 (n = 55) and two-thirds one-way analysis of variance (ANOVA) was performed (Sigma of the tags in 2011 (n = 24) had a nominal delay of 165 s and Plot version 12.3, SYSTAT software) to determine statistical a battery life of 751 d. The remaining tags in the 2011 releases significance. (n = 12) were delay-start tags that also have a nominal delay of Spawning movements.—We evaluated the movements of 165 s but that will not be active until 775 d posttagging. There- sonic-tagged Burbot that were age 3 or older during the spawn- fore, these delay-start tags were not intended to be detected ing season. Without recapturing individuals we could not con- until summer 2013 and the data from them were not included in firm their spawning condition, but sexual maturity often oc- this movement analysis. In total, we analyzed 109 sonic-tagged curs at 3 years of age in other southern latitudinal populations Burbot for release survival, movements, and habitat use. in Canada (Arndt and Hutchinson 2000; Stapanian and The V9 tags were surgically implanted within the peri- Madenjian 2007). Additionally, data collected from hatchery toneum by methods similar to those described by Neufeld and progeny that were held and observed in captivity indicate that Rust (2009). Burbot were anesthetized with MS-222 (tricaine males and females can mature at sizes as small as 47 cm and methanesulfonate), and a 12-mm incision was made on the lat- 500 g (N. Jensen, ARI, unpublished data) and several of our eral surface approximately two-thirds of the way from the ven- sonic-tagged fish were this size at ages 2 and 3 at the time of

Downloaded by [Department Of Fisheries] at 22:52 25 November 2013 tral midline to the lateral line and midway between the pectoral tagging. girdle and the anal fin. Transmitters and PIT tags were cleaned Peak spawning in the Kootenay River occurs in the first with ethanol and rinsed with distilled water prior to being placed 2 weeks of February (Paragamian 2000; Paragamian and in the peritoneal cavity. The incision was then closed with two Wakkinen 2008). We analyzed the movements of all Burbot or three monofilament synthetic absorbable sutures (Ethicon 3/0 over 3 years old during a period from January 15 to February 1 28 in all 3 years (2010–2012) as representative of the spawn- PDS II, SH-1 22 mm 2 c taper needles). Burbot were monitored at the ARI for 30 d posttagging to ensure physical recovery prior ing period. This study used passive telemetry, and we did not to release. have daily detections of all fish during the spawning period. To fill this void, we manipulated our data set to simulate an ac- Tracking tive telemetry study in which we assigned a daily location for In the Kootenay River and Kootenay Lake there was an estab- each fish. This daily location was the RKM of the last receiver lished array of 73 Vemco (VR2s and VR2Ws) 81-kHz receivers where it was detected. These data were then used to calculate that extends from the border of Idaho with Montana, down- daily movements and areas of high use and to make compar- stream throughout the Kootenay River, in the Goat River, and isons across years in order to determine potential site fidelity throughout Kootenay Lake (Figure 1). The receivers were set and imprinting to release locations. SURVIVAL AND DISPERSAL OF HATCHERY-REARED BURBOT 1675

TABLE 3. Release survival (first month at large) for all releases of Burbot from 2009 to 2011, by release location and age at release.

Release survival at age Release locationa Age 1 Age 2 Age 3 All RKM 259 1/6 (17%) 1/6 (17%) Moyie River 3/?b 4/?b 1/3 (33%) 8/20 (40%) RKM 240 6/7 (86%) 6/7 (86%) RKM 170 6/7 (86%) 6/7 (86%) Boundary Creek 5/?b 5/?b 2/2 (100%) 12/20 (60%) Goat River 11/15 (73%) 22/27 (81%) 6/7 (86%) 39/49 (80%) All 19/31 (61%) 44/66 (67%) 9/12 (75%) 72/109 (66%)

aRKM values refer to the main-stem Kootenay River. bThere were 6 age-1 and 10 age-2 Burbot that were never detected and could not be properly assigned to a release location (which could be either the Moyie River or Boundary Creek). Including these fish, the release survival of age-1 Burbot was 31% and that of age-2 Burbot was 47%.

RESULTS River in Idaho downstream to the Lardeau River Delta in the North Arm of Kootenay Lake, covering a total linear distance of Detections and Survival 235 km. Dispersal differed across release locations and included The survival of the tagged Burbot was very high; 121 both upstream and downstream movements within the 6-month Burbot survived the 30-d posttagging laboratory observation pe- dispersal period. Of all the Burbot included in the dispersal cal- riod and were successfully released at six different release sites culations, 29% (21/72) entered Kootenay Lake at some point, 9 on the Kootenay River. Only 1 Burbot did not survive the initial of which made it all the way to the Lardeau Delta (over 100 km observation period and likely never recovered from anesthesia. [north-to-south distance] across the lake. There were significant Of the 121 Burbot released 109 were evaluated postrelease, as differences in the mean dispersal distances for the six release the remaining 12 were tagged with delay-start tags that were not locations (ANOVA: F4, 68 = 5.89, P < 0.001; Figure 3). When active during our study period. Based on the pilot study analyz- ing the data from the first release in 2009 (Neufeld et al. 2011a), there were no differences in detectability between the two dif- ferent pulse delay tags used in the first release (nominal delays of 120 and 180 s); as a result, only the longer pulse interval tags were used for tagged fish in 2010 and 2011 to ensure the longest battery life possible. Overall, the release survival of all Bur- bot was high; 72 of 109 (66%) were subsequently detected for a minimum of 1 month postrelease by the array of sonic receivers. A more detailed analysis of survival rates during the first month postrelease suggests that release survival differs across release sites and ages at release. The releases at the two sites furthest upstream (the Moyie River and RKM 259 of the main- stem Kootenay River) had the lowest release survival; all other Downloaded by [Department Of Fisheries] at 22:52 25 November 2013 release locations had 60% survival or better (Table 3). All release data were pooled to investigate potential survival differences with age at release. This analysis showed that the age-1 release group had the lowest survival and that survival increased with age (Table 3). In addition to release survival, the length of this study permitted a look at annual survival rates for the first 2 release years, which included all release ages and the three tributary release locations. The first-year survival for both the 2009 and 2010 releases was estimated at 54% (46/85); excluding the mortalities that occurred soon after release, annual survival increases to 78% (46/59).

Dispersal FIGURE 3. Mean dispersal distances (whiskers = 95% CIs) from release There was high overall dispersal, with a mean of 50.5 km (SE, locations for sonic-tagged, hatchery-reared Burbot, by release location. The fish 6.6) for all released Burbot; dispersal extended from the Moyie were released into the Kootenay River from 2009 to 2011. 1676 STEPHENSON ET AL.

FIGURE 4. Mean dispersal distances (whiskers = 95% CIs) from release locations for sonic-tagged, hatchery-reared Burbot, by age at release. FIGURE 5. Mean times (whiskers = 95% CIs) taken to disperse from release tributaries to the main-stem Kootenay River by the sonic-tagged Burbot released in the Goat River, Boundary Creek, and the Moyie River from 2009 to 2011, by we pooled the data from all release locations and stratified them age at release. based on age, they suggested that there are differences in dis- persal distance based on age at release (ANOVA: F2, 70 = 9.95, no data to conclusively evaluate imprinting to release locations. P < 0.001; Figure 4). Given the magnitude of the difference for Although a few Burbot released in the Goat River returned dur- the Burbot released at 1 year old (mean, 7.2 km; SE, 4.0) and the ing the spawning period, even more Goat River releases were older releases (age 2: mean, 65.9 km; SE, 8.9; and age 3: mean, detected elsewhere during the spawning season. Dispersal and 76.9 km; SE, 17.5), further investigations into the movements of subsequent detections near multiple spawning locations, though the younger releases were initiated. Our analysis suggested that not conclusive data for evaluating imprinting, suggest that when released into tributaries, the youngest groups remained in the respective tributaries for on average 1 year postrelease, a significantly longer period than the older releases (ANOVA: F2, 57 = 35.9, P < 0.001; Figure 5).

Spawning Movement

Downloaded by [Department Of Fisheries] at 22:52 25 November 2013 Of our sonic tagged Burbot, 51 individuals were of poten- tial spawning age and were alive with active tags through a minimum of one spawning season; of these 51, 17 were active through two spawning seasons. There were extensive individual daily movements during the spawning season for many of the 51 Burbot, up to 59 km upstream and 33 km downstream. Fur- thermore, when we removed days of no movement and did not differentiate for directionality, the mean daily movement was 10.6 km/d (SE = 0.7). There are three historically known wild Burbot spawning locations in the Kootenay River (Paragamian 2000), includ- ing Ambush Rock (RKM 244.7), the Goat River, and the FIGURE 6. Relative frequency of the daily locations of adult Burbot (age 3+) Lardeau Delta on Kootenay Lake (RKM 18). These three lo- in Kootenay Lake (RKM 18–121) and the Kootenay River (RKM 122–259) cations had high use by sonic-tagged, spawning-age Burbot during all three spawning seasons (2010–2012). The three historical spawning during the spawning season (Figure 6). However, there were locations—the Lardeau Delta, the Goat River, and Ambush Rock—are noted. SURVIVAL AND DISPERSAL OF HATCHERY-REARED BURBOT 1677

spawning site selection by hatchery fish is similar to that of zone of lakes and cooler deep holes in rivers (Pa¨akk¨ onen¨ et al. wild individuals. 2000; Amundsen et al. 2003). Releases of Burbot in the Goat River were completed in the fall, while releases in Boundary Creek and the Moyie River were completed during the summer DISCUSSION months when temperatures were higher (the summer release This telemetry study permitted the first evaluation of annual temperature ranged from 16◦Cto18◦C, the fall release temper- survival for subadult and adult hatchery-reared Burbot released ature from 5◦Cto8◦C). It is possible that the lower survival at the into a historical range. The annual survival for the first year Moyie River and Boundary Creek resulted from differences in at large (54–78%) was comparable to the annual survival rate season of release, as the highest mortality often occurs in the first of wild adult Burbot from the broodstock source (Moyie Lake, few days to weeks postrelease (Brown and Day 2002). During which is within the Kootenay River drainage area) and has been this period, while Burbot are adjusting to feeding and predation estimated between 53% (Prince 2007) and 70–80% (Neufeld avoidance in the wild, other stressors such as warmer water tem- 2008). In another southern population, Schram (2000) estimated perature may contribute to higher mortality rates; however, we the total annual survival of adult Burbot in western Lake Su- were not able to suitably evaluate this possibility with our study perior at 57%. The comparably high annual survival suggests design. Future experimental releases should target comparisons that this southern Kootenay system is suitable to support the across all release sites by season. long-term survival of adult and subadult Burbot. Similar to the findings of Neufeld et al. (2011a) and other The postrelease survival of all sonic-tagged, hatchery-reared Burbot telemetry studies (Breeser et al. 1988), dispersal from re- Burbot from Moyie Lake broodstock was high. Additionally, lease sites was extensive for age-2 and age-3 Burbot. In fact, the the postrelease survival at the Moyie River and Boundary Creek maximum daily movements that we recorded were greater than may be underestimated because some released fish may have those in any other study (e.g., Breeser et al. 1988; Paragamian survived during the study period but remained undetected in et al. 2005); fish traveled up to 59 km/d, which may indicate that these tributaries lacking receivers. Previous telemetry work in Burbot are stronger swimmers than often thought. With disper- the Kootenay system focused on larger-scale movements, and sal throughout the river and downstream to the Lardeau Delta as a result receivers were positioned to maximize coverage area (RKM 18), these data suggest that a minimal number of release (up to 7 km apart in the river and 13 km in the lake). The only locations may be adequate to allow released Burbot access to all tributary in this study area with sonic receivers was the Goat available habitats. However, our analysis suggested that age-1 River, and these receivers were only deployed with the first Burbot do not disperse as far or as quickly as the older Burbot. release of sonic-tagged, hatchery-reared Burbot in 2009. The For example, the majority of age-1 Burbot released into the Goat results from a pilot study with the first release of age-2 and -3 River in 2010 never left the Goat River, whereas the age-2 and Burbot suggested that all Burbot moved out of tributaries quickly age-3 Burbot released into the Goat River in 2009 all left the (Neufeld et al. 2011a). Therefore, due to sufficient coverage in river within 1–9 d (Neufeld et al. 2011a). the main stem and the assumed quick dispersal from release The lower dispersal rates associated with younger hatchery- locations, receivers were not deployed in other release tributaries reared Burbot could have large implications for release strate- (Boundary Creek and the Moyie River). However, data from the gies. Based on our data for age-2 + Burbot we expect that 2010 and 2011 releases suggested that the majority of age-1 their movement and dispersal will increase as they get older, Burbot released into tributaries remained there for an extended but there is a possibility that these Burbot will not disperse at length of time postrelease; this is evident from the Goat River all. Slavik et al. (2005) investigated the home ranges of riverine receivers and the first detections of the other age-1 tributary Burbot and found that they can be quite small (mean, ∼7km),

Downloaded by [Department Of Fisheries] at 22:52 25 November 2013 releases in the main stem over 1 year postrelease. Perhaps this with larger migrations during the spawning season. If younger is related to an ontogenetic niche shift in food resource use; releases tend to remain close to their release sites, more re- Amundsen (2003) found that Burbot are largely piscivorous but lease locations would be needed to have Burbot throughout the that Burbot smaller than 300 mm also had large amounts of Kootenay Burbot’s historical range. Furthermore, if Burbot do zoobenthos in their diet. It is possible that tributaries provide a disperse farther as they get older, there could be implications for better food source for these smaller age-1 Burbot than for the spawning site selection, with potential imprinting from lengthy large age-2 + Burbot. residence near their release location. If, in fact, there were lower survival rates for the Burbot re- The movements of age-3 + Burbot during the spawning leased at the Moyie River and Boundary Creek release sites, period were extensive, and telemetry detections (as indicators our data suggest that this was due to season of release. Burbot of habitat use) suggest that there was use of three known are a stenothermal species that inhabit large cool rivers of north spawning locations (Ambush Rock, the Goat River, and the temperate regions, preferring temperatures below 14◦C(Edsall Lardeau Delta). Our data suggest that hatchery-reared Burbot et al. 1993; Pa¨akk¨ onen¨ and Marjomaki¨ 2000; Hofmann and have contributed to wild spawning, but our evaluation of Fischer 2002). In summer, Burbot are less active than in the fall imprinting from release locations was inconclusive. Imprinting and winter months and can be found primarily in the profundal has been observed for Burbot in other systems, in which 1678 STEPHENSON ET AL.

tagged wild Burbot demonstrated fidelity to specific rivers and Arndt, S. K., and J. Hutchinson. 2000. Characteristics of Burbot spawning in a estuaries (Hedin 1983; Hudd and Lehtonen 1987). Although tributary of Columbia Lake, British Columbia, over a four year period. Pages our evaluation of imprinting for age-2 and -3 releases was 48–60 in V.L. Paragamian and D. W. Willis, editors. Burbot: biology, ecology, and management. American Fisheries Society, Fisheries Management Section inconclusive, imprinting may be more likely to occur for Burbot Publication 1, Spokane, Washington. released at age 1 or younger, as these fish do not disperse as Breeser, S. W., F. D. Stearns, M. W. Smith, R. L. West, and J. B. Reynolds. quickly. This is a subject that warrants further evaluation. 1988. Observations of movement and habitat preferences of Burbot in an Overall, this evaluation of telemetry data for hatchery-reared Alaskan glacial river system. Transactions of the American Fisheries Society Burbot supports a positive outlook on the current aquaculture- 117:506–509. Brown, C., and R. Day. 2002. The future of stock enhancements: lessons related rehabilitation efforts for the Kootenay Burbot popu- for hatchery practice from conservation biology. Fish and Fisheries 3: lation. As found in the first analysis of the telemetry data 79–94. from hatchery-reared Burbot released into the Kootenay River Edsall, T. A., G. W. Kennedy, and W. H. Horns. 1993. Distribution, abundance, (Neufeld et al. 2011a), the progeny of lake-origin broodstock and resting microhabitats on Julian’s Reef, southwestern Lake Michigan. had high survival and dispersed throughout the system, uti- Transactions of the American Fisheries Society 122:560–574. Hedin, J. 1983. Seasonal spawning migrations of the Burbot, (Lota lota L.)ina lizing the entire range of available habitat (both riverine and coastal stream of the northern Bothnian Sea. Fauna Norrlandica 6:1–9. lacustrine). This finding of behavioral plasticity in Burbot is Hofmann, N., and P. Fischer. 2002. Temperature preferences and critical ther- supported by the current study, where the use of the Moyie Lake mal limits of Burbot: implications for habitat selection and ontogenetic Burbot population as a brood source has proven to be successful habitat shift. Transactions of the American Fisheries Society 131:1164– for Kootenay River reintroduction efforts. Utilizing telemetry to 1172. Hudd, R., and H. Lehtonen. 1987. Migrations and home ranges of natural and the track dispersal and survival of hatchery-reared Burbot has transplanted Burbot (Lota lota) off the coast of Finland. Pages 201–206 in also proven to be successful. Supporting mark–recapture efforts, S. O. Kullander and B. Fernholm, editors. Proceedings of the fifth congress of this project provides a passive, data-rich method of evaluating European ichthyologists. Swedish Museum of Natural History, Stockholm. Burbot release strategies. With the current low recapture results Ireland, S. C., and P. N. Perry. 2008. Burbot restoration in the Kootenai River for hatchery Burbot in the Kootenay system, continued teleme- basin: using agency, tribal, and community collaboration to develop and implement a conservation strategy. Pages 251–256 in V. L. Paragamian and try studies are crucial to evaluating the methodology and success D. H. Bennett, editors. Burbot: ecology, management, and culture. American of releasing hatchery-reared Burbot into the Kootenay River. Fisheries Society, Symposium 59, Bethesda, Maryland. KVRI (Kootenai Valley Resource Initiative) Burbot Committee. 2005. Kootenai River/Kootenay Lake Burbot conservation strategy. Kootenai Tribe of Idaho, ACKNOWLEDGMENTS Bonner’s Ferry. We would like to thank the Kootenai Tribe of Idaho (KTOI) Neufeld, M. D. 2005. White Sturgeon and Burbot recovery progress in British Columbia, 2004–05. Ministry of Environment Report, Nelson, British and the Idaho Department of Fish and Game (IDFG) for spon- Columbia. soring this investigation and for Burbot rehabilitation efforts Neufeld, M. D. 2008. Moyie Lake Burbot: population assessment 2007. Ministry in collaboration with the Kootenai Valley Resource Initiative. of Environment Report, Nelson, British Columbia. We would like to thank Christin Davis of the British Columbia Neufeld, M. D., K. D. Cain, N. R. Jensen, S. C. Ireland, and V. L. Paragamian. Ministry of Forests, Lands and Natural Resource Operations; 2011a. Movement of lake-origin Burbot reared in a hatchery environment and released into a large river drainage. North American Journal of Fisheries Don Miller of Kootenay Wildlife Services Ltd.; Jack Siple, Eric Management 31:56–62. Wagner, and Chris Lewandowski of the KTOI; and Corie Laude Neufeld, M. D., C. A. Davis, K. D. Cain, N. R. Jensen, S. C. Ireland, and of IDFG for field and logistical support. We would like to thank C. C. Lewandowski. 2011b. Evaluation of methods for the collection and the many people at the University of Idaho Aquaculture Re- fertilization of Burbot eggs from a wild stock for conservation aquaculture search Institute who are responsible for the continued success operations. Journal of Applied Ichthyology 27:9–15. Neufeld, M. D., V. L. Paragamian, S. Ireland, P. Anders, K. Cain, and N. Jensen. Downloaded by [Department Of Fisheries] at 22:52 25 November 2013 of spawning and rearing Burbot. Paul Anders of Cramer Fish 2009. Kootenay River Burbot–5-year operational research plan (2006–2010). Sciences reviewed an earlier version of this report. The Bon- Report prepared for the Kootenai Valley Resource Initiative and Kootenai neville Power Administration provided funding for this project Tribe of Idaho, working draft, Bonner’s Ferry, Idaho. through the Northwest Power and Conservation Council’s Fish Neufeld, M. D., and P. Rust. 2009. Using passive sonic telemetry meth- and Wildlife Program, in cooperation with the IDFG and KTOI. ods to evaluate dispersal and subsequent movements of hatchery-reared White Sturgeon in the Kootenay River. Journal of Applied Ichthyology 25: 27–33. Pa¨akk¨ onen,¨ J. -P. J., and T. J. Marjomaki.¨ 2000. Feeding of Burbot, Lota lota, REFERENCES at different temperatures. Environmental Biology of Fishes 58:109–112. Ahrens, R., and J. Korman. 2002. What happened to the West Arm Burbot Paragamian, V. L. 2000. The effects of varying flows on Burbot spawning stock in Kootenay Lake? Use of an age-structured population model to de- migrations in the Kootenai River, Idaho and British Columbia, Canada, after termine the possible causes of recruitment failure. Report prepared for the construction of Libby Dam. Pages 111–123 in V. L. Paragamian and D. British Columbia Ministry of Water, Land, and Air Protection, the Habitat W. Willis, editors. Burbot: biology, ecology, and management. American Conservation Trust Fund, and the Bonneville Power Administration, Nelson. Fisheries Society, Fisheries Management Section Publication 1, Spokane, Amundsen, P., T. Bøhn, O. A. Popova, F. J. Staldvik, Y. S. Reshetnikov, N. A. Washington. Kashulin, and A. A. Lukin. 2003. Ontogenetic niche shifts and resource Paragamian, V.L., R. R. Hardy, and B. B. Gunderman. 2005. Effects of regulated partitioning in a subarctic piscivore fish guild. Hydrobiologia 497:109–119. discharge on Burbot migration. Journal of Fish Biology 66:1199–1213. SURVIVAL AND DISPERSAL OF HATCHERY-REARED BURBOT 1679

Paragamian, V.L., and V.D. Wakkinen. 2008. Seasonal movement and the inter- Schram, S. T. 2000. Seasonal movement and mortality estimated of Burbot in action of temperature and discharge on Burbot in the Kootenai River, Idaho, Wisconsin waters of western Lake Superior. Pages 90–95 in V. L. Paragamian USA, and British Columbia, Canada. Pages 55–77 in V. L. Paragamian and and D. W. Willis, editors. Burbot: biology, ecology, and management. D. H. Bennett, editors. Burbot: ecology, management, and culture. American American Fisheries Society, Fisheries Management Section Publication 1, Fisheries Society, Symposium 59, Bethesda, Maryland. Spokane, Washington. Paragamian, V. L., V. Whitman, J. Hammond, and H. Andrusak. 2000. Col- Slavık, O., L. Bartos,ˇ and D. Mattas. 2005. Does stream morphology pre- lapse of the Burbot fisheries in Kootenay Lake, British Columbia Canada, dict the home range size in Burbot? Environmental Biology of Fishes 74: and the Kootenai River, Idaho, USA, post-Libby Dam. Pages 155–164 in 89–98. V. L. Paragamian and D. W. Willis, editors. Burbot: biology, ecology, and Stapanian, M. A., and C. P. Madenjian. 2007. Evidence that Lake Trout served management. American Fisheries Society, Fisheries Management Section as a buffer against Sea Lamprey predation on Burbot in Lake Erie. North Publication 1, Spokane, Washington. American Journal of Fisheries Management 27:238–245. Partridge, F. 1980. Subproject IV: river and stream investigations study VI: Stapanian, M. A., V. L. Paragamian, C. P. Madenjian, J. R. Jackson, J. Kootenai River fisheries investigations. Period covered: 1 March 1979 to 29 Lappalainen, M. J. Evenson, and M. D. Neufeld. 2010. Worldwide status February 1980. Idaho Department of Fish and Game, Federal Aid in Sport of Burbot and conservation measures. Fish and Fisheries 11:34–56. Fish Restoration, Project F-73-R-2, Job Performance Report, Bonner’s Ferry. Stapanian, M. A., L. D. Witzel, and A. Cook. 2010. Recruitment of Burbot (Lota Prince, A. 2007. East Kootenay Burbot population assessment. Westslope lota L.) in Lake Erie: an empirical modeling approach. Ecology of Freshwater Fisheries, Cranbrook, British Columbia. Fish 19:326–337. Downloaded by [Department Of Fisheries] at 22:52 25 November 2013 This article was downloaded by: [Department Of Fisheries] On: 25 November 2013, At: 22:53 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Transactions of the American Fisheries Society Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/utaf20 Effects of Stocking Density on Survival and Yield of North American Burbot Reared under Semi-Intensive Conditions James M. Barron a b , Nathan R. Jensen c a , Paul J. Anders a c , Joshua P. Egan a c , Susan C. Ireland d & Kenneth D. Cain a e a Department of Fish and Wildlife Resources , University of Idaho , Post Office Box 441136, Moscow , Idaho , 83844-1136 , USA b U.S. Fish and Wildlife Service , Abernathy Fish Technology Center , 1440 Abernathy Creek Road, Longview , Washington , 98632 , USA c Cramer Fish Sciences , 317 West 6th Street, Suite 204, Moscow , Idaho , 83843 , USA d Kootenai Tribe of Idaho, Post Office Box 1269 , Bonners Ferry , Idaho , 83805 , USA e National Centre for Marine Conservation and Resource Sustainability , University of Tasmania , Locked Bag 1370, Launceston , Tasmania , 7250 , Australia Published online: 20 Nov 2013.

To cite this article: James M. Barron , Nathan R. Jensen , Paul J. Anders , Joshua P. Egan , Susan C. Ireland & Kenneth D. Cain (2013) Effects of Stocking Density on Survival and Yield of North American Burbot Reared under Semi-Intensive Conditions, Transactions of the American Fisheries Society, 142:6, 1680-1687, DOI: 10.1080/00028487.2013.788557 To link to this article: http://dx.doi.org/10.1080/00028487.2013.788557

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions Transactions of the American Fisheries Society 142:1680–1687, 2013 C American Fisheries Society 2013 ISSN: 0002-8487 print / 1548-8659 online DOI: 10.1080/00028487.2013.788557

SPECIAL SECTION: BURBOT

Effects of Stocking Density on Survival and Yield of North American Burbot Reared under Semi-Intensive Conditions

James M. Barron Department of Fish and Wildlife Resources, University of Idaho, Post Office Box 441136, Moscow, Idaho 83844-1136, USA; and U.S. Fish and Wildlife Service, Abernathy Fish Technology Center, 1440 Abernathy Creek Road, Longview, Washington 98632, USA Nathan R. Jensen Cramer Fish Sciences, 317 West 6th Street, Suite 204, Moscow, Idaho 83843, USA; and Department of Fish and Wildlife Resources and Aquaculture Research Institute, University of Idaho, Post Office Box 441136, Moscow, Idaho 83844-1136, USA Paul J. Anders and Joshua P. Egan Cramer Fish Sciences, 317 West 6th Street, Suite 204, Moscow, Idaho 83843, USA; and Department of Fish and Wildlife Resources, University of Idaho, Post Office Box 441136, Moscow, Idaho 83844-1136, USA Susan C. Ireland Kootenai Tribe of Idaho, Post Office Box 1269, Bonners Ferry, Idaho 83805, USA Kenneth D. Cain* Department of Fish and Wildlife Resources and Aquaculture Research Institute, University of Idaho, Post Office Box 441136, Moscow, Idaho 83844-1136, USA; and National Centre for Marine Conservation and Resource Sustainability, University of Tasmania, Locked Bag 1370, Launceston, Tasmania 7250, Australia

Abstract The effects of six stocking densities on the survival and yield of larval Burbot Lota lota in a semi-intensive culture setting were investigated over a 3-year period. A stocking initiation trial indicated that a stocking date of at least 45 d Downloaded by [Department Of Fisheries] at 22:53 25 November 2013 after the first exogenous feeding (DPEF) would yield surviving juveniles after a 108-d semi-intensive culture period. Following this, stocking density was investigated, and larval Burbot were stocked into in-ground outdoor tanks 45 DPEF at densities of 50, 100, 150, 200, 250, and 300 larvae/m2. Tanks were harvested after 65 d, and the trials were repeated over two consecutive years. At harvest, the mean TL of fish ranged from 41 to 68 mm and the mean weight from 0.5 to 2.1 g over both years. Survival ranged from 1.0% to 12.7%, with lower stocking densities exhibiting higher survival. An exponential decay model revealed a significant influence of stocking density on survival, with 50 larvae/m2 being predicted to provide the highest survival and the maximum yield being predicted to occur at 100/m2.These results indicate that a stocking density of 100 larvae/m2 should not be exceeded under the conditions described in this study. This experiment demonstrated that semi-intensive culture strategies can be successfully adapted for North American Burbot. Relative to other culture methods, this semi-intensive approach may represent a less labor-intensive and less costly method of efficiently producing Burbot for conservation or commercial production programs.

*Corresponding author: [email protected] Received September 4, 2012; accepted March 14, 2013

1680 SEMI-INTENSIVE BURBOT CULTURE 1681

The Burbot Lota lota is the only true freshwater gadiform When managing a culture pond, whether semi-intensive or fish. A cold-adapted species, Burbot have a circumpolar distri- extensive, several critical decisions must be made, including bution, generally occurring north of 40◦N (Van Houdt et al. when to stock, how many individuals to stock, and when to har- 2003). Presently, two genetically distinct Burbot subspecies vest. Successful methods for semi-intensive and extensive cul- have been classified: the Eurasian subspecies, Lota lota lota, ture of Burbot require knowledge of when larvae can effectively which ranges from Europe to the Great Slave Lake in North be transitioned from an intensive culture system to an alternative America, and the North American subspecies, Lota lota macu- rearing system. Additionally, targeting optimal larval stocking losa, which is found south of the Great Slave Lake (Van Houdt densities may enhance culture success. An understanding of et al. 2003). Burbot populations have declined across parts of how long the pond culture phase can be successfully extended their geographic range, making aquaculture an attractive tool is also critical to ensuring consistent survival and production. for restoring imperiled populations (Maitland and Lyle 1990, Extensive culture of Walleyes Sander vitreus and Striped Bass 1996; Keith and Allardi 1996; Argent et al. 2000; Arndt and Morone saxatilus focused on stocking densities from 10 to 60 Hutchinson 2000; Paragamian 2000; Stapanian et al. 2011). larvae/m2, which resulted in survival ranging from 0.7% to 96% The development of culture techniques has only recently begun (Geiger 1983; Geiger and Parker 1985; Fox and Flowers 1990; for both Eurasian and North American Burbot (Wolnicki et al. Culver et al. 1993; Qin et al. 1995). The inherent variability of 2001, 2002; Harzevili et al. 2003, 2004; Jensen et al. 2008; natural culture systems can make identification of the effects of Zarski˙ et al. 2009; Trebelsi et al. 2011; Wocher et al. 2011; stocking density on survival and yield challenging. Barron et al. 2012). Also, successful egg incubation and early Zooplankton concentrations are also of key importance dur- larval rearing methods have been developed for North American ing pond culture when zooplankton is the intended food source. Burbot (Jensen et al. 2008). Zooplankton abundance and density are difficult to control, as Newly hatched Burbot are difficult to culture due to their del- they are dictated by the interaction of multiple factors, includ- icate nature, small size at hatch (3.0–4.0 mm TL), and live feed ing predation, temperature, light intensity, water quality, and requirement for at least 5 weeks following alimentary tract de- nutrient availability (Qin et al. 1995). Pond culture systems are velopment (Jensen et al. 2008). Burbot pass through a larval life often fertilized to increase food availability and fish produc- stage, and at the onset of exogenous feeding the digestive system tion (Geiger 1983; Fox et al. 1992; Qin et al. 1995). In first- and body morphology do not resemble those of an adult fish. feeding Walleyes, zooplankton densities greater than 49 organ- Further development is required before larval Burbot become a isms /L at stocking were sufficient to prevent starvation mortality diminutive version of the adult, at which time they are consid- (Johnston and Mathias 1993). Ideally, the zooplankton present ered to be juveniles (Støttrup and McEvoy 2003:11–12). Burbot will be suitable for consumption by the larval fish and provide reach the juvenile stage by approximately 30 mm TL (Ryder adequate nutrition. and Pessendorfer 1992; Fischer 1999; McPhail and Paragamian The present study was conducted to identify the larval Burbot 2000). stocking densities that maximize growth, survival, and yield by Due to the challenges faced in the larval rearing of Burbot, modeling and subsequently testing the influence of stocking a natural approach such as extensive or semi-intensive culture density in a series of replicated trials. The specific objectives may be advantageous for this species. Extensive culture can were to determine (1) a stocking time that yields survivors to the be defined as rearing in static or low-water-exchange ponds juvenile life stage, (2) acceptable water quality parameters and that rely on natural food sources (Piper et al. 1982:480). Semi- zooplankton densities for rearing Burbot larvae to the juvenile intensive culture is similar but with additional intervention such life stage in a semi-intensive culture system, and (3) optimal as aeration or supplemental feedings (Sumagaysay-Chavoso and stocking densities.

Downloaded by [Department Of Fisheries] at 22:53 25 November 2013 San Diego-McGlone 2003). Labor and associated costs may be reduced in extensive and semi-intensive culture systems by offerring larvae live prey beyond brine shrimp Artemia spp. This may allow larval Burbot to reach the juvenile life stage without METHODS the need to transition to artificial diets. Stocking initiation trial.—All research was conducted at the Although extensive and semi-intensive culture systems that University of Idaho’s Aquaculture Research Institute (UI–ARI), rely on natural zooplankton as a food source may enhance using Burbot larvae from a larger ongoing restoration project. Burbot larviculture, such systems remain largely undeveloped. A stocking initiation trial was conducted to determine when to A few observations from pond culture of Eurasian Burbot have stock larval Burbot into ponds following an initial period of in- reported limited success (Harsanyi´ and Aschenbrenner 1992; tensive culture. In this trial, six circular fiberglass tanks (3.7 m Stipekˇ 1992; Kainz and Gollmann 1996; Wolnicki et al. 1999, in diameter, 0.9 m deep) were sunk into the ground to simulate a all cited by Wolnicki et al. 2001). However, techniques for pond small natural pond environment. The tanks were initially filled culture of the North American Burbot have not been reported. to a depth of 60 cm (6,400 L per tank) with municipal water Burbot are a candidate for natural rearing as an alternative to treated with 100 g of sodium thiosulfate to neutralize the chlo- intensive culture due to the challenging nature of Burbot larvi- rine. Water depths were measured weekly, and water was added culture and the high fecundity of this species. as needed to maintain the target depth. All tanks were covered 1682 BARRON ET AL.

with a shade-cloth-lined canopy to reduce solar radiation. A 15- At approximately 40 d prior to stocking Burbot larvae, each cm air diffuser (Sweetwater, Aquatic Ecosystems, Inc., Apopka, tank was seeded with cyclopoid copepods to establish densities Florida) was placed in the center of each tank to provide aera- of 3.9 and 1.4 organsims/L during 2009 and 2010, respectively. tion and circulation, defining this as semi-intensive culture. No Remnant D. magna in each tank persisted after their introduc- supplemental nutrient enhancement strategies were employed. tion in the stocking initiation trial. These remnants appeared Water temperature was recorded via Onset stowaway data log- to provide a sufficient seed population, as cladoceran densities gers (Onset Computer Corporation, Pocasset, Massachusetts) 1 d prior to starting the trials were 56.7 ± 28.7 and 66.7 ± 26.2 placed on the bottom of two of the tanks selected at random organisms/L (mean ± SD) in 2009 and 2010, respectively. for the duration of the culture period. Data loggers recorded the Zooplankton densities and water quality were measured weekly temperature every 2 h. Temperature readings were then averaged beginning 1 d before Burbot stocking, totaling 10 sampling over the culture period to derive mean water temperature. Fresh- points per tank over the duration of each trial. Aeration was water rotifers (Sach’s Aquaculture, Inc., St. Augustine, Florida) increased at least 1 h before sampling to increase mixing. Three were added to each tank 6 d prior to the first stocking treat- 1-L samples were collected from the center of each tank weekly. ment to establish a density of 10 organisms/L. At 62 d after the The zooplankton in each sample were concentrated with a final stocking treatment, daphnia Daphnia magna were added 64-µm filter and categorized as cladocerans, copepods, or other. to each tank to establish a density of 19.5 organisms/L. Daph- The zooplankton counts from the three samples were averaged nia magna were chosen because the body width of a newborn to produce a representative estimate of zooplankton density at (approximately 300–400 µm; Ranta et al. 1993) should allow each sampling point. The zooplankton density and water quality for consumption by a larva that has been consuming Artemia values measured at all sampling points over the duration of the nauplii, which are approximately 180 µm wide (Knutsen and trial were averaged for each tank. Tilseth 1985), for over 2 weeks. Additionally, the relatively large Water temperature, pH, alkalinity, nitrite, and ammonia were size of the adults (body width near 2,000 µm; Ranta et al. 1993) measured using an FF1A Aquaculture Test Kit (Hach Company, could make them a viable prey item for an extended culture Loveland, Colorado), and dissolved oxygen was measured with period. a YSI Y55 Dissolved Oxygen Meter (YSI, Inc., Yellow Springs, Three temporal stocking treatments (15, 30, and 45 d after Ohio). Tank depth was measured weekly, and if necessary water first exogenous feeding [DPEF]) were randomly assigned to was added to maintain a depth of 80 cm. the tanks, with two replicates per treatment. Fish stocked at One of six stocking density treatments (50, 100, 150, 200, 15 and 30 DPEF were offered the marine rotifer Branchionus 250, and 300 larvae/m2) were randomly assigned to each of plicatilis as a food source during intensive culture. Fish stocked six semi-intensive culture tanks. This was repeated in consec- at 45 DPEF were initially reared for 30 d on the B. plicatilis utive years to create replication within each density. Surface followed by co-feeding with Artemia franciscana for the final area–based densities were used rather than volumetric-based 15 d prior to stocking to provide a more substantial prey item densities, as the former are commonly used in pond culture op- for the growing larvae. A stocking density of 475 larvae/m2 was erations. Each tank’s surface area was 10.5 m2. Burbot larvae used in all treatments in this trial. The mean ± SD TLs (mm) were stocked after 45 d on live feeds under intensive culture con- at stocking were 4.0 ± 0.1, 4.9 ± 0.4, and 6.3 ± 1.1 for 15-, 30-, ditions. This corresponded to stocking dates of May 22, 2009, and 45-DPEF treatments, respectively. and May 21, 2010. All tanks were harvested 138 d after stocking the 15-DPEF During the intensive-culture period prior to stocking, larvae treatment. The tanks were siphoned to a depth of approx- were first offered the marine rotifer B. plicatilis, followed by A. imately 15 cm, after which all fish were easily visualized franciscana after 30 d. Larval Burbot were randomly selected

Downloaded by [Department Of Fisheries] at 22:53 25 November 2013 and captured by dip net. All extracted water went through a for each tank from a single group of fish with mean ± SD (n = 3.0-mm mesh net to ensure the capture of juvenile Burbot. The 30) TLs of 6.4 ± 0.8 mm in 2009 and 5.4 ± 0.4 mm in 2010. weight and TL of each surviving Burbot were measured and At 65 d poststocking, all tanks were harvested as previously recorded. Fulton’s condition factor (K) was calculated for each described. The total number of survivors from each tank was fish using the equation K = 100 × weight (g)/[TL (cm)]3 (Fulton recorded, along with TL, weight, and Fulton’s condition factor. 1904). Statistical analysis.—Data analysis was performed using Stocking density trials.—Density trials were conducted to SAS 9.2 (SAS Institute, Inc., Cary, North Carolina). Statisti- determine an optimal stocking density, acceptable levels of wa- cal significance was defined by P < 0.05 for all comparisons. ter quality parameters, and zooplankton densities for rearing To account for differences between years, the square roots of Burbot larvae to the juvenile life stage in a semi-intensive cul- the mean total zooplankton density (organisms/L) for each tank ture system. Two-replicate trials were conducted from late May sampled on the day prior to larvae stocking were compared using through late July in 2009 and 2010. The same tanks at the UI– a one-way analysis of variance (ANOVA). The total zooplank- ARI described in the stocking initiation trial were used during ton density prior to the stocking of Burbot was used because it this investigation. Tank depths were increased to a depth of was expected that the larval stocking density would influence 80 cm, providing a volume of 8,400 L per tank. zooplankton density after stocking. The square root was taken SEMI-INTENSIVE BURBOT CULTURE 1683

TABLE 1. Mean ± SE water quality values (ranges in parentheses) for the density trials, as determined from all of the individual weekly measurements.

Year 2009 2010 Temperature (◦C) 15.6 ± 0.3 (8.6–20.5) 13.9 ± 0.3 (9.6–18.5) Dissolved oxygen (mg/L) 8.17 ± 0.15 (6.38–11.02) 7.95 ± 0.14 (6.11–9.98) pH 8.2 ± 0.04 (7.6–8.5) 8.0 ± 0.02 (7.6–8.4) Alkalinity (mg/L) 129.4 ± 1.2 (119.7–153.9) 162.7 ± 1.7 (136.8–188.1) Un-ionized ammonia (mg/L) 0.009 ± 0.001 (0–0.04) 0.003 ± 0.001 (0–0.01)

to convert the count-based data to a continuous variable in or- weight, and K could not be measured for these treatments. der to meet the assumptions of the ANOVA. The mean water Water temperature for the duration of the culture period was temperature (◦C) over the entire culture period was also com- 13.9 ± 0.8◦C (range, 3.8–21.8◦C). pared between years by ANOVA. Relationships between stock- ing density and survival, mean TL, mean weight, and mean K Stocking Density Trials were assessed by calculating Pearson correlation coefficients Water quality.—The mean values for all water quality param- (r). The results are reported as means ± SDs across tanks unless eters that were measured on a weekly basis during each year, otherwise noted. along with the range of extremes we encountered, are found To further investigate the relationships between Burbot stock- in Table 1. The mean water temperature across tanks over the ing density and survival, the binomial proportion of survival at entire culture period was significantly higher in 2009 than in harvest was regressed on stocking density (larvae/m2)usinga 2010 (P = 0.0025; Figure 1). Nitrite was not detected in 2009 model of exponential decay, namely, or 2010. Zooplankton.—Table 2 shows the mean zooplankton den- −β × = β × ( 1 X) + sities and the range of values observed for during each trial, yˆx 0 loge C;(1) along with the values prior to Burbot stocking and the overall composition of the zooplankton population, for both 2009 and yˆx = the predicted value of survival (number of juveniles surviving/larvae stocked) at a given stocking density (X); 2010. No significant difference in total zooplankton density was P = β = survival as stocking density approaches 0; detected between 2009 and 2010 ( 0.59) on the day before 0 the trials began (i.e., prior to any influence from Burbot stock- β1 = the rate of exponential decay; X = the stocking density (number of larvae stocked/m2); ing). Other zooplankton parameters were not analyzed, as they C = the horizontal limit of survival as it approaches 0. were influenced by the density of the Burbot stocked in each tank. Survival and yield.—Of the 11,025 fish that were stocked in This model provided the best fit, as survival was observed to the density trials each year, a total of 194 fish from all tanks decline exponentially as stocking density increased. The fol- survived 65 d to harvest in 2009, compared with 365 fish in lowing equation was used to estimate yield (Y) in juveniles harvested/m2 for given values of X:

Y = yˆ × X (2)

Downloaded by [Department Of Fisheries] at 22:53 25 November 2013 x x

where Yx is the predicted yield (number of juveniles 2 harvested/m ) at a given stocking density X and yˆx is as de- fined above.

RESULTS Stocking Initiation Trial Fish surviving at harvest were present only in tanks stocked after 45 d on live feeds. Survival for this treatment averaged 0.22 ± 0.23% over a period of 108 d in the semi-intensive culture tank. In the 45-DPEF treatment, TL at harvest was 102 ± 6 mm. Mean weight was 6.5 ± 1.8 g, and K was 0.60 ± 0.06. No sur- FIGURE 1. Mean water temperature across all tanks at each sampling date vival was observed in the 15- and 30-DPEF treatments, so TL, throughout the Burbot culture periods in 2009 and 2010; error bars = SDs. 1684 BARRON ET AL.

TABLE 2. Zooplankton densities (organisms/L) 1 d prior to the beginning of the trials and over the trial periods. The values are the means ± SEs (ranges in parentheses) of all the weekly measurements across all tanks for the year. The density of seed organisms is represented by the density of cladocerans and copepods. Mean compositions (%) of the zooplankton population are also shown for the major groups.

Year 2009 2010 Prestocking zooplankton density 63 ± 13 (29–104) 73 ± 12 (32–106) Overall zooplankton density 194 ± 29 (27–1,315) 153 ± 13 (32–339) Cladoceran and copepod density 134 ± 86 (5–620) 142 ± 30 (29–317) Cladocerans (%) 63.6 83.1 Copepods (%) 5.2 10.2 Other aquatic organisms (%) 31.4 6.7

2010. The survival and yield values at each density are found intensive culture and live feeding were required for at least 45 d in Table 3. The Pearson correlation coefficient for survival by prior to stocking. The initial zooplankton densities in both den- stocking density was −0.70, justifying further modeling of this sity trials were comparable to the level of 49 organisms/L recom- relationship (P = 0.0119). mended to prevent starvation in first-feeding Walleyes (Johnston In response, an exponential-decay model was selected be- and Mathias 1993). Zooplankton densities increased for the first cause it best fit the data. All terms in the model were significant, 3 weeks of the density trials, indicating that predation pres- including β0 (estimate ± SE = 0.1854 ± 0.0259; P < 0.001), sure did not exceed the zooplankton’s capacity to repopulate. β1 (0.0135 ± 0.0016; P < 0.001), and C (0.0080 ± 0.0019; The zooplankton densities observed during these trials (≥27 P = 0.0013). The predicted values indicated that survival would organisms/L) appeared to provide sufficient prey for Burbot to be highest (10.2%) at a stocking density of 50 larvae/m2, while survive the culture period. The seed organisms of D. magna and juvenile yield would be highest (5.6 juveniles/m2) at a stocking cyclopoid copepods successfully colonized the culture tanks and density of 100 larvae/m2 (Table 3). The model also revealed that increased in number for the first 3–4 weeks of culture. However, survival decreases as stocking density increases (Figure 2). densities began to decrease midway through the culture period Growth.—No correlation was found between stocking den- (Figure 3). This was likely a result of increased predation by the sity and mean TL (r = 0.03). At harvest, mean TL ranged from surviving Burbot. Although both seed organisms were detected, 51 to 68 mm in 2009 and from 41 to 59 mm in 2010. No clear the abundance of copepods (5.2–10.2% of total zooplankton) relationship was detected between stocking density and mean was far lower than that of cladocerans (63.6–83.1% of total weight (r = 0.25). Final mean weight ranged from 0.8 to 2.1 g in zooplankton). Burbot larvae and juveniles (TL ≤ 30 mm) have 2009 and from 0.5 to 1.6 g in 2010. No correlation was detected been shown to prey on both copepods and Daphnia in the natural between stocking density and K (r = 0.19). Mean K at harvest environment, but have a strong preference for copepods, which ranged from 0.56 to 0.65 in 2009 and from 0.67 to 0.84 in 2010. may have contributed to the low copepod abundance observed (Ryder and Pesendorfer 1992; Ghan and Sprules 1993). DISCUSSION Our water quality parameters appeared to be within the range Stocking Burbot larvae into semi-intensive culture systems reported during intensive culture of larval and juvenile Burbot was shown to result in the production of juveniles in each trial (Wolnicki et al. 2001, 2002; Harzevili et al. 2004; Zarskietal.˙ described here. Although the percentage survival was extremely 2009). A parameter of particular concern was un-ionized am-

Downloaded by [Department Of Fisheries] at 22:53 25 November 2013 low in the stocking initiation trial, it provided an effective start- monia; trace levels were first detected at the onset of the 4th ing point for the stocking density trials. It was found that initial week of the experiment in both years and rose steadily until

TABLE 3. Observed and predicted survival (%) and yield (juveniles/m2) over the range of stocking densities tested.

Stocking density Observed survival Predicted Observed yield Predicted (larvae/m2) (2009, 2010) survival (2009, 2010) yield 50 4.0, 10.9 10.2 2.0, 5.4 5.1 100 4.5, 12.7 5.6 4.5, 12.7 5.6 150 2.4, 2.5 3.2 3.6, 3.7 4.9 200 1.2, 2.3 2.0 2.5, 4.7 4.1 250 1.0, 1.5 1.4 2.6, 3.8 3.6 300 1.1, 1.5 1.1 3.3, 4.5 3.4 SEMI-INTENSIVE BURBOT CULTURE 1685

with the trends we observed regarding temperature, growth, and survival between years. The temperatures recorded during the study remained within the range shown to yield survival under intensive culture in Burbot larvae and juveniles (Wolnicki et al. 2001, 2002; Harzevili et al. 2004). Growth and condition were not significantly influenced by stocking density. It was expected that growth and condition would be highest at lower densities due to reduced competition for limited food resources (Fox and Flowers 1990; Qin et al. 1995). However, this trend was not observed. This may have been due to the inherent variability of natural systems, nonlimit- ing prey abundance at all stocking densities, or the inherent can- nibalism that has been observed for Burbot (Trebelsi et al. 2011). The inverse relationship between Burbot stocking density and survival observed in this study indicates that stocking density influences survival, which affects juvenile yield. However, the causes and timing of mortality during the culture period were not identified due to the inability to intermittently sample. In FIGURE 2. Relationship between the proportion surviving (number Walleye pond culture, stocking densities from 20 to 60 larvae/m2 harvested/number stocked) and stocking density, as determined by an exponential-decay model. Predicted values from the model are represented by showed no effect on survival (Fox and Flowers 1990; Qin et al. the solid line. 1995). The differences between those results and the results of our study may be due to the lower stocking densities used for harvest, with 0.04 mg/L being the highest level detected. In ju- the Walleyes or to differences in feeding ecology between the veniles of a related species, Atlantic cod Gadus morhua, chronic two species. exposure to un-ionized ammonia concentrations of 0.11 mg/L Predicted yield values (as calculated by equation 2), in- for 96 d resulted in reduced growth rates; however, decreased creased as stocking density increased until a density of 100 2 survival was not observed (Foss et al. 2004). Dissolved oxy- larvae/m was reached, at which point yield was estimated to be 2 gen remained above the minimum concentration of 6.0 mg/L 5.6 juveniles/m (Table 3). Lower yield is expected at densities 2 maintained by Zarski˙ et al. (2009) during intensive larval rear- above 100 larvae/m , as the increase in the quantity of initial fish ing of Burbot. The lower mean water temperatures in 2010 does not compensate for the decrease in survival. The observed may have contributed to the survival and growth differences yields support this finding, with the highest yield occurring at a 2 between years. Larval and juvenile Burbot tend to experience stocking density of 100 larvae/m in both years. The observed higher survival at lower temperatures but faster growth at higher survival values also support this stocking density, as the high- temperatures (Wolnicki et al. 2001, 2002; Harzevili et al. 2004; est survival rates occurred at the two lowest stocking densities, 2 Donner and Eckmann 2011; Barron et al. 2012), which agrees with depressed survival at densities above 100 larvae/m . Thus, to maximize juvenile harvest, these data support the use of stock- ing densities that do not exceed 100 larvae/m2. It should also be noted that a stocking density of 100 larvae/m2 was not the most efficient use of larvae, as survival is predicted to be higher at a 2 Downloaded by [Department Of Fisheries] at 22:53 25 November 2013 stocking density of 50 larvae/m . This study demonstrated the feasibility of semi-intensive Burbot culture. Under the conditions tested, it was shown that on average 5.6 juveniles/m2 should be produced if the initial stock- ing density is 100 larvae/m2. Burbot of the size range harvested in this study could directly contribute to conservation programs in which fish would be tagged and immediately released fol- lowing capture. The K values provide a reference by which to compare Burbot juveniles in the future. Further research should focus on validating these estimates in larger ponds and better defining predator–prey dynamics in relation to possible nutri- ent enhancement strategies. This study provides a solid starting FIGURE 3. Mean cladoceran plus copepod densities across all tanks (n = 6) point for using semi-intensive culture to enhance the production at each sampling date throughout the Burbot culture periods in 2009 and 2010; of Burbot, which may prove especially useful for conservation error bars = SEs. and restoration efforts. 1686 BARRON ET AL.

ACKNOWLEDGMENTS Geiger, J. G., and N. C. Parker. 1985. Survey of Striped Bass hatchery manage- This project was funded in part by the U.S. Fish and Wildlife ment in the southeastern United States. Progressive Fish-Culturist 47:1–13. Service (grant 14330-7-H067); we offer special thanks to Ray Ghan, D., and W. G. Sprules. 1993. Diet, prey selection and growth of larval and juvenile Burbot Lota lota (L.). Journal of Fish Biology 42:47–64. Jones for program funding and coordination. This project was Harsanyi,´ A., and P.Aschenbrenner. 1992. Die Rutte Lota lota (Linnaeus, 1758): also supported by the Kootenai Tribe of Idaho (KTOI) and the Biologie und Aufzucht. [The Burbot Lota lota (Linnaeus, 1758): biology and Bonneville Power Administration (project 198806400, contracts rearing.] Fischer und Teichwirt 10:372–376. 37267 and 46821). We extend our deepest gratitude to the KTOI, Harzevili, A. S., D. De Charleroy, J. Auwerx, J. Van Slycken, P. Dhert, the British Columbia Ministry of Lands, Forests, and Natural and P. Sorgeloos. 2003. Larvae rearing of Burbot (Lota lota)usingBra- chionus calyciflorus rotifer as starter food. Journal of Applied Ichthyology 19: Resource Operations, the Idaho Department of Fish and Game 84–87. (IDFG), Cramer Fish Sciences, the U.S. Fish and Wildlife Ser- Harzevili, A. S., I. Dooremont, I. Vught, J. Auwerx, P. Quataert, and D. De vice, and the University of Idaho Aquaculture Research Institute Charleroy. 2004. First feeding of Burbot, Lota lota larvae under different (UI–ARI), as this work would not have been possible without temperatures and light conditions. Aquaculture Research 35:49–55. the valuable collaborations between these agencies. We thank Jensen, N., S. C. Ireland, J. T. Siple, S. R. Williams, and K. D. Cain. 2008. Evaluation of egg incubation methods and larval feeding regimes for North UI–ARI staff and Scott Williams for installing the experimen- American Burbot. North American Journal of Aquaculture 70:162–170. tal in-ground tanks. Additional thanks to ARI staff members Johnston, T. A., and J. A. Mathias. 1993. Mortality of first-feeding postlarval who assisted during this study, especially Chris Thornton. We Walleye (Stizostedion vitreum) in culture ponds. Canadian Journal of Fish also thank Vaughn Paragamian of the IDFG for his involvement Aquatic Sciences 50:1835–1843. in Burbot conservation. We also extend our gratitude to the Kainz, E., and H. P. Gollmann. 1996. Laichgewinnung, Erbrutung¨ und erste Aufzuchtversuche bei Aalrutten (Lota lota). [Successful spawning, brooding, anonymous reviewers and the associate editor whose thought- and first attempts at rearing Burbot (Lota lota).] Osterreichs¨ Fischerei 49:154– ful advice greatly improved this manuscript. Finally, we thank 160. Bahman Shafii and William Price of the University of Idaho for Keith, P., and J. Allardi. 1996. Endangered freshwater fish: the situation in statistical advice and guidance. France. Pages 35–54 in A. Kirchover and D. Hefti, editors. Conserva- tion of endangered freshwater fish in Europe. Birkhauser¨ Verlag, Basel, Switzerland. REFERENCES Knutsen, G. M., and S. Tilseth. 1985. Growth, development, and feeding success Argent, D. G., R. F. Carline, and J. R. Staufer Jr. 2000. A method to identify and of Atlantic Cod larvae Gadus morhua related to egg size. Transactions of the conserve rare fishes in Pennsylvania. Journal of the Pennsylvania Academy American Fisheries Society 114:507–511. of Science 74:3–12. Maitland, P. S., and A. A. Lyle. 1990. Practical conservation of British fishes: Arndt, S., and J. Hutchinson. 2000. Characteristics of a tributary-spawning pop- current action on six declining species. Journal of Fish Biology 37(Supple- ulation of Burbot from Columbia Lake. Pages 48–60 in V. L. Paragamian and ment A):255–256. D. W. Willis, editors. Burbot: biology, ecology, and management. American Maitland, P. S., and A. A. Lyle. 1996. Threatened freshwater fishes of Great Fisheries Society, Fisheries Management Section Publication 1, Spokane, Britain. Pages 9–21 in A. Kirchover and D. Hefti, editors. Conservation of Washington. endangered freshwater fish in Europe. Birkha user¨ Verlag, Basel, Switzerland. Barron, J. M., N. R. Jensen, P. J. Anders, J. P. Egan, S. C. Ireland, and K. D. McPhail, J. D., and V. L. Paragamian. 2000. Burbot biology and life history. Cain. 2012. Effects of temperature on the intensive culture performance of Pages 11–23 in V. L. Paragamian and D. W. Willis, editors. Burbot: biology, larval and juvenile North American Burbot (Lota lota maculosa). Aquaculture ecology and management. American Fisheries Society, Fisheries Manage- 364/365:67–73. ment Section Publication 1, Spokane, Washington. Culver, D. A., S. P.Madon, and J. Qin. 1993. Percid pond production techniques: Paragamian, V. L. 2000. The effects of variable discharges on Burbot spawning timing, enrichment, and stocking density manipulation. Journal of Applied migrations in the Kootenai River, Idaho, USA, and British Columbia, Canada. Aquaculture 2:9–31. Pages 111–123 in V.L. Paragamian and D. W. Willis, editors. Burbot: biology, Donner, M. T., and R. Eckmann. 2011. Diel vertical migration of larval and ecology, and management. American Fisheries Society, Fisheries Manage- early-juvenile Burbot optimizes survival and growth in a deep, pre-alpine ment Section Publication 1, Spokane, Washington. lake. Freshwater Biology 56:916–925. Piper,R.G.,I.B.McElwain,L.E.Orme,J.P.McCranen,L.G.Fowler,andJ.R. Downloaded by [Department Of Fisheries] at 22:53 25 November 2013 Fischer, P. 1999. Otolith microstructure during the pelagic, settlement, and Leonard. 1982. Fish hatchery management. U.S. Fish and Wildlife Service, benthic phases in Burbot. Journal of Fish Biology 54:1231–43. Washington, D.C. Foss, A., S. I. Siikavuopio, B. Sæther, and T. H. Evensen. 2004. Effect of Qin, J., S. P. Madon, and D. A. Culver. 1995. Effect of larval Walleye (Stizoste- chronic ammonia exposure on growth in juvenile Atlantic Cod. Aquaculture dion vitrum) and fertilization on the plankton community: implications for 237:179–189. larval fish culture. Aquaculture 130:51–65. Fox, M. G., and D. D. Flowers. 1990. Effects of fish density on growth, survival Ranta, E., J. Bengtsson, and J. McManus. 1993. Growth, size and shape of and food consumption by juvenile Walleyes in rearing ponds. Transactions Daphnia longispina, D. magna and D. pulex. Annales Zoologici Fennici of the American Fisheries Society 119:112–121. 30:299–311. Fox, M. G., D. D. Flowers, and C. Watters. 1992. The effect of supplementary Ryder, R. A., and J. Pesendorfer. 1992. Food, growth, habitat, and community inorganic fertilization on juvenile Walleye (Stizostedion vitreum) reared in interactions of young-of-the-year Burbot, Lota lota L., in a Precambrian organically fertilized ponds. Aquaculture 106:27–40. Shield lake. Hydrobiologia 243/244:211–227. Fulton, T. W. 1904. The rate of growth of fishes. Pages 141–241 in Twenty- Stapanian, M. A., V. L. Paragamian, C. P. Madenjian, J. R. Jackson, J. Lap- second annual report of the Fishery Board for Scotland, being for the year palainen, M. J. Evenson, and M. D. Neufeld. 2011. Worldwide status of 1903. Part III—scientific investigations. James Hedderwick and Sons, Glas- Burbot and conservation measures. Fish and Fisheries 11:34–56. gow, UK. Stipek,ˇ J. 1992. Erfahrungen bei der Aufzucht der Rutte (Lota lota L.) in der Geiger, J. G., 1983. Zooplankton production and manipulation in Striped Bass Tschechoslowakei. [Experiences with the rearing of Burbot (Lota lota L.) in rearing ponds. Aquaculture 35:331–351. Czechoslovakia.] Fischer und Teichwirt 10:376–379. SEMI-INTENSIVE BURBOT CULTURE 1687

Støttrup, G. J., and L. A. McEvoy, editors 2003. Live feeds in marine aquacul- Wolnicki, J., R. Kaminski, and L. Myszkowski. 2002. Temperature-influenced ture. Blackwell Scientific Publications, Oxford, UK. growth and survival of Burbot Lota lota (L.) larvae fed live food under Sumagaysay-Chavoso, N. S., and M. L. San Diego-McGlone. 2003. Water controlled conditions. Archives of Polish Fisheries 10:109–113. quality and holding capacity of intensive and semi-intensive Milkfish (Chanos Wolnicki, J., M. Kleszcz, R. Kaminski,˜ M. Korwin-Kossakowski, and L. chanos) ponds. Aquaculture 219:413–429. Myszkowski. 1999. Burbot: a new species in Polish aquaculture-chosen as- Trebelsi, A., J. Gardeur, F. Teletchea, and P. Fontaine. 2011. Effects of 12 pects of spawning, rearing fry in ponds and controlled larvae rearing. Pages factors on Burbot Lota lota (L., 1758) weaning performances using fractional 99–105 in A. Wołos, editor. IV Krajowa konferencja rybackich uzytkownik˙ ow´ factorial design experiment. Aquaculture 316:104–110. jezior. [Title: IV national conference of users fishing lakes.] Instytut Rybactwa Van Houdt, J. K., B. Hellemans, and F. A. M. Volckaert. 2003. Phylogenetic re- Sr´ odl´ adowego im. St. Sakowicza, Olsztyn, Poland. (in Polish). lationships among Palearctic and Nearctic Burbot (Lota lota): Pleistocene ex- Wolnicki, J., L. Myszkowski, and R. Kaminski.´ 2001. The influence of water tinctions and recolonization. Molecular Phylogenetics and Evolution 29:599– temperature on the growth, survival, condition factor and biological quality 612. of juvenile Burbot, Lota Lota (L.). Archives of Polish Fisheries 9:79–86. Wocher H., A. Harsanyi,´ and F. J. Schwarz. 2011. Husbandry conditions in Zarski,˙ D., W. Sasinowski, D. Kucharczyk, M. Kwiatkowski, S. Krejszeff, and Burbot (Lota lota L.): impact of shelter availability and stocking density on K. Targonska.´ 2009. Mass initial rearing of Burbot Lota lota (L.) larvae under growth and behavior. Aquaculture 315:340–347. controlled conditions. Polish Journal of Natural Sciences 24:76–84. Downloaded by [Department Of Fisheries] at 22:53 25 November 2013 This article was downloaded by: [Department Of Fisheries] On: 25 November 2013, At: 22:53 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Transactions of the American Fisheries Society Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/utaf20 Artificial Marker Selection and Subsequent Tagging Evaluations with Juvenile Burbot Neil K. Ashton a , Susan C. Ireland b & Kenneth D. Cain c a Aquaculture Research Institute, Department of Fish and Wildlife Sciences , University of Idaho , 875 Perimeter Drive, Mail Stop 2260, Moscow , Idaho , 83844-2260 , USA b Kootenai Tribe of Idaho , Post Office Box 1269, Bonners Ferry , Idaho , 83805 , USA c Aquaculture Research Institute, Department of Fish and Wildlife Sciences , University of Idaho , 875 Perimeter Drive, Mail Stop 1136, Moscow , Idaho , 83844-1136 , USA Published online: 20 Nov 2013.

To cite this article: Neil K. Ashton , Susan C. Ireland & Kenneth D. Cain (2013) Artificial Marker Selection and Subsequent Tagging Evaluations with Juvenile Burbot, Transactions of the American Fisheries Society, 142:6, 1688-1698, DOI: 10.1080/00028487.2013.788558 To link to this article: http://dx.doi.org/10.1080/00028487.2013.788558

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions Transactions of the American Fisheries Society 142:1688–1698, 2013 C American Fisheries Society 2013 ISSN: 0002-8487 print / 1548-8659 online DOI: 10.1080/00028487.2013.788558

SPECIAL SECTION: BURBOT

Artificial Marker Selection and Subsequent Tagging Evaluations with Juvenile Burbot

Neil K. Ashton Aquaculture Research Institute, Department of Fish and Wildlife Sciences, University of Idaho, 875 Perimeter Drive, Mail Stop 2260, Moscow, Idaho 83844-2260, USA Susan C. Ireland Kootenai Tribe of Idaho, Post Office Box 1269, Bonners Ferry, Idaho 83805, USA Kenneth D. Cain* Aquaculture Research Institute, Department of Fish and Wildlife Sciences, University of Idaho, 875 Perimeter Drive, Mail Stop 1136, Moscow, Idaho 83844-1136, USA

Abstract A conservation program developed by regional stakeholders incorporates stock enhancement as one of several approaches to restore an imperiled Burbot Lota lota population native to Idaho and British Columbia. Tagging juvenile fish is pivotal to stock enhancement monitoring; however, limited information is currently available on marks or tags applied to Burbot. We identified six criteria to guide artificial marker selection that are specific to imperiled juvenile fish. A short-term experiment with age-0 Burbot (65–92 mm TL) tested fin clips, freeze brands, visible implant elastomer, passive integrated transponders, and an unmarked control group. At 4 weeks posttagging, no significant differences were found between marking treatments with respect to fish survival (100%), absolute growth rate (0.15 ± 0.06 mm/d), specific growth rate (0.55 ± 0.32 g·g−1·d−1), or condition factor (0.64 ± 0.05). Mean tag retention ranged from 88% to 100%, and no significant differences were detected between treatments. Recognition of dorsal freeze brands differed significantly between two independent tag assessments. Overall, we found no adverse short-term effects and high tag retention in this preliminary trial of seven artificial marks applied to hatchery-reared Burbot.

Downloaded by [Department Of Fisheries] at 22:53 25 November 2013 Marking or tagging fish for identification upon recapture Guy et al. 1996; Hammer and Blankenship 2001; Thorsteinsson is one of the most important tools in fisheries management 2002; ISRP 2009; Skalski et al. 2009). Biologists need to make and research (Hilborn et al. 1990). Information regarding stock informed decisions regarding which marks or tags best com- identity, movements and migration, abundance, age and growth, plement specific monitoring objectives. The selection process mortality, behavior, and stocking success can all be gleaned from can be difficult, given that many biological, physical, environ- fish marking (McFarlane et al. 1990). An assessment of suitable mental, and anthropogenic factors can affect how a mark per- markers for Burbot Lota lota has not been published despite forms in fish. Marker selection can be further complicated when growing concerns about Burbot conservation worldwide (Stapa- the species of interest is imperiled, threatened, or endangered nian et al. 2010). Previous fish-tagging reviews have formed a and has no extensive history of being monitored. Constraints consensus that all markers—natural or artificial—have advan- placed on fish sampling can restrict options for marking; poor tages and disadvantages, i.e., no single marker fulfills all tagging catch rates may limit opportunities for fish recapture and mark objectives (Wydoski and Emery 1983; McFarlane et al. 1990; recovery.

*Corresponding author: [email protected] Received September 4, 2012; accepted March 13, 2013

1688 MARKER SELECTION AND TAGGING EVALUATIONS 1689

The native Burbot population in the lower Kootenai River and calcein-stained fin rays) as well as natural markers (e.g., ge- is considered functionally extirpated as a result of physical al- netic and geochemical) made our list of potential tags suitable terations to the river and overfishing (Paragamian et al. 2000; for fingerling Burbot but were not included in this present study. Ahrens and Korman 2002; Anders et al. 2002; Paragamian Investigators have used short-term trials (∼1monthorless)to et al. 2008; Paragamian and Hansen 2009). Recovery efforts assess the potential effects of fin clipping (McNeil and Crossman are under way for this population and revolve around a con- 1979; Conover and Sheehan 1999), freeze branding (Conover servation program developed by regional stakeholders (KVRI and Sheehan 1999; Evrard 2005), VIE tagging (Bruyndoncx Burbot Committee 2005). This program incorporates stock en- et al. 2002; Reeves and Buckmeier 2009; Soula et al. 2012), hancement as one of several approaches to restore the Burbot and PIT tagging (Prentice et al. 1990; Bruyndoncx et al. 2002; population to its historical levels (Jensen et al. 2008; Neufeld Tatara 2009; Burdick 2011) on the growth and survival of small et al. 2011; Paragamian et al. 2011; Barron et al. 2012). Aqua- fish. Certain size ranges and species of fish respond differently culture production of Burbot is developing in other regions of to stress caused by tagging processes specific to each of these the world in response to population declines, extirpations, and markers (the tags themselves appear less likely to factor into fish commercial interests (Jensen et al. 2008; Vught et al. 2008; health). Therefore, we felt that a posttagging evaluation period Stapanian et al. 2010; Wocher et al. 2010; Zarski˙ et al. 2010; of 1 month should be suitable for a prototypic assessment of the Jensen et al. 2011; Paragamian and Hansen 2011; Trabelsi et al. effects of artificial marking on juvenile Burbot survival, growth, 2011). Paragamian and Hansen (2011) estimated that 110,000– and condition. 900,000 juvenile Burbot per year may need to be stocked into The specific objectives of our study were to (1) select artifi- the lower Kootenai River for the population to rebuild within cial markers with a likelihood of meeting criteria 1–4 for imper- 25 years. Affordable and effective markers that are nonlethally iled juvenile fish, (2) conduct a short-term trial to verify which detected in juvenile Burbot will need to be identified as stocking of the selected markers meet criteria 1–3 in fingerling Burbot, and monitoring efforts increase. Artificial markers have several (3) compare the advantages and disadvantages of each tested distinct advantages over some natural markers (e.g., genetic or marker with respect to criterion 4 and field applicability, and geochemical) with respect to rapid implementation. Artificial (4) evaluate our methods and provide recommendations for fu- marks can be detected in the field and are typically less expen- ture research that validates specific markers in terms of long- sive to decode (Hammer and Blankenship 2001; ISRP 2009; term tag retention and mass-marking feasibility in fingerling Skalski et al. 2009). They can be easily acquired or applied in a Burbot. variety of types, and there is ample literature on adapting artifi- cial marking techniques to novel and unique species of interest, such as Burbot. Artificial markers can also serve as a stopgap METHODS measure that enables monitoring and evaluation until more ideal Experimental design.—A 28-d study was initiated at the Uni- markers (e.g., genetic ones) are fully developed and validated. versity of Idaho’s Aquaculture Research Institute in October of The process of incorporating contemporary marking tech- 2010. Eight marking treatments were tested for their effect on niques into novel Burbot conservation programs should be juvenile Burbot survival, growth, condition, and mark reten- guided by criteria that differentiate the suitability of potential tion. The artificial markers studied included fin clips, freeze tags. We identified six marker criteria after factoring in conser- brands, VIE, and PIT tags. Age-0 Burbot were obtained via vation goals that include protecting an imperiled fish stock and production methods described by Jensen et al. (2008) and size- monitoring the release of hatchery-reared juveniles. An effective graded 14 d prior to experimentation to discourage cannibalism mark should (1) have no impact on fish survival, growth, ecol- (Barron et al. 2012). A total of 240 Burbot (65–92 mm TL)

Downloaded by [Department Of Fisheries] at 22:53 25 November 2013 ogy, or capture by predation (McDonald et al. (2003) referred to were randomly assigned to 24 polyethylene tanks (31 cm × these as major assumptions in mark–recapture models used 17 cm × 11 cm) with lids at a stocking density of 10 fish to estimate abundance and population growth), (2) be afford- per tank. This stocking density was conservative in compari- able to apply and detect by nonlethal sampling, (3) be feasible son with the higher levels maintained for optimal growth and to apply to fingerlings, (4) retain consistent and stable charac- hatchery production of Burbot fingerlings (N. K. Ashton and teristics, (5) be feasible to apply to fry or larvae, and (6) provide A. C. Cianciotto, University of Idaho, unpublished data). Previ- individual fish identification. No markers were expected to meet ous attempts to culture Burbot in small tanks resulted in satis- all six criteria, but those that failed to meet criterion 1, 2, or factory growth and survival (N. K. Ashton, unpublished data); 3 based on findings from other investigators were considered juveniles naturally inhabit small spaces under rocks, debris, and unsuitable for our evaluations with juvenile Burbot. Follow- other cover (Lawler 1963; Boag 1989; Ryder and Pesendorfer ing a review of the fish-tagging literature, we decided that fin 1992; McPhail and Paragamian 2000). clips, freeze brands, visible implant elastomer (VIE), and pas- The water volume in each tank was 1.5 L, with a water sive integrated transponder (PIT) tags had a strong likelihood flow of 0.80 ± 0.04 L·min−1·tank−1 (mean ± SD) from a of meeting criteria 1–4 in juvenile fish and should be tested in flow-through system supplying dechlorinated municipal water. fingerling Burbot. Other artificial markers (e.g., coded wire tags Water temperature (15–16◦C), dissolved oxygen (>6.0 mg/L), 1690 ASHTON ET AL.

pH (7.7–7.9), and NH3-N (<0.05 mg/L) remained at acceptable lowered by hand onto the branding tip, a 3-mm diameter hole levels for juvenile Burbot aquaculture over the study’s duration. in the gurney bottom guided contact of clean, dry Burbot skin Overhead lighting consisted of daylight spectrum compact flu- with cryo-cooled aluminum at the desired branding location. orescent bulbs (Commercial Electric, Model EDXO-19) with The branding tip was quickly wiped clean with a paper towel a light intensity of 11 ± 3 lx at the water surface (MW 700 and allowed to re-chill to cryo-temperatures for 5 s between portable lux meter; Milwaukee, Rocky Mount, North Carolina). each freeze brand application. The photoperiod was timer-controlled, providing 12 h light : VIE treatments.—Visible implant elastomer (Northwest 12 h dark cycles. Fish were maintained on Otohime C1 diet Marine Technologies, Shaw Island, Washington) is a liquid plas- (Reed Mariculture, Campbell, California) at a feeding rate of tic that cures into a pliable solid after being injected. This elas- 1% of wet body weight per day. tomer can also fluoresce in the presence of deep violet light, Juveniles were anesthetized in buffered 50-mg/L tricaine adding greater mark visibility. The VIE marks were admin- methanesulfonate (Tricaine-S; Western Chemical, Ferndale, istered with a handheld 0.3-mL tuberculin syringe (29-gauge Washington) prior to tagging. Total anesthesia time was lim- needle) coupled with a manual elastomer injector (NMT). In ited to 14 min, allowing 8 min to complete batch tagging of all accordance with the manufacturer’s directions, small quantities 10 fish in each tank. Fish were similarly anesthetized for sam- of the elastomer and hardening agent were mixed in the tuber- ples of TL, wet weight, and mark retention at the conclusion of culin syringes 24 h prior to tagging. Droplets were injected onto the study. Tagging was performed by a single person to mini- a petri dish at room temperature to ensure that proper curing mize tagger skill as a confounding variable in the study. At the occurred after mixing. Syringes were then placed in a −30◦C end of the trial, marks were read independently by the tagger freezer to delay curing until tagging. In accordance with the and a second person less familiar with the marking treatments injection techniques described by Frederick (1997), fluorescent to account for potential tag assessment bias. green elastomer was implanted subcutaneously either along the Fin clip treatments.—Excision of the anterior dorsal fin was right margin of the anterior dorsal fin base (treatment VIED) or chosen as a marking treatment (FCD), since this fin has a seem- into periocular tissue posterior to the right eye (treatment VIEE). ingly negligible locomotive or behavioral function in Burbot. A deep violet (wavelength, 405 nm) flashlight (VI Light; NMT) Excision of a single pectoral fin was selected as a second treat- was used to improve visualization of tags whenever the unaided ment (FCP) based on previous work showing that complete human eye failed to detect them. removal of one pectoral fin has no apparent deleterious effects PIT treatment.—As a precaution, juvenile Burbot were sur- on survival or growth in other fish (Churchill 1963; Thompson gically inserted with PIT tags (9.0 mm × 2.05 mm, 134.2 kHz; et al. 2005; Wagner et al. 2009). The filament-like pelvic fin may Destron Fearing, St. Paul, Minnesota) into the peritoneum (treat- have been easier to excise, but we chose not to remove this fin ment PIT) via a three-step process: (1) skin puncture using a because it possesses olfactory and tactile cues that assist with tuberculin syringe with 29-gauge needle, (2) small incision us- foraging (Hinkens and Cochran 1988). ing a no. 11 scalpel blade, and (3) tag insertion using forceps. Freeze brand treatments.—Freeze brands were applied for The skin puncture was located 5 mm posterior to the pelvic arch 2 s to juvenile Burbot either dorsally (treatment FBD) above the along the linea alba and improved the precision of scalpel blade anterior lateral line or ventrally (treatment FBV) along the linea penetration and control of incision lengths to 3–5 mm. After PIT alba medial to the pelvic arch and anus. Raleigh et al. (1973) tags were inserted into the incision, the tag was massaged until suggested a branding time of 2 s to prevent excessive damage it was fully enveloped within the peritoneum. Incisions were to the epidermis of the fish. Our branding apparatus consisted left unsutured following the methods described by Archdeacon of a 6-mm-diameter × 200-mm-long aluminum rod with a cir- et al. (2009). All surgical instruments were placed in 70% iso-

Downloaded by [Department Of Fisheries] at 22:53 25 November 2013 cular branding tip machined to 2.5 mm diameter. The bottom propyl alcohol and wiped clean with a new Kimwipe prior to two-thirds portion of the rod was continuously immersed in liq- each PIT implantation. The PIT tags were immersed in 70% iso- uid nitrogen contained within a 1-L Nalgene Dewar flask. A propyl alcohol for 10 min and verified with a Power Tracker V polystyrene foam lid with a center hole secured the aluminum reader (AVID Microchip ID Systems, Norco, California) prior branding rod and insulated the liquid nitrogen. Both salmonids to insertion. and centrarchids have been successfully branded using liquid Control treatment.—The control group (treatment CON) nitrogen at −196◦C (Raleigh et al. 1973). Anesthetized fish consisted of untagged juveniles exposed to anesthesia and han- were preblotted with absorbent paper to eliminate excess mois- dling stress comparable to the levels experienced in all other ture (Bryant et al. 1990) and restrained in a V-shaped, felt-lined tagging treatments. No sham tags, saline injections, or simu- plastic gurney. Small, medium, and large gurneys were con- lated marks were used in the control group, since a variety of structed to accommodate the size range of Burbot used in the artificial tag types were evaluated in this study. experiment and to minimize inconsistencies in brand application Data analyses.—Means ± SDs were derived for the fol- and mark quality—a frequent complaint with freeze branding lowing dependent variables: TL (mm), wet weight (W [g]), (Coutant 1972; Raleigh et al. 1973; Smith 1973; Hammer and absolute growth rate (AGRTL/W [mm/d] or [g/d]; Ricker Blankenship 2001; Skalski et al. 2009). As fish and gurney were 1975), specific growth rate (SGR [g·g1·d−1]or[%/d]; MARKER SELECTION AND TAGGING EVALUATIONS 1691

Ricker 1975), the coefficient of variation in TL (CV [%]), condition factor (K; Ricker 1975), and percent mark retention. The mean value of a dependent variable from all fish in each tank was considered a treatment replicate. The eight marking treatments were replicated in triplicate with randomized tank se- lection and fish stocking. The following equations define some of the dependent variables:

AGRTL,W = (x2 − x1)/(t2 − t1), where x = TL or W; t = study day SGR = 100 × (log W − log W )/(t − t ) e 2 e 1 2 1 FIGURE 1. Mean TLs (+ SD) of age-0 Burbot on sampling days 0 and 28, CV = 100 × SD/TL by marking treatment (see Table 1 for abbreviations). For a given treatment, K = 105 × W/TL3 different lowercase letters denote significant differences between sampling days (two-way ANOVA followed by Tukey’s HSD; P < 0.05). % mark retention = 100 × marks retained/total marks applied 1.5 mm, respectively. A two-way ANOVA detected significant Test statistics are presented with degrees of freedom in sub- changes (F = 30.01, P < 0.01) in wet weight between sam- < 1, 32 scripts and significance defined at P 0.05 for all statistical pling days 0 and 28 (Figure 2); however, post hoc tests failed to analyses. Two-way analysis of variance (ANOVA) followed detect significant changes associated with any specific marking by Tukey’s honestly significant difference (HSD) post hoc treatment. There were no significant differences between treat- tests (SAS, Proc GLM, version 9; SAS Institute, Cary, North ments at the start (F7, 16 = 0.46, P = 0.85) or end (F7, 16 = 0.42, Carolina) were performed on dependent variables TL, W,CV, P = 0.88) of the experiment. The pooled wet weights on study and K for comparisons between marking treatment and sam- days 0 and 28 were 3.1 ± 0.2 and 3.6 ± 0.4 g, respectively. pling day. One-way ANOVAs followed by Tukey’s HSD tests The pooled length–weight relationship for all fish on study were performed on all dependent variables for comparisons be- day 0 was described by the equation log10W = 2.75·log10TL – tween marking treatments. The length–weight relationships of 2 4.71 (F1, 22 = 57.66, P < 0.01, R = 0.72). The length–weight juvenile Burbot were described by linear regression analyses equation calculated at the end of the study was log10W = (Prism 5; GraphPad Software, La Jolla, California) performed 2 3.72·log10TL – 6.56 (F1, 22 = 15.94, P < 0.01, R = 0.42). The on log10 transformations of mean TL and W. The general linear length–weight relationship of fish did not change significantly model assumptions of homogeneity of variance and normal- (slope: F1, 44 = 1.10, P = 0.30; y-intercept: F1, 45 = 0.04, P = ity of data were verified with Levene’s tests and linear trends 0.85) between study days 0 and 28. in normal probability plots of residuals (SAS, version 9). Pro- A Kruskal–Wallis test failed to detect any significant portional mark retention data were arcsine transformed to meet differences (H = 7.0, P = 0.43) in the mean absolute growth normality assumptions (Zar 1999). Two-way ANOVAs followed rates in TL between marking treatments over the study duration. by Tukey’s HSD tests were performed on arcsine-transformed A one-way ANOVA test also found no significant differences mean percent tag retention for comparisons between marking (F7, 16 = 0.50, P = 0.82) in the absolute growth rates in wet treatments and tag assessors. Nonparametric Kruskal–Wallis weight between treatments over the study duration. The pooled H-tests (Prism 5) were performed on dependent variable data absolute growth rates for TL and W were 0.15 ± 0.06 mm/d that were nonnormally distributed or lacked homogeneity of Downloaded by [Department Of Fisheries] at 22:53 25 November 2013 and 0.02 ± 0.01 g/d, respectively. A similar test of SGR variance.

RESULTS Survival, Growth, and Condition Juvenile Burbot survival was 100% in all treatments over the study duration. A two-way ANOVA detected significant changes (F1, 32 = 66.88, P < 0.01) in TL between sampling days 0 and 28 (Figure 1); also, Tukey’s HSD post hoc tests found significant changes in TL for marking treatments FBD, FCD, FCP, and VIED. There were no significant differences in TL between treatments at the start (F7, 16 = 0.67, P = 0.69) or = = end (F7, 16 0.63, P 0.73) of the experiment. The pooled total FIGURE 2. Mean wet weights (+ SD) of age-0 Burbot on sampling days 0 lengths on study days 0 and 28 were 78.3 ± 1.9 and 82.5 ± and 28, by marking treatment (see Table 1 for abbreviations). 1692 ASHTON ET AL.

failed to detect significant differences (F7, 16 = 0.39, P = 0.90) 0.0% retention but noted substantial mark fading. The mean between treatments, and the pooled specific growth rate was observation of FBD retention by both assessors was 88.3 ± 0.55 ± 0.32 g·g−1·d−1 for the study duration. 16.5%, which was not significantly different from those of the A two-way ANOVA detected significant changes (F1, 32 = other marking treatments. Tag retention did not differ (F6, 14 = 5.61, P = 0.02) in the coefficient of variation in TL between 0.69, P < 0.66) between treatments according to assessor 1, sampling days 0 and 28; however, post hoc tests failed to but mark retention was significantly different (F6, 14 = 3.89, P detect significant changes associated with any specific marking < 0.02) between treatments according to assessor 2. Post hoc treatment. There were no significant differences in CV between tests of the mark retentions recorded by assessor 2 found a treatments at the start (F7, 16 = 1.11, P = 0.41) or end (F7, 16 = significant difference (P < 0.01) between the FBD and VIEE 0.30, P = 0.95) of the study; the respective pooled CVs were treatments. 6.2 ± 1.6 and 5.2 ± 1.2%. No significant changes occurred K in between sampling days 0 and 28. Condition factor did not Fish Tagging and Mark Characteristics differ between marking treatments on study day 0 (F = 0.65, 7, 16 It was difficult to fin-clip the smallest juvenile Burbot P = 0.71) or day 28 (F = 0.70, P = 0.67); the respective 7, 16 (<75 mm TL) using the standard technique (i.e., holding the pooled condition factors were 0.64 ± 0.03 and 0.64 ± 0.05. fish out of the water with one hand while excising the fin with scissors held in the other hand). The tiny, imbedded scales and Tag Retention copious skin mucus of juvenile Burbot make them difficult to Two independent assessors read the tags from the seven mark- grasp by hand. Burbot skin and body characteristics also caused ing treatments at 28 d posttagging (Table 1). From highest to the small, delicate fins to stick to the body when removed from lowest, the mean percent tag retentions were as follows: VIEE = water. Our solution was to keep anesthetized juveniles in a shal- 100.0 ± 0%, FBV = 98.3 ± 2.4%, FCP = 98.3 ± 2.4%, low pan of water while forceps were used to grasp the free- PIT = 96.7 ± 0%, VIED = 93.3 ± 4.7%, FCD = 91.7 ± floating fin before excising. The standard technique could prob- 2.4%, and FBD = 88.3 ± 16.5%. A one-way ANOVA per- ably be used for fin-clipping larger fingerlings (>90 mm TL). formed on arcsine-transformed mean percent tag retention data Complete removal of either the anterior dorsal or right pectoral narrowly failed to detect significant differences (F6, 14 = 2.49, fin resulted in wounds that fully healed over the study duration. P = 0.08) between marking treatments. The pooled tag retention Incomplete fin removal resulted in noticeable regeneration of for all marking treatments was 95.2 ± 5.8% at 28 d posttag- dorsal anterior fins and negligible regeneration of pectoral fins ging. A two-way ANOVA performed on arcsine-transformed at 4 weeks posttagging. percent mark retention data detected significant differences be- The freeze-branded skin of juvenile Burbot was immediately tween tagging treatments (F6, 28 = 2.67, P = 0.04) and assessors white after branding and gradually turned translucent after sev- (F1, 28 = 10.61, P < 0.01), but no significant interaction between eral days due to pigment loss. At the anterior dorsal branding these factors (F6, 28 = 2.12, P = 0.08). Post hoc tests found a location, this loss of pigmentation resulted in a brand that ap- significant difference (P < 0.01) between assessors with regard peared light grey. However, it was difficult to distinguish this to the recognition of FBD marks. The assessor who was less brand from natural mottling of juvenile Burbot. Fading of dor- familiar with the marking treatments (assessor 2) observed only sal freeze brands was noticeable at 1 month posttagging as 76.7 ± 5.8% retention of the FBD marks. The assessor who regenerated skin replaced the scar tissue. Freeze-branding the also tagged the fish for this study (assessor 1) observed 100 ± belly of juvenile Burbot resulted in a translucent mark, and the

TABLE 1. Results of independent assessments of mark retention in fingerling Burbot at 28 d posttagging. Marking treatments are as follows: FBD = freeze Downloaded by [Department Of Fisheries] at 22:53 25 November 2013 brands, dorsal; FBV = freeze brands, ventral; FCD = excisions, anterior dorsal fins; FCP = excisions, pectoral fins; PIT = passive integrated transponder tags; VIED = VIE injections, base of anterior dorsal fins; and VIEE = VIE injections, posterior to right eye; see text for more details. For each assessor, different lowercase letters denote significant differences in percent tag retention between marking treatments (two-way ANOVA followed by Tukey’s HSD; P < 0.05).

Assessor 1 Assessor 2 Marking treatment n Tag No tag % retention n Tag No tag % retention FBD 30 30 0 100 z 30 23 7 77 y FBV 30 30 0 100 z 30 29 1 97 yz FCD 30 28 2 93 z 30 27 3 90 yz FCP 30 30 0 100 z 30 29 1 97 yz PIT 30 29 1 97 z 30 29 1 97 yz VIED 30 29 1 97 z 30 27 3 90 yz VIEE 30 30 0 100 z 30 30 0 100 z MARKER SELECTION AND TAGGING EVALUATIONS 1693

internal organs were sometimes visible due to depigmentation. fingerlings could be pertinent to the tagging of age-0 Burbot in The ventral marks became more opaque at 28 d posttagging but the fall season for most populations within the Burbot’s native were still much more distinct than the dorsal marks. Anesthesia distribution. in combination with gurneys aided in restraining fish movement We observed growth (in terms of AGR and SGR) that was during branding, but complete immobilization of fish was not less than the rates reported for Burbot of similar sizes at age in always realized. The gurneys improved the precision of brand other investigations (Miller 1970; Ryder and Pesendorfer 1992; applications to the belly and dorsal skin of small fish (<75 mm Taylor 1998; Fischer 2000; Kjellman and Eloranta 2002; Bar- TL); however, gurneys were probably unnecessary for the suc- ron 2012; Ashton, unpublished data). The condition factors in cessful branding of larger fingerlings (>90 mm TL), given the our study overlapped with the reported values for some wild greater surface area to accommodate a 2.5-mm brand at both populations (Hofmann and Fischer 2003; Bennetta and Janz anatomical locations. 2007) but were below the levels calculated for other similarly- The VIE tag characteristics were inconsistent at the periocu- sized Burbot (Miller 1970; Baxter 1998; Fischer 2000; Barron lar and dorsal fin base implant locations. Mottled pigmentation 2012). We concluded that fish grew slowly in our short-term at the dorsal injection site partially obscured some tags, but vi- trial and that their condition factors were below the levels that sualization was improved by the use of a VI Light (NMT). The would have been attained during optimal rearing. Age-0 Bur- periocular injection site had less pigment and therefore fewer bot of similar sizes at age have grown at faster rates under instances of partially obscured VIE tags. Variability in the depth very similar experimental conditions (i.e., same tanks, stocking of subcutaneous injection and elastomer tag size was noticed for density, water flows, temperature, and feed ration) but with a the dorsal fin base and periocular body locations, but this was different year-class and revised weaning strategy (Ashton, un- not tested. published data). The slow growth observed in our study was The PIT incisions were not completely healed at 28 d post- likely due to incomplete weaning to a commercial diet prior tagging. In one instance, a juvenile Burbot was observed with to starting the trial. The coefficient of variation in TL calcu- half of a PIT tag protruding from a partially healed incision. All lated from our study fish was substantially less than that re- PIT tagged fish were characterized by a translucent scar approx- ported by Barron (2012) and was likely affected by two factors: imately 5 mm in length along the linea alba. The PIT tagging (1) the effectiveness of grading the fish prior to starting the of Burbot as small as 65 mm TL was accomplished without trial and (2) the slow relative growth of our study fish. The mortality. restricted feed ration (1% of wet body weight per day) that we selected was meant to simulate the reduced feeding and growth expected from the posttagging release of hatchery-adapted fin- DISCUSSION gerlings into a natural river (Johnson et al. 1996). However, We concluded that all seven artificial marks could be ap- one drawback to holding the juvenile Burbot on this mainte- plied to fingerling Burbot with no adverse short-term effects on nance ration is that it may have limited the effects of growth survival, growth, or condition. This conclusion is supported by on mark retention. Topics for future studies could include (1) the following study results: (1) 100% fish survival regardless of the long-term effects of fin removal on Burbot foraging, health, marking treatment, (2) modest growth in all treatments over the and ecology, (2) the impact of externally visible VIE marks on study period, and (3) no significant differences between tagged the behavior of marked individuals or that of fish interacting and untagged fish in terms of survival, growth, or condition. with them, and (3) the stresses and challenges experienced by There is limited data on the size and growth of the wild ju- hatchery-adapted Burbot released into a natural river. venile Burbot inhabiting the lower Kootenai River; however, None of the seven artificial markers tested in this study posed

Downloaded by [Department Of Fisheries] at 22:53 25 November 2013 such information is available for other Burbot populations. The major concerns regarding affordable marking or nonlethal tag length–weight relationship (log10W = 2.75·log10TL – 4.71) cal- detection. Each tested mark was nonlethally recovered either by culated at the beginning of our study closely matched the re- visual identification (fin clips, freeze brands, VIE) or passive lationship (log10W = 2.58·log10TL – 4.30) described by Han- electronic detection (PIT). All tested marks were considered to son and Qadri (1980) for Burbot of similar age inhabiting the be affordable in terms of marking and tag detection (Guy et al. Ottawa River, Ontario. We also recorded TLs and weights that 1996; Hammer and Blankenship 2001; Skalski et al. 2009). overlapped with the reported size ranges of similarly-aged Bur- Although the cost of a PIT tag (∼$2 per tag at bulk price; S. bot from other regions (Clemens 1951; Miller 1970; Bailey C. Ireland, Kootenai Tribe of Idaho, personal communication) 1972; Ryder and Pesendorfer 1992; Fischer and Eckmann 1997; can exceed the expense of the other tested marks by up to 100- Baxter 1998; Taylor 1998, 2001, 2002; Fischer 2000; Kjellman fold (Hammer and Blankenship 2001), we considered this a and Eloranta 2002; Hofmann and Fischer 2003; Bennetta and reasonable price for the information gained from individual fish Janz 2007). Despite the Holarctic expanse that separates these identification. study areas, the size of age-0 Burbot following the first summer The excellent fish survival and high overall mark retention of growth appears to be fairly uniform. Thus, our conclusions observed in our study suggest that all tested markers could be regarding the application of artificial marks to hatchery-reared applied to hatchery-reared fingerling Burbot. Age-0 fingerlings 1694 ASHTON ET AL.

typically reach 60−150 mm TL after the first summer of growth would likely improve if larger fingerling Burbot (>90 mm TL) in a hatchery setting (Barron et al. 2012; Ashton, unpublished were fin-clipped. data), and that should probably be the target size for stocking. In terms of mark consistency and recognition, dorsal freeze Hatchery releases of smaller fish (juveniles, fry, or larvae) in brands were the most problematic marker in our study. Mottled other species have been shown to result in diminished survival to pigmentation and rapid skin regeneration interfered with freeze adulthood (Kampa and Hatzenbeler 2009; Steffensen et al. 2010; brand mark recognition at the dorsal body location. Conversely, Tipping 2011). The stocking of fingerlings has been practiced ventral freeze brands were observed to show excellent read- by supplementation programs seeking to minimize fish propaga- ability and mark retention. High variability in the persistence tion expenses, deter long-term hatchery-induced domestication, and quality of marks has been the most common complaint of and restore a fish population to historical levels (Mahnken et al. freeze brands (Coutant 1972; Raleigh et al. 1973; Smith 1973; 1982; Storck and Newman 1988; Dickhoff et al. 1995; Tipping Hammer and Blankenship 2001; Skalski et al. 2009). Dorsal et al. 1995; Dehaan et al. 2008). It was evident from our study skin is the most frequently reported branding location in other that all of the artificial marks tested were difficult to apply in fish (Sorensen et al. 1983; Myers and Iwamoto 1986; Bryant the smallest Burbot (<75 mm TL). Since the mean dependent et al. 1990; Knight 1990; Haines et al. 1998; Conover and variable from a tank was chosen as the experimental unit and not Sheehan 1999; Evrard 2005), but ventral skin surfaces have individual fish, none of the analyses we performed could test the also been successfully branded (Raleigh et al. 1973; Dando and effect of initial fish size on posttagging growth or mark reten- Ling 1980; Evrard 2005). Properly controlled freeze-branding tion. The likelihood of such an effect may increase with longer procedures have putatively produced stable marks (Raleigh et al. study durations (e.g., one or more years); thus, any long-term 1973; Conover and Sheehan 1999). We found it slightly bother- investigation of these markers should incorporate an experimen- some trying to maintain mark consistency in both freeze brand tal design that enables analysis of individual fish or incorporates treatments, particularly with smaller fish (<75 mm TL). Per- initial fish size as a treatment factor. sistent removal of frozen condensation on the branding tip and We observed variability in some mark characteristics and blotting excess moisture and mucus from fish was necessary for determined that tag assessment bias was a significant factor mark consistency. The use of freeze branding as a mass-marking that should be considered in future studies incorporating the technique seems more plausible with larger fingerling Burbot recognition of fin clips, freeze brands, and VIE tags applied to (>90 mm TL) and without the use of gurneys. Burbot. Other major factors that are likely to affect tag retention Poststudy observations determined that dorsal freeze brand- include fish growth and age (Raleigh et al. 1973; Guy et al. 1996; ing times could be extended to 7 s without noticeable harm to ju- Skalski and Griswold 2006; Knudsen et al. 2009; Skalski et al. venile Burbot. Conversely, ventral freeze branding times longer 2009; Soula et al. 2012). In our preliminary study, the short trial than 2 s resulted in branding wounds that became infected (most duration, slow fish growth, and inability to track individual fish likely due to the thin epidermis of the belly). Raleigh et al. (1973) limited the domain of inferences regarding tag loss. Any future recommended a branding time of 2 s in three species of centrar- studies involving the application of these markers to juvenile chids and three species of trout. Longer branding times resulted Burbot should take this into consideration. in a blurred brand or a loss of dermal tissues over the brand area The difficulties in recognizing anterior dorsal fin clips in our that produced open sores. However, branding times up to 4−6s study were mostly due to fin regeneration resulting from in- have been reported without harmful effects in the American complete fin excision—the tagger found that it was difficult to Eel Anguilla rostrata (Sorensen et al. 1983), Razorback Sucker verify fin removal in smaller juvenile Burbot (<75 mm TL). Xyrauchen texanus, Colorado Pikeminnow Ptychocheilus lu- Incomplete fin clips are typically ephemeral marks (Hammer cius (Haines et al. 1998), and Roach Rutilus rutilus (Evrard

Downloaded by [Department Of Fisheries] at 22:53 25 November 2013 and Blankenship 2001) and regeneration of fins can be a dis- 2005). Since the Burbot is a predominately benthic species, a tinct disadvantage with some species (Moring 1990). The subtle recognizable dorsal freeze brand could have advantages over a difference between the short anterior and continuous posterior ventral brand for in situ tag reading. Any long-term study of dorsal fins of juvenile Burbot may have also factored into di- dorsal freeze branding in juvenile Burbot should implement an minished recognition of dorsal fin clips. Underreporting of fin extended branding time (4−7 s) to improve mark permanency. clips due to mark recognition failure is a commonly reported For VIE tagging of juvenile Burbot, we observed that fish pig- problem (Guy et al. 1996). Incomplete removal of pectoral fins mentation and the depth of the VIE injection were two factors did sporadically occur during our marking efforts; however, requiring close scrutiny by the tagger. Since less pigmentation regeneration of this fin was slow in juvenile Burbot and pec- typically occurred at the periocular target site in juvenile Burbot, toral fin clips were generally easy to recognize. Other investi- we generally found this location to be more favorable to tag con- gators have found that fin regeneration time can be dependent sistency. Variability in the depth of injection was also generally upon several factors: the degree of fin removal, fin type, fish less of an issue at the periocular location. Frederick (1997) size, and species (McNeil and Crossman 1979; Johnson and reported that VIE tag visibility in reef fishes was negatively Ugedal 1988; Coombs et al. 1990; Moring 1990; Zymonas and impacted by injection depth and the presence of overlying skin McMahon 2006). Fin removal and subsequent mark recognition pigmentation. We observed a similar effect in our study and MARKER SELECTION AND TAGGING EVALUATIONS 1695

found that substantial tagging skill was necessary to produce ity, we recommend that a long-term PIT tag study be performed consistent and stable VIE marks (see Soula et al. 2012 for a re- with fingerling Burbot that evaluates several tag locations and view). Our use of the “tattoo” technique described by Frederick implantation techniques. (1997) occasionally produced highly visible VIE tags in darkly Each artificial mark that we tested has the advantages of pigmented dorsal skin of fingerling Burbot; unfortunately, we greater affordability, more rapid implementation, and more im- found the reproducibility of this technique to be a challenge. mediate detection in the field than some natural marks (e.g., The restricted space for VIE injection at the periocular loca- genetic or geochemical). The primary disadvantages of the ar- tion was problematic in small Burbot (<75 mm TL), but the tificial marks we tested center on their long-term retention and translucent and thin tissue precluded the need for a tattooing feasibility for mass-marking; undoubtedly, the efficacy of these technique. Phillips and Fries (2009) reported that dorsal and marks in juvenile Burbot hinges on further research in these ventral VIE marks were retained equally well in small Fountain two areas. If restoration of the lower Kootenai River Burbot Darters Etheostoma fonticola; however, the dorsal location was population requires that 110,000–900,000 juveniles be stocked described as having better visibility for in situ tag reading. We per year (Paragamian and Hansen 2011), then the development reached a similar conclusion with Burbot; dorsal VIE tags were of affordable, effective mass-marking techniques will be neces- easier to see in unanesthetized juveniles that were swimming or sary to monitor and evaluate stock enhancement success. Our resting in our experimental tanks. tagging study was a prototypic evaluation with imperiled juve- The postocular adipose eyelid in salmonids is transparent nile Burbot that implicitly required investigation of conservative and has been a popular and successful implant location for techniques for fin clipping, freeze branding, and PIT tagging. VIE tags (see Northwest Marine Technology 2008 and refer- We recommend that future research on Burbot tagging focus on ences therein). Juvenile Burbot lack sufficient adipose eyelid techniques that enable a high rate of marking without violating tissue for VIE tagging; we would describe the periocular lo- criteria 1–4. cation as translucent, lightly pigmented, and resembling cheek Removal of the anterior dorsal fin could be an effective tissue. Any future studies evaluating the long-term efficacy of single-batch mark that differentiates hatchery Burbot from wild VIE tags in juvenile Burbot should incorporate a greater variety fish. This fin clip could also serve as an external secondary mark of VIE colors and body locations for tagging. The visibility of for coded wire tags or other internal marks. The primary util- VIE tags could be improved by injection into translucent tissues ity of anterior dorsal fin removal would be field identification located in the snout and each pectoral or pelvic fin base. Post- by anglers, commercial fishermen, or samplers using traditional study observations with larger Burbot fingerlings (>120 mm fishing gear. Identification of anterior dorsal fin clips in juve- TL) found that VIE could be injected into transparent tissue niles will require capture, anesthetization, and close inspection between the pectoral fin rays. The efficacy of VIE tags may of the fish; however, not all of these steps may be necessary for also be significantly affected by tag color (see Soula et al. 2012 mark identification in adults. We feel it would be overly inva- for a review). For example, fluorescent red, orange, or pink sive to remove the pectoral fin in all hatchery-reared Burbot. may contrast with Burbot skin coloration and improve VIE tag The pectoral fin likely has greater functional importance than visibility. the anterior dorsal fin in terms of swimming, feeding behavior, The retention of PIT tags was high in our study fish de- and perhaps mating rituals. A more suitable field application for spite the fact that the incisions were only partially healed at pectoral fin clips might be small-scale mark–recapture investi- 28 d posttagging. Numerous factors can influence PIT tag re- gations. tention, including anatomical placement, seasonal fish behavior The utility of freeze brands is basically synonymous with that (e.g., spawning), fish size, and differences among fish species of fin clips; however, implementation of mass marking could

Downloaded by [Department Of Fisheries] at 22:53 25 November 2013 (Clugston 1996; Guy et al. 1996; Baras et al. 1999; Buzby and be easier with freeze brands. Bryant et al. (1990) developed an Deegan 1999). Considerable variation in PIT tag retention has easy-to-use field method for freeze-branding salmonids and also been reported for a variety of fish. Zaroban and Anglea (2010) achieved a respectable marking rate (∼300 fish/h). Recognition observed a 38.8% loss rate of dorsal PIT tags and a 2.5% loss of dorsal and ventral freeze brands on Burbot in the field will rate of body-cavity PIT tags in Shorthead Sculpins Cottus con- likely require capture, anesthetization, and close inspection of fusus. Alternatively, Dieterman and Hoxmeier (2009) observed fish. 56% retention when PIT tags were injected into the body cavity The VIE tag can be useful in a single- or multiple-batch (i.e., and 95% when they were placed into the dorsal musculature multiple colors and tag locations) marking system that identi- of Brown Trout Salmo trutta. Skalski et al. (2009) reported fies a hatchery fish along with providing additional information that early studies of PIT tags showed high tag retention (e.g., (e.g., year-class or release site). A dorsal VIE tag could be ad- Prentice et al. 1990) but that recent work has estimated tag losses vantageous for underwater field sampling of juvenile Burbot of up to 15% in returning adult Chinook Salmon Oncorhynchus (e.g., snorkel surveys at night) or in situ observations of tagged tshawytscha (e.g., Knudsen et al. 2009) and up to 30% in juve- fish in tanks. A periocular VIE tag may be better suited for stud- nile Brown Trout smaller than 57 mm at tagging (e.g., Acolas ies requiring greater tag stability. Identification of VIE tags in et al. 2007). In response to this reported tag retention variabil- the field may or may not require capture, anesthetization, and 1696 ASHTON ET AL.

close inspection of fish; we found that tag visibility was highly the body cavity of juvenile Brown Trout (Salmo trutta). Fisheries Research dependent on tagger skill and the pigmentation of individual 86:280–284. Burbot. Ahrens, R., and J. Korman. 2002. What happened to the West Arm Burbot stock in Kootenay Lake? Use of an age-structured population model to de- A PIT implant would be a superlative artificial tag that en- termine the possible causes for recruitment failure. Report prepared for the ables individual identification, passive detection, and telemetry British Columbia Ministry of Water, Land, and Air Protection, the Habitat of Burbot. Portable PIT tag antennas that are attached to sub- Conservation Trust Fund, and the Bonneville Power Administration, Nelson. mersible poles have been particularly useful for identifying fish Anders, P. J., D. L. Richards, and M. S. Powell. 2002. The first endangered that seek refuge in substrate or for noninvasive detection of White Sturgeon population: repercussions in an altered large-river–floodplain ecosystem. Pages 67–82 in W. Van Winkle, P. J. Anders, D. H. Secor, and imperiled, threatened, or endangered fish (Cucherousset et al. D. A. Dixon, editors. Biology, management, and protection of North 2005; Sloat and Baker 2011). Major drawbacks to PIT tag use for American sturgeon. American Fisheries Society, Symposium 28, Bethesda, Burbot population monitoring are the higher costs and need for Maryland. specialized electronic tag detection equipment not commonly Archdeacon, T. P., W. J. Remshardt, and T. L. Knecht. 2009. Comparison of two used by anglers or fishermen. methods for implanting passive integrated transponders in Rio Grande Silvery Minnow. North American Journal of Fisheries Management 29:346–351. The application of other artificial markers (e.g., coded wire Bailey, M. M. 1972. Age, growth, reproduction, and food of the Burbot, Lota tags and calcein staining of fin rays) as well as natural mark- lota (Linnaeus), in southwestern Lake Superior. Transactions of the American ers (e.g., genetic and geochemical signatures in fin rays) also Fisheries Society 4:667–674. need to be studied in juvenile Burbot. Some—or all—of these Baras, E., L. Westerloppe, C. Melard,` J. C. Philippart, and V. Benech.` 1999. markers could enable identification of hatchery Burbot released Evaluation of implantation procedures for PIT-tagging juvenile Nile Tilapia. North American Journal of Aquaculture 61:246–251. as larvae or fry (criterion 5). Development of a parentage-based Barron, J. M., N. R. Jensen, P. J. Anders, J. P. Egan, S. C. Ireland, and K. D. tag for Burbot from either microsatellites or single-nucleotide Cain. 2012. Effects of temperature on the intensive culture performance of polymorphisms is particularly interesting because these mark- larval and juvenile North American Burbot (Lota lota maculosa). Aquaculture ers could identify family relationships, population connectivity, 364–365:67–73. and individuals (criterion 6) by nonlethal tissue sampling. Baxter, J. S. 1998. Summary of juvenile Burbot sampling in the upper Columbia River catchment, 1997. Report to the Columbia Basin Fish and Wildlife This preliminary study of artificial markers found that fin Compensation Program, Nelson, British Columbia. clips, freeze brands, VIE, and PIT tags can be safely applied to Bennetta, P. M., and D. Janz. 2007. Bioenergetics and growth of young-of the- fingerling Burbot. Our findings open the way for research on year Northern Pike (Esox lucius) and Burbot (Lota lota) exposed to metal long-term retention, mass-marking feasibility, and field appli- mining effluent. Ecotoxicology and Environmental Safety 68:1–12. cability. Future investigations that address the aforementioned Boag, T. D. 1989. Growth and fecundity of Burbot, Lota lota L., in two Alberta lakes. Master’s thesis. University of Alberta, Edmonton. uncertainties and recommendations discussed in our study will Bruyndoncx, L., G. Knaepkens, W. Meeus, L. Bervoets, and M. Eens. 2002. The be integral to the validation of specific markers as effective tools evaluation of passive integrated transponder (PIT) tags and visible implant for monitoring Burbot populations across the Holarctic ecozone. elastomer (VIE) marks as new marking techniques for the bullhead. Journal of Fish Biology 60:260–262. Bryant, M. D., C. A. Dolloff, P. E. Porter, and B. E. Wright. 1990. Freeze ACKNOWLEDGMENTS branding with CO2: an effective and easy-to-use field method to mark fish. Pages 30–35 in N. C. Parker, A. E. Giorgi, R. C. Heidinger, D. B. Jester Jr., Funding for this work was provided by the Kootenai Tribe E. D. Prince, and G. A. Winans, editors. Fish-marking techniques. American of Idaho and the Bonneville Power Administration (contracts Fisheries Society, Symposium 7, Bethesda, Maryland. 46821, 57791; project 198806400). We thank the Kootenai Val- Burdick, S. M. 2011. Tag loss and short-term mortality associated with pas- ley Resource Initiative committee members for their continuing sive integrated transponder tagging of juvenile Lost River suckers. North support of Burbot conservation in Idaho and British Columbia. American Journal of Fisheries Management 31:1088–1092. Buzby, K., and L. Deegan. 1999. Retention of anchor and passive integrated Downloaded by [Department Of Fisheries] at 22:53 25 November 2013 We appreciate the assistance of staff at the University of Idaho’s transponder tags by Arctic Grayling. North American Journal of Fisheries Aquaculture Research Institute (UI–ARI), particularly Nathan Management 19:1147–1150. Jensen, John Foltz, Ryan Johnson, and Joshua Egan. We also Churchill, W. S. 1963. The effect of fin removal on survival, growth, and vulnera- thank James Barron for constructive advice during the plan- bility to capture of stocked Walleye fingerlings. Transactions of the American ning of the experiment and for volunteering as a tag assessor. Fisheries Society 92:298–300. Clemens, H. P. 1951. The growth of the Burbot, Lota lota maculosa (LeSueur), We gratefully acknowledge the staff from the British Columbia in Lake Erie. Transactions of the American Fisheries Society 80:163–173. Ministry of Forests, Lands, and Natural Resource Operations, Clugston, J. P. 1996. Retention of T-bar anchor tags and passive integrated the Kootenai Tribe of Idaho, and the Idaho Department of Fish transponder tags by Gulf Sturgeons. North American Journal of Fisheries and Game for their assistance with spawning wild adult Bur- Management 16:682–685. bot and transporting fertilized eggs to the UI–ARI for juvenile Conover, G. A., and R. J. Sheehan. 1999. Survival, growth, and mark persistence in juvenile Black Crappies marked with fin clips, freeze brands, or oxytetra- Burbot propagation. cycline. North American Journal of Fisheries Management 19:824–827. Coombs, K. A., J. K. Bailey, C. M. Herbinger, and G. W. Friars. 1990. Evaluation of various external marking techniques for Atlantic Salmon. Pages 142–146 REFERENCES in N. C. Parker, A. E. Giorgi, R. C. Heidinger, D. B. Jester Jr., E. D. Prince, and Acolas, M. L., J. M. Roussel, J. M. Lebel, and J. L. Bagliniere.` 2007. Laboratory G. A. Winans, editors. Fish-marking techniques. American Fisheries Society, experiment on survival, growth, and tag retention following PIT injection into Symposium 7, Bethesda, Maryland. MARKER SELECTION AND TAGGING EVALUATIONS 1697

Coutant, G. C. 1972. Successful cold branding of nonsalmonids. Progressive Johnson, B. O., and O. Ugedal. 1988. Effects of different kinds of fin-clipping on Fish-Culturist 29:175–176. over-winter survival and growth of fingerling Brown Trout, Salmo trutta L., Cucherousset, J., J. M. Roussel, R. Keeler, R. A. Cunjak, and R. Stump. 2005. stocked in small streams in Norway. Aquaculture and Fisheries Management The use of two new portable 12-mm PIT tag detectors to track small fish in 19:305–311. shallow streams. North American Journal of Fisheries Management 25:270– Johnson, J. H., J. F. McKeon, and D. S. Dropkin. 1996. Comparative diets of 274. hatchery and wild Atlantic Salmon smolts in the Merrimack River. North Dando, P.R., and R. Ling. 1980. Freeze-branding of flatfish: flounder Platichthys American Journal of Fisheries Management 16:440–444. flesus, and plaice, Pleuronectes platessa. Journal of the Marine Biological Kampa, J. M., and G. R. Hatzenbeler. 2009. Survival and growth of Walleye Association of the United Kingdom 60:741–748. fingerlings stocked at two sizes in 24 Wisconsin lakes. North American DeHaan, P. W., G. R. Jordan, and W. R. Ardren. 2008. Use of genetic tags Journal of Fisheries Management 29:996–1000. to identify captive bred Pallid Sturgeon (Scaphirhynchus albus) in the wild: Kjellman, J., and A. Eloranta. 2002. Field estimations of temperature-dependent improving abundance estimates for an endangered species. Conservation Ge- processes: case growth of young Burbot. Hydrobiologia 481:187–192. netics 9:691–697. Knight, A. E. 1990. Cold-branding techniques for estimating Atlantic Salmon Dickhoff, W. W., B. R. Beckman, D. A. Larson, C. Mahnken, C. B. Schreck, C. parr densities. Pages 36–37 in N. C. Parker, A. E. Giorgi, R. C. Heidinger, Sharp, and W. S. Zaugg. 1995. Quality assessment of hatchery-reared spring D. B. Jester Jr., E. D. Prince, and G. A. Winans, editors. Fish-marking tech- Chinook Salmon smolts in the Columbia River basin. Pages 292–302 in niques. American Fisheries Society, Symposium 7, Bethesda, Maryland. H. L. Schramm Jr. and R. G. Piper, editors. Uses and effects of cultured Knudsen, C., M. Johnston, S. Schroder, W. Bosch, D. Fast, and C. Strom. fishes in aquatic ecosystems. American Fisheries Society, Symposium 15, 2009. Effects of passive integrated transponder tags on smolt-to-adult recruit Bethesda, Maryland. survival, growth, and behavior of hatchery spring Chinook Salmon. North Dieterman, D. J., and R. J. H. Hoxmeier. 2009. Instream evaluation of passive American Journal of Fisheries Management 29:658–669. integrated transponder retention in Brook Trout and Brown Trout: effects of KVRI (Kootenai Valley Resource Initiative) Burbot Committee. 2005. Kootenai season, anatomical placement, and fish length. North American Journal of River/Kootenay Lake Burbot conservation strategy. Prepared by the Kootenai Fisheries Management 29:109–115. Tribe of Idaho and S. P. Cramer and Associates, Bonners Ferry. Evrard, G. 2005. Use and effect of freeze branding on roach (Rutilus rutilus L.). Lawler, G. H. 1963. The biology and taxonomy of the Burbot, Lota lota,in Bulletin Francais de la Peche et de la Pisciculture 374:35–42. Heming Lake, Manitoba. Journal of the Fisheries Research Board of Canada Fischer P, 2000. Test of competitive interactions for space between the two ben- 29:417–433. thic fish species, Burbot (Lota lota L.) and Stone Loach (Barbatula barbatula Mahnken, C., E. Prentice, W. Waknitz, M. G. Sims, and J. Williams. 1982. L.). Environmental Biology of Fishes 58:439–446. The application of recent smoltification research to public hatchery releases: Fischer, P., and R. Eckmann. 1997. Spatial distribution of littoral fish species in an assessment of size/time requirements for Columbia River hatchery coho Lake Constance, Germany. Archiv fuer Hydrobiologie 140:91–116. (Oncorhynchus kisutch). Aquaculture 28:251–268. Frederick, J. L. 1997. Evaluation of fluorescent elastomer injection as a method McDonald, T. L., S. C. Amstrup, and B. F. J. Manly. 2003. Tag loss can bias for marking small fish. Bulletin of Marine Science 61:399–408. Jolly–Seber capture–recapture estimates. Wildlife Society Bulletin 31:814– Guy, C. S., H. L. Blankenship, and L. A. Nielsen. 1996. Tagging and mark- 822. ing. Pages 353–383 in B. R. Murphy and D. W. Willis, editors. Fisheries McFarlane, G. A., R. S. Wydoski, and E. D. Prince. 1990. Historical review techniques, 2nd edition. American Fisheries Society, Bethesda, Maryland. of the development of external tags and marks. Pages 9–29 in N. C. Parker, Haines, G. B., S. H. Severson, and T. Modde. 1998. Evaluation of Razorback A. E. Giorgi, R. C. Heidinger, D. B. Jester Jr., E. D. Prince, and G. A. Winans, Sucker and Colorado Squawfish batch marking technique. Progressive Fish- editors. Fish-marking techniques. American Fisheries Society, Symposium Culturist 60:272–275. 7, Bethesda, Maryland. Hammer, S. A., and H. L. Blankenship. 2001. Cost comparison of marks, tags, McNeil, F. I., and E. J. Crossman. 1979. Fin clips in the evaluation of stocking and mark-with tag combinations used in salmonid research. North American programs for Muskellunge, Esox masquinongy. Transactions of the American Journal of Aquaculture 63:171–178. Fisheries Society 108:335–343. Hanson, J. M., and S. Qadri. 1980. Morphology and diet of young-of-the-year McPhail, J. D., and V. L. Paragamian. 2000. Burbot biology and life history. Burbot, Lota lota, in the Ottawa River. Canadian Field-Naturalist 94:311– Pages 11–23 in V. L. Paragamian and D. W. Willis, editors. Burbot: biology, 314. ecology, and management. American Fisheries Society, Fisheries Manage- Hilborn, R., C. J. Walters, and D. B. Jester. 1990. Value of fish marking in fish- ment Section Publication 1, Spokane, Washington. eries management. Pages 5–7 in N. C. Parker, A. E. Giorgi, R. C. Heidinger, Miller, D. D. 1970. A life history study of Burbot in Boysen Reservoir, Ring D. B. Jester Jr., E. D. Prince, and G. A. Winans, editors. Fish-marking tech- Lake, and Trail Lake, part II. Wyoming Game and Fish Commission, Federal Downloaded by [Department Of Fisheries] at 22:53 25 November 2013 niques. American Fisheries Society, Symposium 7, Bethesda, Maryland. Aid in Fish Restoration, Project F-41-R-2, Completion Report, Lander. Hinkens, E., and P.A. Cochran. 1988. Taste buds on pelvic fin rays of the Burbot, Moring, J. R. 1990. Marking and tagging intertidal fishes: review of techniques. Lota lota (L.). Journal of Fish Biology 32:975. Pages 109–116 in N. C. Parker, A. E. Giorgi, R. C. Heidinger, D. B. Jester Jr., Hofmann, N., and P. Fischer. 2003. Impact of temperature on food intake and E. D. Prince, and G. A. Winans, editors. Fish-marking techniques. American growth in juvenile Burbot. Journal of Fish Biology 63:1295–1305. Fisheries Society, Symposium 7, Bethesda, Maryland. ISRP/ISAB (Independent Scientific Review Panel/Independent Scientific Ad- Myers, J. M., and R. N. Iwamoto. 1986. Evaluation of thermal and chemical visory Board). 2009. Tagging report. A comprehensive review of Columbia marking techniques for tilapia. Progressive Fish-Culturist 48:288–289. River basin fish tagging technologies and programs. Northwest Power and Neufeld, M. D., C. A. Davis, K. D. Cain, N. R. Jensen, S. C. Ireland and C. Conservation Council, ISRP/ISAB 2009-1, Portland, Oregon. Lewandowski. 2011. Evaluation of methods for the collection and fertilization Jensen, N. R., P. J. Anders, C. A. Hoffman, L. S. Porter, S. C. Ireland, and of Burbot eggs from a wild stock for conservation aquaculture operations. K. D. Cain. 2011. Performance and macronutrient composition of age-0 Journal of Applied Ichthyology 27:9–15. Burbot fed four diet treatments. North American Journal of Aquaculture Northwest Marine Technology. 2008. Visible implant elastomer tag project man- 73:360–368. ual: guidelines on planning and conducting projects using VIE and associated Jensen, N. R., S. C. Ireland, J. T. Siple, S. R. Williams, and K. D. Cain. equipment. Northwest Marine Technology, Shaw Island, Washington. Avail- 2008. Evaluation of egg incubation methods and larval feeding regimes for able: http://www.nmt.us/support/appnotes/ape06.pdf. (September 2012). North American Burbot. North American Journal of Aquaculture 70:162– Paragamian, V. L. 2000. The effects of variable discharges on Burbot spawning 170. migrations in the Kootenai River, Idaho, USA, and British Columbia, Canada. 1698 ASHTON ET AL.

Pages 111–123 in V.L. Paragamian and D. W. Willis, editors. Burbot: biology, Steffensen, K. D., L. A. Powell, and J. D. Koch. 2010. Assessment of hatchery- ecology, and management. American Fisheries Society, Fisheries Manage- reared Pallid Sturgeon survival in the lower Missouri River. North American ment Section Publication 1, Spokane, Washington. Journal of Fisheries Management 30:671–678. Paragamian, V. L., and M. J. Hansen. 2009. Rehabilitation needs for Burbot Storck, T., and D. Newman. 1988. Effects of size at stocking on survival and in the Kootenai River, Idaho, USA, and British Columbia, Canada. North harvest of Channel Catfish. North American Journal of Fisheries Management American Journal of Fisheries Management 29:768–777. 8:98–101. Paragamian, V.L., and M. J. Hansen. 2011. Stocking for rehabilitation of Burbot Tatara, C. P. 2009. Size at implantation affects growth of juvenile steelhead im- in the Kootenai River, Idaho, USA and British Columbia, Canada. Journal of planted with 12-mm passive integrated transponders. North American Journal Applied Ichthyology 27:22–26. of Fisheries Management 29:417–422. Paragamian, V. L., C. Laude, K. D. Cain, J. M. Barron, and N. R. Jensen. 2011. Taylor, J. 1998. Progress report for a study on larval and juvenile Bur- A novel experiment of rearing Burbot larvae in cages. Journal of Applied bot in the upper Columbia River (1997). University of British Columbia, Ichthyology 27:16–21. Vancouver. Paragamian, V. L., B. J. Pyper, M. J. Daigneault, R. C. P. Beamesderfer, and Taylor, J. L. 2001. The early life history and ecology of Columbia Lake Burbot. S. C. Ireland. 2008. Population dynamics and extinction risk of Burbot in the Master’s thesis. University of British Columbia, Vancouver. Kootenai River, Idaho, USA and British Columbia, Canada. Pages 213–234 in Taylor, J. L. 2002. Juvenile Burbot sampling in Columbia and Windermere V. L. Paragamian and D. Bennett, editors. Burbot: ecology, management, and lakes, summer 2002. Columbia Basin Fish and Wildlife Compensation Pro- culture. American Fisheries Society, Symposium 59, Bethesda, Maryland. gram, Nelson, British Columbia. Available: http://www.fwcpcolumbia.ca/ Phillips, C. T., and J. N. Fries. 2009. An evaluation of visible implant elas- version2/reports/index.php. (September 2013). tomer for marking the federally listed Fountain Darter and the San Marcos Thompson, J. M., P. S. Hirethota, and B. T. Eggold. 2005. A comparison of salamander. North American Journal of Fisheries Management 29:529–532. elastomer marks and fin clips as marking techniques for Walleye. North Prentice, E. F., T. A. Flagg, and C. S. McCutcheon. 1990. Feasibility of using American Journal of Fisheries Management 25:308–315. implantable passive integrated transponder (PIT) tags in salmonids. Pages Thorsteinsson, V. 2002. Tagging methods for stock assessment and research 317–322 in N. C. Parker, A. E. Giorgi, R. C. Heidinger, D. B. Jester Jr., in fisheries. Marine Research Institute, Report of Concerted Action, FAIR E. D. Prince, and G. A. Winans, editors. Fish-marking techniques. American CT.96.1394, Reykjavik, Iceland. Fisheries Society, Symposium 7, Bethesda, Maryland. Tipping J. M., R. V. Cooper, J. B. Byrne, and T. H. Johnson. 1995. Length Raleigh, R. F., J. B. McLaren, and D. R. Graff. 1973. Effects of topical location, and condition factor of migrating and nonmigrating hatchery-reared winter branding techniques and changes in time on recognition of cold brands in steelhead smolts. Progressive Fish-Culturist 57:120–123. centrarchid and salmonid fish. Transactions of the American Fisheries Society Tipping, J. M. 2011. Effect of juvenile length on Chinook Salmon survivals at 102:637–641. four hatcheries in Washington State. North American Journal of Aquaculture Reeves, K. S., and D. L. Buckmeier. 2009. Mortality, predation, and tag visibility 73:164–167. of fish marked with visible implant elastomer tags. North American Journal Trabelsi, A., J. Gardeur, F. Teletchea, and P. Fontaine. 2011. Effects of 12 of Fisheries Management 29:323–329. factors on Burbot Lota lota (L., 1758) weaning performances using fractional Ricker, W. E. 1975. Computation and interpretation of biological statistics of factorial design experiment. Aquaculture 316:104–110. fish populations. Fisheries Research Board of Canada Bulletin 191. Vught, I., A. S. Harzevili, J. Auwerx, and D. De Charleroy. 2008. Aspects of Ryder, R. A., and J. Pesendorfer. 1992. Food, growth, habitat, and community reproduction and larviculture of Burbot under hatchery conditions. Pages interactions of young-of-the-year Burbot, Lota lota (L.), in a Precambrian 167–178 in V. L. Paragamian and D. Bennett, editors. Burbot: ecology, man- Shield lake. Hydrobiologia 243/244:211–227. agement, and culture. American Fisheries Society, Symposium 59, Bethesda, Skalski, J. R., R. A. Buchanan, and J. Griswold. 2009. Review of marking Maryland. methods and release recapture designs for estimating the survival of very Wagner, C. P., L. M. Einfalt, A. B. Scimone, and D. H. Wahl. 2009. small fish: examples from the assessment of salmonid fry survival. Reviews Effects of fin clipping on the foraging behavior and growth of age-0 in Fisheries Science 17:391–401. Muskellunge. North American Journal of Fisheries Management 29:1644– Skalski, J., and J. Griswold. 2006. Design and analysis of salmonid tagging 1652. studies in the Columbia Basin, volume XXI. A summary of methods for Wocher, H. 2010. Uberpr¨ ufung¨ der fischart Rutte (Lota lota L.) fur¨ zwecke conducting salmonid fry mark–recapture studies for estimating survival in der speisefischproduktion. [Validation of the Burbot (Lota lota L.) for pur- tributaries. Bonneville Power Administration, Project 198910700, Report poses of aquaculture production.] Abschlussbericht des Pilotprojektes II DOE/BP-00025091-2, Portland, Oregon. (2007–2010), Bezirk Niederbayern, Fachberatung fur¨ Fischerei. LOTAqua- Sloat, M. R., and P. F. Baker. 2011. Estimating habitat-specific abundances of Hatchery, Uberlingen,¨ Germany. (In German.) Downloaded by [Department Of Fisheries] at 22:53 25 November 2013 PIT-tagged juvenile salmonids using mobile antennas: a comparison with Wydoski, R., and L. Emery. 1983. Tagging and marking. Pages 215–237 in L. A. standard electrofishing techniques in a small stream. North American Journal Nielsen and D. L. Johnson, editors. Fisheries techniques. American Fisheries of Fisheries Management 31:986–993. Society, Bethesda, Maryland. Smith, J. R. 1973. Branding Chinook, Coho, and Sockeye salmon fry with hot Zar, J. H. 1999. Biostatistical analysis. Prentice-Hall, Upper Saddle River, New and cold metal tools. Progressive Fish-Culturist 35:94–96. Jersey. Sorensen, P. W., M. Bianchini, and H. E. Winn. 1983. Individually marking Zaroban, D. W., and S. M. Anglea. 2010. Efficacy of using passive integrated American Eels by freeze branding. Progressive Fish-Culturist 45:62–63. transponder technology to track individual Shorthead Sculpins. Western Soula, M., A. Navarro, S. Hildebrandt, M. J. Zamorano, J. Roo, C. M. North American Naturalist 70:218–223. Hernandez-Cruz,´ and J. M. Afonso. 2012. Evaluation of VIE (visible implant Zarski,˙ D., D. Kucharczyk, W. Sasinowski, K. Targonska,´ and A. Mamcarz. elastomer) and PIT (passive integrated transponder) physical tagging systems 2010. The influence of temperature on successful reproductions of Burbot, for the identification of Red Porgy fingerlings (Pagrus pagrus). Aquaculture Lota lota (L.) under hatchery conditions. Polish Journal of Natural Sciences International 20:571–583. 25:93–105. Stapanian, M. A., V. L. Paragamian, C. P. Madenjian, J. R. Jackson, J. Zymonas, N. D., and T. E. McMahon. 2006. Effect of pelvic fin ray removal Lappalainen, M. J. Evenson, and M. D. Neufeld. 2010. Worldwide status on survival and growth of Bull Trout. North American Journal of Fisheries of Burbot and conservation measures. Fish and Fisheries 11:34–56. Management 26:953–959. This article was downloaded by: [Department Of Fisheries] On: 25 November 2013, At: 22:54 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Transactions of the American Fisheries Society Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/utaf20 Hearing Sensitivity of the Burbot Peter A. Cott a b , Tom A. Johnston b c , John M. Gunn b & Dennis M. Higgs d a Department of Fisheries and Oceans , 301, 5205-50th Avenue, Yellowknife , Northwest Territories , X1A 1E2 , Canada b Cooperative Freshwater Ecology Unit, Vale Living with Lakes Center , Laurentian University , 935 Ramsey Lake Road, Sudbury , Ontario , P3E 2C6 , Canada c Ontario Ministry of Natural Resources , 935 Ramsey Lake Road, Sudbury , Ontario , P3E 2C6 , Canada d Department of Biological Sciences, University of Windsor , 401 Sunset Avenue , Windsor , Ontario , N9B 3P4 , USA Published online: 20 Nov 2013.

To cite this article: Peter A. Cott , Tom A. Johnston , John M. Gunn & Dennis M. Higgs (2013) Hearing Sensitivity of the Burbot, Transactions of the American Fisheries Society, 142:6, 1699-1704, DOI: 10.1080/00028487.2013.788559 To link to this article: http://dx.doi.org/10.1080/00028487.2013.788559

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions Transactions of the American Fisheries Society 142:1699–1704, 2013 C American Fisheries Society 2013 ISSN: 0002-8487 print / 1548-8659 online DOI: 10.1080/00028487.2013.788559

SPECIAL SECTION: BURBOT

Hearing Sensitivity of the Burbot

Peter A. Cott* Department of Fisheries and Oceans, 301, 5205-50th Avenue, Yellowknife, Northwest Territories X1A 1E2, Canada; and Cooperative Freshwater Ecology Unit, Vale Living with Lakes Center, Laurentian University, 935 Ramsey Lake Road, Sudbury, Ontario P3E 2C6, Canada Tom A. Johnston Ontario Ministry of Natural Resources, 935 Ramsey Lake Road, Sudbury, Ontario P3E 2C6, Canada; and Cooperative Freshwater Ecology Unit, Vale Living with Lakes Center, Laurentian University, 935 Ramsey Lake Road, Sudbury, Ontario P3E 2C6, Canada John M. Gunn Cooperative Freshwater Ecology Unit, Vale Living with Lakes Center, Laurentian University, 935 Ramsey Lake Road, Sudbury, Ontario P3E 2C6, USA Dennis M. Higgs Department of Biological Sciences, University of Windsor, 401 Sunset Avenue, Windsor, Ontario N9B 3P4, USA

Abstract Acoustic communication is central to the reproductive ecology of many fish species, particularly when conditions prevent the use of visual mating cues. The Burbot Lota lota is a freshwater codfish that spawns in a light-limited, under-ice environment. Both sexes possess swim bladder muscles, suggesting that both sexes engage in vocalization and that auditory cues are important to their mating system, but research on acoustic communication has been very limited in this species. In the current study we assessed the hearing sensitivity of Burbot from different size-classes. Burbot hearing was found to be more sensitive in juveniles than in adults, but across size-classes it was most sensitive at lower frequencies, which is similar to results with other codfishes and corresponds to the sounds produced by gadoids. Anthropogenic noise has the potential to disturb fish. The information gained in this study can be useful in assessing the impact of such noise, particularly under ice cover when Burbot are spawning. Further research

Downloaded by [Department Of Fisheries] at 22:54 25 November 2013 is required to determine whether winter-based resource development activities that generate under-ice noise are disruptive to Burbot communication and reproduction.

Sound, as it relates to animals, can be separated into two to increase (Popper and Hawkins 2012). Aquatic animals, in- broad categories; sounds that are of interest are typically called cluding many fish, rely on sound production and reception to “signals,” and sounds that interfere with signal detection are interact with conspecifics and their environment (Popper and generally referred to as “noise” (Fay 2012). The effects of Hawkins 2012), and environmental noise has the potential to noise on biota in aquatic environments are a growing con- block this important sensory modality. Fish have a wider ar- cern around the world, as anthropogenic noise sources continue ray of sound production mechanisms than amphibians, reptiles,

*Corresponding author: [email protected] Received September 5, 2012; accepted March 13, 2013 1699 1700 COTT ET AL.

birds, and mammals (Bass et al. 2008), and their hearing ability duction and reception a prime sensory system for transmitting varies greatly among species (Popper and Hawkins 2012). Fish mating cues. produce sounds mainly in social contexts (Hawkins 1993) and, The current study investigates the acoustic ecology of Burbot as in other types of fauna, sound can be an important com- by directly assessing their hearing ability. It was hypothesized ponent of the reproductive ecology (Andersson 1994), being that Burbot use acoustic signaling in association with their re- used for courtship and to signal reproductive readiness as well production, that they can hear well, and that their hearing is as for agonistic, territorial, and alarm signals (Hawkins 1993; within the sound energy range of their spawning vocalizations. Kasumyan 2009). Investigations of sound production and recep- As acoustic signaling may be a part of the mating system of tion are therefore vital to fully understanding the reproductive Burbot, an ontogenetic shift in hearing sensitivity (increasing ecology of many fish species and the impacts of anthropogenic from juvenile to adult) was predicted. sounds on fish ecology. In northern regions, resource development is often conducted One mechanism that fish use to produce sound is the rapid in the winter, using ice roads to reach areas that would be oth- contraction of striated muscles located dorsolaterally on the erwise inaccessible (Cott et al. 2008). The construction of and swim bladder (Hawkins and Amorin 2000; Parmentier et al. transportation across these roads can produce substantial noise 2006). These striated muscles are the fastest contracting muscles in underwater environments (Mann et al. 2009a). The com- known among vertebrates (Parmentier et al. 2006). The swim bination of the under-ice spawning period of Burbot and the bladders of fish are the earliest vocalization organs in verte- anthropogenic noise prevalent in their habitat at that time of brates, having evolved before the larynx (amphibians, reptiles, year makes it critical to better understand the sensory ecology and mammals) and the syrinx (birds) (Bass et al. 2008). The of this species. Gaining a better understanding of the reproduc- fishes in the family Gadidae (the codfishes), a highly important tive ecology of Burbot will also further our understanding of the family commercially, are known to possess these swim bladder mating systems of the codfishes in general, many of which, like muscles (hereafter referred to as drumming muscles) and to be that of the Burbot, are poorly studied. able to use them to produce sounds during the mating season (Hawkins 1993; Kasumyan 2009). Within this group, some fish are known to vocalize but others are not (Hawkins and Ras- METHODS mussen 1978). They may have a simple call, as is the case with The hearing ability of Burbot was assessed in the laboratory Atlantic Cod Gadus morhua, where the vocalization is limited to based on auditory evoked potentials (AEPs), a technique that a low-frequency grunt (Brawn 1961; Hawkins and Rasmussen measures the whole-brain electrical response to auditory stim- 1978; Rowe and Hutchings 2006) or hum (Rowe and Hutchings uli (Corwin et al. 1982; Mann et al. 2007). Adult Burbot were 2006); or, like Haddock Melanogrammus aeglefinus, they may collected in September 2009 from Windy Lake in northeastern produce a more complex mixture of grunts, knocks, hums, and Ontario (46◦36N, 81◦27W) using baited longlines, as per Cott buzzes (Hawkins and Amorin 2000; Hawkins and Rasmussen et al. (2011). Juvenile Burbot were collected from shallow water 1978). Not all codfishes possess drumming muscles, nor are all along rocky shorelines in Windy Lake using a Smith-Root Type of those with drumming muscles known to vocalize. However, VII backpack electrofishing unit. All fish were transported to the all the gadoid fish known to vocalize do possess drumming mus- laboratory in aerated holding tanks (adults and juveniles sepa- cles, making the presence of drumming muscles a good indicator rated) filled with lake water. Auditory testing was conducted of potential vocal capability (Hawkins and Rasmussen 1978). in a 1.7-m-long tank placed in a sound-attenuating chamber The Burbot Lota lota is one of four species within the Loti- (vocalbooths.com) with ambient noise levels below 85 dB re nae subfamily of the Gadidae, the others being the European 1 µPa (Wright et al. 2005; Belanger et al. 2010); fish were com-

Downloaded by [Department Of Fisheries] at 22:54 25 November 2013 Ling Molva molva, the Blue Ling Molva dypterygia, and the pletely submerged at least 5 cm below the water surface. Sound Cusk Brosme brosme (Cohen et al. 1990). The Burbot is the stimuli (tone bursts) were delivered via an underwater speaker only freshwater member of the cod family (Cohen et al. 1990) (UW-30; Lubell Laboratories) situated 75 cm from the fish. and is one of the most widely distributed freshwater fish species A stainless steel electrode (Rochester Electromedical) was in- in the world (Van Houdt et al. 2005), occurring in lakes, rivers, serted under the skin on the dorsal midline of the head above the and streams throughout the Northern Hemisphere (Scott and brain stem, with the reference and ground electrodes positioned Crossman 1973; Cohen et al. 1990; McPhail and Paragamian as per Mann et al. (2007). The tone bursts were generated in 2000; Stapanian et al. 2010). Very little is known about the re- SigGen (Tucker-Davis Technologies [TDT]) software and pre- productive ecology of Burbot (McPhail and Paragamian 2000), sented through a TDT System 3 evoked potential workstation. largely because they have the unusual trait of spawning at night Each tone burst was 10 ms in duration with a 2-ms rise–fall time in midwinter, often under ice cover (Scott and Crossman 1973; gated through a Hanning window. The tone bursts were pre- McPhail and Paragamian 2000). The fact that Burbot are not sex- sented in alternate phases (90◦ and 270◦) with 200 presentations ually dimorphic in appearance (McPhail 2007; Cott 2013; Cott of each phase and the responses averaged to reduce stimulus arti- et al. 2013), possess drumming muscles, and spawn at night in an fact, resulting in 400 traces being averaged for each combination otherwise low-light environment potentially makes sound pro- of sound level and frequency. The responses were collected and BURBOT HEARING 1701

TABLE 1. Summary statistics for the Burbot used in auditory evoked hearing potential assessment.

TL (mm) Total weight (g) Age (year) Size-class (g) n Mean SD Range Mean SD Range Mean SD Range <10 5 92 8 80–102 4.5 1.2 2.6–5.8 0 0 10–40 3 160 26 136–188 23.4 10.6 14.5–35.2 1 1 >500 3 495 39 457–534 709.5 197.7 566.4–935.1 12.7 3.5 0.9–16

averaged in BioSig through the TDT evoked potential worksta- (GLM procedure); differences were considered significant at tion and stored for offline assessment of the auditory thresholds α ≤ 0.5. for each frequency. Burbot were subjected to sounds at frequen- cies of 100, 200, 400, 800, and 1,600 Hz with an initial sound pressure level (SPL) of 140 dB re 1 µPa rms. Auditory thresh- RESULTS olds were determined by visual analysis of the resultant auditory There was evidence of an ontogenetic shift in the hearing waveforms as the decibel levels at which the neural response was ability of Burbot; however, the direction of the shift was con- not above the background level (Wright et al. 2005; Mann et al. trary to what was predicted, the smallest Burbot having the 2007). To assess the potential for ontogenetic shifts in hearing greatest hearing sensitivity. For all Burbot hearing was most ability, AEP was conducted on three different size-classes of sensitive in the lower frequency range, with decreasing sen- Burbot (Table 1), with the hearing ability being analyzed at each sitivities as frequency increased (Table 2; Figure 1). At 100, frequency, as per Popper et al. (2005). The Burbot were mea- 200, and 400 Hz, the smallest size-class of Burbot (juveniles sured to the nearest millimeter, gram, and year (sagittal otolith). <10 g) had the greatest hearing sensitivity. At 100 and 200 Hz, To facilitate comparisons with previous work, the data are pre- adult Burbot had the second most sensitive hearing (ANOVA: sented in both pressure units and acceleration units. Accelera- F2, 8 = 12.6, P = 0.003; and F2, 8 = 16.0, P = 0.002, re- tion values were obtained by placing a triaxial accelerometer spectively; Tukey’s test: P < 0.05). At 400 Hz, hearing sen- (Bruel¨ and Kjær; Type 4524 modified for underwater use) on sitivity decreased with increasing size-class (ANOVA: F2, 7 = the fish testing platform after the trials were completed. Anal- 32.2, P = 0.0003; Tukey’s test: P < 0.05). There were no sig- yses of variance (ANOVA) were performed with SAS version nificant differences in auditory sensitivity among Burbot size- 9.2 (SAS 2009) using size-class as the class variable and the classes at 800 and 1,600 Hz (ANOVA: F2, 7 = 0.02, P = 0.98; SPL at a given frequency as the dependant variable, followed by and F1, 2 = 0.48, P = 0.56, respectively). Hearing ability was a Tukey’s honestly significant different comparison of means not detected among the smallest Burbot at 1,600 Hz. Adult

TABLE 2. Auditory evoked potentials for three size-classes of Burbot. Abbreviations are as follows: SPL = sound pressure level, PA = particle acceleration, NR = no response detected, * = no response detected from one Burbot, ** = no response detected from two Burbot.

SPL (dB re 1 µPa) PA(dBre1m·s−2) Size-class n Frequency (Hz) Mean (SD) Range Mean (SD) Range < − − − Downloaded by [Department Of Fisheries] at 22:54 25 November 2013 10 g 5 100 114 (7.4) 105–125 59.3 (7.2) 68.0– 48.6 5 200 118 (4.5) 115–125 −58.8 (4.3) −61.7–−52.2 5 400 128 (4.5) 125–135 −50.6 (3.9) −53.3–−44.5 5 800 147 (2.7) 145–150 −33.7 (2.1) −35.3–−31.4 4 1600 NR NR NR NR 10–40 g 3 100 138.3 (7.6) 130–145 −35.6 (7.4) −43.7–−29.2 3 200 138.3 (5.8) 135–145 −39.5 (5.5) −42.7–−33.2 3 400 143.3 (2.9) 140–145 −37.2 (2.5) −40.1–−35.7 3 800 146.7 (2.9) 145–150 −34.0 (2.3) −35.3–−31.4 3 1600 146.7 (10.4) 135–155 −30.0 (9.7) −40.9–−22.3 >500 g 3 100 131.7 (5.8) 125–135 −42.1 (5.6) −48.6–−38.9 3 200 131.7 (5.8) 125–135 −45.8 (5.5) −52.2–−42.7 3* 400 152.5 (3.5) 150–NR −29.1 (3.1) −31.3–NR 3* 800 147.5 (10.6) 140–NR −33.3 (8.3) −39.2–NR 3** 1600 155–NR 155–NR −22.3 −22.3–NR 1702 COTT ET AL.

160 and decrease as they grow (Belanger et al. 2010), as was the case with the Burbot in this study. Ontogenetic differences in 150 habitat use between juveniles and adults offer a possible ex-

Pa) planation for the direction of the ontogenetic shift in hearing µ 140 ability. Juveniles are found in shallow water near the shore- line and strongly associated with cover (Ryder and Pesendorfer 1992; Stewart and Watkinson 2004), as they are readily preyed 130 upon by a variety of piscivorous fishes (Ryder and Pesendor- fer 1992; Amundsen et al. 2003). Perhaps young Burbot need 120 more sensitive hearing to avoid predation. In contrast, adult Burbot generally reside in the offshore benthic zone (Scott and

Threshold (dB re 1 110 Crossman 1973; Stewart and Watkinson 2004), and would be at little risk of predation. Fabricius (1954) observed young-of-the- 100 year Burbot “fighting vigorously” over their respective hiding 0 200 400 600 800 1000 1200 1400 1600 spots in aquaria. Such aggressive intraspecific behavior may Acoustic stimulus (Hz) be accompanied by agonistic calls. Some codfishes, such as the Lythe (also known as the European Pollack) Pollachius pol- FIGURE 1. Auditory evoked potentials for three size-classes of Burbot col- lachius, are known to vocalize as both juveniles and adults while lected from Windy Lake in September 2009. The symbols represent means and the whiskers represent SEs. Black circles denote juvenile Burbot of <10 g, some species only vocalize as juveniles. For instance, juvenile triangles juvenile Burbot from 10 to 20 g, and squares adult Burbot (>500 g). Pollock Pollachius virens have drumming muscles but these are absent in adults (Hawkins and Rasmussen 1978). Assessing hearing was most sensitive at 100 and 200 Hz (mean = 131.7 dB whether young Burbot vocalize is a topic for future study. re 1 µPa). The only other documented hearing threshold for Burbot was reported by Mann et al. (2007), where the hearing potential for a single 38-mm (0.3-g), young-of-the-year Burbot was assessed. DISCUSSION Given its size (less than the smallest Burbot assessed in our The current study is the first to document and describe the study), this Burbot would probably have only recently switched hearing ability of adult Burbot and one of only two (see Mann from a pelagic to a benthic existence (Ryder and Pesendorfer et al. 2007) that have examined hearing in this species at all. 1992; McPhail and Paragamian 2000). The AEP of the Burbot The hearing thresholds for adult Burbot were lowest at the fre- in Mann et al. (2007) was more sensitive than those of the quencies of their sound production (Cott 2013) and similar to Burbot examined in the current study. This ontogenetic trend the hearing abilities of other gadoid fishes (Chapman 1973; of decreased hearing sensitivity is not unprecedented among Chapman and Hawkins 1973; Mann et al. 2009b). The physio- the gadoid fishes. Walleye Pollock Gadus chalcogrammus have logical hearing thresholds derived from auditory tests performed lower hearing sensitivity with increasing age despite having a in laboratories (where the ambient noise levels are lower than in larger inner ear and, like Burbot, all age-classes have better natural environments) should be seen as estimates of the poten- hearing at lower frequencies (Mann et al. 2009b). tial hearing ability of the species; actual hearing abilities will The under-ice soundscape at the time of Burbot spawning vary among environments (Belanger et al. 2010). The hearing vocalizations would be expected to be relatively quiet (a 1-min measured for European Ling and other codfishes by Chapman period measured 96.5 ± 5.1 [mean ± SD] dB re 1 µPa rms

Downloaded by [Department Of Fisheries] at 22:54 25 November 2013 (1973) and Chapman and Hawkins (1973) was more sensitive SPL over the frequency range 10–8,000 Hz; Cott 2013). With than that found for Burbot. This may be the result of differences the absence of wave and storm noises, ice cracks would be in hearing ability among these fishes or differences in the ex- one of the only forms of natural noise (Mann et al. 2009a). In perimental approach or technology used in the studies. Caution northern areas, resource development is often conducted dur- should be exercised when comparing the hearing sensitivities of ing winter over frozen lakes and rivers because of the ability fishes among studies. to make ice roads to otherwise inaccessible and remote areas What was unexpected was that the juvenile Burbot would (Cott et al. 2008; Mann et al. 2009a). These activities gener- have more sensitive hearing than the adults. It has been spec- ate sounds that infiltrate under-ice environments, changing the ulated that juvenile fish have less sensitive hearing than adult soundscapes of aquatic ecosystems. Such sounds have the po- fish because auditory systems develop during maturation (Pop- tential to disturb fish (Mann et al. 2009a; Cott et al. 2012). Ice per et al. 2005). Another possible explanation for this is that cracks can be quite loud, with peak pressure levels often exceed- the electrode is closer to the neural tissue in the smaller fish ing a peak SLP of 165 dB re 1 µPa (Cott 2013), but the differ- and thus gives a stronger response that is more easily de- ence between natural noises like ice cracks and anthropogenic tected. However, for species such as Round Goby Neogobius noises is that the former is transient and the latter persist (Mann melanostomus, hearing can be most sensitive in young fish et al. 2009a). Persistent anthropogenic noises that are above the BURBOT HEARING 1703

normal ambient conditions of the environment can raise the hear- Cott, P. A., T. A. Johnston, and J. M. Gunn. 2011. Food web position of Burbot ing threshold requirement for signal detection by a given species relative to Lake Trout, Northern Pike, and Lake Whitefish in four sub-Arctic and may render some low-level sounds undetectable (Hawkins boreal lakes. Journal of Applied Ichthyology 27:49–56. Cott, P. A., T. A. Johnston, and J. M. Gunn. 2013. Life history stability among 1993; Fay 2012). Because Burbot spawn in the winter, under-ice burbot populations across environmental gradients. Transactions of the Amer- noise produced by industrial activities may disturb them directly ican Fisheries Society. or mask their spawning vocalizations, reducing their ability to Cott, P. A., D. A. Mann, D. M. Higgs, T. A. Johnston, and J. M. Gunn. 2012. find each other and possibly disrupting their spawning activ- Assessing disturbance from under-ice noise on fishes in boreal lakes. Pages ities. Finstad and Nordeide (2004) suggested that noise from 363–366 in A. N. Popper and A. Hawkins, editors. Effects of noise on aquatic life. Advances in experimental medicine and biology. Springer, New York. commercial fishing activities disturbs the spawning behavior of Cott, P. A., P. K. Sibley, A. M. Gordon, R. A. Bodaly, K. H. Mills, W. M. Atlantic Cod. Such disruptions may interfere with mate selec- Somers, and G. A. Fillatre. 2008. The effects of water withdrawal from ice- tion, spawning success, or egg viability (Rowe and Hutchings covered lakes on oxygen, temperature and fish. Journal of the American Water 2006). Anthropogenic noise also has the potential to physically Resources Association 44:328–342. harm fish (Popper and Hawkins 2012). For example, noise from Fabricius, E. 1954. Aquarium observations on the spawning behaviour of the Burbot, Lota vulgaris L. Report of the Institute of Freshwater Research 35:51– air gun–generated seismic activities has been shown to cause 57. temporary (Popper et al. 2005) or permanent (McCauley et al. Fay, R. R. 2012. Listening to noise. Pages 129–134 in A. N. Popper and A. 2003) hearing loss in fish, depending on the species and sever- Hawkins, editors. The effects of noise on aquatic life. Springer, Advances in ity of exposure. Understanding the hearing abilities quantified Experimental Medicine and Biology 730, New York. in the current study is vital to assessing the impacts of anthro- Finstad, J. L., and J. T. Nordeide. 2004. Acoustic repertoire of spawning cod, Gadus morhua. Environmental Biology of Fishes 70:427–433. pogenic noise issues in the light of increasing northern ice-based Hawkins, A. D. 1993. Underwater sound and fish behaviour. Pages 129–169 in development, particularly for a winter-spawning fish. T. J. Pitcher, editor. Behaviour of teleost fishes, 2nd edition. Chapman and Hall, London. Hawkins, A. D., and M. C. Amorin. 2000. Spawning sounds of the male Had- ACKNOWLEDGMENTS dock, Melanogrammus aeglefinus. Environmental Biology of Fishes 59:29– We thank Carl Jorgensen for field assistance and Jim Reist 41. and Jeff Hutchings for review of earlier drafts. Critical peer re- Hawkins, A. D., and K. J. Rasmussen. 1978. The calls of gadoid fish. Journal view provided by David Jude, Martin Stapanian, and two anony- of the Marine Biological Association of the United Kingdom 58:891–911. mous reviewers was highly constructive. This work was funded Kasumyan, A. O. 2009. Acoustic signaling in fish. Journal of Ichthyology 49:963–1020. by Natural Resources Canada through the Program of Energy, Mann, D. A., P. A. Cott, B. W. Hanna, and A. N. Popper. 2007. Hearing in Research, and Development and under the guidance of Animal eight species of northern Canadian freshwater fishes. Journal of Fish Biology Use Protocol FWI-ACC-2009-063. 69:1–12. Mann, D., P. A. Cott, and B. Horne. 2009a. Under-ice noise generated from diamond exploration in a Canadian sub-arctic lake and potential impacts on REFERENCES fishes. Journal of the Acoustical Society of America 126:2215–2222. Amundsen, P.-A., T. Bøhn, O. A. Popova, F. J. Staldvik, Y. S. Reshetnikov, Mann, D. A., C. D. Wilson, J. Song, and A. N. Popper. 2009b. Hearing sensi- N. A. Kashulin, and A. A. Lukin. 2003. Ontogenetic niche shifts and resource tivity of the Walleye Pollock. Transactions of the American Fisheries Society partitioning in a subarctic piscivore fish guild. Hydrobiologia 497:109–119. 138:1000–1008. Andersson, M. 1994. Sexual selection. Princeton University Press, Princeton, McCauley, R. D., J. Fewtrell, and A. N. Popper. 2003. High intensity anthro- New Jersey. pogenic sound damages fish ears. Journal of Acoustical Society of America Bass, A. H., E. H. Gilland, and R. Baker. 2008. Evolutionary origins for social 113:638–642. vocalization in a vertebrate hindbrain-spinal compartment. Science 321:417– McPhail, J. D. 2007. The freshwater fishes of British Columbia. University of 421. Alberta Press, Edmonton. Belanger, A. J., I. Bobeica, and D. M. Higgs. 2010. The effect of stimulus type McPhail, J. D., and V. L. Paragamian. 2000. Burbot biology and life history. Downloaded by [Department Of Fisheries] at 22:54 25 November 2013 and background noise on hearing abilities of the Round Goby Neogobius Pages 11–23 in V. L. Paragamian and D. W. Willis, editors. Burbot: biology, melanostomus. Journal of Fish Biology 77:1488–1504. ecology, and management. American Fisheries Society, Fisheries Manage- Brawn, V. M. 1961. Sound production by the cod (Gadus callarias L.). Be- ment Section Publication 1, Spokane, Washington. haviour 18:239–255. Parmentier, E., J.-P. Lagardere,` J.-B. Braquegnier, P. Vandewalle, and M. L. Chapman, C. J. 1973. Field studies of hearing in teleost fish. Helgolaender Fine. 2006. Sound production mechanism in a carapid fish: first exam- Wissenschaftliche Meeresuntersuchungen 24:371–390. ple with a slow sonic muscle. Journal of Experimental Biology 209: Chapman, C. J., and A. D. Hawkins. 1973. A field study of hearing in the cod, 2952–2906. Gadus morhua L. Journal of Comparative Physiology A 85:147–167. Popper, A. N., and A. Hawkins, editors. 2012. The effects of noise on aquatic Cohen, D. M., T. Inada, T. Iwamato, and N. Scialabba. 1990. FAO species life. Springer, Advances in Experimental Medicine and Biology 730, New catalogue, version 10: gadiform fishes of the world (Order Gadiformes). York. Food and Agriculture Organization of the United Nations, Rome. Popper, A. N., M. E. Smith, B. W. Hanna, A. O. MacGillivray, M. E. Austin, Corwin, J. T., T. H. Bullock, and J. Schweitzer. 1982. The auditory brainstem and D. A. Mann. 2005. Effects of exposure to seismic airgun use on hearing response in five vertebrate classes. Electroencephalography and Clinical Neu- of three fish species. Journal of the Acoustical Society of America 117:3958– rophysiology 54:629–641. 3971. Cott, P. A. 2013. Life history and reproductive ecology of a mid-winter spawner: Rowe, S., and J. A. Hutchings. 2006. Sound production by Atlantic Cod the Burbot (Lota lota). Doctoral dissertation. Laurentian University, Greater during spawning. Transactions of the American Fisheries Society 135: Sudbury, Ontario. 529–538. 1704 COTT ET AL.

Ryder, R. A., and J. Pesendorfer. 1992. Food, growth, habitat, and community Stewart, K. W., and D. A. Watkinson. 2004. The freshwater fishes of Manitoba, interactions of young-of-the-year Burbot, Lota lota L., in a Precambrian volume 276. University of Manitoba Press, Winnipeg. Shield lake. Hydrobiologia 243/244:211–227. Van Houdt, J. K. J., L. de Cleyn, A. Perretti, and F. A. M. Volckaert. 2005. SAS. 2009. SAS OnlineDoc, version 9.2. SAS Institute, Cary, North Carolina. A mitogenic view on the evolutionary history of the Holarctic freshwater Scott, W. B., and E. J. Crossman. 1973. Freshwater fishes of Canada. Fisheries gadoid, Burbot (Lota lota). Molecular Ecology 14:2445–2457. Research Board of Canada, Ottawa. Wright, K. J., D. M. Higgs, A. J. Belanger, and J. M. Leis. 2005. Auditory and Stapanian, M. A., V. L. Paragamian, C. P. Madenjian, J. R. Jackson, J. Lap- olfactory abilities of pre-settlement larvae and post-settlement juveniles of palainen, M. J. Evenson, and M. D. Neufeld. 2010. Worldwide status of a coral reef damselfish (Pisces: Pomacentridae). Marine Biology 147:1425– Burbot and conservation measures. Fish and Fisheries 11:34–56. 1434. Downloaded by [Department Of Fisheries] at 22:54 25 November 2013 This article was downloaded by: [Department Of Fisheries] On: 25 November 2013, At: 22:54 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Transactions of the American Fisheries Society Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/utaf20 Variability in Burbot Cohort Abundance at Juvenile and Adult Stages in Columbia Lake, British Columbia Josh L. Taylor a & Steven K. A. Arndt b a Environmental Dynamics, Inc. , Unit 208A-2520 Bowen Road, Nanaimo , British Columbia , V9T 3L3 , Canada b Ministry of Forests, Lands and Natural Resource Operations , Fish and Wildlife Compensation Program–Columbia Basin , 401-333 Victoria Street, Nelson , British Columbia , V1L 4K3 , Canada Published online: 20 Nov 2013.

To cite this article: Josh L. Taylor & Steven K. A. Arndt (2013) Variability in Burbot Cohort Abundance at Juvenile and Adult Stages in Columbia Lake, British Columbia, Transactions of the American Fisheries Society, 142:6, 1705-1715, DOI: 10.1080/00028487.2013.774292 To link to this article: http://dx.doi.org/10.1080/00028487.2013.774292

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions Transactions of the American Fisheries Society 142:1705–1715, 2013 C American Fisheries Society 2013 ISSN: 0002-8487 print / 1548-8659 online DOI: 10.1080/00028487.2013.774292

SPECIAL SECTION: BURBOT

Variability in Burbot Cohort Abundance at Juvenile and Adult Stages in Columbia Lake, British Columbia

Josh L. Taylor Environmental Dynamics, Inc., Unit 208A-2520 Bowen Road, Nanaimo, British Columbia V9T 3L3, Canada Steven K. A. Arndt* Ministry of Forests, Lands and Natural Resource Operations, Fish and Wildlife Compensation Program–Columbia Basin, 401-333 Victoria Street, Nelson, British Columbia V1L 4K3, Canada

Abstract This study examined changes in the abundance of juvenile (age-0 and age-1) and adult Burbot Lota lota in Columbia Lake, British Columbia, for the 1991–1999 cohorts. The objectives were to quantify the degree of variation in cohort abundance at different life stages and investigate the timing of recruitment limitation. Adult spawner abundance and age composition were monitored at a tributary spawning site from 1996 to 2001. Juvenile cohort abundance was estimated from 1997 to 1999, providing age-0 abundance indices for the 1997–1999 cohorts and age-1 indices for the 1996–1998 cohorts. The number of tributary spawners declined from about 1,500 in 1996 and 1997 to 86 in 1999 and then rebounded to 995 by 2001. Adult length frequency and age composition showed that this fluctuation reflected periodic influxes of strong cohorts that dominated the spawning population. The strongest new cohort (1999) observed at the tributary in 2001 came from the smallest number of tributary spawners. Substantial differences in the juvenile abundance of cohorts spawned from similar numbers of adults (1996, 1997) suggested a density-independent effect at the egg or larval stage. The 1999 cohort, however, was more abundant than several other cohorts at the adult stage but not at age 0, suggesting a second limiting period after the first growing season. Possible explanations for the exceptional survival of the 1999 cohort include a temporary expansion of interstitial habitat caused by unusually high water levels and low abundance of older Burbot. As large fluctuations in cohort abundance appear to be characteristic of some Burbot populations, inclusion of age composition is recommended for population assessments. Downloaded by [Department Of Fisheries] at 22:54 25 November 2013 In the upper Columbia River system, Burbot Lota lota are among these factors may have contributed to these declines near the southern limit of their western North American dis- (McPhail 1997; Cohen et al. 2000; Fisher 2000; Paragamian tribution. Since the 1960s, some populations in the drainage, et al. 2000; Ahrens and Korman 2002; Stapanian et al. 2010a). namely, those in Kootenay Lake and the Kootenay River, have On a worldwide basis the species remains abundant throughout collapsed (Paragamian et al. 2000; Ahrens and Korman 2002). much of its natural range, but there are many populations in Consequently, the lower Kootenay population is red-listed (clas- decline or that have been extirpated (Stapanian et al. 2010a). sified as endangered or threatened) in British Columbia (British Such population declines, as well as concern about the preser- Columbia Conservation Data Centre 2012) and is designated vation of biodiversity, are causing increased interest in Burbot as a species of special concern in Idaho (Ireland and Perry management. 2008). Impoundments, changes in temperature or nutrient load- One impediment to rational Burbot management is the ing, changes in climate, dyking, overfishing, or interactions meager data on early life stages in relation to recruitment

*Corresponding author: [email protected] Received September 4, 2012; accepted January 30, 2013 1705 1706 TAYLOR AND ARNDT

variability. Recruitment variation tends to increase as population size is reduced, thereby reducing population stability (Bedding- ton and May 1977; Myers 2001). For a population depressed by fishing pressure, fluctuations in recruitment success can cause or accelerate stock collapse (Gjoesaeter 1995; Smolders et al. 2000). Recruitment indices are commonly used in freshwater and marine fisheries assessments to study variability in recruit- ment (Sundby et al. 1989; de Lafontaine et al. 1992). Studies of this type may be especially important in managing Burbot populations in the Columbia system because the influence of en- vironmental stochasticity on recruitment is often greatest near the edge of a species’ range (Myers 2001). Understanding nat- ural or environmentally induced recruitment variation is also important for predicting the potential effects of climate change (Jackson et al. 2008; Worthington et al. 2011). Burbot are the only truly freshwater cod (Gadidae; Nelson 1994), and as a freshwater species with a typically marine life history (i.e., late-winter spawning time, enormous fecundity, and extended larval period; McPhail and Paragamian 2000), Burbot may have especially high recruitment variability. Pepin and My- ers (1991) found that recruitment variability in marine fish is correlated with the size difference at hatching and metamorpho- sis. They argue that this difference is a measure of the duration of the larval stage. Extending this argument to freshwater fish suggests that recruitment variability in Burbot, with an approxi- mately 3.5-month larval period (Taylor 2001), should be higher than in other freshwater fish, which generally have shorter larval periods. A weak link between higher fecundity and increased recruitment variability has also been suggested for other species (Rickman et al. 2000; Myers 2001). Houde (1994) suggests that weather-driven variation in early mortality is greater for fish in FIGURE 1. Map of Columbia Lake showing the locations of the spawning tributary with the counting weir and the 12 juvenile sampling sites (dots). freshwater environments than in marine environments because freshwater nursery areas tend to be relatively small, shallow, and vulnerable to sudden changes in temperature, pH, oxygen level, The lake is mesotrophic and too shallow to form a summer and chemical constituency. In spite of these reasons to expect thermocline. From 1997 to 1999, peak offshore summer tem- a large effect of environmental stochasticity on Burbot larval peratures at the surface ranged from 19◦Cto24◦C (J. L. Taylor, survival, there have been few attempts to relate the abundance unpublished data). The two major inlets to Columbia Lake are of Burbot at age 0 or age 1 to subsequent adult abundance. Dutch Creek at the north end of the lake and a short, unnamed In this study, our goal was to document the differences in spring-fed stream at the south end (Figure 1). The Columbia

Downloaded by [Department Of Fisheries] at 22:54 25 November 2013 cohort abundance over several years at the juvenile and adult River exits at the north end of the lake and enters Winder- life stages for Burbot in Columbia Lake, British Columbia, to mere Lake about 15 km downstream. Burbot spawning occurs provide insight into the timing of recruitment regulation and the in the unnamed tributary and probably under the ice in the lake. factors influencing survival in a shallow lake near the southern Lake spawning is implied by the presence during the spawning edge of the species’ distribution. Our analyses were restricted season of gravid and recently spent fish in the winter fishery, to age-0, age-1, and adult abundance indices encompassing the which primarily occurs at the north end of the lake, over 10 km 1991–1999 cohorts. from the tributary spawning site (S. K. A. Arndt, unpublished data). The dominant substrate of Columbia Lake, including its METHODS shoreline, is fine silt. Gravel and rocky substrates are rare, and Study site.—Columbia Lake, the source of the Columbia human activities, such as the building of a railway (Figure 1) and River, is located in the Kootenay region of southeastern British jetties, provide most of the gravel and rocky substrate in the lake. Columbia at 50◦15N, 115◦50W at an elevation of 809 m. It has This artificial substrate rarely extends much more than a meter a surface area of 25.7 km2, a length of approximately 13 km, below the high water mark and, due to an average water level a mean depth 2.9 m, and a maximum depth of 6 m (Figure 1). fluctuation of about 1 m, is mostly dry from late fall through BURBOT COHORT VARIABILITY 1707

early spring. Further description of the study site is included in Arndt and Hutchinson (2000) and Taylor (2001). Juvenile sampling.—The relative abundance of four cohorts of juvenile Burbot (1996–1999) was estimated by electrofishing shoreline transects between June 28 and August 1 in 1997, 1998, and 1999. This allowed estimates of age-0 density for each of these cohorts and age-1 density for the 1996–1998 cohorts. Each transect was a 27.5-m stretch of shoreline sampled from the lake’s edge out to either 1 m depth or 27.5 m from shore, whichever came first. The sampling gear was a DC backpack electrofisher with 20-cm-diameter anode, a pulse frequency of 100 Hz, and a current of 0.5 A. The relative abundance index for each transect was defined as the combined Burbot catch from two passes separated by 15 min. All electrofishing was performed by the same person wearing polarized sunglasses and only when light and wind conditions allowed adequate visibility. FIGURE 2. Length frequency histogram for 230 juvenile Burbot caught at 12 Forty transects were sampled in 1997. These sites were uni- transects in Columbia Lake from 1997 to 1999. Note that the summer length formly distributed and represented all of the shoreline habitats distributions of age-0 and older Burbot do not overlap. around the lake. A random sample of 12 transects from the orig- inal 40 transects was resampled in 1998 and 1999. Additional sampling were assigned an age of either 0 or older based on opportunistic electrofishing aimed at larger and generally older length alone (Figure 2). Assignment of age to Burbot too large juveniles from 1997 to 1999 yielded supplementary size-at-age to be age 0 was based, if possible, on the number of annuli on information. Data from 128 of these larger juveniles were in- their sagittal otoliths. cluded in estimating the population growth curve. Sagittal otoliths were obtained from Burbot in creel surveys Adult sampling.—Spawning Burbot were captured from 1996 (81 in the winter of 1995–1996, 91 in the winter of 1996–1997, to 2001 at a weir and trap (bar mesh size, 2.5 cm) in the unnamed and 4 in the winter of 2000–2001), from the tributary weir (2 tributary between the lake and the main area of spring outflows. in 1996, 10 in 1997, and 31 in 2001), and from opportunistic The weir was installed no later than January 25 and maintained shoreline electrofishing samples (128). All otoliths were first until at least February 16. This allowed the capture of most, if not wetted with mineral oil and examined whole against a black all, fish moving upstream. Captured Burbot were measured (TL background with a dissecting microscope. Those showing three and weight), checked for tags from previous spawning events, or more annuli were split and reexamined with oblique light. tagged if they had no previous tag, and then released. Spaghetti- Tightly spaced annuli formed at older ages were often more type FLOY tags were applied in 1996 and passive integrated distinct on split otoliths. The first annulus was verified using transponder (PIT) tags were applied in subsequent years. Further length frequency data (Figure 2). Hyaline (clear) growth was details are provided by Arndt and Hutchinson (2000). Since we seen at the edge of the otolith during the summer. Opaque edge did not apply secondary marks to the fish, we cannot estimate growth occurred during the winter. The ages of a subsample the rates of tag loss; however, a Montana study found retention of 20 otoliths were independently confirmed by another experi- rates of about 75% for Burbot of similar size using PIT tags in enced individual (G. Carder, Salmon Arm, British Columbia). the same location (cheek muscle) as in this study (J. Deshazer, The ages from the creel, spawning tributary, and shoreline

Downloaded by [Department Of Fisheries] at 22:54 25 November 2013 Montana Fish and Game, personal communication). We also samples were combined to estimate the growth curve through obtained length measurements and otoliths from angler-caught the juvenile and adult life stages. The model of growth used was Burbot during the winters of 1995–1996 to 2000–2001, except the von Bertalanffy growth function (VBGF), that angler-caught Burbot were not sampled for length during the winter of 1999–2000 or for otoliths during the winters of 1997– Lt = L∞{1 − exp[−K (t − t0)]}, (1) 1998 to 1999–2000. The sex of individuals was determined by the presence of expressed gametes when examined at the where Lt is length at time t, L∞ is the mean asymptotic weir (Arndt and Hutchinson 2000) or by dissection for creeled length, K is a constant, and t0 is the hypothetical age at which fish. the fish had zero length (Moreau 1987). Time is measured Age and growth.—The hatching date was assumed to be in years. The parameters in the VBGF were estimated us- March 15. This date was calculated using an approximate fertil- ing the nonlinear regression platform of JMP statistical soft- ization date of early February (Arndt and Hutchinson 2000) and ware (version 3; SAS Institute, Inc.). Heteroscedasticity was an incubation period of almost 1.5 months (Taylor and McPhail corrected using iteratively reweighted least-squares regression 2000). Since the lengths of age-0 and age-1 Burbot did not over- (Neter et al. 1996). A variance function was used to obtain the lap until the end of August, Burbot caught in shoreline transect weights for the iteratively reweighted least-squares process. The 1708 TAYLOR AND ARNDT

variance function was estimated by squaring the predicted values RESULTS from a regression of absolute residuals against the reciprocal of age. Growth Pattern Variation in cohort abundance.—Variation in cohort abun- For Columbia Lake Burbot, 39 cm appears to be the thresh- dance was assessed using changes in juvenile abundance from old above which fish begin to recruit to the ice fishery and no 1997 to 1999, changes in spawner abundance and length fre- longer inhabit shoreline habitats during the day (Figure 3). This quency distribution in the tributary from 1996 to 2001, and the threshold was crossed between ages 2 and 3 for most fish (Fig- age composition of adults sampled from the fishery and the ure 3). Consequently, between ages 2 and 3, Burbot from the weir. Juvenile abundance was estimated for the 1996, 1997, creel sample were larger than Burbot from the shoreline sam- 1998, and 1999 cohorts using the relative abundance at the 12 ples. Younger spawners at the tributary were also larger than shoreline transects. A repeated-measures analysis of variance fish of the same age sampled from the shoreline (ages 2 and (RMANOVA) was used to compare age-0 abundance among 3; Figure 3), indicating that the faster-growing individuals of a years. This was an appropriate technique because the same 12 cohort also spawned at an earlier age. As a consequence of the biased size of the Burbot in the sample sites were sampled each year. A log10(Y + 1) trans- formation was used to reduce the skewness and heteroscedas- shoreline, creel, and spawner samples over the age range of ticity of the observed age-0 abundances. Multiple comparisons 22.5–48 months, Burbot within this age range were not included were done using Tukey’s honestly significantly different (HSD) when fitting the VBGF. The estimated VBGF was test. The potential effect of differences in sampling dates be- tween years was examined by plotting the residuals from the Lt = 63.5 {1 − exp[−0.294(t + 0.018)]}. (2) RMANOVA model against sampling date. As data transforma- tion did not sufficiently reduce the skewness and heteroscedas- The modeled lengths at ages 1, 5, and 10 years of age were ticity of the observed age-1 abundances, a nonparametric Fried- 16.4, 49.0, and 60.2 cm, respectively (Figure 3). Mean asymp- man’s rank test was used (instead of RMANOVA) when com- totic length was estimated as 63.5 cm, with a 95% CI of 61.3– paring age-1 abundances between years. The critical level for 66.0 cm. This agrees with growth observations on fish recaptured all tests was set at 0.05. as long as 6 years after tagging (Arndt, unpublished data). No Cohort-specific contributions to the spawner counts from difference in the growth pattern of males and females could be 1996 to 2001 were estimated by apportioning the total returns in distinguished. each year to individual cohorts based on available age and size data. For example, comparison of the size and age distribution of the 35 adult Burbot aged in the winter of 2001 with the 80 size distribution of the 995 spawners counted in 2001 suggested that approximately 40% of these spawners (i.e., 400 fish), were 70 from the 1999 cohort. Cohort returns for each year were summed to obtain an estimated total contribution to tributary spawning 60 for the years 1996–2001. This was estimated for all cohorts encountered in the aged samples from 1991 to 1999. The timing 50 of factors limiting cohort abundance was then examined by comparing the initial numbers of tributary spawners with the 40 relative abundance at age 0 and age 1 (shoreline transects) and

Downloaded by [Department Of Fisheries] at 22:54 25 November 2013 subsequent spawner returns to the tributary for the 1996–1999 30

Total length (cm) shoreline cohorts. spawning inlet Since adult Burbot were not sampled for otoliths during the 20 winters of 1997–1998 through 1999–2000, the age distribution creel of the spawning population during these three winters could 10 growth pattern not be estimated directly. It was still possible, however, to in- fer spawner age distribution during these three winters using 0 the interannual changes in spawner length frequency distribu- 01234567891011 tion, the observed population growth pattern, and the directly Age (years) estimated adult age distributions (i.e., using otoliths) of the pre- ceding and following years. The lack of directly aged fish could FIGURE 3. Observed and predicted lengths at age for Burbot sampled from make spawner age determination less precise for these years, the shoreline, a spawning population, and the ice fishery of Columbia Lake. The horizontal line indicates a threshold of about 39 cm at which Burbot begin but since these three winters had the lowest spawner returns the to recruit to the fishery. The curved line plots the estimated von Bertalanffy effect would be small when evaluating relative strength across growth function Lt = 63.5 {1–exp[−0.294(t + 0.018)]}. See text for further all cohorts. explanation. BURBOT COHORT VARIABILITY 1709

TABLE 1. Number of Burbot spawners, mean TL, and percent with tags from previous spawning for 6 years of weir operation on a tributary of Columbia Lake. The data for 1996–1999 are from Arndt and Hutchinson (2000).

Year 1996 1997 1998 1999 2000 2001 Number of spawners >1,365a 1,487 745 86 152 995 Mean TL (cm) 45.4 47.9 50.0 46.0 45.1 42.4 Previously tagged (%) 8.0 41.1 55.8 17.8 5.3

aThis value is an underestimate because some fish evaded upstream capture by burrow- ing under the weir prior to February 6; downstream captures and average residence time suggest that the total run exceeded 1,500 in 1996 (Arndt, unpublished data).

to 1998, the length distributions were unimodal and the mean and modal lengths increased each year. The length distributions became bimodal in 1999 and then trimodal in 2000, with mean length decreasing each year after 1998. The burbot sampled for otoliths during the winters of 1995– 1996, 1996–1997, and 2000–2001 came from cohorts between 1987 and 1999 (Figure 6). The length distributions of angler- harvested Burbot were almost identical to those of tributary- = FIGURE 4. Mean (whiskers SEs) log abundance of four cohorts of Burbot sampled spawners for Burbot greater than the 39-cm thresh- from Columbia Lake as sampled at 12 shoreline transects in 1997, 1998, and 1999. Lines connect the estimates for the 1997 and 1998 cohorts because they old for vulnerability to harvest, indicating similar age structure were sampled at both age 0 and age 1. throughout the lake (Figure 5). A high proportion (82%) of the adult Burbot aged in the winters of 1995–1996 and 1996– 1997 came from the 1991 and 1993 cohorts (Figure 6). Simi- Juvenile Cohort Abundance lar length and age distributions suggest that these two cohorts Large variations in year-class strength at the juvenile life dominated both the fishery and tributary spawning returns from stage were apparent from the transect surveys (Figure 4). The 1996 to 1999. Even during the winter of 2000–2001, 20% of mean (untransformed) abundance per transect for age-0 Burbot aged adult Burbot still came from the 1991 and 1993 cohorts from the 1997, 1998, and 1999 cohorts was 11.0, 2.7, and 1.7, (Figure 6). respectively. Thus, abundance varied by a factor of 6.5 between New modes in the spawner length distributions of 1999 and 1997 (the year of the highest estimate) and 1999 (the year of the 2000 indicate the entrance of two new cohorts to the tributary lowest estimate). The mean abundance per transect for age-1 spawning population (Figure 5). Although otolith samples were Burbot from the 1996, 1997, and 1998 cohorts was 0, 1.5, and not available for these years, these new peaks were assigned 0.5, respectively. to the 1997 cohort (first showing in 1999) and the 1998 cohort The difference in the abundance of age-0 Burbot among years (first showing in 2000) because the modal body length for the new peak in each year (about 275–300 mm) corresponded to the was statistically significant (RMANOVA; F2, 22 = 10.9, P = 0.0005), with no apparent trend in the residuals when plotted average body length at age 2 (Figure 3), which is a common age against sampling date. A Tukey’s HSD test suggested that age-0 at first spawning for Columbia Lake Burbot. Size-at-age data Downloaded by [Department Of Fisheries] at 22:54 25 November 2013 abundance was significantly larger in 1997 than in 1998 or 1999, for the Burbot aged in 2001 (Figure 6) confirmed that the 1997 but there was no statistical difference between 1998 and 1999. and 1998 cohorts were stronger than other cohorts since 1993. The difference in abundance of age-1 Burbot among years was Despite the appearance of these two new cohorts, the spawner also statistically significant (Freidman’s rank test; χ2 = 9.3, count at the weir was still only 152 in 2000 (Table 1). However, df = 2, P = 0.01), with the 1996 cohort, in particular, much in 2001 there was a large increase in spawner returns, and the smaller than the 1997 cohort (Figure 4). spawner length distribution was dominated by one mode at the lower end (Figure 5). Otoliths indicated that this mode was Adult Cohort Abundance and Spawner Returns composed primarily of age-2 individuals from the 1999 cohort The number of spawners captured at the tributary varied by (Figure 6). As the strong 1999 cohort, along with the moderately more than an order of magnitude over the 6-year period of strong 1997 and 1998 cohorts, entered the spawning population, monitoring, decreasing from approximately 1,500 in the first the proportion of spawners with tags from spawning runs in 2 years to a low of 86 in 1999 and then increasing back to previous years dropped correspondingly from 55.8% in 1999 to 995 by 2001 (Table 1). In addition to the changes in spawner 17.8% in 2000 and then to 5.3% in 2001 (Table 1). abundance, there were trends in length frequency distributions In summary, age and length frequency data for adult Burbot that clearly reflected cohort dominance (Figure 5). From 1996 showed strong to very strong recruitment for the 1991, 1993, 1710 TAYLOR AND ARNDT

FIGURE 6. Length-frequency distributions of all Burbot sampled from the spawning tributary (lines) superimposed upon the age and length compositions of Burbot sampled for otoliths from the winter fishery and spawning tributary for Columbia Lake during the winters of 1995–1996, 1996–1997, and 2000–2001. Cohorts from 1987 to 1999 were represented in the aged samples.

Stock–Recruitment Relationship Comparison of stock size (tributary spawner count) with fu- ture abundance at the juvenile and adult stages is possible for four cohorts. At age 0, there is some support for a positive relationship to stock size, as indicated by tributary spawner count compared with shoreline abundance for the 1997, 1998, Downloaded by [Department Of Fisheries] at 22:54 25 November 2013 FIGURE 5. Length-frequency distributions of Burbot from a spawning tribu- and 1999 cohorts (Figure 7; lower panel). However, there were tary (solid lines) and the winter fishery (dashed lines) in Columbia Lake from 1996 to 2001. Only four Burbot were measured for the 2001 creel, and there no clear relationships between initial tributary spawner escape- was a high proportion of released fish. Creel data were not available for 2000. ment and either age-1 abundance or future adult spawner returns (Figure 7; middle and top panels). In fact, the 1999 cohort that returned in very high numbers at age 2 was from the lowest and 1999 cohorts (Table 2). The 1997 and 1998 cohorts were count of spawners in the 6 years of weir operation (Table 2). moderate in abundance, but the 1992, 1994, and 1996 cohorts Adult returns for the 1996–1998 cohorts (Figure 7; top panel) were detected only in very low numbers and the 1995 cohort correspond more closely to the pattern of age-1 abundance (Fig- was not detected at all (Table 2). These large variations in cohort ure 7; middle panel) than to either age-0 abundance or initial abundance corresponded with two large peaks in spawner abun- tributary spawner counts (Figure 7; lower panel). dance as strong cohorts reached maturity (Table 1). Tributary spawner counts decreased as the initial strong cohorts (1991 DISCUSSION and 1993) got older and increased again as new strong cohorts In a previous study on Columbia Lake, Arndt and Hutchin- reached maturity. son (2000) documented a decline in tributary-spawning Burbot BURBOT COHORT VARIABILITY 1711

TABLE 2. Number of Burbot spawners in a Columbia Lake tributary and subsequent returns by cohort. The number returning to spawn was estimated from available age and length data, summing spawning returns from 1996 to 2001. Note that returns of the 1991 and 1992 cohorts are likely underestimated because returns at younger ages were not monitored, whereas those of the 1997– 1999 cohorts are underestimated because returns were not monitored at older ages. See text and footnotes for further explanation.

Number of Subsequent tributary returns Cohort Year/cohort spawners by cohort strength 1991 Unknown 1,250a Very strong 1992 Unknown 150 Weak 1993 Unknown 2,250 Very strong 1994 Unknown 150 Weak 1995 Unknown ∼0 Not detected 1996 1,500 50 Weak 1997 1,487 300 Moderate 1998 745 200 Moderate 1999 86 400b Strong

aReturns after 1996 (i.e., ≥5-year-old fish only). bReturns in 2001 (age-2 fish only because this was the last year of weir operation).

that corresponded with a length distribution shift, suggesting that a strong year-class was decreasing in abundance as it aged. This study extends the period of observation, showing that the spawning run rebounded to nearly 1,000 fish only 2 years after a decline to less than 100. Age data from this study support the contention that fluctuations in spawner abundance (by an order of magnitude) were driven by periodic influxes of strong and weak cohorts. Population decline during the intervening years was a result of mortality in the dominant cohorts. Simi- FIGURE 7. Tributary spawner counts versus relative abundances at age 0, age lar patterns of cohort dominance have been reported for Lake 1, and among subsequent adult returns for the 1996–1999 Burbot cohorts in Columbia Lake. Mean age-0 and age-1 abundances [log10(fish/transect + 1)] Michigan (Bruesewitz 1990), Lake Opeongo in Ontario (Hack- were sampled at the same 12 shoreline transects from 1997 to 1999. Relative ney 1973), and some Alaskan lakes (Lafferty and Vincent-Lang cohort abundance as adult returns was inferred from the changes in spawner 1991; Taube et al. 2000). In Tolsona Lake, Alaska, Taube et al. abundance, length frequency distribution, and age composition over the period (2000) estimated that annual recruits varied from 0% to 200% 1996–2001. of the adult population over a 13-year period. Indices of abundance at age 0 and age 1 allow an evaluation 1991 and 1993 cohorts, even though spawner abundance trends

Downloaded by [Department Of Fisheries] at 22:54 25 November 2013 of when population regulation may have occurred. Variation in (described below) suggest that a large decrease and increase in the recruitment success of fish is commonly attributed to the returns over a period of only 3 years is very unlikely. influence of environmental fluctuations on density-independent One density-independent factor that could determine Bur- mortality during the egg and larval stages (Myers and Cadigan bot recruitment success early in life is the narrow temperature 1993; Frank and Leggett 1994). In this study, differences in the tolerance of Burbot eggs. Burbot egg survival peaks at about abundance of cohorts spawned from similar numbers of adults 3◦C, and early embryos die at temperatures above 6◦C (Taylor provide some support for an influence of density-independent and McPhail 2000). Although the average daily water temper- factors on Columbia Lake Burbot. For example, the 1996 cohort ature measured at the weir (200 m or more downstream of was less abundant than the 1997 cohort at both the age-1 and where most spawning occurred) was frequently at or just be- adult life stages despite there being similar numbers of tributary low this lethal limit, it was no higher in 1996 than in 1997 spawners in 1996 and 1997. Unfortunately, there were no data (Arndt and Hutchinson 2000). Therefore, the effect of tribu- to determine whether this large difference in year-class strength tary temperature on egg survival does not appear to explain was evident by age 0. The age composition of the adult popula- the greater juvenile recruitment success of the 1997 cohort rel- tion provides additional support for density independence. For ative to that of the 1996 cohort. In Lake Erie, Stapanian et al. example, the 1992 cohort was less abundant than the adjacent (2010b) found an association between Burbot recruitment and an 1712 TAYLOR AND ARNDT

index of the number of days with optimal water temperature for competition among juveniles for crevices, thereby contributing spawning and development, if combined in a model with Yellow to unusually high survival during that period. A link between Perch Perca flavescens abundance. However, when considered juvenile survival and available shelter agrees with expectations alone, the temperature index was not supported as a predictor from the behavioral and bioenergetics studies cited above. We of recruitment. suggest further research to test this hypothesis in lakes and Cohort strength does not appear to be closely related to reservoirs with limited interstitial habitat and fluctuating water spawner abundance in Columbia Lake, as has been observed levels. for the West Arm of Kootenay Lake (Ahrens and Korman 2002) The lower relative abundance of age-0 Burbot during 1999 and Lake Erie (Stapanian et al. 2010b). In this study, the lowest than in 1997 and 1998 does not appear to be related to a decrease tributary spawner count was in 1999, yet the 1999 cohort ap- in capture efficiency caused by higher water levels in 1999. Al- peared to be the strongest since 1993 based on 2001 tributary though the amount of rocky habitat suitable for age-0 Burbot returns. This strong return of mature fish combined with the (predominantly cobble) that occurred deeper than the maximum observation that the age-0 abundance of the 1999 cohort was sampling depth of 1 m would have increased at higher lake lev- significantly less than that of the 1997 cohort (and similar to els, the most suitable habitat for age-0 Burbot was always more that of the 1998 cohort) suggests that a second period of re- abundant at shallow depths, occurring predominantly at depths cruitment limitation occurs after the first summer in Columbia less than 1 m during all 3 years. There was also no evidence that Lake. The approximate correspondence between age-1 abun- the increased availability of suitable habitat in 1999 increased dance and spawner returns for the 1996–1998 cohorts provides age-0 growth and thereby allowed the 1999 cohort to grow faster further support for a recruitment bottleneck after age 0. The and migrate offshore prior to index sampling. Contrarily, the av- potential for small-scale control of fish recruitment during the erage length of age-0 Burbot captured during index sampling in juvenile life stage has been noted in other studies (Sissenwine 1999 (6.1 cm) was approximately the same as or less than that 1984; de Lafontaine et al. 1992; Walters and Juanes 1993), and in 1997 (6.2 cm) and 1998 (7.7 cm). there is support for the importance of juvenile stages in deter- Intercohort predation is another factor that could contribute mining recruitment success in Gadids (Houde 1987; Mehl 1989; to recruitment fluctuations. The primary source of fish mortal- Sundby et al. 1989; Bogstad et al. 1994; Tupper and Boutilier ity during the juvenile stage is usually predation (Sissenwine 1995a, 1995b; Hussy et al. 1997). 1984; Houde 1987; Tupper and Boutilier 1995a). Cannibalism Juvenile Burbot in Columbia Lake (Taylor 2001) and else- occurs in Burbot (Chen 1969; Jackson et al. 2008) and other where (Lawler 1963; Boag 1989; Ryder and Pesendorfer 1992; Gadids (Mehl 1989; Bogstad et al. 1994; Scott and Brown 1998). Fischer and Eckman 1997) live in interstitial spaces in rocky The absence of strong cohorts from 1994 to 1998 and the low substrate. Tupper and Boutilier (1995a) demonstrated that the spawner returns in 1999 and 2000 suggest that encounters be- postsettlement survival of Atlantic cod Gadus morhua was high- tween young of the 1999 cohort and older Burbot probably were est in structurally complex habitats. They attributed this to the much less frequent than for the two previous cohorts. In our data increased availability of shelter and decreased efficiency of spanning nine cohorts, two strong cohorts never occurred in se- predators in these habitats. The shelter provided by interstitial quential years. The influence of juvenile predation by other fish spaces in larger rock substrate is likely of similar importance species is also possible. Other piscivorous species present in the to juvenile Burbot, as evidenced by behavioral and physiolog- lake include Bull Trout Salvelinus confluentus, Rainbow Trout ical studies (Fischer and Ohl¨ 2005; Hirsch and Fischer 2008). Oncorhynchus mykiss, and Northern Pikeminnow Ptychocheilus Lack of adequate shelter can substantially increase metabolic oregonensis (Taylor 2001). rates while decreasing foraging efficiency and growth (Fischer Jackson et al. (2008) investigated a similar shallow lake near

Downloaded by [Department Of Fisheries] at 22:54 25 November 2013 2000a, 2000b). the southern edge of the Burbot’s range in New Yorkusing a data Fischer and Ohl¨ (2005) found significant conspecific com- set spanning four decades. They found a high degree of variation petition for larger shelters as lake level decreased in autumn in Burbot catches at the larval and adult stages. Adult catches in Lake Constance. Likewise, in Columbia Lake the potential were negatively correlated with August water temperature, and for competition among juvenile Burbot for interstitial spaces is both bioenergetic analyses and field measurements suggested a greatest from late fall to early spring, when water levels are at period of reduced feeding and energy loss at temperatures over their lowest, with most of the rocky habitat out of water. In this 20◦C. Higher summer water temperatures, therefore, were im- study, the water level in Columbia Lake was periodically mea- plicated in population decline, although the exact mechanisms sured from July 1997 to November 1999. The water level was were not known. We did not have sufficient data to investigate a about 33 cm higher in late August 1999 than it was in late Au- summer temperature relationship in Columbia Lake; however, gust 1997 and 1998. In fact, it was not until late November 1999 temperatures exceeding 20◦C were recorded during the study that the lake level dropped to that measured in late August of the period and could have affected recruitment success, especially previous 2 years (Taylor, unpublished data). These higher water in combination with low water levels. Summers with high water levels during the fall and winter of 1999–2000 submerged more temperatures may be associated with drought periods that cause coarse substrate in the littoral zone, which may have reduced a decline in water levels. BURBOT COHORT VARIABILITY 1713

The shoreline density of age-0 Burbot in Columbia Lake Columbia Lake, the population tends to have large interannual (mean of 400 fish/km in 1997 to 62 fish/km in 1999) is high fluctuations in abundance and cohort survival. If stock assess- compared with that of surveyed lakes in eastern North America. ment must be conducted over short time periods, age composi- Carl (2000) used similar backpack electrofishing methods to tion sampling should be included. Populations with large fluc- survey age-0 Burbot along randomly chosen stretches of shore- tuations in recruitment (indicated by cohort dominance in the line in ten southern Ontario lakes. Shoreline abundance aver- age composition) are likely to be less stable. Studies to monitor aged about 100 fish/km in five lakes without Ciscoes Coregonus Burbot recruitment variation should include an index of abun- artedi and less than 1 fish/km in five lakes with Ciscoes. This dance from after the first growing season. In Columbia Lake, pattern suggested a demographic bottleneck during the pelagic recruitment of cohorts to the spawning population appeared to larval life stage of Burbot due to predation by Ciscoes. None be more closely associated with juvenile abundance at age 1 of the lakes sampled in the Ontario study had densities as high than at age 0, suggesting the importance of regulation during as 400 fish/km. Thus, the larval period for Burbot in Columbia the juvenile life stage. The results from this study also hint that Lake may not be as severe in limiting recruitment as in some key factors regulating the survival of juvenile Burbot include lakes with different fish communities. Kokanee Oncorhynchus the availability of crevices to provide shelter habitat (which can nerka, a pelagic planktivore, have been captured in Columbia vary with water levels) as well as intercohort predation. Lake during their fall spawning season, but they are not abun- dant and may not be present during the larval period. A potential ACKNOWLEDGMENTS predator present during the larval period is the Peamouth My- This study was funded by the Fish and Wildlife Compensa- locheilus caurinus. In two other southeastern British Columbia tion Program–Columbia Basin, a joint initiative of BC Hydro, lakes, shoreline electrofishing in August and early September the Province of British Columbia, and Fisheries and Oceans has failed to capture any juvenile Burbot (Neufeld and Spence Canada that was created to help sustain and enhance fish and 2009; Westslope Fisheries 2009), even though substantial num- wildlife populations in the Canadian portion of the Columbia bers of adults are captured by other methods. It is possible that River basin. Provincial employees Bill Westover and Jay Ham- juveniles occupy depths greater than 1 m in lakes in which mond initiated the study. Don McPhail of the University of interstitial habitat is available in deeper water. British Columbia provided knowledge and guidance through- To some extent, our estimates of spawning stock abundance out. Martin Stapanian, Steve Martell, Vaughn Paragamian, and assume that Burbot return to their natal locations to spawn. two other reviewers provided constructive comments on earlier Although we do not have radiotelemetry data to demonstrate drafts. Field assistance was provided by Joyce Hutchinson, Pe- spawning site fidelity for the studied population, a winter creel ter Mylechreest, Harald Manson, Tara Fleming, Mark Phillpotts, survey conducted near the north end of the lake between 1995 Larry Taylor, Derek Whyte, Mark Stevenson, Brian Reesor, Ed and 2001 showed a very low proportion of tributary-tagged Russel, Kevin Heidt, and Colin Spence. fish; only four mature Burbot with tags from previous tributary spawning were found in a sample of 389 fish in spawning con- dition (Arndt, unpublished data). Thus, there is no evidence of REFERENCES a large-scale shift in spawning location. A fairly high degree of Ahrens, R., and J. Korman. 2002. What happened to the West Arm Burbot stock site fidelity for spawning was also suggested by radiotelemetry in Kootenay Lake? Use of an age structured population model to determine 1 the possible causes for recruitment failure. Report to the British Columbia and mark–recapture studies in the Kootenai River (Paragamian Ministry of Water, Land, and Air Protection, the Habitat Conservation Trust and Wakkinen 2008). Further, our interpretations of survival to Fund, and the Bonneville Power Administration, Nelson. spawning are based on age distribution as well as the tributary Arndt, S. K. A., and J. Hutchinson. 2000. Characteristics of Burbot spawning in

Downloaded by [Department Of Fisheries] at 22:54 25 November 2013 counts. The correspondence between the tributary length dis- a tributary of Columbia Lake, British Columbia, over a four year period. Pages tributions and the age distributions estimated largely from the 48–60 in V.L. Paragamian and D. W. Willis, editors. Burbot: biology, ecology, and management. American Fisheries Society, Fisheries Management Section creel survey, as well as the similarity in the length distributions Publication 1, Spokane, Washington. of tributary spawners and creeled fish, support the notion that the British Columbia Conservation Data Centre. 2012. BC species and ecosystems changes in abundance observed at the tributary reflect changes explorer. British Columbia Ministry of Environment, Victoria. Available: in adult cohort abundance occurring in the lake as a whole. If http://a100gov.bc.ca/pub/eswp./ (August 2012). this is the case, comparisons of adult cohorts with juvenile abun- Beddington, J. R., and R. M. May. 1977. Harvesting natural populations in a randomly fluctuating environment. Science 197:463–465. dance indices (which are likely to be influenced by spawning in Boag, T. D. 1989. Growth and fecundity of Burbot, Lota lota L. in two Alberta locations other than the tributary) should also be reasonable. Lakes. Master’s thesis. University of Alberta, Edmonton. Burbot fisheries managers should be aware that single-year Bogstad, B., G. R. Lilly, S. Mehl, O. K. Palsson, and G. Stefansson. 1994. or periodic abundance estimates may be misleading if, as in Cannibalism and year-class strength in (Gadus morphua) in Arcto-boreal ecosystems (Barents Sea, Iceland, and eastern Newfoundland). ICES Marine Science Symposium 198:576–599. 1The term “Kootenai” represents the U.S. spelling, the term “Kootenay” the Bruesewitz, R. E. 1990. Population dynamics and movement of Burbot (Lota Canadian spelling. In this article the former is used when it refers to locations lota) in western Lake Michigan and Green Bay. Master’s thesis. University or entities in the United States, the latter when it refers to locations in Canada. of Wisconsin, Stevens Point. 1714 TAYLOR AND ARNDT

Carl, L. M. 2000. Effects of two planktivores on the fish communities in Pre- Lafferty, B., and D. Vincent-Lang. 1991. Status of lake Burbot stocks in the cambrian Shield lakes in southern Ontario. Pages 81–89 in V. L. Paragamian upper Copper/upper Susitna management area. Alaska Department of Fish and D. W. Willis, editors. Burbot: biology, ecology, and management. Ameri- and Game, Sport Fish Division, Anchorage. can Fisheries Society, Fisheries Management Section Publication 1, Spokane, Lawler, G. H. 1963. The biology and taxonomy of the Burbot, Lota lota,in Washington. Heming Lake, Manitoba. Journal of the Fisheries Research Board of Canada Chen, L. C. 1969. The biology and taxonomy of the Burbot, Lota lota leptura, 29:417–433. in interior Alaska. University of Alaska, Biological Paper 11, Fairbanks. McPhail, J. D. 1997. A review of Burbot (Lota lota) life history and habitat Cohen, S. J., K. A. Miller, A. F. Hamlet, and W. Avis. 2000. Climate change use in relation to compensation and improvement opportunities. Canadian and resource management in the Columbia River basin. Water International Manuscript Report of Fisheries and Aquatic Sciences 2397. 25:253–272. McPhail, J. D., and V. L. Paragamian. 2000. Burbot biology and life history. de Lafontaine, Y., T. Lambert, G. R. Lilly, W. D. McKone, and R. J. Miller. Pages 11–21 in V. L. Paragamian and D. W. Willis, editors. Burbot: biology, 1992. Juvenile life stages: the missing link in fisheries research. Canadian ecology, and management. American Fisheries Society, Fisheries Manage- Technical Report of Fisheries and Aquatic Sciences 1890. ment Section Publication 1, Spokane, Washington. Fischer, P. 2000a. An experimental test of metabolic and behavioural responses Mehl, S. 1989. The northeast Arctic Cod stock’s consumption of commer- of benthic fish species to different types of substrate. Canadian Journal of cially exploited prey species in 1984–1986. Rapports et Proces-Verbaux Fisheries and Aquatic Sciences 57:2336–2344. des Reunions Conseil International pour l’Exploration de la Mer 188: Fischer, P. 2000b. Test of competitive interactions for space between two ben- 185–205. thic fish species, Burbot Lota lota, and Stone Loach Barbatula barbatula. Moreau, J. 1987. Mathematical and biological expression of fish growth in Environmental Biology of Fishes 58:439–446. fishes: recent trends and further developments. Pages 81–113 in C. Summer- Fischer, P., and R. Eckmann. 1997. Seasonal changes in fish abundance, biomass felt and G. E. Hall, editors. The age and growth of fish. Iowa State University and species richness are characteristic for the littoral zone of a large Eu- Press, Ames. ropean lake, Lake Constance, Germany. Archives of Hydrobiology 139: Myers, R. A. 2001. Stock recruitment: generalizations about maximum re- 433–448. productive rate, density dependence, and variability using meta-analytic ap- Fischer, P., and U. Ohl.¨ 2005. Effects of water-level fluctuations on the lit- proaches. ICES Journal of Marine Science 58:937–951. toral benthic fish community in lakes: a mesocosm experiment. Behavioural Myers, R. A., and N. G. Cadigan. 1993. Is juvenile natural mortality in marine Ecology 16:741–746. demersal fish variable? Canadian Journal of Fisheries and Aquatic Science Fisher, S. J. 2000. Early life history observations of Burbot utilizing two 50:1591–1598. Missouri River backwaters. Pages 96–101 in V. L. Paragamian and D. Nelson, J. S. 1994. Fishes of the world, 3rd edition. Wiley, New York. W. Willis, editors. Burbot: biology, ecology, and management. American Neter, J., M. H. Kutner, C. J. Nachsheim, and W. Wasserman. 1996. Applied Fisheries Society, Fisheries Management Section Publication 1, Spokane, linear statistical models, 4th edition. Irwin, Chicago. Washington. Neufeld, M. D., and C. R. Spence. 2009. Moyie Lake Burbot program 2008– Frank, K. T., and W. C. Leggett. 1994. Fisheries ecology in the context of eco- 09: sport fishery, juvenile sampling, and conservation aquaculture spawning logical and evolutionary theory. Annual Review of Ecology and Systematics surveys. British Columbia Ministry of Environment, Nelson. 25:401–422. Paragamian, V. L., and V. D. Wakkinen. 2008. Seasonal movement of Burbot Gjoesaeter, H. 1995. Pelagic fish and the ecological impact of the modern fishing in relation to temperature and discharge in the Kootenai River, Idaho, USA, industry in the Barents Sea. Arctic 48:267–278. and British Columbia, Canada. Pages 55–77 in V. L. Paragamian and D. Hackney, P. A. 1973. Ecology of Burbot (Lota lota) with special reference to its H. Bennett, editors. Burbot: ecology, management, and culture. American role in the Lake Opeongo fish community. Doctoral dissertation. University Fisheries Society, Symposium 59, Bethesda, Maryland. of Toronto, Toronto. Paragamian, V. L., V. Whitman, J. Hammond, and H. Andrusak. 2000. Col- Hirsch, P. E., and P. Fischer. 2008. Interactions between native Burbot (Lota lapse of Burbot fisheries in Kootenai River, Idaho, USA, and Kootenay lota) and the invasive Spinycheek Crayfish (Orconectes limosus) in a large Lake, British Columbia, Canada. Pages 155–164 in V. L. Paragamian and European lake. Canadian Journal of Fisheries and Aquatic Sciences 65:2636– D. W. Willis, editors. Burbot: biology, ecology, and management. American 2643. Fisheries Society, Fisheries Management Section Publication 1, Spokane, Houde, E. D. 1987. Fish early life dynamics and recruitment variability. Pages Washington. 17–29 in R. D. Hoyt, editor. 10th Annual larval fish conference. American Pepin, P., and R. A. Myers. 1991. Significance of egg and larval size to recruit- Fisheries Society, Symposium 2, Bethesda, Maryland. ment variability of temperate marine fish. Canadian Journal of Fisheries and Houde, E. D. 1994. Differences between marine and freshwater fish lar- Aquatic Sciences 48:1820–1828. Downloaded by [Department Of Fisheries] at 22:54 25 November 2013 vae: implications for recruitment. ICES Journal of Marine Science 51: Rickman, S., N. Dulvy, S. Jennings, and J. Reynolds. 2000. Recruitment varia- 91–97. tion related to fecundity in marine fishes. Canadian Journal of Fisheries and Hussy, K., M. A. St. John, and U. Bottcher. 1997. Food resource utilization Aquatic Sciences 54:950–968. by juvenile Baltic Cod Gadus morhua: a mechanism potentially influencing Ryder, R. A., and J. Pesendorfer. 1992. Food, growth, habitat, and community recruitment success at the juvenile stage? Marine Ecology Progress Series interactions of young-of-the-year Burbot, Lota lota (L), in a Precambrian 155:199–208. Shield lake. Hydrobiologia 243/244:211–227. Ireland, S. C., and P. N. Perry. 2008. Burbot restoration in the Kootenai River Scott, M., and J. A. Brown. 1998. Diel foraging cycles and interactions basin: using agency, tribal, and community collaboration to develop and among juvenile Atlantic Cod (Gadus morhua) at a nearshore site in New- implement a conservation strategy. Pages 251–256 in V. L. Paragamian and foundland. Canadian Journal of Fisheries and Aquatic Science 55:1307– D. H. Bennett, editors. Burbot: ecology, management, and culture. American 1316. Fisheries Society, Symposium 59, Bethesda, Maryland. Sissenwine, M. P. 1984. Why do fish populations vary? Pages 59–94 in Jackson, J. R., A. J. VanDeValk, J. L. Forney, B. F. Lantry, T. E. Brooking, and R. M. May, editor. Exploitation of marine communities. Springer-Verlag, L. G. Rudstam. 2008. Long-term trends in Burbot abundance in Oneida Lake, Berlin. New York: life at the southern edge of the range in an era of climate change. Smolders, A. J., G. van der Velde, J. G. Roelofs, and M. Guerrero Hiza. 2000. Pages 131–152 in V. L. Paragamian and D. H. Bennett, editors. Burbot: El Nino caused collapse of the Sabala Fishery (Prochilodus lineatus, Pisces: ecology, management, and culture. American Fisheries Society, Symposium Prochilodontidae) in a South American River. Naturwissenschaften 87:30– 59, Bethesda, Maryland. 32. BURBOT COHORT VARIABILITY 1715

Stapanian, M. A., V. L. Paragamian, C. P. Madenjian, J. R. Jackson, J. Pages 30–37 in V. L. Paragamian and D. W. Willis, editors. Burbut: biology, Lappalainen, M. J. Evenson, and M. D. Neufeld. 2010a. Worldwide sta- ecology, and management. American Fisheries Society, Fisheries Manage- tus of Burbot and conservation measures. Fish and Fisheries 11:34– ment Section Publication 1, Spokane, Washington. 56. Tupper, M., and R. G. Boutilier. 1995a. Effects of habitat on settlement, growth, Stapanian, M. A., L. D. Witzel, and A. Cook. 2010b. Recruitment of Burbot and post settlement survival of Atlantic Cod (Gadus morhua). Canadian (Lota lota) in Lake Erie: an empirical modeling approach. Ecology of Fresh- Journal of Fisheries and Aquatic Science 52:1834–1841. water Fish 19:326–337. Tupper, M., and R. G. Boutilier. 1995b. Size and priority at settlement deter- Sundby, S., H. Bjorke, A. V. Soldal, and S. Olsen. 1989. Mortality rates during mine growth and competitive success of newly settled Atlantic Cod. Marine the early life stages and year-class strength of northeast Arctic Cod. Rapports Ecology Progress Series 118:295–300. et Proces-Verbaux des Reunions Conseil International pour l’Exploration de Walters W. J., and F. Juanes. 1993. Recruitment limitation as a consequence la Mer 191:351–358. of natural selection for use of restricted feeding habitats and predation risk Taube, T. T., L. J. Perry-Pake, and D. R. Bernard. 2000. Stock assessment and taking by juvenile fishes. Canadian Journal of Fisheries and Aquatic Sciences biological characteristics of Burbot in Tolsona Lake, 1999 and Lake Louise, 50:2058–2070. 1995–1996, 1999. Alaska Department of Fish and Game, Fisheries Data Westslope Fisheries. 2009. Duncan Reservoir Burbot monitoring annual report Series 00-40, Anchorage. 2008–09 (interim report 1). British Columbia Hydro, Report DDMMON-11, Taylor, J. L. 2001. The early life history and ecology of Columbia Lake Burbot. Vancouver. Master’s thesis. University of British Columbia, Vancouver. Worthington, T., P. Kemp, and P. E. Osbourne. 2011. Factors affecting the Taylor, J. L., and J. D. McPhail. 2000. Temperature, development and behaviour population viability of the Burbot, Lota lota. Fisheries Management and in the early life history of Burbot from Columbia Lake, British Columbia. Ecology 18:322–332. Downloaded by [Department Of Fisheries] at 22:54 25 November 2013 This article was downloaded by: [Department Of Fisheries] On: 25 November 2013, At: 22:55 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Transactions of the American Fisheries Society Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/utaf20 Feeding Ecology of Pelagic Larval Burbot in Northern Lake Huron, Michigan Ellen M. George a , Edward F. Roseman a , Bruce M. Davis a & Timothy P. O’Brien a a U.S. Geological Survey, Great Lakes Science Center , 1451 Green Road, Ann Arbor , Michigan , 48105 Published online: 20 Nov 2013.

To cite this article: Ellen M. George , Edward F. Roseman , Bruce M. Davis & Timothy P. O’Brien (2013) Feeding Ecology of Pelagic Larval Burbot in Northern Lake Huron, Michigan, Transactions of the American Fisheries Society, 142:6, 1716-1723, DOI: 10.1080/00028487.2013.788561 To link to this article: http://dx.doi.org/10.1080/00028487.2013.788561

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions Transactions of the American Fisheries Society 142:1716–1723, 2013 American Fisheries Society 2013 ISSN: 0002-8487 print / 1548-8659 online DOI: 10.1080/00028487.2013.788561

SPECIAL SECTION: BURBOT

Feeding Ecology of Pelagic Larval Burbot in Northern Lake Huron, Michigan

Ellen M. George, Edward F. Roseman,* Bruce M. Davis, and Timothy P. O’Brien U.S. Geological Survey, Great Lakes Science Center, 1451 Green Road, Ann Arbor, Michigan 48105

Abstract Burbot Lota lota are a key demersal piscivore across the Laurentian Great Lakes whose populations have declined by about 90% in recent decades. Larval Burbot typically hatch in the early spring and rely on abundant crustacean zooplankton prey. We examined the stomach contents of larval Burbot from inshore (≤15 m) and offshore sites (37 and 91 m) in northern Lake Huron, Michigan. Concurrent zooplankton vertical tows at the same sites showed that the prey community was dominated by calanoid copepods, dreissenid mussel veligers, and rotifers. Burbot consumed mostly cyclopoid copepods, followed by copepod nauplii and calanoid copepods. Chesson’s index of selectivity was calculated and compared among sites and months for individual Burbot. According to this index, larval Burbot exhibited positive selection for cyclopoid copepods and copepod nauplii and negative selection for calanoid copepods, cladocerans, rotifers, and dreissenid veligers. This selectivity was consistent across sites and throughout the sampling period. Burbot displayed little variation in their prey preferences during the larval stage, which suggests that the recent shifts in zooplankton abundance due to the invasion of the predatory zooplankter Bythotrephes longimanus and competition from invasive Rainbow Smelt Osmerus mordax could negatively impact larval Burbot populations.

Burbot Lota lota are a key demersal piscivore across the small rotifers before targeting copepod nauplii (Vatcha 1990; Laurentian Great Lakes that were once abundant in the offshore Ghan and Sprules 1993; McPhail and Paragamian 2000). How- waters of Lake Huron. Since the mid-1960s Burbot populations ever, several studies have found that larval Burbot avoid ro- have declined due to overfishing and predation by invasive sea tifers altogether and instead begin feeding immediately on cope- lampreys Petromyzon marinus (Berst and Spangler 1972), and pod nauplii or adults (Ryder and Pesendorfer 1992; Wang and the abundance of the species is currently estimated to be about Appenzeller 1998). This may be because nauplii are more nu- 10% of its former value (Bence and Mohr 2008; Stapanian tritious than rotifers: Vught et al. (2008) found that Burbot et al. 2008). In Lake Huron, Burbot are known to spawn at that were fed crustaceans for the first few weeks of growth Downloaded by [Department Of Fisheries] at 22:55 25 November 2013 nearshore locations and in tributaries. The larvae are pelagic and exhibited fewer skeletal deformities than Burbot fed rotifers. can be collected from April through late summer (O’Gorman These first few days of exogenous feeding are critical to Bur- 1983; Roseman and O’Brien 2013). Despite the importance of bot survival, as Burbot require large amounts of nutritious prey early life history ecology to fisheries recruitment (Cushing 1975; (Jensen et al. 2011) and must quickly locate a reliable patch Houde 1987, 1989), information about the larval stage of Burbot of prey and feed before they starve (Miller et al. 1988). They in the Great Lakes is sparse. begin consuming cyclopoid copepods as soon as they are able, Burbot rely on abundant planktonic crustacean prey dur- and several studies have noted that larvae will select for the ing the early stages of larval development. After hatching in largest prey size they can ingest (Link 1996; Hardy et al. 2008). early spring, Burbot feed endogenously before migrating to the However, Ghan and Sprules (1993) observed that while this surface 11–23 d posthatch (Fischer 1999). They then shift to is true for smaller Burbot, more developed Burbot larvae ap- exogenous feeding, perhaps beginning with phytoplankton or pear to lose their preference for larger prey and will continue to

*Corresponding author: [email protected] Received September 7, 2012; accepted March 14, 2013

1716 FEEDING ECOLOGY OF PELAGIC LARVAL BURBOT 1717

consume prey of all sizes opportunistically. They also found that Burbot do not attempt to feed on spined rotifers and cladocer- ans. Instead, Burbot exhibit a strong selectivity for cyclopoid copepods, perhaps because calanoid copepods have a stronger escape response (Confer and Blades 1975; Wright and O’Brien 1984) and daphnids are not as visible or require greater handling time. In general, Burbot avoid ingesting any prey that is difficult to capture or requires time to handle and consume (Confer and Lake 1987; Confer and O’Bryan 1989). Burbot prey selectivity appears to be consistent across depths and habitats (Ghan and Sprules 1993). The zooplankton community in Lake Huron has changed substantially in the last decade. Barbiero et al. (2009) noted a 90% decrease in nonpredatory cladoceran and cyclopoid cope- pod biomass in Lake Huron since 2002. This decline may be due to increased predation pressure by the invasive planktivore Bythotrephes longimanus (Bunnell et al. 2011) in conjunction with a decline in phytoplankton biomass following the introduc- tion of dreissenid mussels (Nalepa et al. 2007; Barbiero et al. FIGURE 1. Map showing the locations of the inshore (1–15 m deep [circles]) 2009). Exotic species that become as pervasive as B. longimanus and offshore (37 m deep [stars] and 91 m deep [triangles]) larval fish sampling can drastically alter the availability of resources in an ecosys- stations in northern Lake Huron during 2007. tem (Crooks 2002), possibly influencing the feeding behavior of planktivores such as larval Burbot. In addition, invasive fish species such as Rainbow Smelt Osmerus mordax that prey on in 2007. The shoreline and bottom substrates differed between zooplankton during their larval stage can directly compete with the inshore areas at Hammond Bay and De Tour: Hammond Bay native fish for food resources during early life history stages. was dominated by sand with a gradually sloping depth contour, In 2007 we took part in a larger 5-month project examining while De Tour was dominated by large boulders and cobbles the roles of invasive invertebrates in the recent declines in Lake except at tributary mouths, where fluvial sand deposits were Huron prey fish populations (Savino 2009). Larval Burbot were found. collected at every site throughout the sampling period and were Zooplankton sampling.—Vertical hauls for crustacean zoo- thus chosen as our focal species for the study presented here. plankton were made at dusk or at night. Inshore sites were sam- Abundance and distribution data for larval Burbot showed that pled with a conical 0.5-m-diameter, 153-µm-mesh zooplankton while some Burbot larvae were collected at all of the sites sam- net. The net was lowered to 1 m above the lake bottom before be- pled in northern Lake Huron, their densities were highest at sites ing retrieved at approximately 0.5 m/s. Offshore sites were sam- near De Tour, Michigan (Roseman and O’Brien 2013). Addi- pled with a conical 0.5-m-diameter, 62-µm-mesh net lowered to tionally, newly hatched yolk sac larvae persisted in collections 20 m below the surface, then retrieved at approximately 0.5 m/s. through July. Our objective was to examine the feeding ecology Although two different mesh sizes were used, the zooplankton of larval Burbot in the wake of the drastic shift in zooplankton species of interest to this study have an approximate minimum community composition in Lake Huron and to consider the im- size of 0.2 mm and therefore should be adequately sampled by

Downloaded by [Department Of Fisheries] at 22:55 25 November 2013 pact of invasive species on Burbot feeding during their early life both sizes of mesh. The exception is rotifers, which are regularly history. Selectivity indices and diet analysis metrics were com- smaller than 0.2 mm and could therefore be inadequately sam- pared among sites and months to determine trends in feeding pled at inshore sites. The volume of water sampled by the net behavior. As larval Burbot’s preference for cyclopoid copepods was estimated by means of a flowmeter mounted in the mouth of has been shown to be constant in spite of changing prey den- the net. Two replicate tows were done at each site. Inshore sites sities (Ghan and Sprules 1993), we expected Burbot to show a were sampled from April to June, whereas offshore sites were preference for cyclopoid copepods in Lake Huron despite the sampled from May to July. Samples were bathed for 2–5 min shift in the composition of the zooplankton community. with antacid to narcotize the organisms before they were fixed in a 5% solution of formalin. The zooplankton were processed and identified following protocols described by Barbiero and METHODS Tuchman (2004). A more detailed description of the processing Study area.—This study was conducted at inshore (depth = of these samples can be found in Bunnell et al. (2011). 1.5–15 m) and offshore habitats (depth = 37 and 91 m) at Larval Burbot sampling.—Samples of larval fish were col- two ports (De Tour and Hammond Bay) in Michigan waters of lected from inshore sites at De Tour and Hammond Bay (Fig- northern Lake Huron (Figure 1) during the spring and summer ure 1) beginning in mid-April as soon as ice-out occurred and 1718 GEORGE ET AL.

the waters were safely navigable with a small vessel (7 m hull samples described above. Chesson’s α is defined as length). Daytime collections were made at both the Hammond   Bay and De Tour inshore sites biweekly through the end of June. ri α =  pi , At inshore sites, collections of pelagic larval fish were made us- i n ri 2 inga2-m framed neuston net fitted with 500-µm-mesh netting. 1 pi The neuston net was towed in the upper 2 m of the water column α = = for approximately 5 min at a speed of 7 km/h. Offshore collec- where i the selectivity coefficient of prey type i, ri the = tions were made at night from a large research vessel (30 m hull proportion of prey type i in the diet, pi the proportion of prey = length) once per month during May, June, and July 2007 at two type i in the environment, and n the number of available prey α > sites (37 and 91 m depths) at Hammond Bay and De Tour (Fig- types. Positive selection is indicated when i (1/n), neutral α = α < ure 1). Night collections began one-half hour after sunset and selection when i (1/n), and negative selection when i were concluded by midnight. Because the neuston net was not (1/n). Student’s t-tests were used to compare selectivity between suitable for use with the large research vessel, offshore larval prey types within months. fish samples were collected using a 1.0-m-diameter, 656-µm- mesh conical net towed off the port side for 5–10 min at an RESULTS approximate speed of 5 km/h. Between two and four replicate samples were collected on each sampling night. For all larval Zooplankton Community Composition fish collection gear, the volume of water sampled by the net was Macrozooplankton were defined as cyclopoid and calanoid estimated by means of a flowmeter mounted in the mouth of copepods, cladocerans, and miscellaneous zooplankton, includ- the net. The neuston nets typically sampled 500 m3 of water in ing B. longimanus; microzooplankton were defined as mussel a 5-min tow, whereas the conical nets strained approximately veligers and copepod nauplii. Zooplankton abundance gener- 750 m3 in 10 min. All larval fish samples were preserved and ally increased over time at both inshore and offshore sites. stored in a 95% solution of ethanol. Very low macrozooplankton abundances (305 individuals/m3) Diet analysis.—In the laboratory, preserved ichthyoplankton were observed at inshore sites in April and at offshore sites samples were sorted and identified to species according to Auer in May, the first sampling months for those sites, respectively. (1982). The TLs of the larval Burbot were measured to the We observed higher abundances of both macro- and micro- nearest 0.01 mm using a digital compound microscope imag- zooplankton at inshore sites than at offshore sites in May and ing system. The stomachs were then excised from the anterior June (Figure 2). Cyclopoid copepods and cladocerans were opening of the esophagus to the posterior end of the stomach. more abundant inshore, while calanoid copepods and nauplii The prey items found were measured to the nearest 0.01 mm and were more abundant offshore. The macrozooplankton commu- categorized as calanoid copepods, cyclopoid copepods, copepod nity was dominated by miscellaneous species at inshore sites nauplii, cladocerans, rotifers, or dreissenid veligers. Identifica- (maximum of 7,144 individuals/m3 in June, which includes B. tion of zooplankton followed Balcer et al. (1984). The lengths longimanus and other zooplankton species not found in Burbot of the prey items were defined as the distance from the top of the diets), followed by cyclopoid copepods (maximum of 3,853 head to the tip of the tail for nauplii, cladocerans, and rotifers; individuals/m3 in June). Calanoid copepods were the most abun- the distance from the front of the head to the base of the caudal dant group at offshore sites in all sampling months (maximum of rami for copepods (Culver et al. 1985); and greatest diameter 6,664 individuals/m3 in July), followed by cladocerans (maxi- for dreissenid veligers. mum of 2,575 individuals/m3 in July). Cyclopoid copepods com- Data analysis.—Because we had small sample sizes for some prised a relatively small proportion of the entire zooplankton 3 Downloaded by [Department Of Fisheries] at 22:55 25 November 2013 locations, we pooled all diet data in our analyses of the overall community (maximum of 1,817 individuals/m [1.8%] in June). trends in diet. We calculated the mean number of prey items Dreissenid veligers dominated the microzooplankton commu- consumed and the percent frequency of occurrence by number nity at inshore sites in June (52,929 individuals/m3) and off- for each prey type for all samples combined (Bowen 1996). shore sites in July (33,282 individuals/m3). Copepod nauplii We used Spearman’s ranked correlation (Rs) to assess the re- were the dominant microzooplankton species during May and lationships between the length of larval Burbot and the length June at offshore sites (19,184 individuals/m3 in May, 25,653 and number of prey consumed. Differences in the numbers of individuals/m3 in June) and during April and May at inshore prey consumed, the frequency of occurrence among prey types, sites (1,776 individuals/m3 in April, 10,999 individuals/m3 in and the prevalence of empty stomachs were analyzed using a May). In addition, copepod nauplii were more abundant than all one-way analysis of variance (ANOVA; SAS 1990). macrozooplankton types combined at offshore sites (maximum Chesson’s (1983) index of selectivity (α), as defined by of 25,653 individuals/m3 in June, compared with a total macro- Vanderploeg and Scavia (1979), was calculated for the indi- zooplankton abundance of 8,032 individuals/m3). Rotifers were vidual Burbot whose stomachs contained prey items. The nu- also found in extremely high abundance at offshore sites, com- merical proportions of zooplankton prey abundance at each site prising over 80% of the entire zooplankton community in July per individual month were obtained from our 2007 zooplankton (maximum of 261,629 individuals/m3). However, they are not FEEDING ECOLOGY OF PELAGIC LARVAL BURBOT 1719

FIGURE 2. Abundance of micro- and macrozooplankton at inshore and offshore sites in Lake Huron from April to July 2007. The inshore sites were sampled from April to June and the offshore sites from May to July. Note the differences in the scales of the y-axes.

included in further zooplankton community analyses or the se- lectivity index for two reasons: first, rotifers were most likely inadequately sampled by the larger mesh net at inshore sites, re- sulting in a perceived extreme abundance of rotifers at offshore sites as opposed to their perceived near absence at inshore sites; Downloaded by [Department Of Fisheries] at 22:55 25 November 2013 second, the results of the larval Burbot diet analysis outlined below suggest that Burbot do not regularly consume rotifers in Lake Huron (only one Burbot was observed to have eaten a single rotifer).

Diet Composition On average, larval Burbot consumed more cyclopoid cope- pods than other prey (ANOVA; P = 0.007), with such copepods occurring in 43% of all Burbot stomachs (Figure 3). Copepod nauplii were the second most consumed zooplankton, occurring FIGURE 3. Frequency of occurrence (bars) and mean number per stomach in 26% of stomachs, followed by calanoid copepods (21%). (diamonds) of each prey type consumed by larval Burbot. The frequency of Dreissenid veligers and rotifers occurred in less than 10% of occurrence includes fish whose stomachs were empty; the mean number per stomachs examined. Cladocerans and rotifers were extremely stomach includes only fish whose stomachs contained prey items. 1720 GEORGE ET AL.

TABLE 1. Collection week, site, depth (D), sample size (N), mean TL, number of empty stomachs (Emp.), and mean numbers of prey for the prey types found in pelagic larval Burbot diets in northern Lake Huron during 2007. The values in parentheses are SDs. Prey types are as follows: Clad = cladocerans, Cal = calanoid copepods, Cyc = cyclopoid copepods, Rot = rotifers, Naup = nauplii, and Vel = veligers.

Prey type Week Site D (m) N TL (mm) Emp. Clad Cal Cyc Rot Naup Vel Apr 17 De Tour <15 1 4.0 1 0 0 0 0 0 0 May15 DeTour <15 30 4.7 (0.5) 27 0 0 0.03 (0.18) 0 0.10 (0.40) 0 May 24 Ham. Bay 37 1 6.7 0 0 0 0 0 8.00 0 De Tour 37 1 6.3 0 0 0 3.00 1.00 0 0 Jun 14 De Tour <15 5 6.3 (1.3) 1 0 0 2.00 (1.87) 0 0 0 Jun 26 De Tour <15 2 6.9 (1.8) 1 0 0 0 0 1.00 (1.41) 0 Ham. Bay 37 3 6.9 (1.2) 1 0 0 0.67 (1.15) 0 0.33 (0.58) 0 Ham. Bay 91 35 8.8 (2.2) 14 0 1.94 (3.92) 3.80 (10.76) 0 1.14 (3.14) 0 De Tour 37 10 7.9 (1.7) 6 0 0 4.50 (6.49) 0 1.60 (3.06) 0 De Tour 91 9 9.2 (1.7) 1 0 10.89 (10.95) 7.33 (7.34) 0 2.22 (4.54) 0 Jul 23 Ham. Bay 37 1 8.5 0 0 0 9.00 0 17.00 0 Ham. Bay 91 8 10.4 (2.6) 0 0.13 (0.35) 0.88 (1.81) 48.75 (39.46) 0 3.88 (6.88) 5.00 (12.20) De Tour 37 4 9.2 (2.0) 1 0 0.25 (0.50) 9.50 (10.66) 0 0 5.00 (8.12) De Tour 91 14 9.6 (2.1) 0 0 1.4 (1.9) 31.71 (31.96) 0 10.36 (12.10) 2.57 (3.48)

rare: only a single cladoceran (Bosmina sp.) and a single rotifer Selectivity Index (Keratella sp.) were consumed, occurring in separate stomachs The Burbot at both inshore and offshore sites exhibited (Table 1). Many Burbot stomachs were empty or contained di- positive selection for both cyclopoid copepods and copepod gested matter that was unidentifiable. Significantly more empty nauplii (Table 2). Negative selection for calanoid copepods, stomachs were observed in larval Burbot collected from shal- cladocerans, and veligers was exhibited at all sites throughout low inshore sites (<15 m) than in fish collected from deeper the year. The selection for cyclopoid copepods was not (37 and 91 m) offshore sites (79% versus 27%; ANOVA: P = significantly greater than that for nauplii (P = 0.57) except 0.014). When we considered only Burbot that contained identi- during July at offshore sites, where the selection for nauplii fiable prey, they contained on average 16.1 cyclopoid copepods was negative (P < 0.001). Burbot showed a significantly per stomach, confirming cyclopoid copepods as the dominant higher preference for nauplii at offshore sites during May prey source (ANOVA: P = 0.004). Burbot consumed on average and June than in July (P = 0.006). Selectivity did not change 3.8 copepod nauplii, 2.7 calanoid copepods, and 1.6 dreissenid significantly over time for any other prey type (P > 0.05). veligers. Cladocerans and rotifers were not a substantive food source, with less than 0.1 being found per stomach for both prey types (Table 1). DISCUSSION The Burbot in northern Lake Huron exhibited a dietary pref-

Downloaded by [Department Of Fisheries] at 22:55 25 November 2013 erence for cyclopoid copepods and nauplii, results that are con- Prey Size versus Burbot Length sistent with previous findings in a variety of other systems Burbot found offshore were significantly larger than Burbot (Ghan and Sprules 1993; Wang and Appenzeller 1998; Probst found inshore (Student’s t-test: P < 0.001; Table 1). We found no and Eckmann 2009). This selection pattern persisted through- statistically significant relationship between the length of larval out the sampling season despite the high abundance of other Burbot and the size of the prey consumed (Rs < 0.3; P > 0.25 zooplankton types, such as cladocerans and calanoid copepods. for all prey types; Figure 4). In fact, some small Burbot larvae The negative selection of calanoid copepods is most likely due to (4–5 mm) ate large prey items, including cyclopoid copepods of their ability to evade predation better than cyclopoid copepods lengths greater than 1 mm. Larger Burbot continued to consume (Confer and Blades 1975; Wright and O’Brien 1984). Also, small prey, with some large specimens of 13–15 mm consuming because Burbot feed at dusk (Martin et al. 2011), low light con- small copepod nauplii and dreissenid veligers. However, there ditions may make sparsely pigmented prey such as cladocerans was a significant positive relationship between Burbot size and less visible than the heavily pigmented cyclopoid copepods. the number of prey items eaten (Rs = 0.72274; P < 0.0001). The avoidance of cladocerans may also be due to competition Yolk sac larvae and small larvae (<6.0 mm) occurred in samples with other predators: European Perch (also known as Eurasian late in the summer (Figure 5). Perch) Perca fluviatilis larvae have been shown to competitively FEEDING ECOLOGY OF PELAGIC LARVAL BURBOT 1721

FIGURE 5. Lengths of larval Burbot collected in Lake Huron from inshore (April–June) and offshore sites (May–July) during 2007, by collection date. FIGURE 4. Relationships between larval Burbot length and prey length, by prey type. mussels, suggesting that these species are affecting the feeding opportunities of larval Burbot through indirect mechanisms. A possible direct competitor with larval Burbot is invasive exclude Burbot from feeding on cladocerans in Lake Constance Rainbow Smelt. Rainbow Smelt larvae were the most abun- (Probst and Eckmann 2009) and may encourage Burbot to move dant ichthyoplankton at all offshore sites and De Tour inshore deeper into the water column and feed on copepods in order to sites in 2007 (Roseman and O’Brien 2013) and have been avoid competition. Probst and Eckmann (2009) propose that found to select for copepods while in the larval stage (Evans the positive selection of copepods by Burbot is a strategy to and Loftus 1987). This could put them in direct competition avoid competition with perch, resulting in the negative selection with Burbot for cyclopoid copepods: indeed, Rainbow Smelt of cladocerans by Burbot. A probable analog to the European have been shown to competitively exclude other planktivorous perch in Lake Huron is the invasive crustacean B. longimanus, species (Willis and Magnuson 2006), possibly to the point of which is a dominant planktivore in that lake. Bythotrephes longi- extirpation, as hypothesized by Hrabik et al. (1998) for Cis- manus is responsible for 78% of all plankton consumption in coes Coregonus artedi in Sparkling Lake, Wisconsin. Rainbow Lake Huron’s main basin, and at times its consumption exceeds Smelt have been described as “voracious” consumers (Schneider plankton production (Bunnell et al. 2011). As B. longimanus and Leach 1977), and their introduction has been linked to the most likely prefer cladoceran prey (Vanderploeg et al. 1993; recruitment failure of populations of several species, includ- Schultz and Yurista 1999;), they would occupy a niche in the ing that of Burbot in Elliott Lake, Ontario (Gray and Maraldo Lake Huron food web similar to that of the European Perch 1982; Evans and Loftus 1987). In addition, adult Rainbow Smelt in Lake Constance. Although B. longimanus and larval Burbot have been observed feeding on young Burbot (Evans and Loftus may not compete directly for their preferred prey types, this 1987), which may further affect Burbot recovery. does not necessarily mean that feeding by B. longimanus has no The crepuscular feeding behavior of larval Burbot may also

Downloaded by [Department Of Fisheries] at 22:55 25 November 2013 impact on the larval Burbot population. As noted in Barbiero explain the higher number of empty stomachs at inshore sites. et al. (2009), the biomass of cyclopoid copepods has also de- These sites were sampled during the day, when Burbot were clined as a result of the invasion of B. longimanus and dreissenid likely not feeding, whereas the offshore sites were sampled

TABLE 2. Selectivity index values (αi) for the prey types found in larval Burbot collected at inshore and offshore sites in northern Lake Huron in 2007. Positive selection (asterisks) is indicated when αi > 1/n,wheren = the number of prey types; neutral selection is indicated when αi = 1/n and negative selection when αi < 1/n. The letter N represents the number of Burbot sampled; see Table 1 for prey type abbreviations. Location Month N 1/n Cyc Cal Naup Clad Vel Inshore May 3 0.11 0.33* 0.00 0.67* 0.00 0.00 Jun 5 0.11 0.77* 0.00 0.23* 0.00 0.00 Offshore May 2 0.14 0.50* 0.00 0.50* 0.00 0.00 Jun 35 0.14 0.66* 0.12 0.20* 0.00 0.00 Jul 26 0.14 0.85* 0.01 0.02 <0.01 0.11 1722 GEORGE ET AL.

at night, when larval Burbot are known to move toward the bow Smelt for the preferred cyclopoid prey could also restrict surface to prey on zooplankton (Wang and Appenzeller 1998; Burbot growth and survival. A more comprehensive understand- Martin et al. 2011). Roseman and O’Brien (2013) suggest that ing of the effects of invasive species on larval Burbot and other Burbot move from inshore to offshore sites as they age, which planktivorous ichthyoplankton is needed to fully evaluate how is supported by our finding of significantly smaller Burbot at the Burbot’s early life history is being affected by new species inshore sites. at multiple trophic levels. A limitation of this study was the disparity between the time of day at which sampling occurred at inshore and offshore sites. This study was conducted as part of a larger project examining ACKNOWLEDGMENTS the role of invasive species on fish populations in Lake Huron, Bryon Daley, Erick Larson, Erik McDonald, Phil “Phlipp” and sampling times and locations were chosen to take advantage Pepper, and Edward O. Roseman assisted with field collections of existing sampling regimes. Because Burbot feed at dusk and and laboratory processing. Jeffrey S. Schaeffer provided statisti- have been observed to make slight diel migrations (Wang and cal assistance. Patricia Thompson and Jenny Sutherland assisted Appenzeller 1998; Martin et al. 2011), they may be more sus- with editing and formatting. This project was funded by the U.S. ceptible to capture at night. We would suggest sampling larval Environmental Protection Agency Project DW-14-94816701-0 Burbot exclusively at night to capture fish during feeding and and the U.S. Geological Survey, Great Lakes Science Center. postfeeding periods. This is contribution number 1733 of the Great Lakes Science Another limitation of our sampling was the disparity in the Center. depths at which offshore larval fish and zooplankton were col- lected. For our selectivity analyses we assumed that the zoo- REFERENCES plankton community represented in the 20-m integrated vertical Auer, N. A., editor. 1982. Identification of larval fishes of the Great Lakes zooplankton hauls was representative of the community found basin with emphasis on the Lake Michigan drainage. Great Lakes Fisheries in the upper 2 m of the water column where larval fish were Commission, Special Publication 82-3, Ann Arbor, Michigan. collected. Further, Burbot have been shown to occur at higher Balcer, M. D., N. L. Korda, and S. I. Dodson. 1984. Zooplankton of the Great densities at greater depths (5–10 m) in Lake Constance (Wang Lakes: a guide to the identification and ecology of the common crustacean and Appenzeller 1998) and Lake Michigan (Martin et al. 2011). species. University of Wisconsin Press, Madison. Barbiero, R. P., M. Balcer, D. C. Rockwell, and M. L. Tuchman. 2009. Recent An analysis of the relative vertical distribution and diet prefer- shifts in the crustacean zooplankton community of Lake Huron. Canadian ences of Burbot and their potential competitors could provide Journal of Fisheries and Aquatic Sciences 66:816–828. more insight into the habitat preference and feeding ecology Barbiero, R. P., and M. L. Tuchman. 2004. Changes in the crustacean com- of larval Burbot in Lake Huron and explain why the Burbot in munities of Lakes Michigan, Huron, and Erie following the invasion of the our study preyed almost exclusively on cyclopoid copepods and predatory cladoceran Bythotrephes longimanus. Canadian Journal of Fish- eries and Aquatic Sciences 61:2111–2125. nauplii in the upper 2 m of the water column. Bence, J. R., and L. C. Mohr. 2008. The state of Lake Huron in 2004. Great Burbot are recognized as an important top predator in Lake Lakes Fishery Commission, Special Report 08-1, Ann Arbor, Michigan. Huron, accounting for roughly 15% of consumption by preda- Berst, A. H., and G. R. Spangler. 1972. Lake Huron: effects of exploitation, tors in 1999 and therefore having substantial ecological impor- introductions, and eutrophication on the salmonid community. Journal of the tance to the lake (Bence and Mohr 2008). However, the Burbot Fisheries Research Board of Canada 29:877–887. Bowen, S. 1996. Quantitative description of the diet. Pages 513–532 in B. R. population has been declining in recent decades, and much of Murphy and D. W. Willis, editors. Fisheries techniques. American Fisheries the focus has centered on predation by invasive sea lampreys Society, Bethesda, Maryland. on adult Burbot (Berst and Spangler 1972) and the collapse of Bunnell, D. B., B. M. Davis, D. M. Warner, M. A. Chriscinske, and E. F.

Downloaded by [Department Of Fisheries] at 22:55 25 November 2013 the prey fish community due to the nearly complete loss of in- Roseman. 2011. Planktivory in the changing Lake Huron zooplankton com- vasive Alewives Alosa pseudoharengus after 2003 (Riley et al. munity: Bythotrephes consumption exceeds that of Mysis and fish. Freshwater Biology 56:1281–1296. 2008; Roseman and Riley 2009). Alewives were the dominant Chesson, J. 1983. The estimation and analysis of preference and its relationship prey in Lake Huron and a staple of piscivore diets (Diana 1990; to foraging models. Ecology 64:1297–1304. O’Gorman et al. 2012). Because early life history stages are Confer, J. L., and P. I. Blades. 1975. Omnivorous zooplankton and planktivorous integral to a species’ success, a complete understanding of the fish. Limnology and Oceanography 20:571–579. ontogenetic feeding ecology, growth, and movement of larval Confer, J. L., and G. J. Lake. 1987. Influence of prey type on growth of young Yellow Perch (Perca flavescens). Canadian Journal of Fisheries and Aquatic Burbot could provide insight into the factors that limit Burbot Sciences 44:2028–2033. recovery. The positive selection for cyclopoid copepods and Confer, J. L., and L. M. O’Bryan. 1989. Changes in prey rank and preference by copepod nauplii by larval Burbot in our study and the negative young planktivores for short-term and long-term ingestion periods. Canadian selection for other, more abundant prey types suggest that Bur- Journal of Fisheries and Aquatic Sciences 46:1026–1032. bot are relatively inflexible in their prey choices at the larval Crooks, J. A. 2002. Characterizing ecosystem-level consequences of biological invasions: the role of ecosystem engineers. Oikos 97:153–166. stage. If the zooplankton community continues to shift and de- Culver, D. A., M. M. Boucherle, D. J. Bean, and J. W. Fletcher. 1985. Biomass of cline due to B. longimanus predation, this could cause a decline freshwater crustacean zooplankton from length-weight regressions. Canadian in the abundance of preferred prey. Competition with Rain- Journal of Fisheries and Aquatic Sciences 42:1380–1390. FEEDING ECOLOGY OF PELAGIC LARVAL BURBOT 1723

Cushing, D. H. 1975. Plankton production and year class strength in fish popu- management: a binational perspective, 2nd edition. Michigan State University lations: an update of the match/mismatch hypothesis. Pages 249–293 in D. H. Press, East Lansing. Cushing, editor. Marine ecology and fisheries. Cambridge University Press, Probst, W., and R. Eckmann. 2009. Diet overlap between young-of-the-year Cambridge, UK. perch, Perca fluviatilis, and Burbot, Lota lota L., during early life-history Diana, J. S. 1990. Food habits of angler-caught salmonines in western Lake stages. Ecology of Freshwater Fish 18:527–537. Huron. Journal of Great Lakes Research 16:271–278. Riley, S. C., E. F. Roseman, S. J. Nichols, T. P. O’Brien, C. S. Kiley, and Evans, D. O., and D. H. Loftus. 1987. Colonization of inland lakes in the Great J. S. Schaeffer. 2008. Deepwater demersal fish community collapse in Lake Lakes region by Rainbow Smelt, Osmerus mordax: their freshwater niche Huron. Transactions of the American Fisheries Society 137:1879–1890. and effects on indigenous fishes. Canadian Journal of Fisheries and Aquatic Roseman, E. F., and T. P. O’Brien. 2013. Inshore-offshore distribution of lar- Sciences 44:249–266. val fish in northern Lake Huron. Journal of Aquatic Ecosystem Health and Fischer, P. 1999. Otolith microstructure during the pelagic, settlement and ben- Management 16:211–221. thic phases in Burbot. Journal of Fish Biology 54:1231–1243. Roseman, E. F., and S. C. Riley. 2009. Biomass of deepwater demersal for- Ghan, D., and W. G. Sprules. 1993. Diet, prey selection and growth of larval age fishes in Lake Huron, 1994–2007: implications for offshore predators. and juvenile Burbot Lota lota (L.). Journal of Fish Biology 42:47–64. Aquatic Ecosystem Health and Management 12:29–36. Gray, J. E., and D. Maraldo. 1982. A historical look at the fisheries of Saddle, Ryder, R. A., and J. Pesendorfer. 1992. Food, growth, habitat, and community Sill, Goulais, Gong, Megisan, Astonish, Flack and Elliott lakes. Ontario interactions of young of the year Burbot, Lota lota L., in a Precambrian shield Ministry of Natural Resources, Algoma Fisheries Assessment Unit, Sault St. lake. Hydrobiologia 243:211–227. Marie. SAS. 1990. SAS/STAT user’s guide, version 6, 4th edition. SAS Institute, Cary, Hardy, R., V. L. Paragamian, and M. D. Neufeld. 2008. Zooplankton commu- North Carolina. nities and Burbot relative abundance of some oligotrophic lakes of Idaho, Savino, J. 2009. The role of invasive species in Lake Huron food web disruptions. USA and British Columbia, Canada. Pages 79–89 in V. L. Paragamian and Great Lakes National Program Office, Research Completion Report prepared D. H. Bennett, editors. Burbot: ecology, management, and culture. American for U.S. Environment Protection Agency, Chicago. Fisheries Society, Symposium 59, Bethesda, Maryland. Schneider, J. C., and J. H. Leach. 1977. Walleye (Stizostedion vitreum vitreum) Houde, E. D. 1987. Fish early life dynamics and recruitment variability. Pages fluctuations in the Great Lakes and possible causes, 1800–1975. Journal of 17–29 in R. D. Hoyt, editor. 10th Annual larval fish conference. American the Fisheries Research Board of Canada 34:1878–1889. Fisheries Society, Symposium 2, Bethesda, Maryland. Schultz, K. L., and P. M. Yurista. 1999. Implications of an invertebrate preda- Houde, E. D. 1989. Subtleties and episodes in the early life of fishes. Journal of tor’s (Bythotrephes cederstroemi), atypical effects on a pelagic zooplankton Fish Biology 35:29–38. community. Hydrobiologia 380:179–193. Hrabik, T. R., J. J. Magnuson, and A. S. McLain. 1998. Predicting the effects Stapanian, M. A., C. P. Madenjian, C. R. Bronte, M. P. Ebener, B. F. Lantry, and of Rainbow Smelt on native fishes in small lakes: evidence from long-term J. D. Stockwell. 2008. Status of Burbot populations in the Laurentian Great research on two lakes. Canadian Journal of Fisheries and Aquatic Sciences Lakes. Pages 111–130 in V.L. Paragamian and D. H. Bennett, editors. Burbot: 55:1364–1371 ecology, management, and culture. American Fisheries Society, Symposium Jensen, N. R., P. J. Anders, C. A. Hoffman, L. S. Porter, S. C. Ireland, and 59, Bethesda, Maryland. K. D. Cain. 2011. Performance and macronutrient composition of age-0 Vanderploeg, H. A., J. R. Leibig, and M. Omai. 1993. Bythotrephes preda- Burbot fed four diet treatments. North American Journal of Aquaculture 73: tion on Great Lakes’ zooplankton measured by an in situ method: implica- 360–368. tions for zooplankton community structure. Archives fur Hydrobiologie 127: Link, J. 1996. Capture probabilities of Lake Superior zooplankton by an obligate 1–8. planktivorous fish—the lake herring. Transactions of the American Fisheries Vanderploeg, H. A., and D. Scavia. 1979. Calculation and use of selectivity Society 125:139–142. coefficients of feeding: zooplankton grazing. Ecological Modelling 7:135– Martin, B., S. Czesny, and D. Wahl. 2011. Vertical distributions of larval fish in 149. pelagic waters of southwest Lake Michigan: implications for growth, survival Vatcha, R. 1990. The food spectrum and growth of Burbot (Lota lota L.) fry and dispersal. Journal of Great Lakes Research 37:279–288. in experimental conditions. Bulletin of the Vodnany Research Institute of McPhail, J. D., and V. L. Paragamian. 2000. Burbot biology and life history. Fishery and Hydrobiology 26:14–19. (In Czech with English summary.) Pages 11–23 in V. L. Paragamian and D. W. Willis, editors. Burbot: biology, Vught, I., A. Harzevilli, J. Auwerx, and D. de Charleroy. 2008. Aspects of repro- ecology, and management. American Fisheries Society, Fisheries Manage- duction and larviculture of Burbot under hatchery conditions. Pages 167–178 ment Section Publication 1, Spokane, Washington. in V. L. Paragamian and D. H. Bennet, editors. Burbot: ecology, manage- Miller, T. J., L. B. Crowder, J. A. Rice, and E. A. Marschall. 1988. Larval ment and culture. American Fisheries Society, Symposium 59, Bethesda, Downloaded by [Department Of Fisheries] at 22:55 25 November 2013 size and recruitment mechanisms in fishes: toward a conceptual framework. Maryland. Canadian Journal of Fisheries and Aquatic Sciences 45:1657–1670. Wang, N., and A. Appenzeller. 1998. Abundance, depth distribution, diet com- Nalepa, T. F., D. L. Fanslow, S. A. Pothoven, A. J. Foley III, and G. A. Lang. position and growth of perch (Perca flavescens) and Burbot (Lota lota) larvae 2007. Long-term trends in benthic macroinvertebrate populations in Lake and juveniles in the pelagic zone of Lake Constance. Ecology of Freshwater Huron over the past four decades. Journal of Great Lakes Research 33:421– Fish 7:176–183. 436. Willis, T. V., and J. J. Magnuson. 2006. Response of fish communities in five O’Gorman, R. 1983. Distribution and abundance of larval fish in the nearshore north temperate lakes to exotic species and climate. Limnology and Oceanog- waters of western Lake Huron. Journal of Great Lakes Research 9:14–22. raphy 51:2808–2820. O’Gorman, R., C. P. Madenjian, E. Roseman, and M. Bur. 2012. Alewife in the Wright, D. I., and W. J. O’Brien. 1984. The development and field test of Great Lakes: old invader-new millennium. Pages 705–732 in W. W. Taylor, a practical model of the planktivorous feeding of White Crappie (Pomoxis A. J. Lynch, and N. J. Leonard, editors. Great Lakes fisheries policy and annularis). Ecological Monographs 54:65–98. This article was downloaded by: [Department Of Fisheries] On: 25 November 2013, At: 22:56 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Transactions of the American Fisheries Society Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/utaf20 Temporal Changes and Sexual Differences in Spatial Distribution of Burbot in Lake Erie Martin A. Stapanian a , Larry D. Witzel b & Andy Cook c a U.S. Geological Survey , 6100 Columbus Avenue, Sandusky , Ohio , 44870 , USA b Ontario Ministry of Natural Resources , 1 Passmore Street, Port Dover , Ontario , N0A 1N0 , Canada c Ontario Ministry of Natural Resources , 320 Milo Road, Wheatley , Ontario , N0P 2P0 , Canada Published online: 20 Nov 2013.

To cite this article: Martin A. Stapanian , Larry D. Witzel & Andy Cook (2013) Temporal Changes and Sexual Differences in Spatial Distribution of Burbot in Lake Erie, Transactions of the American Fisheries Society, 142:6, 1724-1732, DOI: 10.1080/00028487.2013.795191 To link to this article: http://dx.doi.org/10.1080/00028487.2013.795191

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions Transactions of the American Fisheries Society 142:1724–1732, 2013 C American Fisheries Society 2013 ISSN: 0002-8487 print / 1548-8659 online DOI: 10.1080/00028487.2013.795191

SPECIAL SECTION: BURBOT

Temporal Changes and Sexual Differences in Spatial Distribution of Burbot in Lake Erie

Martin A. Stapanian* U.S. Geological Survey, 6100 Columbus Avenue, Sandusky, Ohio 44870, USA Larry D. Witzel Ontario Ministry of Natural Resources, 1 Passmore Street, Port Dover, Ontario N0A 1N0, Canada Andy Cook Ontario Ministry of Natural Resources, 320 Milo Road, Wheatley, Ontario N0P 2P0, Canada

Abstract We used GIS mapping techniques to examine capture data for Burbot Lota lota from annual gill-net surveys in Canadian waters of Lake Erie during late August and September 1994–2011. Adult males were captured over a larger area (3–17% for ≥20% maximum yearly catch [MYC]) than adult females. More males than females were caught in the gill nets in 14 of the 15 study years. Collectively, these results support a hypothesis of greater activity by adult males during summer, when Burbot are actively feeding. The area of capture contracted by more than 60% (for ≥20% MYC) for both sexes during the time period, which is consistent with the documented decrease of the Burbot population in the lake. The sex ratio (females : males) varied over the time series but declined steadily from 0.97 in 2001 to 0.59 in 2011. The overlap in the capture areas of adult males and females was scale dependent. The depth distribution at which adult Burbot were caught did not change over the time series, and there was no difference in the median depths (about 30 m) at which adult male and female Burbot were caught. The last results are consistent with the Burbot’s reliance on coldwater habitats. Additional research is recommended, including telemetry to describe daily and seasonal movements and assessment of gender bias in active and passive capture gear.

The Burbot Lota lota is one of two species of freshwater (Smith 1968). Burbot occur in all of the Great Lakes, but in fish that has a circumpolar range (McPhail and Lindsey 1970). Lake Erie the species is generally restricted to the colder and

Downloaded by [Department Of Fisheries] at 22:56 25 November 2013 Worldwide, many Burbot populations are threatened or endan- deeper eastern portion of the lake during thermal stratification gered or have been extirpated, mainly due to pollution, habitat (Scott and Crossman 1973; Trautman 1981; CWTG 2012). change, and the adverse effects of invasive species (Stapanian Burbot populations recovered during 1960–2000 in Lakes et al. 2010a). The management of Burbot stocks is poorly un- Michigan, Huron, and Erie after near extirpation in those lakes derstood worldwide, and the species is usually ignored in fish (Stapanian et al. 2006, 2008). The main reason for the recovery management and conservation programs. Even basic life history in Lake Erie, following increases in water quality, was control information, including their spatial distribution in large lakes, of Sea Lampreys Petromyzon marinus (Stapanian et al. 2006). is lacking. This is due in part to the species’ unpopularity as The Burbot population in Lake Erie declined after about 2001 a commercial and sport fish in much of its range (Paragamian (Stapanian et al. 2008, 2010b). This decline was due in large part 2000; Stapanian et al. 2010a). to decreased recruitment and reproductive success (Stapanian Burbot and Lake Trout Salvelinus namaycush are the two et al. 2010b), which are common characteristics of declining native coldwater piscivores of the Laurentian Great Lakes and threatened Burbot populations (Stapanian et al. 2010a).

*Corresponding author: [email protected] Received August 31, 2012; accepted April 8, 2013

1724 SPATIAL DISTRIBUTION OF BURBOT 1725

Modeling results indicate that the low recruitment of Burbot TABLE 1. Summary of Burbot captured during annual Ontario Partnership is associated with increased predation on Burbot eggs and fry Gillnet surveys. Data for 1995 were excluded because fewer than 5 adults of Perca flavescens each sex were captured. No surveys were conducted in 1996 and 1997. Age data by Yellow Perch and declines in the number for 1997–2011 were obtained from Burbot captured in Ontario waters during of days in which water temperatures were optimal for Burbot the annual Coldwater Task Group Gillnet Survey in the eastern basin of Lake spawning and egg development (Stapanian et al. 2010b). The Erie (CWTG 2012; USGS and OMNR, unpublished data). effects of this decline in abundance on the spatial distribution of Burbot in the lake are unknown. Survey Total Adult Adult Gill-net Mean age The Burbot in Lake Erie exhibited a switch in diet from year Burbot males females lifts (years) predominantly Rainbow Smelt Osmerus mordax (particularly 1994 43 10 6 28 age-1 and older fish) from the 1980s to 2000 to predominantly 1995 Round Goby Neogobius melanostomus by 2005 (Madenjian 1996 et al. 2011; Stapanian et al. 2011). By 2007, Burbot exhib- 1997 5.2 ited predatory control of the invasive Round Goby in offshore 1998 309 127 158 39 4.8 > habitats (depth, 20 m; Madenjian et al. 2011). Nighttime hy- 1999 244 130 96 39 5.1 droacoustics and trawls indicate that age-0 Rainbow Smelt oc- 2000 265 126 96 48 5.6 cur in the epilimnion, whereas age-1 and older Rainbow Smelt 2001 207 96 93 48 5.9 are captured mostly at or below the thermocline when Lake 2002 208 112 95 50 7.0 Erie is stratified (P. Kocovsky, U.S. Geological Survey [USGS], 2003 278 143 127 50 8.0 personal communication). Round Goby and Burbot are mostly 2004 171 97 74 49 9.1 benthic (Charlebois et al. 1997; McPhail and Paragamian 2000). 2005 145 86 58 50 8.6 Stapanian et al. (2011) hypothesized that this switch in diet 2006 102 61 35 46 9.9 would result in lower foraging costs for Burbot. Perhaps 2007 69 39 29 25 10.1 the switch is associated with changes in depth and spatial 2008 50 24 23 49 10.0 distributions. 2009 51 28 17 49 10.3 Telemetry and toxicology studies (e.g., Lucas et al. 1993; 2010 17 10 5 47 12.5 Altimiras et al. 1996; Hutchings and Gerber 2002; Madenjian 2011 44 27 16 50 13.7 et al. 2011 and references therein) suggest that adult male pis- civorous fish are more active than adult females. The greater activity by males contributes to a higher rate of energy expen- done during late August and September. In brief, the ONPT data diture in males (e.g., Beamish 1964; Smith and Brown 1983; are derived from 25–50 sites fished in the east basin (Table 1). Cooke 2004; Madenjian et al. 2011). If adult male Burbot are Nets were set at random locations among three depth strata (0– more active than adult females during summer, when Burbot are 15, 15–30, and >30 m), and the number of sites within each actively feeding, it is logical to predict that in gill-net surveys stratum was proportional to its area within the basin. The gill conducted during summer (1) proportionately more adult males nets consisted of 1.8-m × 380-m panels of graded monofil- will be caught than adult females and (2) adult males will be ament mesh, with 14 mesh sizes ranging from 32 to 152 mm caught over a broader area than adult females. stretched measure. Mesh sizes ≥57 mm were composed of two In this paper we test these two predictions. We quantify the 15.25-m panels, and those less than 57 mm were single pan- area of capture, sex ratios, and spatial overlap of adult male els 15.25 m in length. Although the ONPT sets nets in other and female Burbot in Lake Erie and test whether a switch in parts of the water column, we only considered the data for nets

Downloaded by [Department Of Fisheries] at 22:56 25 November 2013 diet to Round Goby was temporally associated with a change that were set on-bottom for approximately 24 h. We limited our in depth at capture for Burbot. Our study should provide in- study to ONPT because it is the only survey of coldwater fishes sight into the differences between sexes and temporal changes in the eastern basin of Lake Erie that is not restricted to wa- in spatial distribution by a top predator in coldwater systems. ters below the thermocline (sensu CWTG 2012). The sex and Further, identifying differences and temporal changes in spatial maturity of the Burbot captured were determined by examining distribution may be useful for devising management plans for reproductive structures. Age data for 1997–2011 were obtained Burbot populations that are being restored or rehabilitated (e.g., from Burbot captured in Ontario waters during the annual Cold- Paragamian et al. 2005, 2008; Ireland and Perry 2008). water Task Group Gill-Net Survey in the eastern basin of Lake Erie (CWTG 2012; USGS and Ontario Ministry of Natural Re- sources [OMNR], unpublished data). METHODS Spatial distribution methods.—Although the ONPT survey Study area and field methods.—Our study area included the began in 1989, we examined ONPT data for 1994–2011. No Canadian waters of the eastern basin of Lake Erie (Figure 1). surveys were conducted during 1996 and 1997. Further, we The data for this study were obtained from a subset of the annual restricted our analysis to include years in which all standard Ontario Partnership Gillnetting Surveys (ONPT; OMNR 2012) gear was employed and years in which at least five adult females 1726 STAPANIAN ET AL.

FIGURE 1. Bathymetric map of eastern Lake Erie. The long peninsula approximately 35 km south of Port Dover is Long Point peninsula. Long Point Bay is between Long Point peninsula and Port Dover.

and five adult males were captured. Based on these criteria, the effect or drift to interpolate the Burbot catch by sex for a defined data for 1995 were excluded. The data were then divided into area of eastern Lake Erie encompassing all of the gill-net loca- three 5-year periods: 1994–2001, 2002–2006, and 2007–2011. tions sampled since 1994 using SURFER contouring and surface Ecologically, these periods represented (1) a period in which mapping software (version 10.7; Golden Software, Inc. 2012). the Burbot population was increasing and there were no Round The number of gill-net sites sampled in a given year ranged Goby in the diet, (2) a period in which Burbot were first abundant from 25 in 2007 to 50 in 2002, 2003, 2005, and 2011(Table 1),

Downloaded by [Department Of Fisheries] at 22:56 25 November 2013 and then declined and there were Round Goby in the diet, and resulting in the following total numbers of sample observations (3) a period in which the Burbot population was small and per analysis period: 245 (1994–2001), 220 (2002–2006), and Round Goby were the main prey item (Stapanian et al. 2011; 202 (2007–2011). CWTG 2012). Several individual survey design grids were sampled more We used index gill-net catch data to map the Burbot distribu- than once per analysis period and generated multiple %MYC tion as a percent of the maximum yearly catch (%MYC) by sex; observations in close spatial proximity. A median (i.e., 50%) we chose this metric in order to normalize the effects of year- MYC was taken as the point observation for samples situated class strength. By design, gill nets were fished near the center inside X (easting) and Y (northing) tolerances of 1,350 m and of randomly selected 2.5-min (latitude) × 2.5-min (longitude) 1,800 m, respectively. The filtering of duplicate %MYC data was sample grids. The latitude–longitude coordinates (NAD83) of done during kriging operations in SURFER (Golden Software, the gill-net sampling locations derived from the Wide Area Aug- Inc. 2012). Interpolated estimates of %MYC for a 725,353-ha mentation System–enabled GPS instrumentation onboard the grid area at 250-m × 250-m resolution was visually represented survey vessels were converted to UTM (Zone 17N) coordinates in contour maps. We also examined the distribution of male for mapping purposes. We applied an ordinary point-method and female Burbot by separately mapping and estimating the kriging algorithm with a linear variogram and without a nugget planar surface area of gridded regions of %MYC at three range SPATIAL DISTRIBUTION OF BURBOT 1727

levels: >1%, >20%, and >50% MYC. We reasoned that >1% MYC represented the overall distribution of adult Burbot, >20% MYC was a more conservative measure of Burbot distribution, and >50% MYC was an approximation of the greatest Burbot area density. Contour maps of male and female Burbot for the three 5-year periods were produced and examined for temporal and gender-based differences in spatial distribution. The planar surface area of Burbot distributions and area of overlap between sexes for each 5-year period were estimated using SURFER version 10.7 (Golden Software, Inc. 2012). The sex ratio was defined as the number of females per male captured during each survey year. We used a sign test to determine whether sex ratios <1 occurred more often than expected by chance (rejection level, P < 0.05). We examined the general trends in the sex ratio as a function of mean age and the number of adults captured per gill-net lift with linear FIGURE 2. Catch per gill-net lift of adult Burbot and ratio of females to males in Lake Erie during 1994–2011 (survey years 1995–1997 were not included). regression. We calculated the percentage change in planar area The dashed line represents a sex ratio of 1:1. [Figure available in color online.] for both sexes and the overlap during the three 5-year periods. Changes in the depth distributions during the time series were examined by calculating the median and the upper (75%) and The sex ratio (females : males) varied considerably over the lower (25%) depth quartiles of the ≥20% and ≥1% MYC. time series, but it was less than 1 in all survey years except 1998 (Figure 2). Thus, significantly more males were captured than females over the study period (sign test; N = 15, P < 0.01). During 2001–2011, the sex ratio declined steadily from 0.97 RESULTS to 0.59. This decline occurred at the same time as a decrease The catch per lift of adult Burbot declined from 7.3 in 1998 to in number of Burbot caught and during a steady increase in the 0.86 in 2011 (Figure 2). Using the CWTG data (CWTG 2012), mean age of the population (Table 1). Over the time series, there the mean age of the population increased from 5.24 years in was a positive association between the sex ratio and the number 1997 to 13.67 in 2011 (Table 1). The planar area decreased over of adults per gill-net lift (regression; N = 15, r2 = 0.51, P = time for both sexes (Table 2; Figures 3, 4). For males, the total 0.003). In contrast, there was a negative association between the planar area decreased by 17% for ≥1% of the maximum yearly sex ratio and the mean age of the population (regression; N = catch, by 61% for ≥20% of the maximum yearly catch, and by 14, r2 = 0.45, P = 0.008). 73% for ≥50% of the maximum yearly catch between 1994– The area in which males were collected during the study pe- 2001 and 2007–2011. For females, these decreases were 41, 66, riod was slightly greater than the area in which females were and 77%, respectively. This general shrinkage of capture area collected (Table 2; Figures 3, 4). In seven of the nine combi- was consistent with the decreasing population size. nations of time period and %MYC, the planar area for males

TABLE 2. Planar areas and percent overlap in capture areas of adult Burbot in Lake Erie during three time periods. Total planar area = 3,526.5 km2. Abbreviations are as follows: %MYC = percentage of the maximum yearly catch, %TPA = percentage of the total mapped planar area, %FPA = percentage of the female planar = Downloaded by [Department Of Fisheries] at 22:56 25 November 2013 area, and %MPA percentage of the male planar area.

Overlap as Female area Male area Overlap Female area Male area Period %MYC (km2)(km2)(km2) (%TPA) (%TPA) %TPA %FPA %MPA 1994–2001 ≥50 143.63 216.73 40.05 4.1 6.1 1.1 27.9 18.5 ≥20 1,012.98 1,069.33 753.81 28.7 30.3 21.4 74.4 70.5 ≥1 2,430.11 2,217.36 2,132.17 68.9 62.9 60.5 87.7 96.2 2002–2006 ≥50 63.14 36.29 1.09 1.8 1.0 0.0 1.7 3.0 ≥20 706.81 726.17 484.43 20.0 20.6 13.7 68.5 66.7 ≥1 2,020.35 2,173.60 1,955.87 57.3 61.6 55.5 96.8 90.0 2007–2011 ≥50 32.73 58.45 0.00 0.9 1.7 0.0 0.0 0.0 ≥20 346.83 419.01 123.74 9.8 11.9 3.5 35.7 29.5 ≥1 1,430.08 1,843.37 1,199.26 40.6 52.3 34.0 83.9 65.1 1728 STAPANIAN ET AL.

FIGURE 3. Spatial distributions of the percentage of maximum yearly catch of adult female Burbot during the study’s three time periods. FIGURE 4. Spatial distributions of the percentage of maximum yearly catch of adult male Burbot during the study’s three time periods.

(as a percentage of the entire mapped planar area) ranged from In contrast, the areas of ≥50% MYC did not correspond to a

Downloaded by [Department Of Fisheries] at 22:56 25 November 2013 0.7% (in 2007–2011; ≥50% MYC) to 11.7% (in 2007–2011; particular area or depth contour. ≥1% MYC) greater than that for females. The planar area of The depth distribution did not change during the study period, the distribution for all three%MYC categories was highest for and males and females were not separated by depth (Figure 7). the period 1994–2001 and lowest for the period 2007–2011. The median depths for both sexes were 29–32 m for 1% MYC The amount of overlap in planar area between sexes was during the time series. For 20% MYC, the median depth was 29– dependent on scale and time period. Overlap decreased over the 32 m in the two later time periods. Although the median depths time period for the ≥1%, ≥20%, and ≥50% MYCs (Table 2; for males (39 m) and females (34 m) during 1994–2001 were Figures 5, 6). In each time period, the overlap was greatest somewhat greater at 20% MYC, the overlaps in the distribu- for ≥1% MYC (range, 34.0–60.5%) and least for ≥50% MYC tions made these differences statistically insignificant. Further, (range, 0–1.1%). The greatest amount of overlap during the Burbot were not captured closer to shore during the study pe- time series corresponded to those areas that exhibited the riod. Both sexes were caught closest to the north coastline dur- greatest bathymetric variation and near the 40-m depth contour ing 1994–2001 and farthest away during 2007–2011 (Figures 3, (Figures 1, 5, 6). The areas south and east of Long Point toward 4). This last result should be interpreted with caution, how- the Canadian–U.S. border exhibited the greatest bathymetry ever, because fewer Burbot were captured during the last time and had considerable overlap in planar area between sexes. period. SPATIAL DISTRIBUTION OF BURBOT 1729

FIGURE 5. Spatial distributions of the overlap (≥1% of maximum yearly ≥ catch) between adult male and female Burbot during the three study periods. FIGURE 6. Spatial distributions of the overlap ( 20% of maximum yearly [Figure available in color online.] catch) between adult male and female Burbot during the three study periods. [Figure available in color online.]

DISCUSSION males; the higher energy expenditure rate was due, at least in The results strongly supported our prediction that more adult part, to higher activity in males.

Downloaded by [Department Of Fisheries] at 22:56 25 November 2013 males would be caught than adult females in summer gill-net An alternative hypothesis for the patterns observed in this surveys. In addition, our prediction that adult males would be study is there were more adult male Burbot than adult female more broadly distributed spatially than adult females was gen- Burbot in the study area. A bottom trawl survey targeting Burbot erally supported by our data. Rudstam et al. (1984) identified was conducted in the study area in the summer and early fall of two components of the probability of a fish’s being captured in 2007 (Madenjian et al. 2011). The female : male ratio for that a gill net: (1) the probability that it will encounter the gill net survey was 1.63:1, compared with 0.74:1 for the Ontario Part- and (2) the probability that it will be retained in the net after nership gillnetting survey and 0.60:1 for the CWTG gillnetting it is caught. A more active fish would be expected to have a survey in Ontario waters of eastern Lake Erie in that same year higher probability of encountering a gill net than a less active (Madenjian et al. 2011; OMNR and USGS, unpublished data). fish. Thus, higher activity in males would explain the higher Although limited to 1 year, the results do not support this alter- catch of males in gill nets. Our results are supported by the native hypothesis. It is possible, however, that active gear (e.g., findings of Madenjian et al. (2013), who observed higher poly- bottom trawls) also produce sex ratios of Burbot that are biased. chlorinated biphenyl concentrations in male Burbot from Lake For example, because male Burbot are more active they may be Erie than in female Burbot. These higher concentrations were able to avoid bottom trawls better than females. Burbot are a attributed to males expending energy at a higher rate than fe- comparatively slow-swimming fish and have comparatively low 1730 STAPANIAN ET AL.

collected during August and September over 15 years, and from an aging Burbot population. Coldwater Task Group gill nets also set in Ontario waters of Lake Erie during August cap- tured more male Burbot than females in 9 out of 10 study years during 2002–2011 (USGS and OMNR, unpublished data). In contrast, during 1997–2001, when the mean age of the Burbot captured was much less (4.8–5.1 years versus 7.0–13.7 years during 2002–2011), more females were captured in three out of five years (USGS and OMNR, unpublished data). Males captured during October–November outnumbered females by approximately 3:1 (n = 223) in trap nets in Arrow Lakes Reser- voir, British Columbia in 2008 (Glova et al. 2009) and by 2.3:1 (n = 60) in 2009 (Glova et al. 2010). The Burbot captured (n = 73) in hoop nets in the Kootenai River in Idaho and British Columbia during October 2000–March 2001 were 60% males, 15% females, and 25% unknown (i.e., the fish were examined but their sex could not be determined; Kozfkay and Paragamian 2002). In contrast, the sex ratios of Burbot captured in gill nets in northern Lake Michigan in May 2002–2012 did not exhibit this pattern (C. Madenjian, USGS, personal communication) and was nearly 1:1 overall. Nearly all (98%) of the Burbot captured (n = 133) in May were in postspawning condition. Greater adult female mortality could also result in low female : male ratios in captured Lake Erie Burbot. Sea Lampreys are the only wild predators of adult Burbot in Lake Erie. According to Coldwater Task Group data collected in Ontario waters of Lake Erie during 1994–2011, fresh (types A1–A4; King 1980) Sea Lamprey wounds were only recorded for 10 adult Burbot, all after 1998 (USGS and OMNR, unpublished data). Of these, nine were males and the overall wounding rate was less than 1%. Although the data are limited, these results do not support the contention of greater adult female mortality in Lake Erie, at least due to Sea Lamprey predation. Although our results generally support the predictions of greater spatial distribution and greater “catchability” of adult FIGURE 7. Burbot depth distributions (≥1% and ≥20% of the maximum males in gill nets, they do not constitute proof that males are yearly catch) during the three study periods. [Figure available in color online.] more active. Further, greater activity may not necessarily result in greater spatial range. The males may be more active than the females while inhabiting an equal or smaller spatial area. For

Downloaded by [Department Of Fisheries] at 22:56 25 November 2013 endurance (Jones et al. 1974; Vokoun and Watrous 2009). How- example, the benthic habit and abundance of Round Goby in ever, neither Jones et al. (1974) nor Vokoun and Watrous (2009) Lake Erie may provide sufficiently low foraging and handling reported a sex difference in swimming speed or endurance in costs to Burbot (Stapanian et al. 2011) that a wider foraging Burbot. range is not necessary to satisfy their metabolic requirements. The sex ratios of the Burbot captured in passive gear may However, a shift to a diet of primarily Round Goby was probably be affected by season and type of gear. The activity of adult not the primary cause for recent increases in size at age among male Burbot in particular may be quite different during sum- the Burbot in the lake. We recommend telemetry studies to mer, fall, and early winter than during late winter and spring. quantify the ranges and physiological studies to quantify the For example, Burbot spawning typically occurs during Febru- metabolism of individual fish. ary. Thus, a sexual difference in the probability of encountering Our results are restricted to a comparatively small annual capture gear may vary by season. Previous studies involving period (late August–September). However, it represents a pe- passive gear indicate that male Burbot appear to be substan- riod of active feeding for Burbot in Lake Erie (Stapanian et al. tially more active than females during the summer, fall, and 2011), which was appropriate for testing our predictions. Pre- early winter months, whereas the sex difference in activity ap- vious analyses (CWTG 2012) based on gill-net surveys in Lake pears to be less pronounced during the spring. Our data were Erie during 2007–2011 indicated that there was no difference SPATIAL DISTRIBUTION OF BURBOT 1731

between male and female Burbot or among survey years in ACKNOWLEDGMENTS the proportion of total (wet) prey weight consumed of either We thank the participants in the Ontario Partnership Gillnet- Round Goby or Rainbow Smelt. Further, the interaction be- ting Program and the technicians from the Ontario Ministry of tween the sex of the Burbot and survey year was not significant. Natural Resources in Port Dover and Wheatley. P. Kocovsky, These results indicate that any lack of overlap in planar area R. Kraus, C. Madenjian, and three anonymous reviewers pro- between the sexes was not due to their foraging for different vided comments on earlier versions of this paper. The mention prey. of trade, product, or firm names does not imply endorsement by Our results did not indicate a change in the depth distribution the U.S. Government. This article is contribution 1753 of the of Burbot in Lake Erie. As mentioned, it is unclear whether Great Lakes Science Center. the density of Round Goby is actually greater in waters >20 m deep than in shallower waters. Burbot are coldwater fish, and warming temperatures have been shown to be deleterious to REFERENCES Burbot populations (Paragamian et al. 2000; Jackson et al. 2008; Altimiras J., A. D. F. Johnstone, M. C. Lucas, and I. G. Priede. 1996. Sex Stapanian et al. 2010a). Therefore, our results are consistent differences in the heart rate variability spectrum of free-swimming Atlantic with the coldwater habit of Burbot throughout the time period. Salmon (Salmo salar L.) during the spawning season. Physiological Zoology Although Burbot are known to forage nocturnally for periods in 69:770–784. Beamish, F. W. H. 1964. Seasonal changes in the standard rate of oxygen the epilimnetic waters of smaller, inland oligotrophic lakes (Carl consumption of fishes. Canadian Journal of Zoology 42:189–194. 1995), this study does not provide evidence of Burbot straying Carl, L. M. 1995. Sonic tracking of Burbot in Lake Opeongo, Ontario. Trans- from the hypolimnetic waters of eastern Lake Erie. actions of the American Fisheries Society 124:77–83. The proportion of overlap in planar area between males and Caswell, H. 2001. Matrix population models. Sinauer, Sunderland, females was scale dependent and declined over the time series. Massachusetts. Charlebois, P. M., J. E. Marsden, R. G. Goettel, R. K. Wolfe, D. J. Jude, This result is consistent with a decrease in population den- and S. Rudnika. 1997. The Round Goby, Neogobius melanostomus (Pallas), a sity. The capture of a single Burbot at a gill-net site when review of European and North American literature. Illinois-Indiana Sea Grant the population density is very low will be assigned a higher Program and Illinois Natural History Survey, INHS Special Publication 20, percentage of the maximum than when the population density Champaign, Illinois. is high. The decrease in planar area at the ≥1% maximum Cooke, S. J. 2004. Sex-specific differences in cardiovascular performance of a centrarchid fish are only evident during the reproductive period. Functional catch for both sexes over the time period is strong evidence for Ecology 18:398–403. a contraction in the spatial distribution with lower population CWTG (Coldwater Task Group). 2012: Report of the Lake Erie Coldwater density. Task Group, March 2012. Lake Erie Committee of the Great Lakes Fishery The areas surrounding Long Point and along the Canadian– Commission, Ann Arbor, Michigan. Available: http://www.glfc.org/lakecom/ U.S. border consistently showed overlap in planar area between lec/CWTG docs/annual reports/CWTG report 2012. (January 2013). Glova, G., D. Robichaud, and R. Bussanich. 2009. Arrow Lakes Reservoir sexes. These areas contain some of the steepest bathymetries in Burbot life history and habitat use (year 1). Report prepared for BC Hydro Canadian waters of eastern Lake Erie. Collectively, these results by LGL Ltd., CLBMON-31, Vancouver. suggest that habitat heterogeneity is associated with overlap Glova, G., D. Robichaud, E. Plate, and M. Matthews. 2010. Arrow Lakes Reser- in the spatial distributions of adult male and female Burbot. voir Burbot life history and habitat use (year 2). Report prepared for BC However, these areas were not associated with consistently high Hydro by LGL Ltd., CLBMON-31, Vancouver. Golden Software, Inc. 2012. SURFER surface mapping system, version 10.7. catches of Burbot. We recommend further studies to identify Golden Software, Inc., Golden, Colorado. additional key habitat requirements for Burbot. Hutchings, J. A., and L. Gerber. 2002. Sex-biased dispersal in a salmonid fish. Adult sex ratio is one of the key factors in population mod- Proceedings of the Royal Society Series B (Biological Sciences) 269:2487– 2493. Downloaded by [Department Of Fisheries] at 22:56 25 November 2013 eling and demography studies (Caswell 2001). If the mortality of female Burbot is not higher than that of males, our results Ireland, S. C., and P. N. Perry. 2008. Burbot restoration in the Kootenai River basin: using agency, tribal, and community collaboration to develop and suggest that gill-net surveys conducted during summer and fall implement a conservation strategy. Pages 251–256 in V. L. Paragamian and (particularly on comparatively old Burbot populations) will pro- D. H. Bennett, editors. Burbot: ecology, management, and culture. American vide sex ratios that are male biased. In systems in which Burbot Fisheries Society, Symposium 59, Bethesda, Maryland. populations are threatened (reviewed by Stapanian et al. 2010a), Jackson, R. J., A. J. VanDeValk, J. L. Forney, B. F. Lantry, T. E. Brooking, and such results could lead to management or restoration programs L. G. Rudstam. 2008. Long-term trends in Burbot abundance in Oneida Lake, New York: life at the southern edge of the range in an era of climate change. that are suboptimal. In particular, underestimating the number Pages 131–152 in V. L. Paragamian and D. H. Bennett, editors. Burbot: of females might result in a population model that overestimates ecology, management, and culture. American Fisheries Society, Symposium the time to population recovery. However, models that overesti- 59, Bethesda, Maryland. mate the time to population recovery may represent conservative Jones, D. R., J. W. Kiceniuk, and O. S. Bamford. 1974. Evaluation of the estimated recovery times and are probably better than the alter- swimming performance of several fish species from the Mackenzie River. Journal of the Fisheries Research Board of Canada 31:1641–1647. native. As discussed above, however, it is possible that active King, E. L. J. 1980. Classification of Sea Lamprey (Petromyzon marinus) attack gear (e.g., bottom trawls) also provide sex ratios of Burbot that marks on Great Lakes Lake Trout (Salvelinus naymaycush). Canadian Journal are biased. of Fisheries and Aquatic Sciences 37:1989–2006. 1732 STAPANIAN ET AL.

Kozfkay, J. R., and V. L. Paragamian. 2002. Kootenai River fisheries investi- Paragamian, V. L., V. Whitman, J. Hammond, and H. Andrusak. 2000. Col- gation: stock status of Burbot. Idaho Department of Fish and Game, Annual lapse of the Burbot fisheries in Kootenay Lake, British Columbia Canada Report to the Bonneville Power Administration, Project 88-65, Boise. and the Kootenai River, Idaho, USA, post Libby Dam. Pages 155–164 in Lucas, M. C., A. D. F. Johnstone, and I. G. Priede. 1993. Use of physiolog- V. L. Paragamian and D. W. Willis, editors. Burbot: biology, ecology, and ical telemetry as a method of estimating metabolism of fish in the natural management. American Fisheries Society, Fisheries Management Section environment. Transactions of the American Fisheries Society 122:822–833. Publication 1, Spokane, Washington. Madenjian, C. P., M. A. Stapanian, R. R. Rediske, and J. P. O’Keefe. 2013. Sex Rudstam, L. G., J. J. Magnuson, and W. M. Tonn. 1984. Size selectivity of difference in Polychlorinated Biphenyl concentrations of Burbot Lota lota passive fishing gear: a correction for encounter probability applied to gill from Lake Erie. Archives of Environmental Contamination and Toxicology nets. Canadian Journal of Fisheries and Aquatic Sciences 41:1252–1255. 65:300–308. Scott, W. B., and E. J. Crossman. 1973. Freshwater fishes of Canada. Bulletin Madenjian, C. P.,M. A. Stapanian, L. D. Witzel, S. A. Pothoven, D. W. Einhouse, of the Fisheries Research Board of Canada 184. and H. L. Whitford. 2011. Evidence for predatory control of the invasive Smith, S. E. 1968. Species succession and fishery exploitation in the Great Round Goby. Biological Invasions 13:987–1002. Lakes. Journal of the Fisheries Research Board of Canada 25:667–693. McPhail, J. D., and C. C. Lindsey. 1970. Freshwater fishes of northwestern Smith, K. L. Jr., and N. O. Brown. 1983. Oxygen consumption of pelagic Canada and Alaska. Fisheries Research Board of Canada Bulletin 173. juveniles and demersal adults of the deep-sea fish Sebastolobus altivelis McPhail, J. D., and V. L. Paragamian. 2000. Burbot biology and life history. measured at depth. Marine Biology 76:325–332. Pages 11–23 in V. L. Paragamian and D. W. Willis, editors. Burbot: biology, Stapanian, M. A., C. P. Madenjian, C. Bronte, M. Ebener, B. F. Lantry, and J. ecology, and management. American Fisheries Society, Fisheries Manage- Stockwell, 2008. Status of Burbot populations in the Laurentian Great Lakes: ment Section Publication 1, Spokane, Washington. a synthesis. Pages 111–130 in V. L. Paragamian and D. H. Bennett, edi- OMNR (Ontario Ministry of Natural Resources). 2012. 2011 status of major tors. Burbot: ecology, management, and culture. American Fisheries Society, stocks. OMNR, Lake Erie Management Unit, London. Symposium 59, Bethesda, Maryland. Paragamian, V. L. 2000. The effects of variable flows on Burbot spawning mi- Stapanian, M. A., C. P. Madenjian, and L. D. Witzel. 2006. Evidence that grations in the Kootenai River, Idaho, USA, and British Columbia, Canada. Sea Lamprey control led to recovery of the Burbot population in Lake Erie. Pages 111–123 in V.L. Paragamian and D. W. Willis, editors. Burbot: biology, Transactions of the American Fisheries Society 135:1033–1043. ecology, and management. American Fisheries Society, Fisheries Manage- Stapanian, M. A., V. L. Paragamian, C. P. Madenjian, J. R. Jackson, J. ment Section, Special Publication 1, Bethesda, Maryland. Lappalainen, M. J. Evenson, and M. D. Neufeld. 2010a. World-wide sta- Paragamian, V. L., R. Hardy, and B. Gunderman. 2005. Effects of regu- tus of Burbot and conservation measures. Fish and Fisheries 11:34–56. lated discharge on Burbot migration. Journal of Fish Biology 66:1199– Stapanian, M. A., L. D. Witzel, and A. Cook. 2010b. Recruitment of a recently 1213. recovered native piscivore in a large lake: an empirical modeling approach. Paragamian, V. L., B. J. Pyper, M. J. Daigneault, R. C. Beamesderfer, and Ecology of Freshwater Fish 19:326–337. S. C. Ireland. 2008. Population dynamics and extinction risk of Burbot in the Stapanian, M. A., L. D. Witzel, and W. H. Edwards. 2011. Recent changes in Kootenai River, Idaho, USA and British Columbia, Canada. Pages 213–234 in growth of Burbot in Lake Erie. Journal of Applied Ichthyology 27(Supple- V. L. Paragamian and D. H. Bennett, editors. Burbot: ecology, management, ment 1):57–64. and culture. American Fisheries Society, Symposium 59, Bethesda, Maryland. Trautman, M. B. 1981. The fishes of Ohio. Ohio State University Press, Paragamian, V.L., and V.D. Wakkinen. 2008. Seasonal movement and the inter- Columbus. action of temperature and discharge on Burbot in the Kootenai River, Idaho, Vokoun, J. C., and D. C. Watrous. 2009. Determining swim speed performance USA, and British Columbia, Canada. Pages 55–77 in V. L. Paragamian and characteristics for fish passage of Burbot using an experimental flume and D. H. Bennett, editors. Burbot: ecology, management, and culture. American nature-like fishway. University of Connecticut, Department of Environmental Fisheries Society, Symposium 59, Bethesda, Maryland. Protection, Completion Report, Storrs. Downloaded by [Department Of Fisheries] at 22:56 25 November 2013 This article was downloaded by: [Department Of Fisheries] On: 25 November 2013, At: 22:57 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Transactions of the American Fisheries Society Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/utaf20 Burbot Early Life History Strategies in the Great Lakes David J. Jude a , Yu Wang b , Stephen R. Hensler a & John Janssen b a School of Natural Resources , University of Michigan , 440 Church Street, Ann Arbor , Michigan , 48109 , USA b School of Freshwater Sciences , University of Wisconsin–Milwaukee , 600 East Greenfield Avenue, Milwaukee , Wisconsin , 53203 , USA Published online: 20 Nov 2013.

To cite this article: David J. Jude , Yu Wang , Stephen R. Hensler & John Janssen (2013) Burbot Early Life History Strategies in the Great Lakes, Transactions of the American Fisheries Society, 142:6, 1733-1745, DOI: 10.1080/00028487.2013.795192 To link to this article: http://dx.doi.org/10.1080/00028487.2013.795192

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions Transactions of the American Fisheries Society 142:1733–1745, 2013 C American Fisheries Society 2013 ISSN: 0002-8487 print / 1548-8659 online DOI: 10.1080/00028487.2013.795192

SPECIAL SECTION: BURBOT

Burbot Early Life History Strategies in the Great Lakes

David J. Jude* School of Natural Resources, University of Michigan, 440 Church Street, Ann Arbor, Michigan 48109, USA Yu Wang School of Freshwater Sciences, University of Wisconsin–Milwaukee, 600 East Greenfield Avenue, Milwaukee, Wisconsin 53203, USA Stephen R. Hensler School of Natural Resources, University of Michigan, 440 Church Street, Ann Arbor, Michigan 48109, USA John Janssen School of Freshwater Sciences, University of Wisconsin–Milwaukee, 600 East Greenfield Avenue, Milwaukee, Wisconsin 53203, USA

Abstract Burbot Lota lota exhibit four previously known reproductive strategies in the Great Lakes region. In this paper we review those strategies and provide evidence for a fifth one—delayed deepwater spawning. The four known, shallow-water strategies are as follows: (1) spawning by self-sustaining, landlocked populations, (2) spawning in tributaries in winter and the exit of larvae to a Great Lake, (3) spawning by residents in a spawning stream with access to a Great Lake, and (4) spawning on unconsolidated and rocky areas in shallow water in winter in the lake proper. Resident, landlocked populations exist in some Michigan and Wisconsin rivers (e.g., the Muskegon River in Michigan). The evidence for winter tributary spawning is the appearance of newly hatched Burbot in the St. Marys and Bark rivers during April–June. Evidence for Burbot juveniles leaving spawning streams is U.S. Fish and Wildlife Service tributory mouth trap data. The evidence for winter nearshore spawning comes from power plant monthly entrainment studies (Mansfield et al. 1983). Our proposed fifth strategy is spring and summer spawning at deep reefs, where there is probably cobble or boulder habitat. Our evidence comes from midlake reefs in Lake Michigan and offshore areas of Lake Huron: (1) we collected adult Burbot at midlake reefs in Lake Michigan, (2) we collected many Burbot larvae (many of which were newly hatched) from Lakes Michigan and Huron in June–August, and (3) we

Downloaded by [Department Of Fisheries] at 22:57 25 November 2013 collected a Burbot egg in a PONAR grab in mid-July from 73 m in southern Lake Huron. An important question remains, namely, which life history strategy provides the highest recruitment success for this species. It may be that adaptability ensures the survival of this important, top-predator fish during periods of crisis (e.g., encounters with dams, Sea Lamprey Petromyzon marinus predation).

Burbot Lota lota occupy the widest range of depths of Great Lakes. In Lake Baikal (Siberia), their depth range is given as Lakes basin fishes, ranging from anthropogenically and perhaps ranging from tributaries to 200 m (Sideleva 2003). There are historically landlocked populations in small streams to fish oc- two other Great Lakes fish with such a wide range in depths: the cupying depths of at least 300 m (Boyer et al. 1989) in the Great Slimy Sculpin Cottus cognatus, which inhabits cold streams and

*Corresponding author: [email protected] Received September 10, 2012; accepted April 6, 2013 1733 1734 JUDE ET AL.

the Great Lakes but whose densities diminish with depth, and and recruitment strategy of larval fish is influenced by water Lake Trout Salvelinus namaycush, which is found from great temperature (Cushing 1982; Steele 1985; Cardinale et al. 2008; depths within the Great Lakes to some lakes inland. This range Kallasvuo et al. 2010), which in turn affects the productivity of of depth and habitat means that their life histories must be suit- zooplankton prey and larval fish predators, such as Alewives able for diverse ecological conditions. The early life histories Alosa pseudoharengus (Madenjian et al. 2002). In addition, of Burbot differ dramatically from those of most other fishes, currents can impact the success of larval fishes by transport- since Burbot eggs and larvae are very small and semibuoyant, a ing them to areas that are optimal (or not) for survival (e.g., strategy consistent with that of all other members of the Gadidae Yellow Perch Perca flavescens; Beletsky et al. 2007). As far as family, all of which are marine. However, some species, such as currently recorded, Burbot exhibit four early life history strate- the Atlantic Tomcod Microgadus tomcod, are anadromous. gies: (1) adults come from the Great Lakes or major rivers to Burbot are important ecologically as keystone predators. spawn in tributaries in winter (fish larvae are then transported They are top predators in all five Great Lakes and therefore im- from the tributaries to a Great Lake; Scott and Crossman 1973; portant ecologically for the roles they play as consumers of prey Jude et al. 1998); (2) adults from the Great Lakes spawn in trib- and potential competitors with other predators like Lake Trout utaries in winter, but the larvae remain there and may or may (Martin and Olver 1980). Recently, Burbot have been switch- not return to a Great Lake as juveniles (D. J. Jude, unpublished ing from native prey to Round Goby Neogobius melanostomus data); (3) populations that are landlocked (with maximum sizes on shallow reefs (Hensler et al. 2008; Jacobs et al. 2010). Since much smaller than those of Great Lakes fish) are self-sustaining, Burbot eat sculpins (Cottus spp. and Deepwater Sculpin Myoxo- spawning where they occur (Dixon and Vokoun 2009; Seelbach cephalus thompsonii; Fratt et al. 1997), which are egg predators and Wiley 1997); and (4) adults spawn on rocky and unconsoli- (Wojcik et al. 1982), and Burbot are known Lake Trout egg dated substrates in a Great Lake in winter (Scott and Crossman and fry predators (Stauffer and Wagner 1979; Jones et al. 1995; 1973). In this paper we review the known life history strategies Janssen et al. 2006), their abundance may impact Lake Trout of Burbot. We explore the spring distribution of Burbot dur- egg survival on spawning reefs (Janssen et al. 2007). Unlike ing thermal bar formation and stratification on Lake Michigan the other deepwater top predator, Lake Trout, Burbot were not midlake reefs, examine potential differences between the east largely extirpated due to Sea Lamprey Petromyzon marinus pre- and west sides of Lake Michigan as tributary sources of Burbot dation in the Great Lakes in the 1950s and this may be related to larvae, and document a new spawning behavior in hypolimnetic their spawning in diverse habitats. Burbot are known to spawn waters in spring and summer. in rivers (Bjorn 1939; Jude et al. 1998), in the Great Lakes (December–January: Lake Michigan; March–April: Lake Erie; and February–March: Lake Superior), and in Canadian inland METHODS lakes as well as Lake Constance, Germany. Prior to the present Because Burbot are generally not target species for fishery study it has been assumed that they spawn from January to management, collections usually do not specifically target them; March (Auer et al. 1982; Donner and Eckmann 2011). How- as a result, our specimens were largely bycatch obtained during ever, in Lake Constance spawning can be delayed until the end other projects or from general surveys. of May, when water temperatures rise to 5.6◦C (Fischer 1999). Density of adult Burbot at Lake Michigan’s midlake reef Because the Great Lakes are slow to warm, it is possible that complex.—Adult Burbot are common to abundant among boul- spawning in deepwater, should it occur, is also delayed. ders at two subreefs of the midlake reef complex (MLRC): Burbot eggs are small (1.3–1.8 mm) and are broadcast ran- Sheboygan Reef (about 43◦20.5N, −87◦09W) and East Reef domly at night by adults that form writhing masses during re- (about 43◦01N, −87◦21W) in southern Lake Michigan (see

Downloaded by [Department Of Fisheries] at 22:57 25 November 2013 production (Scott and Crossman 1973). Eggs are semibuoyant Figure 1 and Janssen et al. 2006 for map). Our only attempt at when first spawned, then become demersal. The larvae hatch quantification was at East Reef as a byproduct of searching for at around 3–4 mm (Auer 1982), rise to the surface to fill their and collecting Lake Trout fry using a remotely operated vehicle air bladders, and have to initiate feeding 6–9 d posthatching (ROV) modified for electroshocking Lake Trout fry (described (Fischer 1999). In the Great Lakes larval Burbot undergo a diel in Janssen et al. 2006). The sampling date was August 6, 2009, vertical migration (Oyadomari and Auer 2004), presumably to and the ROV had recently been outfitted with a GPS tracking avoid predation, to pursue migrating prey, and/or for energetic system that allowed mapping. optimization (Donner and Eckmann 2011). They are initially Offshore larval fish and egg collections.—A 1-m × 2-m pelagic, then go through a settlement phase at around 21 mm or neuston net with 500-µm mesh was used in duplicate at less at about 68 d; they settle to the bottom, then migrate along several standard U.S. Environmental Protection Agency (EPA) the profundal zone (16 m maximum depth for Oneida Lake monitoring stations in Lake Michigan (Figure 1; Barbiero and and 250 m for Lake Constance) toward shore, where they pre- Tuchman 2001, 2004, 2009); the net was deployed in the top 2 m sumably stay (Clady 1976; Fischer 1999; Hoffman and Fischer of water during 2007 and 2008 during day and night, depending 2001). on arrival time at each station. The net was towed at 1–2 m/s Water temperature and the currents in the Great Lakes are and was equipped with a Rigosha flowmeter. Samples were likely critical factors that affect Burbot survival. The spawning preserved in ethanol. Fish larvae were removed from the sample BURBOT EARLY LIFE HISTORY STRATEGIES 1735

FIGURE 1. Map of Lakes Michigan, Huron, and Superior showing the standard EPA monitoring stations (plus signs and triangles), the various towns and sites mentioned in the text, and the location of the midlake reef complex in southern Lake Michigan.

in the laboratory, counted, measured to the nearest 0.1 mm, a 1.4-m-wide × 2-m-deep Tucker trawl equipped with 500-µm and identified following Auer et al. (1982). Larval fish were mesh (Wang et al. 2012). Sampling was conducted on April also saved from the routine EPA monitoring of zooplankton 23 and May 3, 4, and 7, 2007; April 17, 21, and 28 and May and crustaceans Mysis spp. during surveys on Lakes Michigan, 5, 2008; and April 28 and May 4 and 14, 2009. The proce- Downloaded by [Department Of Fisheries] at 22:57 25 November 2013 Huron, and Superior during 2007–2008. Briefly, zooplankton dures of Rice et al. (1987), Nash and Geffen (1991), Geffen and and rotifers were collected at about 20 stations per lake Nash (1992), and Dettmers et al. (2005) were followed. As the (Figure 1; and see Barbiero and Tuchman 2001, 2002, 2004) thermal bar moved offshore, so did the transects and depths sam- using a 153-µm-mesh net (100 m deep, vertical tow) and pled (the nearshore transect depth increased from 17 to 37 m; 63-µm-mesh net (20 m deep, vertical tow), respectively, at the offshore depths increased from 39 to 78 m). We used a step- each station. Samples were preserved in a 10% formaldehyde wise, oblique tow starting at 10 m deep, with a 1-m rise every solution. Mysis tows were only conducted at night using a 1- 3 min (30- min tow). Samples were stored in 95% ethanol. Fish m × 1-m, 500-µm-mesh net that narrowed to a panel of 250 µm larvae were removed, identified, and measured to the nearest on the bottom. The net was lowered to within 3–5 m of the millimeter. If the sample size was >100, a subsample was taken bottom and towed vertically to sample the entire water column. using a Folsom plankton splitter, such that a sample of 100 or During 2007–2009, sampling was also conducted at night less was processed and TLs measured for the length-frequency along transects in Lake Michigan on both sides of the thermal analysis. A General Oceanics flowmeter mounted in the center bar off Milwaukee, Wisconsin (in a variety of locations from of the net allowed calculation of water volume and larval fish 43◦05N, −87◦50Wto43◦05N, −87◦42W; Figure 1) using densities. 1736 JUDE ET AL.

At the MLRC stations on Lake Michigan, sampling was and 5 m as well as the surface for 2 min each (for some stations conducted during the day at two locations with similar that were <20 m we modified the protocol to sample the entire bottom depths—coastal (55–62 m deep; 43◦01.3404N, water column). Samples were preserved in ethanol. In addition, −87◦42.4820Wto43◦01.6125N, −87◦43.1299W) and East a standard PONAR grab sampler was deployed at the same 15 Reef (55–60 m deep; 43◦01.5703N, −87◦21.1581W)—from stations to sample fish eggs. 2007 to 2009 (Figure 1). Reef locations were about 27 km Spawning and hatching date calculations for larval Bur- offshore of the coastal location. East Reef, which is composed bot.—We calculated the probable spawning and hatching dates of bedrock with a cobble and sand veneer, rises from a depth for four size-classes (4, 9, 10, and 15 mm) of larval Burbot col- of 100 m or more (see Janssen et al. 2006 for the bathymetry lected in mid-July and mid-August to estimate spawning times of its western face). The target species was Bloater Coregonus during nonwinter months, which have not been documented for hoyi larvae, which are near the surface during the day. We used the Great Lakes. Spawning date back-calculations were based the same Tucker trawl and sampling steps and procedures as on the findings of a 32-d incubation period at 4◦C(Jager¨ et al. used with the thermal bar transect sampling (Wang et al. 2012). 1981) and a 30–40-d incubation period at 4–5◦C (Donner and Sampling was conducted on six dates: June 25 and July 6, Eckmann 2011); there was a 70-d incubation period for Bur- 2007; June 23 and July 1, 2008; and June 18 and July 14, 2009. bot eggs if they were in 0–1.5◦C temperatures (McCrimmon Two-way analysis of variance (ANOVA) was used with date 1959; Muth 1973), which would occur during winter in most and location as factors to compare the larval fish density and rivers and nearshore zones of the Great Lakes. We choose an size of Burbot between these two locations. incubation period of 35 d at 4–5◦C as a compromise value. If During July 9–12, 2012, a 0.75-m-diameter, 571-µm-mesh spawning occurred during stratification, the water temperatures net equipped with a Rigosha flowmeter was deployed at 15 on the bottom would be expected to be near 4–5◦C. We mea- sites throughout Lake Huron (see Table 1 for GPS coordinates) sured bottom temperatures in Lake Huron on July 12 at 4.3◦C. using a stepped-oblique protocol for 10 min for each tow at an To determine the hatching date, we calculated the number of average speed of 1.3 m/s on the EPA vessel Lake Guardian.The days that a given-sized larva was alive based on length-at-age tow consisted of deploying the net with a depressor to 20, 15,10, data from Lake Constance (Fischer 1999). We back-calculated

TABLE 1. Details of Burbot larvae collected in EPA routine surveys (ZP = zooplankton tows, MY = Mysis tows) and additional fish larvae tows (FL) done during 2007–2008 in Lakes Michigan (MI), Huron (HU), and Superior (SU). Individual lengths of each fish larva caught are given under length.

EPA Sample Station Sample Length Density Date stationa type depth (m) depth (m) (mm) (no./1,000 m3) Spring 17 Apr 2008 MI 34 FL 159 1 9.5 1 18 Apr 2008 MI 47 FL 191 1 6 1 Summer 2 Aug 2007 MI 11 MY 126 124 5,5 16 1 Aug 2007 MI 18 ZP 160 100 5.5 51 5 Aug 2007 HU 61 MY 118 116 5.5, 5.5 17 20 Aug 2007 SU 16 ZP 178 100 15, 8.5 102

Downloaded by [Department Of Fisheries] at 22:57 25 November 2013 4 Aug 2007 HU 32 ZP 83 81 3 63 5 Aug 2007 HU 48 ZP 112 100 4.5 51 HU 54 ZP 124 100 5 51 1 Aug 2008 MI 11 ZP 125 100 6.5 51 2 Aug 2008 MI 27 ZP 104 100 5 51 MI 18 MY 160 157 8 6 MI 27 MY 104 101 7, 8 20 MI 18 MY 160 157 8 6 6 Aug 2008 HU 97 MY 43 40 13.5, 13.5 50 7 Aug 2008 HU 27 MY 56 54 14.5 19 HU 95 MY 68 65 9,9 31 HU 27 MY 55 52 14.5 19 6 Aug 2008 HU 45 ZP 98 96 5, 7.5, 6, 4.5, 4 212 HU 37 ZP 74 71 4 72

aSee Figure 1. BURBOT EARLY LIFE HISTORY STRATEGIES 1737

its hatching date by going back in time that many days from electroshocker. Burbot deep within the layers of the boulders the collection date. Lake Constance is a large, deep oligotrophic were likely missed. The Burbot observed via ROV on Julian’s lake comparable to Lakes Michigan and Huron with respect to Reef (23–41-m deep) in southern Lake Michigan had a similar its thermal cycle. Burbot grew slowly during the first 23 d of density of 1.4/100 m2 (Edsall et al. 1993). life (average calculated value, 0.084 mm/d) but faster thereafter Larval Burbot distribution during thermal bar formation.— (0.57 mm/d). Using this data, a 4–9-mm Burbot larva collected We sampled for larval Burbot off Milwaukee on the western side in July and August in the Great Lakes would be from 1 to 29 d of Lake Michigan (Figure 1) during thermal bar formation in old, while a 10–15-mm larva would be 30–39 d old. spring. Inside the thermal bar, sampling depths increased from 17 to 37 m (2.4–10.0 km from shore) as the bar moved offshore. The depths of the offshore locations increased from 39 to 78 m RESULTS (8.0–18.9 km from shore). Sampling occurred within (>4◦C) ◦ Offshore Larval Fish Collections and outside the bar (<4 C) three to four times during April 17– Burbot larvae were collected in Lakes Superior (small sam- May 14, 2007–2009 to determine whether thermal conditions or ple), Michigan, and Huron during spring and summer 2007– prey densities favored the survival of Burbot larvae. No Burbot 2008; these fish had a mean density of 124/1,000 m3 (Table 1) larvae were found during these studies despite our collecting 11 and ranged in size from 3 to 15 mm (Figure 2). The vertical zoo- samples (5.5 h of sampling effort), suggesting that no spawning plankton and Mysis tows showed that Burbot were present dur- occurred in the local tributaries to Lake Michigan on the western ing both the spring and summer cruise periods. Newly hatched side; no resident Burbot populations exist in these streams either 3–4-mm individuals were present during August, demonstrat- (Becker 1983). ing that spawning had occurred later in the year, a deviation Larval Burbot distribution around the midlake reefs.—We from winter spawning behavior. Based on 90–160 degree-days collected 4,074 Burbot larvae in Tucker trawls during June– to hatch in 4◦C water and the fact that fish larvae may not show July 2007—2009 (length range, 3.4–14 mm) at East Reef and an increment on the otolith for 2 d (Fischer 1999), we estimated the coastal site (Figure 1); many Burbot larvae were newly that these larvae were spawned sometime between May 22 and hatched. We aged eight Burbot larvae that were 3–4 mm long June 19. and found that they averaged3dold.Densities were significantly higher on the reef (6–221/1,000 m3) than in the coastal area (5– 3 Nearshore and Midlake Reef Collections 55/1,000 m ) (Figure 3; F1, 5 = 13.6, P = 0.014). Date was Adult Burbot at the midlake reef complex.—On August 6, marginally significant (F5, 5 = 4.9, P = 0.052). 2009, we extracted 13 adult Burbot from dense, multilayered Total lengths of larval Burbot were compared between coastal boulder cover in an area approximating 1,500 m2. This density locations and the MLRC using two-way ANOVA, with the fac- (about 1/100 m2) should be considered minimal because Burbot tors being sampling date and location (Figure 4). The sam- were not the target species, sampling in the area was not spatially ples from June 25, 2007, were accidentally dehydrated, so we comprehensive, and we could only see Burbot affected by the deleted that date from the length-frequency analysis. Larval Burbot lengths were log10-transformed prior to analysis. The ANOVA showed a statistically significant interaction between date and location (F4, 356 = 3.03, P = 0.018), so the main effects were not interpretable. The post hoc test (Dunnett’s test com- paring the lengths at the two locations by date, for five pairwise comparisons) showed that the MLRC Burbot larvae were sig-

Downloaded by [Department Of Fisheries] at 22:57 25 November 2013 nificantly smaller than the coastal larvae on July 6, 2007 (P < 0.05) and June 23, 2008 (P = 0.01). There were no statistically distinguishable differences in length between sites for the other three dates. While this is a mixed result, it is consistent with larval Burbot originating at or near the MLRC and then drifting away.

Lake Huron Collections Distribution and density of Burbot larvae.—Larval Burbot were collected from Lake Huron during July 9–12, 2012, to determine the spatial and length distributions throughout the lake (Figure 1). From 0 to 14 larval Burbot were collected at FIGURE 2. Length-frequency distribution of larval Burbot collected in Mysis the 15 stations sampled, with the highest density being ob- and zooplankton vertical tows as part of the EPA monitoring program in Lakes served in Georgian Bay (Table 2). Densities ranged from 0 to Michigan, Huron, and Superior, 2007–2008. 255/1,000 m3, while total lengths ranged from 4 to 13 mm. There 1738 JUDE ET AL.

FIGURE 3. Densities of Burbot larvae along the eastern coast (about 50–60 m bottom depth) and on East Reef (midlake reef complex; depth range, 52–83 m) during June–July 2007–2009 in southern Lake Michigan. The two locations are separated by about 27 km, and the bottom depth adjacent to East Reef is >100 m. [Figure available in color online.]

did not seem to be any trends in density across Lake Huron, ex- eggs at 4–5◦C takes approximately 30–40 d (Jager¨ et al. 1981; cept that no Burbot were found in eutrophic Saginaw Bay or Donner and Eckmann 2011), making the spawning date around the shallow station USGS 17 just outside the bay (in contrast June 3–13. to the deeper station in Georgian Bay, which had the highest Burbot lengths and spawning times.—Our length-frequency density of all stations sampled). More Burbot were collected data from the Lake Michigan midlake reefs (Figure 4), together during the day (255/1,000 m3) than at night (59/1,000 m3)at with those from the EPA collections from Lakes Michigan and the Georgian Bay site. Presence of newly hatched fish larvae in Huron (Figure 2) and those from Lake Huron in 2012 (Table 2) the 4-mm range suggested recent hatching by Burbot in Lake during June–August show two cohorts, one composed of newly Huron. These larvae had to have originated from spawning in hatched and young individuals (4–9 mm) and one composed of offshore Lake Huron in early spring, since the incubation of larger individuals (10–15 mm). Based on the incubation times

TABLE 2. Details of Burbot larvae collected during July 9–12, 2012 in Lake Huron in the top 20 m with 10-min tows with a 0.75-m-diameter, 500-µm-mesh plankton net. See Figure 1 and Barbiero et al. (2009) for station descriptions. Station designations: D = day; N = night.

Station GPS coordinates No. Density Size range (mm) HU 37 44.7610◦N −82.7829◦W 8 146 7–11 ◦ − ◦ Downloaded by [Department Of Fisheries] at 22:57 25 November 2013 USGS 40 44.3295 N 83.2255 W 1 18 11 USGS 73 44.3188◦N −82.9192◦W 1 18 11 USGS 64 44.3277◦N −83.0073◦W 4 73 9–12 USGS 17 44.3317◦N −83.2527◦W0 0 SAG BAY 43.9420◦N −83.6237◦W0 0 HU 27 44.1982◦N −82.5040◦W 1 18 13 HU 15 44.0000◦N −82.3500◦W 7 121 5–12 HU 93 44.0993◦N −82.1173◦W2 486 HU 32 44.4532◦N −82.3417◦W 3 55 6–9 G. BAY 1-D 45.3043◦N −81.2399◦W 14 255 4–8 G. BAY 1-N 45.3043◦N −81.2399◦W 5 59 8–11 G.BAY 2-D 45.7600◦N −81.1878◦W 2 70 6–7 HU 48 45.2778◦N −82.4521◦W 2 23 9–11 HU 53 45.4492◦N −82.9159◦W 8 205 7–8 HU 45 45.1333◦N −82.9833◦W 8 180 6–11 BURBOT EARLY LIFE HISTORY STRATEGIES 1739

FIGURE 4. Length frequency histograms for larval Burbot collected at coastal and midlake reef sites on two sample dates/year during 2007–2009 (Figure 1).

Downloaded by [Department Of Fisheries] at 22:57 25 November 2013 FIGURE 5. Sizes of Burbot larvae (4, 9, 10, and 15 mm) collected during mid- at 0–1.5◦C and 4–5◦C, the 4-mm Burbot collected in mid-July July (top panel) and mid-August (bottom panel) and their possible spawning dates (x-axes; spawn-70 = based on an incubation time of 70 d at 0–1.5◦C and mid-August would have been spawned sometime during the [McCrimmon 1959]; spawn 35 = based on an incubation time of 35 d at 4–5◦C period May 5–July 5, while 9-mm Burbot collected during mid- [Jager¨ et al. 1981; Donner and Eckmann 2011]). An age-length key generated August would have been spawned sometime during the period using growth estimates from Fischer (1999) was used to estimate the number May 9–June 13 (Figure 5). Even some of the larger larvae (10– of days from the collection date (triangles) to the hatching date (cross symbols) 15 mm) collected in mid-July, would have been spawned during (see Methods). May 3–11 if the eggs incubated in 4–5◦C water, while all of the 10–15-mm larvae collected in mid-August under either regime Michigan. The water temperature was 0.4◦C (Jude et al. 1979). would have been spawned over the period April 29–June 11. An egg was found in a PONAR sample collected on July 12, 2012, in southern Lake Huron (EPA station HU 32; Table 1), Egg Deposition Nearshore and Offshore in where the depth was 73 m and the water temperature was 4.3◦C. Lakes Michigan and Huron This egg was subsequently excised and the embryo inside was Closest to shore, Burbot eggs were found during January identified as a Burbot. Assuming that this egg was spawned in in the eastern Lake Michigan beach zone near Port Sheldon, and remained in 4◦C water, it would take approximately 30–40 d 1740 JUDE ET AL.

to hatch (Jager¨ et al. 1981; Donner and Eckmann 2011), making (Becker 1983) and Burbot larvae were found around Ozaukee its approximate spawning date June 3–13, 2012. This is evidence County and Haven, Wisconsin (Mansfield et al. 1983), which of spawning offshore in the middle of the southern basin of Lake are about 90 km from Milwaukee, where some of our midlake Huron during times undocumented in the literature. sampling occurred. Burbot eggs were also identified in the diet of three Slimy Another question arising from this study is whether spawning Sculpins by genetic analyses. Fish were sampled in a single Burbot from Lake Michigan ascend rivers like the Ford River trawl haul offshore of Frankfort, Michigan, at 128 m on April (which has no dams), where juvenile Burbot have been found 20, 2010 (J. Londer, personal communication, U.S. Geological 224 km upstream, or whether these juveniles are the result of Survey [USGS], Ann Arbor, Michigan). As Slimy Sculpins are local spawning by nonmigrating residents. Spatial segregation benthic with a small home range, these data also document can lead to speciation. Genetic studies would help to unravel probable Burbot spawning in early spring in deep waters of this conundrum. Lake Michigan. We believe that some juvenile Burbot, especially those close to a Great Lake, employ a second life history strategy, leaving DISCUSSION the tributary in which they were spawned and returning to the lake rather than spending their entire lives in the tributary. Sea Documented Burbot Spawning Strategies Lamprey traps placed in many tributary mouths by the U.S. Fish There appear to be four major life history strategies that Bur- and Wildlife Service (USFWS) during spring catch juvenile Bur- bot use to overcome environmental and biological obstacles to bot (J. Slade and D. Keffer, USFWS, personal communication), survival. First, Burbot spawn during winter in tributaries and which suggests that some juveniles exit tributaries. remain in rivers that often extend >200 km upstream from a In the third strategy, spawning occurs during winter in rivers Great Lake. We also documented with field sampling, distribu- and streams, with larvae exiting in large numbers during April— tion records for Michigan (Bailey et al. 2004) and Wisconsin May. The evidence for this is the appearance of larvae in late- (Becker 1983), and survey records of the Michigan Depart- March–May samples in Lake Michigan near tributaries (Mans- ment of Natural Resources (MDNR) (Bassett and Kramer 2009; field et al. 1983) and the appearance of 4-mm larvae from April Bassett and Herman 2009), and the Michigan Rivers Inventory 26 through May 30 in the St. Marys River (Jude et al. 1998). (Seelbach and Wiley 1997) that juvenile and adult Burbot (gen- Data obtained offshore from the Ford River in 2012 showing erally <300 mm) are widespread throughout the state of Michi- high densities of 4–5-mm Burbot confirm ongoing spawning in gan and in northern Wisconsin, especially around Green Bay. northern Lake Michigan (J. Schaeffer, USGS, personal commu- Michigan DNR Burbot collections from the Ford River made nication). The lack of Burbot in Wisconsin streams corresponds during electroshocking surveys during August 2009 at three with the lack of Burbot larvae from our thermal bar studies, sites near Hendersen (46.10863◦N, 87.84127◦W) showed that which contrasts starkly with the eastern side of Lake Michigan, Burbot ranged from 76 to 229 mm (N = 53; A. Abrahamson, where there are Burbot in streams and high densities of Burbot MDNR, personal communication). This site is approximately larvae in nearshore waters. 224 km from the mouth of the river, which runs into Green In the fourth strategy, spawning occurs during winter in Bay near Escanaba, Michigan. These populations appear to be nearshore and offshore sites (including unconsolidated sites and self-sustaining, and some are landlocked by dams. For exam- those with rocky substrates). The evidence for this includes the ple, on August 4, 2011, we electroshocked 25 juvenile Burbot discovery of Burbot eggs in the nearshore zone of eastern Lake from the Muskegon River near Hersey, Michigan, that were 65– Michigan during January (Jude et al. 1979) and in the stomachs 210 mm long (Jude, unpublished data). Interestingly, this section of Slimy Sculpin captured at 128 m offshore of Frankfort, Michi-

Downloaded by [Department Of Fisheries] at 22:57 25 November 2013 of the river (Osceola County: T17NR9WS11) is isolated by two gan, on April 20, 2010 (J. Londer, personal communication). In dams and happened to have been treated with rotenone by the addition, most published information on Burbot documents this MDNR for fish assessment (those data also show an abundance type of spawning in most of the areas of the Great Lakes where of Burbot in the system during August 5, 1993; MDNR 1993); Burbot occur (McCrimmon 1959; Bailey 1972; Muth 1973; 397 Burbot were collected that ranged from 71 to 270 mm. Scott and Crossman 1973). These sizes (maximum near 300 mm) appear to be typical of Lastly, we documented spawning at nearshore and offshore these landlocked populations. Dixon and Vokoun (2009) stud- sites in the hypolimnion during spring and summer in Lakes ied Burbot in the Hoosatonic River in Connecticut, collecting Michigan and Huron. There is no evidence that any spawning fish from 84 to 356 mm; the fish were mature at 2 years old, by Burbot occurred in tributaries during this time. We based and none >500 mm were seen. Lake Superior fish are mature this new finding on the appearance of large numbers of newly at 280 mm (Bailey 1972), so these fish appear to be mature hatched Burbot on the midlake reefs in late June–August 2007– enough to spawn around 300 mm. In contrast, no juvenile or 2009, which argues for spawning there later in the year than has adult Burbot have been reported in adjacent streams or rivers been documented in the Great Lakes and rivers, e.g., late April– along the western shore in southern Lake Michigan. However, May (Scott and Crossman 1973; Mansfield et al. 1983). We base juveniles and adults were present farther north in Green Bay this deduction on four lines of evidence. First, we routinely see BURBOT EARLY LIFE HISTORY STRATEGIES 1741

and collect adult Burbot on the reefs, as did Edsall et al. (1993). stances. It is unknown whether any of these fish leave their natal Second, we collected 4,074 Burbot larvae in Tucker trawls from tributaries as juveniles or adult fish but, as noted, juveniles have June 18 to July 14, 2007–2009, many of which were newly been captured in Sea Lamprey traps near the mouths of sev- hatched, and more were collected on the reefs than in nearshore eral rivers where sampling was ongoing. In the Great Lakes, waters. Third, these results were corroborated by the appearance winter spawners deposit eggs in nearshore and offshore sandy of many newly hatched Burbot larvae in the EPA’s zooplankton and rocky areas (Scott and Crossman 1973; Jude et al. 1979). and Mysis samples collected during August 2007–2009 at sev- In the nearshore zone there is higher productivity and tempera- eral stations (43–190 m) in Lakes Michigan and Huron. Fourth, tures favoring the survival of larval Burbot, but this advantage the calculation of spawning and hatching times based on pub- presumably comes with a higher mortality rate from predators, lished incubation times (McCrimmon 1959; Jager¨ et al. 1981) especially Alewives (Madenjian et al. 2002). Offshore water and growth rates (Fischer 1999) clearly show that spawning and temperatures are colder and zooplankton prey are usually less hatching occurred later than published winter spawning dates, abundant than nearshore (with the possible exception of the some in June and early July. These spawning times are far reefs, where currents and the elevated structure [sea mount] removed from published spawning times of December–March can concentrate both larvae and zooplankton prey). Does this (Bjorn 1939; Mansfield et al. 1983; Nash and Geffen 1991; increased diversity in spawning improve their adaptability, or Jude et al. 1998). These authors showed that there is a pulse of do the deep reef Burbot become reproductively marooned, with newly hatched Burbot larvae from winter spawning from late larvae hatching too late for optimal survival? Which spawning March through June, with peaks in May and lesser numbers of areas (tributaries, nearshore sandy and rocky reefs, or offshore fish being collected in June. These changes are not due to cli- reefs) produce the most recruits is a question that may be im- mate change, since studies have shown that there have been no portant for managers trying to restore Burbot populations. This long-term temperature trends in Lake Michigan over recent time question should be addressed using genetic analyses or trace (McCormick and Fahnenstiel 1999). Using the 70-d incubation chemical or stable isotope analyses (Campana 1999) of otoliths. temperatures at 0–1.5◦C of McCrimmon (1959) would suggest Burbot larvae collected during the peak mid-May pulse were Dynamics of Temperature-Related Burbot Larvae spawned with a midpoint of March 5. Mansfield et al. (1983) Distribution concluded that 3–4-mm Burbot from northern Lake Michigan With potential larval Burbot sources ranging from tribu- collected in June were spawned in March or early April. taries to extreme depths, there is an interesting question as to Finally, the discovery of a Burbot egg in a PONAR sam- which source of larval Burbot recruits where and how water ple from 73 m in Lake Huron on July 12 corroborated mid- temperatures affect this relationship. Because Burbot inhabit a lake spawning during stratification. Interestingly enough, Bai- wide range of habitats with very different annual temperatures ley (1972) reported that Burbot collected in late January and regimes, it is not surprising that the composite spawning season February from nearshore Lake Superior were spent; however, is quite long, perhaps in excess of half a year. The geographic those collected offshore had not spawned in January and March, variation in the temperature regime has certain surprises. For suggesting spawning later in the year. example, for the East Reef the warmest time of year at its 50-m These combined findings document spawning by Burbot on summit is late November (7–8◦C), when the thermocline de- the midlake reefs and deep offshore sites in spring and summer; scends (Janssen et al. 2007). Consider the fate of a Burbot larva the same probably occurs on other, similar reefs in the Great in a coastal stream flushed into the main basin in spring. For Lakes and provides more reproductive diversity for enhancing the four deepest Great Lakes (Superior, Michigan, Huron, and survival. A similar finding of late spawning (Burbot larvae col- Ontario), coastal water begins warming, partially fed by trib-

Downloaded by [Department Of Fisheries] at 22:57 25 November 2013 lected in August) was made by Donner and Eckmann (2011) utaries, which warm more quickly. The offshore temperature and Probst and Eckmann (2009a) in Lake Constance. This be- is <4◦C, and the nearshore water is physically separated from havior provides yet another adaptation to improve recruitment the coastal water by the sinking 4◦C water (the “thermal bar”). and survival in the face of catastrophic events such as climate The coastal water is both warmer and more fertile than the off- change, the loss of river habitat through the construction of cul- shore water, so the larvae in it are likely to grow faster. The verts (MacPherson et al. 2012) or dams, and the introduction thermal bar transitions first into a “thermal wedge” as the lake of exotic species such as Sea Lampreys and Alewives (Maden- warms; final full-lake stratification is a consequence of local jian et al. 2002). In fact, multiple spawning behaviors and the warming and offshore movement of the warmer coastal water. widespread presence of juveniles throughout Michigan and parts For Lake Michigan, the thermal bar generally is initiated in of Wisconsin may explain why Burbot survived the Sea Lam- April and stratification may not be complete for a month. The prey invasion while Lake Trout were extirpated in all the Great process is most completely reviewed in Mortimer (2004), and Lakes except Lake Superior (Coble et al. 1990). a recent example is in Wang et al. (2012). Note that a Burbot The source of most of the adult Burbot found in the Great larva originating in a tributary prior to the thermal bar period Lakes is unknown. Adult Burbot from tributaries are usually could be either flushed by lake currents well offshore into off- small (<300 mm) but apparently abundant in some circum- shore water in which warming is well delayed or stay nearshore 1742 JUDE ET AL.

and experience much earlier warming within the thermal bar. larvae. Areas at which strong currents intercept areas of rapid The thermal bar/thermal wedge never reached as far as the Lake shallowing (termed “abrupt topographies” in the oceanographic Michigan MLRC, so Burbot larvae emerging in the coastal zone literature) can concentrate both zooplankton and fish (Genin must wait for complete stratification before transport to offshore 2004). Where measured, the currents at the MLRC can be strong deep water. (Gottlieb et al. 1989) and their upwelling is likely responsible for frequent thermal fronts there (Ullman et al. 1998). One mecha- Spatial Differences in Production of Burbot Larvae nism discussed by Genin (2004) involves a behavioral response in Lake Michigan: East versus West by zooplankton, i.e., to swim downward in an upwelling water There appear to be large differences in the distribution and mass to maintain a favored, light-determined depth preference. occurrence of Burbot larvae on the eastern and western sides This could not only concentrate prey for larval Burbot but, be- of Lake Michigan, since none were collected during exten- cause the Burbot migrate vertically, it could concentrate them. sive studies during thermal bar formation in April–May 2007– Houghton et al. (2010) proposed another mechanism that topo- 2009 on the western side, while great numbers of larvae (2– graphically traps vertically migrating Mysis during their dawn 843/1,000 m3) were observed during March 29–June 29 in sam- descent. In Lake Constance larval Burbot undergo a diel verti- ples at power plant sites (Mansfield et al. 1983) and by Nash and cal migration with a range of about 10 m (nocturnal) to 70 m Geffen (1991) during June–July (up to 18/1,000 m3) along the (daytime) (Probst and Eckmann 2009a, 2009b). If this occurs in eastern side. More were found offshore than nearshore. Great the Great Lakes, the drifting larvae could settle and transition numbers of larval Burbot were documented by Mansfield et al. to their benthic phase when their drift encounters the bottom. (1983) on the western side of Lake Michigan at the Haven and Transitioning fish would likely concentrate at about 70 m not Ozaukee sites north of the MLRC. Apparently few or no Bur- only along coasts but also at the MLRC and other offshore bot larvae originate from Wisconsin rivers in the southern part deep reefs such as Six-Fathom Bank and Yankee Shoals in Lake of Lake Michigan; this is substantiated by Becker (1983), who Huron. Interestingly enough, Nash and Geffen (1991) found that provide no records of juvenile or adult Burbot in tributaries Deepwater Sculpin larvae on the eastern side of Lake Michigan into the main basin of Lake Michigan, south of Green Bay. became demersal in around 50–75 m, where adult prey Diporeia However, Burbot larvae are present in sometimes high densities and Mysis were in high abundances. (24,000/1,000 m3; Mansfield et al. 1983) farther north of Mil- waukee in Green Bay tributaries from April 27 to14 June 14. Settlement of Burbot Larvae to the Bottom Some of the large (10–15-mm) Burbot larvae that we found on No Burbot larvae greater than 15 mm were collected dur- the midlake reefs during June 18–July 14 were thought to be ing this study; this is consistent with what Ghan and Sprules spawned during winter, probably originating from these Green (1993) found in Oneida Lake and Miler and Fischer (2004) Bay rivers (e.g., tributaries like the Bark and Ford rivers). These found in Lake Constance, although Clady (1976) reported max- large larvae could also have originated from or near the midlake imum lengths of 20 mm in Oneida Lake studies while Oyado- reefs. In contrast, juvenile and adult Burbot are readily seen dur- mari and Auer (2004) found maximum lengths of 17.3 mm ing scuba dives on the western side (which is more rocky) but in Lake Superior. We know that after hatching larvae emerge not the eastern side (which is mostly sand). Burbot were found and go to the surface to fill their swim bladders and inhabit the in 10–15 m of water along the Door Peninsula north of Sturgeon pelagic zone (Ryder and Pesendorfer 1992). They perform a diel Bay and as shallow as 1 m during early autumn, or collected vertical migration during their 2–3 months in the pelagic zone in gill nets out to at least 35 m in the vicinity of Milwaukee (J. (Wang and Appenzeller 1998; Miler and Fischer 2004), with the Janssen, unpublished data). range of their migration increasing with their size, so that they

Downloaded by [Department Of Fisheries] at 22:57 25 November 2013 were around the 10-m stratum during night feeding (Probst and Effect of Currents on the Distribution of Burbot Larvae Eckmann 2009a) while descending to 70 m during the day We found significantly more larval Burbot on the midlake (Probst and Eckmann 2009b). At some point around 17–21 mm, reefs (90.3%) than off the reefs in the coastal zone (9.7%) dur- they settle to the bottom in the profoundal zone (Miler and ing June and July. A possible reason for this finding is the Fischer 2004) and move along the bottom to the littoral zone probable spawning grounds on the reef and the concentration of (Fischer 1999), usually in August and September (Fischer and fish larvae and zooplankton in the area by currents. For example, Eckmann 1997). Clady (1976) reported that some demersal larval Deepwater Sculpin disappeared from the nearshore wa- Burbot (44–70 mm) in Oneida Lake tended to move into and ter column after stratification, a time when larval Yellow Perch toward the mouth of tributary streams in summer. We have done also moved offshore with the currents (Dettmers et al. 2005). extensive seining, trawling, and scuba diving on both sides of Lakewide current studies (Mortimer 2004; Beletsky et al. 2007) Lake Michigan for the last 40 years and have only seen one clearly show current patterns and the advection of particles from juvenile Burbot in a trawl offshore in deep water on the eastern nearshore to offshore waters. Many hydrodynamic factors likely side. In addition, a few were impinged on a power plant screen operate at the MLRC that either render it a profitable place for a (Thurber and Jude 1985) on the eastern side (which is mostly larval fish to forage during stratification or physically aggregate sand), while moderate numbers of juvenile Burbot have been BURBOT EARLY LIFE HISTORY STRATEGIES 1743

observed at natural and artificial reefs south of Milwaukee on the reproduce if the necessary spawning cues are absent, even if western side (which is more rocky). Some juveniles were also they survive to settlement in an open Great Lake environment. collected on the midlake reefs during ROV work there (Janssen et al. 2006), suggesting that some of the Burbot that become demersal stay on the reef rather than going toward shore. This ACKNOWLEDGMENTS suggests a preference for rocky habitat, which was also observed We thank the Environmental Protection Agency, Great Lakes among inland river populations during electroshocking. National Program Office, the Great Lakes Fishery Trust, and the Great Lakes Fishery Commission for funding the larval fish work. This work was also funded by the University of Evolution of Spawning Strategies Wisconsin Sea Grant Institute under a grant from the National Because Burbot are the only completely freshwater gadid Oceanic and Atmospheric Administration, National Sea Grant fish, the possible evolutionary routes to their having both lentic College Program and the state of Wisconsin. Federal grant num- and lotic spawning, including in deepwater, are of interest. The ber NA06OAR417001, project number R/FI-I. The University Atlantic Tomcod is an anadromous gadid, which spawns during of Michigan School of Natural Resources and Environment increasing spring flows in tidal rivers and thus can be considered funded the collection of Burbot juveniles from the Muskegon a river-pulse spawner. River-pulse spawners include a diverse River. We thank Michael Wiley for help with the Michigan ensemble of fishes that time their spawning to coincide with River Inventory data set and our zooplankton/Mysis technicians spring floods; eggs and/or larvae drift with the current (Bayley for their care in removing larval fish from samples. Erin Burkett 1995). For Atlantic Tomcod the drift is complicated by tidal produced Figure 5 while Lacey Mason created Figure 1; we cycles, but it may approximate an early stage of transition from are indebted. Thanks to the reviewers and Martin Stapanian for the typical marine gadid spawning habit of having drifting eggs critical comments. and larvae. Assuming that riverine spawning by Burbot is the ancestral mode of spawning, it appears that coastal and deep- water spawning is a transition back to the typical gadid marine REFERENCES spawning with drifting eggs being adapted for either river-pulse Auer, N. A., editor. 1982. Identification of larval fish of the Great Lakes Basin spawning or spawning in large lakes with oceanic hydrodynamic with emphasis on the Lake Michigan drainage. Great Lakes Fishery Com- mission, Special Publication 82-3, Ann Arbor, Michigan. properties. The genus Coregonus, which includes river-pulse Bailey, M. M. 1972. Age, growth, reproduction, and food of the Burbot, Lota spawners whose larvae emerge during spring floods (Næsje et al. lota (Linnaeus), in southwestern Lake Superior. Transactions of the American 1995), may offer a similar example in which the deepwater cisco Fisheries Society 101:667–674. ensembles were derived from shallow-water species. Bailey, R. M., W. C. Latta, and G. R. Smith. 2004. An atlas of Michigan fishes Another species that may have undergone an evolution some- with keys and illustrations for their identification. University of Michigan Museum of Zoology Miscellaneous Publication 192. what similar to that of the Burbot is the Fourhorn Sculpin Barbiero, R. P., M. Balcer, D. C. Rockwell, and M. L. Tuchman. 2009. Recent Myoxocephalus quadricornis, which occupies shallow, Arctic, shifts in the crustacean zooplankton community in Lake Huron. Canadian coastal water. The Deepwater Sculpin evolved from the com- Journal of Fisheries and Aquatic Sciences 66:816–828. mon ancestor of Fourhorn Sculpins and Deepwater Sculpins Barbiero, R., and M. Tuchman. 2001. Results from EPA’s biological open wa- and generally lives at depths >70 m in the Great Lakes. Like ter surveillance program in the Laurentian Great Lakes: I. Introduction and phytoplankton. Journal of Great Lakes Research 27:134–154. the Fourhorn Sculpin, the Deepwater Sculpin has pelagic larvae Barbiero, R., and M. Tuchman. 2002. Results from the biological open water (Mansfield et al. 1983; Geffen and Nash 1992), and it might be surveillance program of the Great Lakes 1999. U.S. Environmental Protection that the hydrodynamic conditions that promoted the evolution Agency, Great Lakes National Program Office, EPA-905-R, Chicago.

Downloaded by [Department Of Fisheries] at 22:57 25 November 2013 of pelagic larvae for the many diverse marine species also apply Barbiero, R., and M. Tuchman. 2004. Changes in the crustacean communities to large lakes. The role of hydrodynamics in the dispersal of of Lakes Michigan, Huron, and Erie after the invasion of the predatory clado- ceran Bythotrepes longimanus. Canadian Journal of Fisheries and Aquatic Yellow Perch in Lake Michigan has recently been addressed by Sciences 61:2111–2118. Dettmers et al. (2005) and Beletsky et al. (2007). Bassett, C., and M. Herman. 2009. Whitefish River fishery status report. Recent work on rapid speciation in candelabrum corals Eu- Michigan Department of Natural Resources, Lansing. icea flexuosa suggests that it is worth examining the genetic Bassett, C., and D. Kramer. 2009. Sturgeon River fishery status report. Michigan divergence of deepwater from stream Burbot (Prada and Hell- Department of Natural Resources, Lansing. Bayley, P. B. 1995. Understanding large river–floodplain ecosystems. Bio- berg 2013). In the case of the corals, reefs were colonized by science 45:153–158. “inviable immigrants,” individuals of a sibling coral species Becker, George C. 1983. Fishes of Wisconsin. University of Wisconsin Press, that experienced some survival of larvae after colonizing a reef Madison. deeper than that on which they originated but that were unable Beletsky, D., D. Mason, D. J. Schwab, E. Rutherford, J. Janssen, D. Clapp, and to mature and reproduce. The inviability was likely because J. Dettmers. 2007. Biophysical model of larval Yellow Perch advection and settlement in Lake Michigan. Journal of Great Lakes Research 33:842–866. the appropriate spawning cues were absent at the deeper reef. Bjorn, E. E. 1939. Preliminary observations and experimental study of the For Burbot, larvae originating from coastal streams and derived Ling Lota maculosa (LeSeur), in Wyoming. Transactions of the American from parents that respond to riverine spawning cues may never Fisheries Society 69:192–196. 1744 JUDE ET AL.

Boyer, L. F., R. A. Cooper, D. T. Long, and T. M. Askew. 1989. Burbot (Lota evidence of ecological interaction. Journal of Great Lakes Research 36:312– lota) biogenic sedimentary structures in Lake Superior. Journal of Great 317. Lakes Research 15:174–185. Jager,¨ T., W. Nellen, W. Schofer, and F. Shodjai. 1981. Influence of salinity and Campana, S. 1999. Chemistry and composition of fish otoliths: pathways, mech- temperature on early life stages of Coregonus albula, C. lavaretus, R. rutilus, anisms and applications. Marine Ecology Progress Series 188:263–297. and L. lota. Rapports et Proces-Verbaux de Reunions Conseil International Cardinale, M., J. Hjelm, and M. Casini. 2008. Disentangling the effect of adult pour l’Exploration de la Mer 178:345–348. biomass and temperature on the recruitment dynamics of fishes. Pages 231– Janssen, J., D. Jude, T. Edsall, M. Paddock, N. Wattrus, M. Toneys, and 238 in G. H. Kruse, K. Drinkwater, J. N. Ianelli, J. S. Link, D. L. Stram, V. P. McKee. 2006. Evidence of Lake Trout reproduction at Lake Michi- Westpestad, and D. Woodby, editors. Resiliency of gadid stocks to fishing gan’s mid-lake reef complex: hypotheses regarding the indigenous fish and climate change. Alaska Sea Grant Program, QAK-SG-08-01, Fairbanks. and implications for restoration. Journal of Great Lakes Research 32:749– Clady, M. 1976. Distribution and abundance of larval Ciscoes, Coregonus arte- 763. dii, and Burbot, Lota lota, in Oneida Lake. Journal of Great Lakes Research Janssen, J., J. E. Marsden, C. Bronte, D. Jude, S. Sitar, and F. Goetz. 2007. 2:234–247. Challenges to deepwater reproduction by Lake Trout: pertinence to restoration Coble, D., R. Bruesewitz, T. Fratt, and J. Scheirer. 1990. Lake Trout, Sea in Lake Michigan. Journal of Great Lakes Research 33(Supplement 1):59– Lamprey, and overfishing in the upper Great Lakes: a review and reanalysis. 74. Transactions of the American Fisheries Society 119:985–995. Jones, M. L. K., G. W. Eck, D. O. Evans, M. C. Fabrizio, M. H. Hoff, P. Hudson, Cushing, D. H. 1982. Climate and fisheries. Academic Press, New York. J. Janssen, D. Jude, R. O’Gorman, and J. F. Savino. 1995. Limitations to Dettmers, J. M., J. Janssen, B. Pientka, R. S. Fulford, and D. J. Jude. 2005. Ev- Lake Trout (Salvelinus namaycush) rehabilitation in the Great Lakes imposed idence across multiple scales for offshore transport of Yellow Perch (Perca by biotic interactions occurring at early life stages. Journal of Great Lakes flavescens) larvae in Lake Michigan. Canadian Journal of Fisheries and Research 21(Supplement 1):505–517. Aquatic Sciences 62:2683–2693. Jude, D. J., F. J. Tesar, and H. T. Tin. 1998. Spring distribution and abundance of Dixon, C. J., and J. C. Vokoun. 2009. Burbot resource selection in small streams larval fishes in the St. Marys River, with a note on potential effects of freighter near the southern extent of the species range. Ecology of Freshwater Fish traffic on survival of eggs and larvae. Journal of Great Lakes Research 24:569– 18:234–246. 581. Donner, M. T., and R. Eckmann. 2011. Diel vertical migration of larval and Jude, D., F. Tesar, J. Tomlinson, N. Thurber, T. Miller, J. Dorr III, and G. early juvenile Burbot optimizes survival and growth in a deep, pre-alpine Godun. 1979. Inshore Lake Michigan fish populations near the D. C. Cook lake. Freshwater Biology 56:916–925. Nuclear Plant for preoperational years 1973–1974. Universityof Michigan, Edsall, T., G. Kennedy, and W. Horns. 1993. Distribution, abundance, and Great Lakes Research Division, Special Report 71, Ann Arbor. resting microhabitat of Burbot on Julian’s Reef, southwestern Lake Michigan. Kallasvuo, M., M. Salnen, and A. Lappalainen. 2010. Does the zooplankton Transactions of the American Fisheries Society 122:560–574. prey availability limit the larval habitats of pike in the Baltic Sea? Estuarine, Fischer, P. 1999. Otolith microstructure during the pelagic, settlement and ben- Coastal and Shelf Science 86:148–156. thic phases in Burbot. Journal of Fish Biology 54:1231–1243. MacPherson, L., M. Sullivan, A. Foote, and C. Stevens. 2012. Effects of culverts Fischer, P., and R. Eckmann. 1997. Seasonal changes in fish abundance, biomass, on stream fish assemblages in the Alberta foothills. North American Journal and species richness in the littoral zone of a large European lake, Lake of Fish Management 32:480–490. Constance, Germany. Archiv fuer Hydrobiologia 139:433–448. Madenjian, C. P., G. L. Fahnenstiel, T. H. Johengen, T. F. Nalepa, H. A. Fratt, T. W., D. W. Coble, F. Copes, and R. E. Bruesewitz. 1997. Diet of Burbot Vanderploeg, G. W. Fleischer, P. J. Schneeberger, D. M. Benjamin, E. B. in Green Bay and western Lake Michigan with comparison to other waters. Smith, J. R. Bence, E. S. Rutherford, D. S. Lavis, D. M. Robertson, D. J. Journal of Great Lakes Research 23:1–10. Jude, and M. P. Ebener. 2002. Dynamics of the Lake Michigan food web, Geffen, A. J., and R. D. M. Nash. 1992. The life-history strategy of Deepwater 1970–2000. Canadian Journal of Fisheries and Aquatic Sciences 59:736–753. Sculpin, Myoxocephalus thompsoni (Girard), in Lake Michigan: dispersal Mansfield, P., D. Jude, D. Michaud, D. Brazo, and J. Gulvas. 1983. Distri- and settlement patterns during the first year of life. Journal of Fish Biology bution and abundance of larval Burbot, Lota lota, and Deepwater Sculpin, 41(Supplement B):101–110. Myoxocephalus thompsoni, in Lake Michigan. Transactions of the American Genin, A. 2004. Bio-physical coupling in the formation of zooplankton and fish Fisheries Society 112:162–172. aggregations over abrupt topographies. Journal of Marine Systems 50:3–20. Martin, N., and C. Olver. 1980. The Lake Charr Salvelinus namaycush. Pages Ghan, D., and W. Sprules. 1993. Diet, prey selection, and growth of larval and 206–269 in E. Balon, editor. Charrs: salmonid fishes of the genus Salvelinus. juvenile Burbot Lota lota (L). Journal of Fish Biology 42:47–64. Dr. W. Junk, The Hague, The Netherlands. Gottlieb, E. S., J. Saylor, and G. Miller. 1989. Currents and temperatures ob- McCormick, M., and G. L. Fahnenstiel. 1999. Recent climatic trends in Downloaded by [Department Of Fisheries] at 22:57 25 November 2013 served in Lake Michigan from June 1982 to July 1983. NOAA Technical nearshore water temperatures in the St. Lawrence Great Lakes. Limnology Memorandum ERL GLERL-71. and Oceanography 44:530–540. Hensler, S. R., D. J. Jude, and J. He. 2008. Burbot growth and diets in Lakes McCrimmon, H. R. 1959. Observations on spawning of Burbot in Lake Simcoe, Michigan and Huron: an ongoing shift from native species to Round Gobies. Ontario. Journal of Wildlife Management 23:2347–449. Pages 91–110 in V. L. Paragamian and D. H. Bennett, editors. Burbot: ecol- MDNR (Michigan Department of Natural Resources). 1993. Rotenone survey ogy, management, and culture. American Fisheries Society, Symposium 59, of the Muskegon River in Osceola County (T17N 9W S11) during 5 August Bethesda, Maryland. 1983. Michigan Department of Natural Resources, Fisheries Division Report, Hofmann, N., and P. Fischer. 2001. Seasonal changes in abundance and age Lansing. structure of Burbot Lota lota (L.) and Stone Loach Barbatula barbatula (L.) Miler, O., and P. Fischer. 2004. Distribution and onshore migration behaviour of in the littoral zone of a large pre- alpine lake. Ecology of Freshwater Fish Burbot larvae in Lake Constance, Germany. Journal of Fish Biology 64:176– 10:21–25. 185. Houghton, C. J., C. R. Bronte, R. W. Paddock, and J. Janssen. 2010. Evidence for Mortimer, C. 2004. Lake Michigan in motion: responses of an inland sea to allochthonous prey delivery to Lake Michigan’s mid-lake reef complex: are weather, earth-spin, and human activities. University of Wisconsin Press, deep reefs analogs to oceanic sea mounts? Journal of Great Lakes Research Madison. 36:666–673. Muth, K. 1973. Population dynamics and life history of Burbot Lota lota Jacobs, G. R., C. P. Madenjian, D. B. Bunnell, and J. D. Holuszko. 2010. (Linnaeus), in Lake of the Woods, Minnesota. Doctoral dissertation. Uni- Diet of Lake Trout and Burbot in northern Lake Michigan during spring: versity of Minnesota, Minneapolis. BURBOT EARLY LIFE HISTORY STRATEGIES 1745

Næsje, T., B. Jonssons, and J. Skurdal. 1995. Spring flood: a primary cue for Seelbach, P. W., and M. J. Wiley. 1997. Overview of the Michigan Rivers inven- hatching of river spawning Coregoninae. Canadian Journal of Fisheries and tory project. Michigan Department of Natural Resources, Fisheries Technical Aquatic Sciences 52:2190–2196. Report 97-3, Ann Arbor. Nash, R., and A. Geffen. 1991. Spatial and temporal changes in offshore larval Sideleva, V. G. 2003. The endemic fishes of Lake Baikal. Backhuys, Leiden, fish assemblages in southeastern Lake Michigan. Journal of Great Lakes The Netherlands. Research 17:25–32. Stauffer, T., and W. Wagner. 1979. Fish predation on Lake Trout eggs and fry Oyadomari, J., and N. Auer. 2004. Inshore-offshore distribution of larval fishes in the Great Lakes, 1973–1978. Michigan Department of Natural Resources, in Lake Superior off the western coast of the Keweenaw Peninsula, Michigan. Report 1864, Lansing. Journal of Great Lakes Research 30(Supplement 1):369–384. Steele, J. H. 1985. A comparison of terrestrial and marine ecological systems. Prada, C., and M. E. Hellberg. 2013 Long prereproductive selection and di- Nature 313:355–358. vergence by depth in a Caribbean candelabrum coral. Proceedings of the Thurber, N. J., and D. J. Jude. 1985. Impingement losses at the D. C. Cook National Academy of Sciences of the USA 110:3961–3966. Nuclear Power Plant during 1975–1982 with a discussion of factors responsi- Probst, W. N., and R. Eckmann. 2009a. Diet overlap between young-of-the-year ble and possible impact on local populations. University of Michigan, Great perch, Perca fluviatilis L., and Burbot, Lota lota (L.), during early life-history Lakes Research Division, Special Report 115, Ann Arbor. stages. Ecology of Freshwater Fish 18:527–537. Ullman, D., J. Brown, P. Cornillon, and T. Mavor. 1998. Surface temperature Probst, W. N., and R. Eckmann. 2009b. The influence of light on the diel vertical fronts in the Great Lakes. Journal of Great Lakes Research 24:753–775. migration of young-of-the-year Burbot Lota lota in Lake Constance. Journal Wang, N., and A. Appenzeller 1998. Abundance, depth distribution, diet com- of Fish Biology 74:150–166. position and growth of perch (Perca fluviatilis) and Burbot (Lota lota) larvae Rice, J. A., L. B. Crowder, and M. E. Holey. 1987. Exploration of mechanisms and juveniles in the pelagic zone of Lake Constance. Ecology of Freshwater regulating larval survival in Coregonus hoyi: do otoliths ring true? Transac- Fish 7:176–183. tions of the American Fisheries Society 114:532–539. Wang, Y., T. R. Consi, T. Hansen, and J. Janssen. 2012. The relationship between Ryder, R. A., and J. Pesendorfer. 1992. Food, growth, habitat, and community coastal Mysis diluviana abundance and spring thermal bar dynamics. Journal interactions of young-of-the-year Burbot, Lota lota L., in a Precambrian of Great Lakes Research 38(Supplement 2):68–72. shield lake. Hydrobiologia 243/244:211–227. Wojcik, J., M. Evans, and D. Jude. 1982. Food of Deepwater Sculpin, Myox- Scott, W. B., and E. J. Crossman. 1973. Freshwater fishes of Canada. Fisheries ocephalus thompsoni, from southeastern Lake Michigan. Journal of Great Research Board of Canada Bulletin 184. Lakes Research 12:225–231. Downloaded by [Department Of Fisheries] at 22:57 25 November 2013 This article was downloaded by: [Department Of Fisheries] On: 25 November 2013, At: 22:57 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Transactions of the American Fisheries Society Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/utaf20 Stability in Life History Characteristics among Burbot Populations across Environmental Gradients Peter A. Cott a b , Tom A. Johnston b c & John M. Gunn b a Department of Fisheries and Oceans , 301, 5205 50th Avenue, Yellowknife , Northwest Territories , X1A 1E2 , Canada b Laurentian University, Cooperative Freshwater Ecology Unit , 935 Ramsey Lake Road, Sudbury , Ontario , P3E 2C6 , Canada c Ontario Ministry of Natural Resources, Cooperative Freshwater Ecology Unit , 935 Ramsey Lake Road , Sudbury , Ontario , P3E 2C6 , Canada Published online: 20 Nov 2013.

To cite this article: Peter A. Cott , Tom A. Johnston & John M. Gunn (2013) Stability in Life History Characteristics among Burbot Populations across Environmental Gradients, Transactions of the American Fisheries Society, 142:6, 1746-1756, DOI: 10.1080/00028487.2013.811101 To link to this article: http://dx.doi.org/10.1080/00028487.2013.811101

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions Transactions of the American Fisheries Society 142:1746–1756, 2013 C American Fisheries Society 2013 ISSN: 0002-8487 print / 1548-8659 online DOI: 10.1080/00028487.2013.811101

SPECIAL SECTION: BURBOT

Stability in Life History Characteristics among Burbot Populations across Environmental Gradients

Peter A. Cott* Department of Fisheries and Oceans, 301, 5205 50th Avenue, Yellowknife, Northwest Territories X1A 1E2, Canada; and Laurentian University, Cooperative Freshwater Ecology Unit, 935 Ramsey Lake Road, Sudbury, Ontario P3E 2C6, Canada Tom A. Johnston Ontario Ministry of Natural Resources, Cooperative Freshwater Ecology Unit, 935 Ramsey Lake Road, Sudbury, Ontario P3E 2C6, Canada; and Laurentian University, Cooperative Freshwater Ecology Unit, 935 Ramsey Lake Road, Sudbury, Ontario P3E 2C6, Canada John M. Gunn Laurentian University, Cooperative Freshwater Ecology Unit, 935 Ramsey Lake Road, Sudbury, Ontario P3E 2C6, Canada

Abstract We examined the variation in the growth, morphometry, life history, and reproductive traits of 24 lake-dwelling populations of Burbot Lota lota from across Canada with respect to latitude and lake characteristics as well as the differences in these traits between the sexes across populations. Overall, there was stability in most of the life history traits examined vis-a-vis` the environmental gradients tested (latitude, Secchi depth, lake area, and maximum lake depth). Among populations, spawning dates ranged from the last week of January to mid-March (most commonly within the first three weeks of February), and date of spawning was not strongly related to latitude or lake character- istics. Liver size and gonad size did not vary significantly with latitude or other lake characteristics, and neither gonad size nor egg size were strongly related to maternal size. However, egg size decreased with latitude and increased with lake area. The proportion of females in nonspawning condition within populations was positively related to Secchi depth and lake area, whereas size at age decreased with Secchi depth. Both swim bladder mass and gas gland mass increased with maximum lake depth. With the exception of the relative liver size of fish in spawning condition, there were was no evidence of sexually dimorphic characteristics in Burbot across populations. The monomorphism of gonad size observed in Burbot across populations is unusual among boreal fishes. The geographic variation in Burbot

Downloaded by [Department Of Fisheries] at 22:57 25 November 2013 reproductive ecology observed in this study provides insight into the potential impacts of a changing environment.

The study of life history traits is of fundamental ecological shifts in population structure and dynamics (Rahel 1984; Pers- interest (Cole 1954). Theory suggests that life history patterns son et al. 1991; Holmgren and Appleberg 2000; Post et al. 2000; develop as adaptive strategies to environmental conditions (Cole Olsen et al. 2004; Hutchings et al. 2012). Where an animal re- 1954; Stearns 1976; Roff 1984; Reznick et al. 1990; Winemiller sides along an environmental gradient can have profound in- and Rose 1992). Variation in environmental conditions, be it fluences on key life history attributes such as sex determination natural or anthropogenic, can influence fish life history traits (Lagomarsino and Conover 1993) and overall reproductive strat- such as maturation and growth, and this can translate into broad egy (Leggett and Carscadden 1978). Successful reproduction is

*Corresponding author: [email protected] Received September 4, 2012; accepted May 29, 2013

1746 BURBOT LIFE HISTORY ACROSS GRADIENTS 1747

central to the viability of species, but it is also energetically have on fishes that spawn in the middle of winter? And what taxing to the individual (Cole 1954; Stearns 1976; Andersson other environmental conditions may influence the life history 1994). traits of a midwinter spawner? Unlike with spring- and fall- For fish, reproduction requires trade-offs between spawner spawning freshwater fish species, the influence that environ- survival and gamete production (Cole 1954; Stearns 1976; mental gradients have on a midwinter spawner has yet to be Leggett and Carscadden 1978; Roff 1984; Winemiller and Rose explored. In addition, few studies have examined the influence 1992; Hutchings 1993, 2005). Species with large clutches often of multiple environmental variables on several life history and reproduce in fewer but more synchronous periodic episodes (i.e., reproductive traits (Conover et al. 2009). spawning occurs at the same time each year but with individ- Winter is a difficult time for many lake-dwelling fish due uals spawning over several years) (Winemiller and Rose 1992) to reduced food availability, light, oxygen, and temperature. as a means of optimizing maternal fitness in the face of variable However, despite its harshness, the under-ice environment of environmental conditions (Cole 1954; Stearns 1976; Van Win- winter is quite predictable and stable. Burbot are an interesting kle et al. 1993). The higher costs of reproduction trade off with candidate with which to study the influence of latitude and adult survival and result in earlier maturity, higher mortality, and other environmental variables on life history and reproductive shorter life spans (Hutchings 2005). Periodic (iterative) spawn- traits for a number of reasons. They are the only midwinter- ing allows adult survival over suboptimal conditions, such as spawning freshwater fish in Canada, belonging to a diverse but particularly cold years (Vila-Gipert et al. 2002), by spreading otherwise entirely marine-dwelling family of fishes, the Gadidae reproductive effort over many years to compensate for bad years (Cohen et al. 1990). Burbot have an enormous geographic range, of poor larval or juvenile success and allowing adults to live and spanning 40 degrees of latitude to the Arctic coast in North spawn again (Winemiller and Rose 1992). America (Scott and Crossman 1973; Cohen et al. 1990; Elmer Environmental gradients that are known to influence life his- et al. 2008; Stapanian et al. 2010), and occur in a wide variety tory patterns and the reproductive ecology of fishes include of freshwater ecosystem types and sizes (Scott and Crossman latitude, productivity, and ecosystem size (e.g., lake size). Wa- 1973; McPhail and Paragamian 2000), making a comparison ter temperature and photoperiod change dramatically with lat- of the life history and reproductive traits possible across wide itude, and temperature and photoperiod increase in importance environmental gradients. at higher latitudes (Pankhurst and Porter 2003), making latitude We had two objectives in conducting this research: (1) to an important determinant of reproductive ecology. As latitude determine how the growth, morphometry, body condition, and increases, growing seasons become shorter, spring and sum- reproductive traits of Burbot vary across environmental gra- mer photoperiods increase, and winters become longer. These dients and (2) to determine whether the sex-based differences changes may be associated with reductions in both fecundity observed in an intensively sampled population (Windy Lake, (Leggett and Carscadden 1978) and the duration of the spawn- Ontario; Cott et al. 2013) are evident in other Burbot popu- ing period (Gotelli and Pyron 1991; Conover 1992) as well as lations across a wider geographic area. The underlying goal increased age at maturity due to the reduced growing seasons of this study was to explore patterns and test predictions for (Gotelli and Pyron 1991). It has been suggested that variability selected life history and reproductive traits with respect to envi- in egg and larval survival, year-class strength, and recruitment ronmental conditions across a wide spatial scale, using Burbot also increase with increasing latitude (Leggett and Carscadden populations in boreal lakes as a model. A virtue of using Burbot 1978). The effects of photoperiod and temperature on these life from boreal lakes in Canada is that they are largely unexploited history traits are well documented for many teleost species, but (Stapanian et al. 2010). Exploitation can confound the relation- the mechanisms remain uncertain (Pankhurst and Porter 2003). ships between life history traits and the environment, making

Downloaded by [Department Of Fisheries] at 22:57 25 November 2013 With increased productivity, water clarity can decline and the drivers of change difficult to discern (Olsen et al. 2004). We this can influence the habitat availability and foraging ability of hypothesized that variation in life history traits among popu- fish and thus the overall community composition of the system lations would be primarily related to climate, ecosystem size, (Bergman 1991; Persson et al. 1991). The general relationship and ecosystem productivity. When resources are low, selection between lake area and depth is linear (Vadeboncoeuret al. 2008), pressure favors fewer, larger young (Stearns 1976). Thus, we and lake size is inversely correlated with water clarity (Fee predicted that egg size would increase with latitude to enable et al. 1996; Vadeboncoeur et al. 2008), so the trends seen with fish to produce larger offspring as a means of compensating for productivity gradients will often be opposite of those seen in the shorter growing season in northern systems. lake-size gradients. We predicted that Burbot growth would decline as the av- To better assess the generality of environmental trends in erage annual temperature declines from south to north across life history strategies, a more complete understanding of the their range; this pattern has been observed in other fish species reproductive ecology for a wider range of species is required (e.g., Braaten and Guy 2002). Due to the environmental stabil- (Vila-Gipert et al. 2002). It is well known that latitude influences ity of their midwinter spawning period throughout their range, the reproductive activities of many spring- and fall-spawning lake productivity should have more influence on reproductive fishes occurring in temperate and arctic areas of North America traits than latitude or lake size, with more productive environ- (Scott and Crossman 1973), but what influence does latitude ments allowing for greater growth and gonadal development and 1748 COTT ET AL.

demanding less energy storage. We predicted that growth and tent across a wider range of Burbot populations. We predicted relative gonad size would be positively related to water body that the sex-based differences observed in Windy Lake, such as productivity and that relative liver size would be negatively re- male-biased gonad size and female-biased liver size, would also lated to productivity. In northern areas, lake size can influence be evident in other Burbot populations. food chain length (Post et al. 2000) and community composition (Tessier and Woodruff 2002) due to the increased heterogene- ity of habitats (Rahel 1984). However, the relative size of such METHODS lakes is also inversely related to productivity (Vadeboncoeur Site description and field sampling.—Burbot were sampled et al. 2008); therefore, it was predicted that the trends of traits from lakes across a spatial gradient spanning 17 degrees in relative to lake size would be inverse to those of traits relative latitude from Lake Manitou in central Ontario to Drygeese Lake to productivity. in the Northwest Territories (Table 1; Figure 1). There is limited With respect to lake depth, we predicted that Burbot (a physo- directed harvest of Burbot over this range, but they are part of clistous fish) would have larger swim bladders and gas gland the bycatch of commercial and recreational fishers. Sampling masses in deeper lakes. Gas glands enable fish to regulate buoy- was conducted from mid-January to late March (spanning the ancy through gas exchange between their blood and swim blad- anticipated spawning period of Burbot, and at a time when all der (Alexander 1993). Burbot living in deeper systems would lakes had sufficient ice cover to permit sampling) from 2006 require greater buoyancy regulation capabilities to transition to 2012. For most populations, we attempted to sample over pressure gradients in order to occupy available habitats. multiple dates in order to determine the timing of spawning. These predictions were tested by conducting analyses sepa- However, for some populations we were only able to obtain rately for each sex. In doing so, a secondary objective was met: samples on a single date. Burbot were sampled with overnight to determine whether the sex-related differences observed in sets of baited longlines deployed beneath the ice using a jigger the Windy Lake Burbot population (Cott et al. 2013) are consis- board. For some populations, Burbot were captured by angling

TABLE 1. Locations and physical characteristics of the lakes in which Burbot were sampled. The numbers in parentheses correspond to the lake numbers in Figure 1; YK Bay = Yellowknife Bay. Latitude and longitude are in decimal degrees.

Latitude Longitude Surface Maximum Secchi No. of Lake Provincea (◦N) (◦W) area (km2) depth (m) depth (m) Burbot Alexie (4) NWT 62.67 114.15 4 32 7.5 44 Attawapiskat (6) Ontario 52.30 87.90 281 6 1.5 8 Baptiste (2) NWT 62.70 114.22 4 32 5.5 35 Big Sand (8) Ontario 50.11 94.63 80 62 4.3 24 Chitty (3) NWT 62.70 114.12 3 20 5.0 68 Drygeese (1) NWT 62.73 114.17 6 35 8.2 41 Endikai (20) Ontario 46.59 83.03 6 48 5.8 38 Fox (21) Ontario 46.58 81.75 5 42 4.4 97 Great Slave, YK Bay (5) NWT 62.40 114.33 28,568 614 5.0 268 Kukagami (18) Ontario 46.73 80.55 19 55 8.0 43

Downloaded by [Department Of Fisheries] at 22:57 25 November 2013 Manitou (24) Ontario 45.78 81.98 10 49 8.4 49 Midlothian (12) Ontario 47.91 81.00 4 32 4.6 13 Missinaibi (10) Ontario 48.36 83.69 77 94 4.2 25 Nipigon (9) Ontario 49.83 88.50 4,481 137 6.5 25 Panache (23) Ontario 46.25 81.33 80 55 6.9 133 Peshu (16) Ontario 46.97 83.14 4 51 6.1 55 Round (11) Ontario 48.02 80.04 12 36 1.5 5 Stull (13) Ontario 47.26 80.82 3 34 5.0 13 Temagami (15) Ontario 47.01 80.08 210 110 10.7 113 Temiskaming (14) Ontario 47.45 79.57 295 214 0.9 25 Wakomata (22) Ontario 46.57 83.36 25 73 8.4 13 Wanapitei (17) Ontario 46.75 80.75 133 142 5.8 122 Windy (19) Ontario 46.60 81.44 11 66 5.1 472 Winnipeg, South Basin (7) Manitoba 50.80 96.75 24,387 28 0.8 28

aNWT = Northwest Territories. BURBOT LIFE HISTORY ACROSS GRADIENTS 1749

FIGURE 1. Locations of the lakes in which Burbot were sampled. The numbers correspond to those of the lakes listed in Table 1.

or sampled from the gill-net bycatch of commercial fishermen. evaluated by visual inspection and classified as immature (gonad Burbot were killed with a sharp blow to the back of the head, immature and not developing toward spawning in the sample

Downloaded by [Department Of Fisheries] at 22:57 25 November 2013 packed on ice, and transported to the laboratory for processing. year), green (gonad developing toward spawning in the sample For some populations, Burbot were frozen whole and processed year but gametes not yet free flowing), ripe (mature gametes at a later date. free flowing), spent (gonads recently spawned-out in the sample Laboratory analysis.—Frozen Burbot were thawed at room year), or unknown (unable to determine sexual maturity). Gut temperature prior to processing, but whole ovaries were always contents were removed from each fish before the determination removed from females in a frozen state to ensure that eggs of body mass. Fin lengths were measured along a ruler placed in remained intact prior to freeze-drying (see below). All Burbot the fin–body insertion and held parallel to the longitudinal axis were processed to obtain the following information: sex, state of the fish body. Swim bladders were removed from the dorsal of maturity, TL ( ± 1.0 mm), total body mass ( ± 1.0 g wet), surface of the body cavity using a scalpel and weighed whole. left and right pectoral fin lengths ( ± 1.0 mm), left and right The swim bladders were then everted, and the gas glands were pelvic fin lengths ( ± 1.0 mm), whole gonad mass ( ± 0.1 g wet), scraped from the inner surface and weighed separately. The somatic mass (total body mass minus gonads; ± 1.0 g wet), drumming muscle in Burbot is well fused into the swim bladder liver mass ( ± 0.1 g wet), swim bladder mass ( ± 0.1 g wet), wall, making it difficult to extract, so swim bladder mass was gas gland mass ( ± 0.1 g wet), and eviscerated mass (total body used (with the gas gland scraped off) as a proxy for drumming mass minus gonads and all viscera; ± 1.0 g wet). Maturity was muscle mass. Sagittal otoliths were removed, rubbed clean, and 1750 COTT ET AL.

placed in plastic vials to air dry. For green (unovulated) females, der mass and gas gland mass were expressed as proportions of subsamples (∼20 g) were cut from the midsection of both mature somatic mass. Variables expressed as proportions were arcsine- ovaries, placed in small plastic bags, and frozen at −20◦C. square-root transformed, and other dependent variables were Ovary subsamples were freeze-dried for 7 d, and mean egg loge transformed, as required, to normalize the residuals or lin- size was determined by counting and weighing subsamples of earize the relationships. these ova. Otoliths were prepared following the crack-and-burn To facilitate comparisons between sexes and among popu- technique (Edwards et al. 2011), and ages were determined by lations, all data were adjusted to common representative body counting annuli on the burned surface under a dissecting scope sizes of 1,000 g (where somatic mass was used as the covariate) using reflected light. The outer edge of each otolith was always and 500 mm (where TL was used as the covariate). These ap- counted as a complete annulus because fish were collected near proximated the mean sizes of fish sampled among populations. the time of spawning. In each population, adjusted means (least-square means) were The environmental variables used in our analyses included estimated for each sex using an analysis of covariance (AN- indices of climate, ecosystem size, and water body productivity. COVA) model (GLM procedure) with sex as the class variable We used latitude as our index of climate, as it correlates strongly and body size as the covariate. Differences between sexes in the with mean annual growing degree-days (Johnston and Leggett adjusted population means were tested with paired-comparison 2002) and is thus a good analog to thermal and photoperiod gra- t-tests or Wilcoxon’s signed rank tests (if differences were not dients. Ecosystem size was represented by lake surface area and normally distributed). To determine how these life history traits maximum depth. Water body productivity was used to represent varied with respect to environmental conditions, each dependent food availability, particularly that of zooplankton, which is a key variable was regressed against the environmental variables (lati- component of the diet of Burbot larvae (Ryder and Pesendorfer tude, Secchi depth, lake area, and maximum lake depth) using a 1992; McPhail 2007). We used Secchi depth (water clarity) as multiple-regression model with a stepwise selection procedure our index of productivity, as it is highly and inversely correlated (REG procedure). Data were pooled for the sexes if no signif- with lake productivity (Wetzel 1983) and an established index icant differences were found in the sex-based analyses. Some (e.g., Vander Zanden et al. 1999; Post et al. 2000; Johnston environmental variables were loge transformed to linearize the and Leggett 2002). All environmental data were obtained from relationships, and all environmental variables were converted to aquatic databases maintained by the Ontario Ministry of Natural normal variables (Zi = [Xi – mean]/SD) to reduce scaling ef- Resources and Fisheries and Oceans Canada. fects. Environmental variables were not strongly correlated with Data analysis.—All analyses were performed with SAS ver- each other (correlation analysis; |r| < 0.30, P > 0.15), with the sion 9.2 (SAS 2009) procedures. Burbot samples from 24 pop- exception of lake area and maximum depth (r = 0.66, P < 0.01). ulations were used in our analyses. However, because of differ- We excluded models that contained both of these variables. ences in the timing of collection, not all populations could be used in the analyses of some variables. The mid-spawning date RESULTS (S50) for each population was estimated as the day when 50% of the females spawning in the sample year (classified as green, Sex-Based Differences ripe, or spent) had completed spawning (classified as spent). The von Bertalanffy model fit the growth trajectory well for This was estimated from a logistic model fitted to spawning sta- both sexes of Burbot for most populations. Model parameter tus (0 = green or ripe, 1 = spent) and sampling date (LOGISTIC estimates ranged from 390 to 1,069 mm for maximum length procedure). Where the logistic model could not be fit due to the and from 0.054 to 0.53 for the growth parameter k. Predicted limited temporal coverage of the sampling, the date of spawning mean lengths at ages 7 and 10 were derived from these models

Downloaded by [Department Of Fisheries] at 22:57 25 November 2013 was interpolated or extrapolated from plots of the percentage of and used as indices of the growth rate, as these ages included the spent females versus the sampling date. The percentage of fe- range common to the mature fish sampled from most lakes. The males spawning in the sample year was estimated as the total model fitting procedure did not converge for males or females number of green, ripe, and spent females captured divided by from 2 of the 22 populations, and in these cases we estimated the total number of females captured. This variable was calcu- the mean lengths at age by interpolation from length–age plots. lated using different minimum body size criteria to assess the Across populations, there was no significant difference between prevalence of skipped spawning. Growth was compared among sexes for length at age 7 (paired-comparisons t-test; t = 0.84, n populations using predicted length-at-age estimates from a von = 22, P = 0.41) or age 10 (t = 0.12, n = 22, P = 0.90). Females Bertalanffy growth model (Ricker 1975; Lester et al. 2004) fitted had higher somatic mass at length (paired-comparisons t-test; using nonlinear least-squares (NLIN procedure). The gonado- t = 2.32, n = 21, P = 0.03), but the sexes did not differ in somatic index (GSI) was estimated as the whole gonad mass eviscerated mass at length (t = 0.85, n = 21, P = 0.40). of prespawning Burbot divided by their somatic mass, and the Male and female Burbot did not differ greatly with respect hepatosomatic index (HSI) was estimated as their liver mass to the morphometric characteristics examined. There were no divided by their somatic mass. Similarly, pectoral and pelvic fin significant differences between the sexes in pectoral fin length lengths were expressed as proportions of TL, and swim blad- (paired-comparisons t-test; t = 0.45, n = 21, P = 0.65), pelvic BURBOT LIFE HISTORY ACROSS GRADIENTS 1751

fin length (Wilcoxon’s signed rank test; n = 22, S =−11.5, P = 0.70), swim bladder mass (Wilcoxon’s signed rank test; n = 17, S = 5.5, P = 0.82), or gas gland mass (Wilcoxon’s signed rank test; n = 17, S =−17.5, P = 0.43). Differences in liver size (HSI) depended on the subset of fish examined. When all fish were included in the analysis, HSI did not differ between the sexes (paired-comparisons t-test; t = 1.62, n = 21, P = 0.12), but when the analyses were restricted to adults spawning in the current year, females had slightly larger HSI values (Wilcoxon’s signed rank test; n = 21, S = 61.5, P = 0.03). Gonad size (GSI) did not differ between the sexes (Wilcoxon’s signed rank test; n = 19, S = 17, P = 0.52).

Variation across Environmental Gradients The predicted mean length at age 7 (sexes combined) showed a significant negative relationship with Secchi depth (Figure 2) 2 (regression analysis; F1, 20 = 5.77, P = 0.03, R = 0.22) but no significant relationship with any other environmental variables. A similar outcome was obtained when the predicted length at age 10 was regressed against our suite of environmental vari- ables, though the Secchi depth effect was weaker (F1, 20 = 3.75, P = 0.07, R2 = 0.16). The body condition (predicted somatic mass at 500 mm TL) of female Burbot increased with loge trans- formed latitude (Figure 3a; regression analysis; F1, 21 = 5.27, P = 0.03, R2 = 0.20) but showed no significant relationship with any other environmental variable. In contrast, our stepwise regression analysis indicated that the variation in the condition of males was positively related to the combined effects of loge transformed lake area (partial F1, 17 = 13.0, P < 0.01) and loge transformed latitude (Figure 3a; partial F = 4.50, P = 0.05), 1, 17 FIGURE 3. Relationships between predicted body mass at 500 mm TL and and negatively related to loge transformed maximum depth (par- latitude for 23 Burbot populations from across Canada. Panel (a) utilizes so- tial F1, 17 = 10.0, P = 0.01). These collectively accounted for matic mass, panel (b) eviscerated mass. Mass predictions were based on fitted 55% of the observed variation in male condition. The predicted relationships between loge(mass) and loge(TL). The different symbols represent females (black), males (white), and the sexes combined (gray).

eviscerated mass at 500 mm TL (sexes combined) increased with loge transformed latitude (Figure 3b; regression analysis; 2 F1, 21 = 5.28, P = 0.03, R = 0.20) but showed no significant

Downloaded by [Department Of Fisheries] at 22:57 25 November 2013 relationship with any other environmental variable. Variation with respect to environmental gradients was evi- dent for some morphological traits but not others. No significant relationships with any environmental variables were observed for either pectoral or pelvic fin lengths (regression analysis; P > 0.20). However, stepwise multiple-regression analysis in- dicated that swim bladder mass increased with loge transformed maximum lake depth (Figure 4a; partial F1, 20 = 5.00, P = 0.04) and declined with latitude (partial F1, 20 = 7.85, P = 0.01). Sim- ilarly, stepwise regression indicated a significant positive effect of loge transformed maximum lake depth on gas gland mass (Figure 4b; partial F1, 20 = 9.89, P = 0.01) and a nearly signifi- = = FIGURE 2. Relationship between mean predicted TL at 7 years of age (sexes cant negative effect of latitude (partial F1, 20 3.74, P 0.07). combined) and Secchi depth for 22 Burbot populations from across Canada. The relationship with maximum depth for both swim bladder The predicted TLs were estimated from fitted von Bertalanffy growth models. mass and gas gland mass was strongest for lakes of 20–150 m 1752 COTT ET AL.

FIGURE 5. Relationship between egg size (dry mass of mature ova) and latitude for 15 Burbot populations from across Canada. The symbols are for population means adjusted to a maternal TL of 500 mm.

size was accounted for by latitude (Figure 5; negative effect; par- tial F1, 12 = 9.92, P = 0.01) and lake area (positive effect; partial 2 F1, 12 = 6.14, P = 0.03) (total model R = 0.58). We were able to estimate S50 for 17 of the 23 Burbot popu- lations sampled. There was considerable variability among the southern populations, with dates ranging from late January to mid-March (Figure 6). Two southern populations only 40 km apart had spawning dates separated by 30 d (early February in Windy Lake and early March in Panache Lake). We sampled these two populations in multiple years and confirmed that these spawning dates were temporally stable. The strongest relation- FIGURE 4. Relationships between maximum lake depth and (a) swim bladder mass (with gas gland removed) and (b) gas gland mass for 23 Burbot populations ship observed between S50 and the environmental variables was from across Canada. The symbols are for population means adjusted to a somatic mass of 1,000 g (sexes combined). Note that scale of the lake depth axis is logarithmic.

depth but appeared to break down for the one very shallow and two very deep lakes in our data set (Figure 4).

Downloaded by [Department Of Fisheries] at 22:57 25 November 2013 Except for egg size (mean dry mass), the traits associated with energy storage and reproduction were not strongly related to our environmental variables. We found no significant relation- ship between HSI and any environmental variable, regardless of whether all Burbot or just those spawning in the sampling year were included in the model (regression analysis; P > 0.09). Neither gonad size (GSI) nor egg size were strongly related to maternal size. There was a significant relationship between GSI and maternal size in only 3 of 19 populations (two negative, one positive) for which female GSI was determined, and no signifi- cant relationships between egg size and maternal size in any of the 15 populations for which egg size was measured. We found FIGURE 6. Relationship between the timing of spawning and latitude for 17 no significant relationships between GSI (sexes combined) and Burbot populations from across Canada. The mid-spawn day is the day of the any of our environmental variables (regression analysis; P > year when it was estimated that 50% of the females spawning in the sample year 0.29). In contrast, significant among-population variation in egg had completed spawning (e.g., day 50 = February 19). BURBOT LIFE HISTORY ACROSS GRADIENTS 1753

males. Unlike in the Windy Lake population, in which male Burbot had higher GSI values than females (Cott et al. 2013), the general trend across Burbot populations was for similar gonad weight between sexes. Reproduction is energetically costly (Wong and Jennions 2003; Rideout and Tomkiewicz 2011). Lipids stored in body tis- sues are often used to meet energy demands relating to the repro- duction of iteroparous fishes (Casselman and Schulte-Hostedde 2004; Jonsson and Jonsson 2005; Johnston et al. 2012) and are particularly important to lean fish such as gadids (Marshall et al. 1999; Skjæraasen et al. 2006), including Burbot (Cott 2013). Perhaps the larger livers of female Burbot enable more lipid storage to assist in egg development. However, gonad size remained consistent across populations for all of the gradients examined and was not related to maternal size. These results in- dicate that Burbot are well adapted to winter spawning and are able to adjust their reproductive output in terms of overall gonad FIGURE 7. Relationship between the prevalence of nonspawning females size to a predetermined optimum regardless of the location or and Secchi depth for 22 Burbot populations from across Canada. Nonspawner characteristics of the lakes in which they occur. Environmental prevalence was estimated as the percentage of all females >300 mm TL that conditions are known to select for optimal egg size in fish, with were not spawning in the sampling year. total reproductive effort shifting accordingly via changes in fe- cundity (Sargent et al. 1987). Rather than finding it a handicap, with latitude; S50 increased with latitude, but the effect was not Burbot appear to use the predictability and consistency of their F = statistically significant (Figure 6; regression analysis; 1, 14 otherwise austere under-ice spawning environment to their ad- P = R2 = 3.25, 0.093, 0.19). vantage. Relative to the thermal conditions during spring, sum- The proportion of captured females that were not in spawn- mer, and fall—when almost all other boreal freshwater fishes ing condition was also highly variable and did not appear to be spawn—those in the midwinter spawning period of Burbot are strongly related to latitude. As with S50, there was considerable very stable (1–4◦C) in northern lakes (Wetzel 1983). Striking a variability among the southern populations, with the proportion balance between gonad production and the energy required for of nonspawners ranging from 0% to over 60% of the adult fe- adult survival is a critical life history requirement (Cole 1954; males sampled (Figure 7). When the proportion of nonspawners > Stearns 1976). Skipping spawning years may provide a mech- was estimated for all females 300 mm TL, stepwise regres- anism that enables Burbot in more demanding environments sion indicated that it was positively related to both Secchi depth 2 to replenish the energy required for reproduction and perhaps (Figure 7; F1, 18 = 7.51, P = 0.01, R = 0.28) and lake area 2 growth. (partial F1, 18 = 4.93, P = 0.04, partial R = 0.15). Shifting > It has been suspected that the Burbot in some populations skip the body size criterion to 400 mm TL did not change these spawning seasons (Chen 1969; McPhail and Paragamian 2000), positive trends, though the significance of the effects declined and the data presented here support this hypothesis, showing that F = P = R2 = slightly for both Secchi depth ( 1, 18 4.56, 0.05, some proportion of adult females were in nonspawning condi- F = P = R2 = 0.19) and lake area (partial 1, 18 4.37, 0.05, partial tion during the spawning period in most populations (>60% in 0.16). Downloaded by [Department Of Fisheries] at 22:57 25 November 2013 some populations). However, no significant trend was noted for the percentage of adult females in spawning condition and the DISCUSSION environmental gradients tested. In Atlantic Cod Gadus morhua, Burbot showed some variation in life history traits across the it is thought that low energy reserves at the onset of vitello- environmental gradients examined, but surprisingly some key genesis make individuals in the population skip spawning for indices of reproductive effort (such as HSI) and energy storage that year (Burton et al. 1997; Skjæraasen et al. 2006). Skipping (such as HSI) did not appear to vary significantly with respect to spawning enables fish to retain a portion of their stored energy indices of climate, productivity, or ecosystem size. The strongest to endure adverse conditions and maintain productive potential predictive power we observed was for a model relating egg size (Rideout and Tomkiewicz 2011). to latitude and lake area (R2 = 0.58). Further, no significant Contrary to predictions, egg size decreased with latitude sex-based differences were found across populations, with the and rather increased with lake area. Decreasing egg size with exceptions of higher HSI in females (as in the Windy Lake increasing latitude has been noted for other fishes, such as population; Cott et al. 2013) and higher condition factor in Walleyes Sander vitreus (Johnston and Leggett 2002) and Coho females. However, the higher condition factor of females was Salmon Oncorhynchus kisutch (Fleming and Gross 1990), al- likely influenced, in part, by females’ having larger livers than though the drivers for this phenomenon remain uncertain. The 1754 COTT ET AL.

trends seen in lake size are often opposite to what would be ex- trend for age at maturity and latitude was discovered for Amer- pected in productivity, as lake size is often inversely correlated ican Shad Alosa sapidissima (Leggett and Carscadden 1978), with water clarity (Fee et al. 1996; Vadeboncoeur et al. 2008). and no correlation was found between size at age and latitude This was not the case with our data set, suggesting that the link in an assessment of 21 minnow species (Notropis sp.; Gotelli between larger egg size and lake area must be something other and Pyron 1991). How size at age and growth change along than overall lake productivity. The diet of Burbot larvae is almost latitudinal gradients can vary among species. For example, in exclusively zooplankton (Ryder and Pesendorfer 1992; McPhail a study comparing the life history attributes of river-dwelling 2007). In general, smaller lakes have smaller-sized zooplankton fish along a latitudinal gradient, the growth rate increased with than large, deep lakes (Tessier and Woodruff 2002). Burbot lar- increasing latitude for emerald shiners Notropis atherinoides vae emerging in large lakes would need to be larger in order to but decreased for sauger Sander canadensis (Braaten and Guy feed on larger-sized zooplankton. In the closely related Atlantic 2002). Some have speculated that Burbot grow more in the win- Cod, eggs size, larva size, and mouth gape are correlated, with ter than the summer (Nelson and Paetz 1992), and perhaps this gape directly related to the size of the zooplankton that can be offsets temperature-related latitudinal effects. In our study, there consumed (Knutsen and Tilseth 1985). Also, ice tends to persist was a negative trend for size at age and Secchi depth, indicating, later on large lakes than on small ones; for example, ice-off on not surprisingly, that water body productivity influences growth. Great Slave Lake occurs in early July, over 1 month later than Other factors that were not investigated in this study, such as on the adjacent, 420-ha Alexie Lake. In large lakes, larger eggs interspecific competition for resources, can influence the size would facilitate the longer incubation times required for larval at age of Burbot (Stapanian et al. 2011). The role that selec- emergence to coincide with later post-ice plankton blooms. For tion pressure from interspecific competition plays in Burbot life organisms inhabiting environments with poor conditions, selec- history traits is an area for future research. tion pressure would be expected to favor fewer but larger young Despite the differences in the duration of ice cover from (Stearns 1976), with egg size adjusting accordingly (Sargent south to north, the spawning period for most of the Burbot et al. 1987), and this may offer an explanation for the significant populations sampled was confined to a narrow window within relationship between egg size and lake area observed in this the first 3 weeks of February. The timing of spawning among study. hatchery-raised Burbot has been shown to be highly correlated Physoclistous fishes such as Burbot have gas glands to reg- with temperature. Females reared at naturally occurring temper- ulate their buoyancy, enabling them to transition through pres- ature ranges synchronized their spawning into a 17-d period, as sure gradients (Alexander 1993). As predicted, both the swim opposed to a 2-d period for Burbot reared in an environment in bladder and gas gland masses of Burbot were higher in pop- which the water temperature was abruptly dropped from 6◦Cto ulations in deeper lakes. Other deep-dwelling freshwater fish, 1◦C(Zarski˙ et al. 2010). However, in the present study the great- such as the siscowet form of Lake Trout Salvelinus namaycush, est spread in mid-spawning times was for southern populations have thicker-walled swim bladders than their shallow-water- in relatively close proximity to each other, which would pre- dwelling counterparts, possibly because of atrophy (Eshenroder sumably experience similar temperature conditions. Additional et al. 2008). The heavier swim bladders of Burbot in deeper study would be required to determine which environmental fac- lakes that we observed in this study likely evolved to withstand tors determine the spawning time in these wild populations. the pressure changes associated with migrating though greater It is possible that the circannual reproductive cycle is more pressure gradients. Relatively speaking, deeper systems would endogenously driven for Burbot than for other fishes, mak- offer more available habitat that could be exploited more suc- ing them less dependant on environmental cues. Even for fish cessfully with more effective gas exchange systems. The Burbot whose circannual reproductive clocks are thought to be largely

Downloaded by [Department Of Fisheries] at 22:57 25 November 2013 from the largest lakes sampled in this study (Lake Winnipeg and driven by day length, intrinsic drivers govern their cycles to the Great Slave Lake) appeared to deviate from the positive re- some degree (Randall et al. 1998). The strategies used by extant lationship between gas gland and lake size, possibly because taxa may therefore not be entirely adaptations to their current they represented within-lake stocks occupying the shallower re- environments but rather retained ancestral traits (Coddington gions of these enormous water bodies. Burbot swim bladder and 1988), where genetic programming overrides environmental gas gland mass have also been shown to vary seasonally in one cues (Lagomarsino and Conover 1993). Perhaps, as Cahn (1936) population, being heaviest in the winter months, possibly due to long ago speculated, some of the life history traits of Burbot are the increased depth transitioning required to spawn in shallow holdovers from its marine ancestry, with the timing of repro- water and search for mates (Cott et al. 2013). duction corresponding to that of many marine cods. Although From an investigation of several populations of Burbot the dark, cold, under-ice conditions in which Burbot spawn ap- throughout Quebec, Magnin and Fradette (1977) commented pear to be extreme, the variability of these conditions is low and that size at age is likely latitude dependent, with more northerly animals can adjust accordingly (Winemiller and Rose 1992). populations of Burbot growing more slowly. The results of this In correlation analyses of life history variables for a variety of study do not support their hypothesis, as no trends were detected North American marine and freshwater fishes, marine cods and for size at age and latitude (for ages 7 and 10). Similarly, no Burbot were both categorized as having a periodic life history BURBOT LIFE HISTORY ACROSS GRADIENTS 1755

strategy, with larger size at maturity, episodic or seasonal spawn- Casselman, S. J., and A. I. Schulte-Hostedde. 2004. Reproductive roles predict ing, and large clutches of small eggs (Winemiller and Rose 1992; sexual dimorphism in internal and external morphology of Lake Whitefish, Vila-Gipert et al. 2002). A periodic life history strategy is more Coregonus clupeaformis. Ecology of Freshwater Fish 13:217–222. Chen, L.-C. 1969. The biology and taxonomy of the Burbot, Lota lota leptura, common in temperate and arctic freshwater fish species than in in interior Alaska. Biological Papers of the University of Alaska 11:53. tropical species (Vila-Gipert et al. 2002). For interoparous fish Coddington, J. A. 1988. Cladistic tests of adaptational hypotheses. Cladistics living in unpredictable environments, smaller broods should be 4:3–22. selected as a trade-off for adult fitness, thus minimizing the cost Cohen, D. M., T. Inada, T. Iwamato, and N. Scialabba. 1990. Gadiform fishes of failure should any single reproductive event be unsuccessful of the world (Order Gadiformes). FAO (Food and Agriculture Organization of the United Nations) Fisheries Synopsis 125(10). (Stearns 1976). Cole, L. C. 1954. The population consequences of life history phenomena. Our results indicate that regardless of the system or ge- Quarterly Review of Biology 29:103–137. ographic location of the Burbot population investigated, the Conover, D. O. 1992. Seasonality and the scheduling of life history at different overall gonadal investment did not change. The stability of the latitudes. Journal of Fish Biology 41:161–178. under-ice winter environment enabled Burbot to employ a peri- Conover, D. O., T. A. Duffy, and L. A. Hice. 2009. The covariance between genetic and environmental influences across ecological gradients. Annals of odic spawning strategy with consistency among the populations the New York Academy of Science 1168:100–129. studied. This life history strategy has allowed Burbot to success- Cott, P. A. 2013. Life history and reproductive ecology of a mid-winter spawner: fully occupy an enormous geographic range under a spectrum the Burbot (Lota lota). Doctoral dissertation. Laurentian University, Sudbury, of environmental conditions, with very little variation in their Ontario. life history traits. In fact, Burbot today remain morphologically Cott, P. A., T. A. Johnston, and J. M. Gunn. 2013. Sexual dimorphism in an under-ice spawning fish: the Burbot (Lota lota). Canadian Journal of Zoology similar to the proto-Burbot L. hulai, which first colonized fresh- 91:732–740. water systems over 5 million years ago (Van Houdt 2003; Van Edwards, W. H., M. A. Stapanian, and A. T. Stoneman. 2011. Precision of Houdt et al. 2005). two methods for estimating age from Burbot otoliths. Journal of Applied Ichthyology 27:43–48. Elmer, K. R., J. K. J. Van Houdt, A. Meyer, and F. A. M. Volckaert. 2008. ACKNOWLEDGMENTS Population genetic structure of North American Burbot (Lota lota maculosa) Thanks are extended to Pascale Savage, Rick Walbourne, across the Nearctic and at its contact zone with Eurasian Burbot (Lota lota Sarah Elsasser, Andrew Corston, Angele` Dubuc, Lee Haslam, lota). Canadian Journal of Fisheries and Aquatic Sciences 65:2412–2426. Eshenroder, R. L., E. D. Stevens, M. S. Zimmerman, and C. C. Krueger. 2008. Jason Houle, Tiff-Annie Kenny, Susan Mann, Adam Smits, and Atypical swimbladders of Lake Charr, Salvelinus namaycush, from Great Rex Tang for field and laboratory support; to Carolyn Bakelaar Slave Lake, Northwest Territories, Canada. Environmental Biology of Fishes for map preparation; and to Jim Reist and Jeff Hutchings for 83:91–98. their review of earlier drafts. Thanks to Martin Stapanian and Fee, E. J., R. E. Hecky, S. E. M. Kasian, and D. R. Cruikshank. 1996. Effects of three anonymous reviewers for their constructive critiques of lake size, water clarity, and climatic variability on mixing depths in Canadian Shield lakes. Limnology and Oceanography 41:912–920. this manuscript. Funding and in-kind support were provided Fleming, I. A., and M. R. Gross. 1990. Latitudinal clines: a trade-off between by the Ontario Ministry of Natural Resources, Aquatic Re- egg number and size in Pacific salmon. Ecology 71:1–11. search and Development Section; the Department of Fisheries Gotelli, N. J., and M. Pyron. 1991. Life history variation in North American and Oceans, Western Arctic Area; Natural Resources Canada, freshwater minnows: effects of latitude and phylogeny. Oikos 62:30–40. Program for Energy, Research and Development; Laurentian Holmgren, K., and M. Appleberg. 2000. Size structure of benthic freshwater fish communities in relation to environmental gradients. Journal of Fish Biology University; and the Natural Sciences and Engineering Research 57:1312–1330. Council (Discovery Grant to T.A.J.). Hutchings, J. A. 1993. Adaptive life histories effected by age-specific survival and growth rate. Ecology 74:673–684. Hutchings, J. A. 2005. Life history consequences of overexploitation to pop- Downloaded by [Department Of Fisheries] at 22:57 25 November 2013 REFERENCES ulation recovery in population recovery in northwest Atlantic Cod (Gadus Alexander, R. M. 1993. Buoyancy. Pages 75–97 in D. H. Evans, editor. The morhua). Canadian Journal of Fisheries and Aquatic Sciences 62:824–832. physiology of fishes. CRC Press, Boca Raton, Florida. Hutchings, J. A., R. A. Myers, V. B. Garcia, L. O. Lucifora, and A. Kupari- Andersson, M. 1994. Sexual selection. Princeton University Press, Princeton, nen. 2012. Life-history correlates of extinction risk and recovery potential. New York. Ecological Applications 22:1061–1067. Bergman, E. 1991. Changes in abundance of two percids, Perca fluviatilis and Johnston, T. A., and W. C. Leggett. 2002. Maternal and environmental gradients Gymnocephalus cernuus, and a productivity gradient: relations to feeding in the egg size of an interoparous fish. Ecology 83:1777–1791. strategies and competitive abilities. Canadian Journal of Fisheries and Aquatic Johnston, T. A., D. M.-M. Wong, M. D. Moles, M. D. Wiegand, J. M. Sciences 48:536–545. Casselman, and W. C. Leggett. 2012. Reproductive allocation in exploited Braaten, P. J., and C. S. Guy. 2002. Life history attributes of fishes along the Lake Whitefish (Coregonus clupeaformis) and Walleye (Sander vitreus) pop- latitudinal gradients of the Missouri River. Transactions of the American ulations. Fisheries Research 125/126:225–234. Fisheries Society 131:931–945. Jonsson, B., and N. Jonsson. 2005. Lipid energy reserves influence life-history Burton, M. P. M., R. M. Penney, and S. Biddiscombe. 1997. Time course of decision of Atlantic Salmon (Salmo salar) and Brown Trout (S. trutta)in gametogenesis in northwest Atlantic Cod (Gadus morhua). Canadian Journal fresh water. Ecology of Freshwater Fish 14:296–301. of Fisheries and Aquatic Sciences 54:122–131. Knutsen, G. M., and S. Tilseth. 1985. Growth, development, and feeding success Cahn, A. R. 1936. Observations on the breeding of the lawyer, Lota maculosa. of Atlantic Cod larvae Gadus morhua related to egg size. Transactions of the Copeia 3:136–165. American Fisheries Society 114:507–511. 1756 COTT ET AL.

Lagomarsino, I. V., and D. O. Conover. 1993. Variation in environmental and Ryder, R. A., and J. Pesendorfer. 1992. Food, growth, habitat, and community genotypic sex-determining mechanisms across a latitudinal gradient in the interactions of young-of-the-year Burbot, Lota lota L., in a Precambrian fish, Menidia menidia. Evolution 47:487–494. Shield lake. Hydrobiologia 243/244:211–227. Leggett, W. C., and J. E. Carscadden. 1978. Latitudinal variation in reproductive Sargent, R. C., P. D. Taylor, and M. R. Gross. 1987. Parental care and the characteristics of American Shad (Alosa sapidissima): evidence for popula- evolution of egg size in fishes. American Naturalist 129:32–46. tion specific life history strategies in fish. Journal of the Fisheries Research SAS. 2009. SAS onlinedoc, version 9.2. SAS Institute, Inc., Cary, North Car- Board of Canada 35:1469–1478. olina. Lester, N. P., B. J. Shuter, and P. A. Abrams. 2004. Interpreting the von Berta- Scott, W. B., and E. J. Crossman. 1973. Freshwater fishes of Canada. Fisheries lanffy model of somatic growth in fishes: the cost of reproduction. Proceed- Research Board of Canada, Ottawa. ings of the Royal Society of London B 271:1625–1631. Skjæraasen, J. E., T. Nilsen, and O. S. Kjesbu. 2006. Timing and determination Magnin, E., and C. Fradette. 1977. Croissance et regime alimentaire de la Lotte of potential fecundity in Atlantic Cod (Gadus morhua). Canadian Journal of (Lota lota Linnaeus 1758) dans divers lacs et rivieres` du Quebec.´ [Growth Fisheries and Aquatic Sciences 63:310–320. and diet of Burbot (Lota lota Linnaeus 1758) in various lakes and rivers of Stapanian, M. A., W. H. Edwards, and L. D. Witzel. 2011. Recent changes in Quebec.] La Naturaliste Canadien 104:207–222. Burbot growth in Lake Erie. Journal of Applied Ichthyology 27:57–64. Marshall, T. C., N. A. Yaragina, Y. Lambert, and O. S. Kjesbu. 1999. Total lipid Stapanian, M. A., V. L. Paragamian, C. P. Madenjian, J. R. Jackson, J. Lap- energy as a proxy for total egg production by fish stocks. Nature 402:288–290. palainen, M. J. Evenson, and M. D. Neufeld. 2010. Worldwide status of McPhail, J. D. 2007. The freshwater fishes of British Columbia. University of Burbot and conservation measures. Fish and Fisheries 11:34–56. Alberta Press, Edmonton. Stearns, S. C. 1976. Life history tactics: a review of the ideas. Quarterly Review McPhail, J. D., and V. L. Paragamian. 2000. Burbot biology and life history. of Biology 51:3–47. Pages 11–23 in V. L. Paragamian and D. W. Willis, editors. Burbot: biology, Tessier, A. J., and P. Woodruff. 2002. Cryptic trophic cascade along a gradient ecology, and management. American Fisheries Society, Fisheries Manage- of lake size. Ecology 83:1263–1270. ment Section Publication 1, Spokane, Washington. Vadeboncoeur, Y.,G. Peterson, J. M. Vander Zanden, and J. Kalff. 2008. Benthic Nelson, J. S., and M. J. Paetz. 1992. The fishes of Alberta, 2nd edition. University algal production across lake size gradients: interactions among morphology, of Alberta Press and University of Calgary Press, Edmonton, Alberta. nutrients, and light. Ecology 89:2542–2552. Olsen, E. M., M. Heino, G. R. Lilly, M. J. Morgan, J. Brattey, B. Ernande, and U. Van Houdt, J. 2003. Phylogenetic relationships among Palearctic and Nearctic Dieckmann. 2004. Maturation trends indicative of rapid evolution preceded Burbot (Lota lota): Pleistocene extinctions and recolonization. Molecular the collapse of northern cod. Nature 428:932–935. Phylogenetics and Evolution 29:599–612. Pankhurst, N. W., and M. J. R. Porter. 2003. Cold and dark or warm and Van Houdt, J. K. J., L. de Cleyn, A. Perretti, and F. A. M. Volckaert. 2005. light: variations on the theme of environmental control of reproduction. Fish A mitogenic view on the evolutionary history of the Holarctic freshwater Physiology and Biochemistry 28:385–389. gadoid, Burbot (Lota lota). Molecular Ecology 14:2445–2457. Persson, L., S. Diehl, L. Johansson, G. Andersson, and S. F. Hamrin. 1991. Van Winkle, W., K. A. Rose, K. O. Winemiller, D. L. DeAngelis, S. W. Chris- Shifts in fish communities along the productivity gradient of temperate lakes- tensen, R. G. Otto, and B. J. Shuter. 1993. Linking life history theory, en- patterns and the importance of size-structured interactions. Journal of Fish vironmental setting, and individual-based modeling to compare responses of Biology 38:281–293. different fish species to environmental change. Transactions of the American Post, D. M., M. L. Pace, and N. G. J. Hairstorn. 2000. Ecosystem size determines Fisheries Society 122:459–466. food-chain length. Nature 405:1047–1049. Vander Zanden, M. J., J. M. Casselman, and J. B. Rasmussen. 1999. Stable Rahel, F. 1984. Factors structuring fish assemblages along a bog lake succes- isotope evidence for the food web consequences of species invasions in lakes. sional gradient. Ecology 65:1276–1289. Nature 401:464–467. Randall, C. F., N. R. Bromage, J. Duston, and J. Symes. 1998. Photoperiod- Vila-Gipert, A., R. Moreno-Amich, and E. Garc´ıa-Berthou. 2002. Gradients of induced phase-shifts of the endogenous clock controlling reproduction in the life history variation: an intercontinental comparison of fishes. Reviews in Rainbow Trout: a circannual phase-response curve. Journal of Reproduction Fish Biology and Fisheries 12:417–427. and Fertility 112:399–405. Wetzel, R. G. 1983. Limnology, 2nd edition. Saunders College Publishing, Reznick, D. A., H. Bryga, and J. A. Endler. 1990. Experimentally induced Orlando, Florida. life-history evolution in a natural population. Nature 346:357–359. Winemiller, K. O., and K. A. Rose. 1992. Patterns of life-history diversification Ricker, W. E. 1975. Computation and interpretation of biological statistics of fish in North American fishes: implications for population regulation. Canadian populations. Department of the Environment, Fisheries and Marine Service, Journal of Fisheries and Aquatic Sciences 49:2196–2218. Bulletin 19, Ottawa. Wong, B. B. M., and M. D. Jennions. 2003. Cost influences male mate choice in a Downloaded by [Department Of Fisheries] at 22:57 25 November 2013 Rideout, R. M., and J. Tomkiewicz. 2011. Skipped spawning in fishes: more freshwater fish. Proceedings of the Royal Society of London B 270:S36–S38. common that you might think. Marine and Coastal Fisheries: Dynamics, Zarski,˙ D., D. Kucharczyk, W. Sasinowski, K. Targonska,´ and A. Mamcarz. Management, and Ecosystem Science [online serial] 3:176–189. 2010. The influence of temperature on the successful reproduction of Burbot, Roff, D. A. 1984. The evolution of life history parameters in teleosts. Canadian Lota lota (L.) under hatchery conditions. Polish Journal of Natural Sciences Journal of Fisheries and Aquatic Sciences 41:989–1000. 25:93–105. This article was downloaded by: [Department Of Fisheries] On: 25 November 2013, At: 22:58 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Transactions of the American Fisheries Society Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/utaf20 Ups and Downs of Burbot and Their Predator Lake Trout in Lake Superior, 1953–2011 Owen T. Gorman a & Shawn P. Sitar b a U.S. Geological Survey, Great Lakes Science Center , Lake Superior Biological Station , 2800 Lake Shore Drive East, Ashland , Wisconsin , 54086 , USA b Michigan Department of Natural Resources , Marquette Fisheries Research Station , 484 Cherry Creek Road, Marquette , Michigan , 49855 , USA Published online: 20 Nov 2013.

To cite this article: Owen T. Gorman & Shawn P. Sitar (2013) Ups and Downs of Burbot and Their Predator Lake Trout in Lake Superior, 1953–2011, Transactions of the American Fisheries Society, 142:6, 1757-1772, DOI: 10.1080/00028487.2013.824918 To link to this article: http://dx.doi.org/10.1080/00028487.2013.824918

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions Transactions of the American Fisheries Society 142:1757–1772, 2013 C American Fisheries Society 2013 ISSN: 0002-8487 print / 1548-8659 online DOI: 10.1080/00028487.2013.824918

SPECIAL SECTION: BURBOT

Ups and Downs of Burbot and Their Predator Lake Trout in Lake Superior, 1953–2011

Owen T. Gorman* U.S. Geological Survey, Great Lakes Science Center, Lake Superior Biological Station, 2800 Lake Shore Drive East, Ashland, Wisconsin 54086, USA Shawn P. Sitar Michigan Department of Natural Resources, Marquette Fisheries Research Station, 484 Cherry Creek Road, Marquette, Michigan 49855, USA

Abstract The fish community of Lake Superior has undergone a spectacular cycle of decline and recovery over the past 60 years. A combination of Sea Lamprey Petromyzon marinus depredation and commercial overfishing resulted in severe declines in Lake Trout Salvelinus namaycush, which served as the primary top predator of the community. Burbot Lota lota populations also declined as a result of Sea Lamprey depredation, largely owing to the loss of adult fish. After Sea Lamprey control measures were instituted in the early 1960s, Burbot populations rebounded rapidly but Lake Trout populations recovered more slowly and recovery was not fully evident until the mid-1980s. As Lake Trout populations recovered, Burbot populations began to decline, and predation on small Burbot was identified as the most likely cause. By 2000, Burbot densities had dropped below their nadir in the early 1960s and have continued to decline, with the densities of juveniles and small adults falling below that of large adults. Although Burbot populations are at record lows in Lake Superior, the density of large reproductive adults remains stable and a large reserve of adult Burbot is present in deep offshore waters. The combination of the Burbot’s early maturation, long life span, and high fecundity provides the species with the resiliency to remain a viable member of the Lake Superior fish community into the foreseeable future.

The fish community of Lake Superior has undergone a spec- 1968, 1972; Christie 1974; Lawrie 1978; Smith and Tibbles tacular cycle of decline and recovery over the past 60 years 1980). Subsequent studies have shown that the Burbot popula-

Downloaded by [Department Of Fisheries] at 22:58 25 November 2013 (Lawrie and Rahrer 1972, 1973; Lawrie 1978; Bronte et al. tions in Lake Superior rebounded by the 1970s (Schram et al. 2003; Gorman and Hoff 2009; Gorman 2012). A combination of 2006; Stapanian et al. 2008), but the recovery has not been well Sea Lamprey Petromyzon marinus depredation and commercial documented due to a lack of assessment data prior to 1970. Since overfishing resulted in severe declines in Lake Trout Salvelinus 1990, the abundance of Burbot in Lake Superior has trended namaycush, which served as the primary top predator in the downward and the decline has been attributed to increased pre- Great Lakes (Smith 1968, 1972; Lawrie and Rahrer 1972, 1973; dation by recovered Lake Trout populations (Schram et al. 2006; Christie 1974). Burbot Lota lota, the secondary large native pis- Stapanian et al. 2008). Evidence that adult Lake Trout prey on civore of the Great Lakes (Smith 1968; Christie 1974), which Burbot is found in diet studies (Conner et al. 1993; Peck and overlaps broadly with Lake Trout in habitat use (Selgeby and Sitar 2000; Ray et al. 2007; Sitar et al. 2008; Sitar, in press) Hoff 1996; Gorman et al. 2012b), suffered a parallel decline and food web studies based on stable isotope analysis (Harvey in response to Sea Lamprey depredation and with Lake Trout et al. 2003). Studies conducted in eastern Lake Superior show reached its nadir in Lake Superior in the early 1960s (Smith that Burbot represent a principal component of the diet of adult

*Corresponding author: [email protected] Received October 1, 2012; accepted July 9, 2013 1757 1758 GORMAN AND SITAR

wild lean Lake Trout and siscowets, along with coregonines and Gill-net catch data from 1953 to 2001 were used to elucidate the sculpins (Slimy Sculpin Cottus cognatus, Spoonhead Sculpin changes in Burbot density and size–age structure. Bottom trawl Cottus ricei, and Deepwater Sculpin Myoxocephalus thomp- catch data from 1975 to 2011 were used to elucidate trends in soni) (Peck and Sitar 2000; Sitar et al. 2008; Sitar, in press). the densities of juvenile and adult Burbot. The diets of Lake Diet studies suggest that coregonines larger than 230 mm TL Trout captured from 2001 to 2012 were examined for evidence may have reduced vulnerability to predation due to the prey size of differential predation on smaller Burbot. These analyses were limitation of adult Lake Trout (Conner et al. 1993). Similarly, made in the context of concurrent changes in Lake Trout pop- larger Burbot may be less vulnerable to predation because of the ulations in Lake Superior since 1953 and are summarized in prey size limitation of adult Lake Trout, but the size at which terms of recent changes in the Lake Superior fish community, Burbot become less vulnerable to Lake Trout predation is not Burbot life history, and future trends in Lake Superior Burbot known. Nevertheless, Lake Trout likely affect Burbot popula- populations. tions by predation on younger life stages, with the consequence that the subsequent abundance of adult Burbot is lower. In addi- tion, Lake Trout may compete with Burbot for food, as the two METHODS species show considerable overlap in diets (Schram et al. 2006). Trends in the abundance and sizes of Burbot and adult wild The objective of our study was to address population changes Lake Trout (the latter defined as fish >400 mm TL) in U.S. in Burbot and its primary predator, Lake Trout, in Lake Superior waters of Lake Superior were assessed from catch records from over the 59 years from 1953 to 2011. The earlier year represents gill-net and bottom trawl surveys conducted in 1953 and from a baseline year for the Lake Superior fish community prior to 1958 to 2011 (Figure 1). The records for lean Lake Trout and sis- the rapid expansion of Sea Lamprey populations; the late 1950s cowets were combined into those for “wild Lake Trout” because and early 1960s incorporate the nadir of Lake Trout and Burbot earlier records were inconsistent in discriminating between the populations and coincide with the maximum abundance of Sea two morphs. Hatchery Lake Trout were excluded from the analy- Lampreys; the mid-1960s to mid-1970s include the recovery of sis because we were interested in long-term trends that included Burbot populations following implementation of Sea Lamprey periods when adult hatchery Lake Trout were rare or absent control measures in the early 1960s; the late 1970s to mid-1980s (prior to the early 1960s and after the mid-1980s). Moreover, cover the early and middle stages of the recovery of Lake Trout Burbot represent a small fraction of the diet of hatchery Lake populations; the late 1980s to mid-1990s capture the full recov- Trout (Sitar, in press). With one exception, the gill-net catch ery of Lake Trout populations and the decline of Burbot; and data were from surveys conducted in Michigan waters of Lake the late 1990s to the early 2000s represent the succession of the Superior in 1953 and 1958–2001; the exception was length- fish community toward equilibrium. Our analysis was guided by frequency data for Burbot from 1966 to 1969, which were from the general hypothesis that predation by both Sea Lampreys and Bailey (1972) and obtained from gill-net surveys conducted in Lake Trout contributed to the declines of Burbot populations in Michigan and Wisconsin waters of Lake Superior. Bottom trawl Lake Superior over the past 59 years. From this premise, we pro- catch data were from surveys conducted in nearshore waters ceeded to a series of specific hypotheses addressing the evidence (∼15–80 m deep) throughout U.S. waters (1953, 1978–2011) of predation on Burbot by Lake Trout and Sea Lampreys. and in western Lake Superior and the Apostle Islands region Strong Sea Lamprey depredation of Burbot stocks in the (1953, 1958–1977) (Figure 1). late 1950s should be evidenced by a truncation of the older For all years, multifilament nylon gill nets (2.0–4.5-in [51– individuals in the age distribution, as older (larger) fish are 114-mm] stretch mesh) were set for 24 h on the lake bottom to likely to be the preferred targets of Sea Lampreys (Smith 1968, sample demersal habitat over a 20–188-m depth range between

Downloaded by [Department Of Fisheries] at 22:58 25 November 2013 1972; Smith and Tibbles 1980). After 1962, when control May 1 and September 30 in Michigan waters of Lake Superior measures reduced Sea Lamprey abundance by 90% (Smith and from Keweenaw Bay to Whitefish Point (Figure 1). Gill nets Tibbles 1980), the recovery of the Burbot population should be were constructed of 300-ft × 6-ft (92-m × 1.8-m) panels of evidenced by an increased abundance of younger (smaller) fish multiple meshes tied end-to-end in a random sequence. In some and a recovering age structure, including larger adults. After cases, panel lengths varied from 150 to 600 ft (46–183 m). the recovery of Lake Trout populations and the subsequent Catches were enumerated by individual gill-net panels and tran- increased abundance of large adult Lake Trout, increased scribed to computerized files; catch per unit effort (CPUE) was predation on Burbot populations should be evidenced by the calculated for each species and panel. Trends in the density of truncation of younger fish in the age distribution. This, in turn, Burbot and Lake Trout were expressed as mean CPUE (number should lead to a declining abundance of adults, as increased of fish per 1,000 ft of gill net per night) for the periods 1953, juvenile mortality will result in fewer fish reaching adult size. 1959, 1961–1970, 1971–1978, 1981–1986, and 1994–2000. The These trends should be supported by evidence of differential interaction between the densities of Lake Trout and Burbot was predation on smaller Burbot by adult Lake Trout. assessed by linear regression of the density of Burbot (the de- These hypotheses were addressed by analysis of Lake Supe- pendent variable) on that of wild Lake Trout (the independent rior assessment data dating back to 1953 that have heretofore not variable). The data for 1959 were omitted from the regression been inspected with respect to the trends in Burbot populations. model because heavy Sea Lamprey depredation negated the BURBOT AND LAKE TROUT IN LAKE SUPERIOR 1759

FIGURE 1. Map of Lake Superior showing the locations at which Burbot and Lake Trout were sampled, 1953–2011. Trawl locations shown in the Apostle Islands region (dashed circle) were sampled annually, 1975–2011.

interaction by reducing the densities of both Burbot and wild Trends in abundance of Burbot from the bottom trawl catches Lake Trout to low levels. Kolmogorov–Smirnov (KS) tests were taken lake-wide in 1978–2011 were expressed as mean density used to assess whether the length-frequency distributions of Bur- of fish <301 mm, 301–500 mm, and >500 mm TL. Trends bot derived from gill-net catch data for the periods 1953, 1958– in abundance of Burbot from the bottom trawl catches taken 1965, 1966–1969, 1974–1984, and 1989–2001 represented the in the Apostle Islands in 1975–2011 were expressed as mean same or different populations. The length-frequency distribu- densities of fish in the same three size-classes. Prior to 1975, tions were based on the numbers of Burbot captured in a period Burbot captured in bottom trawls were not measured, so es- arrayed into 20-mm size bins ranging from 160 to 700 mm TL. timates of mean density were for all fish. Length-at-age data Bottom trawls of similar design were used to sample for Burbot collected in Michigan waters of Lake Superior in fish in demersal habitat in nearshore waters. In 1953 and 1998–2011 (Michigan Department of Natural Resources, un- 1958–1973, bottom trawls with 9.8–10.1-m headropes, 11.6– published data) were used to assign fish to the following size– Downloaded by [Department Of Fisheries] at 22:58 25 November 2013 12.2-m footropes, and 13-mm stretch-mesh cod ends were used. age groups: juveniles (<301 mm, ≤6 years of age), small adults From 1974 to 2011, bottom trawls with an 11.9-m headrope, (301–500 mm, 7–9 years), and large adults (>500 mm, 10 years 15.5-m footrope, and 13-mm stretch-mesh cod ends were used. and older). Burbot ages were estimated by the crack-and-burn These trawls sampled the bottom 1.0–1.5 m of the water col- method of age analysis of otoliths (Schreiner and Schram 2001). umn (U.S. Geological Survey, unpublished trawl mensuration Lake trout densities were expressed for fish >400 mm TL, data), commensurate with the demersal stratum sampled by gill which were adults age 7 and older (Gorman et al. 2008, 2012a) nets. Trawls were made at depths ranging from 15 to 100 m dur- that were considered the greatest source of predation on small ing daylight hours for 10–60 min. Most sampling was conducted Burbot (Sitar et al. 2008). The annual mean densities of Burbot during the months of May and June; the exceptions included ad- and Lake Trout from the bottom trawl catch data were trans- ditional samples from July–September in 1953 and 1958–1974. formed by prior-year averaging (i.e., moving averages of the The area swept by each trawl was calculated as the average values from the present year and the previous one) to dampen wingspread as measured by mensuration gear (the width of the the interannual variation without changing the direction of the trawl mouth) multiplied by the distance the trawl was towed. trend. Typically, a reduction of the population SD to ≤70% of Catches from each trawl tow were enumerated by species and the mean by prior-year averaging provided sufficient interannual expressed as the density of fish per hectare. dampening to visually clarify population trends. For Burbot and 1760 GORMAN AND SITAR

Lake Trout, prior-year averaging of the present and previous TABLE 1. Sampling effort in U.S. waters of Lake Superior and the Apostle 2 years (i.e., 3-year moving averages) was sufficient to meet Islands region, 1953–2011. All gill-net surveys were conducted in Michigan waters from Keweenaw Bay to Whitefish Point (management units MI-4 through this goal. MI-7); the Apostle Islands region conforms to Wisconsin management unit WI- Strong year-classes of Burbot were estimated by analysis of 2. Sets, panels, and TL are descriptors of gill-net effort, and tows and area swept the trends in the densities of juvenile and small adult Burbot are descriptors of bottom trawl effort. Figure 1 shows the areas where sampling based on bottom trawl data from U.S. waters and the Apostle was conducted. Islands region. Intervals of higher density separated by inter- No. of vals of lower density were interpreted as the result of alternat- No. of No. of panels/area TL ing stronger and weaker year-classes. Strong year-classes were Period locations sets/tows swept (ha) (m) estimated by subtracting the estimated ages of juveniles and small adults from the years of peak densities. Agreement in Gill nets the predicted strong year-classes generated from juveniles and 1953 20 28 97 9,695 small adults was judged as a validation of year-class estimates. 1959 22 79 538 44,710 Cohen’s weighted K was used to assess agreement and 1961–1970 24 55 321 39,596 Wilcoxon’s signed rank test (W) was used to assess the dif- 1971–1978 23 142 613 68,963 ferences in the arrays of predicted year-classes derived from 1981–1986 17 60 416 38,049 juveniles and small adults. 1994–2000 13 234 289 26,433 Interactions between the trends in the density of adult Lake Total 119 598 2,274 227,447 Trout and Burbot from bottom trawl catches for 1978–2001 in Bottom trawls U.S. waters and 1975–2001 in the Apostle Islands region were 1953a 62 213 297.4 assessed by means of linear regression models, with Lake Trout 1958–1962b 23 230 112.5 as the independent variable and Burbot as the dependent vari- 1963–1977b 13 486 261.4 able. Because Burbot may show a delayed population response 1978–2011a 53 1,644 2,055.7 to changes in the abundance of adult Lake Trout, the densities of Total 151 2,573 2,727.0 Lake Trout were subjected to 0–6-year time lags in comparison with the densities of all Burbot and juvenile, small adult, and a U.S. waters. b large adult Burbot. With the maximum 6-year lag, Lake Trout Apostle Islands. densities in U.S. waters from the 1978–2001 period were com- pared with Burbot densities from the 1984–2007 period. For consumed by Lake Trout was expressed as the mean ratio of the Apostle Islands region, the maximum 6-year lag resulted the length of Burbot consumed to the length of Lake Trout plus in the comparison of Lake Trout densities from the 1975–2001 one SD, multiplied by the mean length of Lake Trout plus one period with Burbot densities from the 1981–2007 period. We SD. This value represents an estimated prey size limitation for interpreted the Burbot size-classes with the highest R2 values the Lake Trout population; Burbot larger than this estimated and significance to be the most likely targets of Lake Trout limitation were judged to be relatively immune to predation. predation, and we interpreted the time lags with the highest R2 values and significance to be the most likely response intervals RESULTS of Burbot populations to Lake Trout predation. Regressions with P < 0.05 were considered indicative of significant interaction. Gill-Net Surveys Significant regressions with R2 ≥ 0.50 were considered to have Over six sampling periods from 1953 to 2000, 227,447 m 2 Downloaded by [Department Of Fisheries] at 22:58 25 November 2013 strong correlations, those with 0.30 ≤ R < 0.50 to have moder- of gill nets were set to characterize the Burbot and Lake Trout ate correlations, and those with 0.15 ≤ R2 < 0.30 to have weak populations in Michigan waters of eastern Lake Superior (Fig- correlations. ure 1; Table 1). The gill-net CPUE of Burbot declined between Lake trout predation on Burbot was assessed by inspection 1953 and 1959, increased sharply over the period 1961–1970, of the stomach contents of wild Lake Trout collected in annual and declined sharply to low levels over the period 1971–2000 gill-net assessments in Michigan waters of Lake Superior at (Figure 2A). The CPUE of Wild Lake Trout declined sharply depths of 24–399 m during 2001–2012 (Michigan Department between 1953 and 1959, began to recover slowly through the of Natural Resources, unpublished data). The length-frequency 1960s and 1970s, and increased sharply after 1978 (Figure 2A). distributions of Burbot taken from the stomachs of 170 wild lean The frequency of occurrence of Burbot and Lake Trout in gill- Lake Trout and 196 siscowets were generated to characterize the net catches generally paralleled the trends in CPUE (Figure 2B). size range of Burbot consumed by Lake Trout. The TLs of wild The early stages of recovery of Lake Trout were evident in the lean Lake Trout and siscowets and the Burbot they consumed 1971–1978 period, as the proportion of panels with adult Lake were compared by linear regression models (with Lake Trout as Trout increased more rapidly than CPUE following the 1959 the independent variable and Burbot as the dependent variable) and 1961–1970 periods. Regression of Burbot density on Lake to assess the relationship between trout size and the size of Trout density showed a significant negative relationship (P < Burbot consumed. Estimated asymptotic prey size of Burbot 0.009; Figure 2C). BURBOT AND LAKE TROUT IN LAKE SUPERIOR 1761

97 538 321 613 416 289 0.2 10.0 1953 lake trout 0.15 8.0 0.1 6.0

A Proporon 0.05 107 4.0

0 2.0 160 200 240 280 360 480 520 560 680 320 400 440 600 640 Mean CPUE (# fish/1000'/night)(# Mean CPUE burbot 0.2 0.0 1953 1959 1961-1970 1971-1978 1981-1986 1994-2000 1958-1965 0.8 0.15 lake trout 0.7 0.1 0.6

0.5 Proporon 0.05 1607 0.4 B 0 0.3 680 240 320 440 480 560 600 640 200 280 400 520 160 360 0.2 0.2 1966-1969 0.1 burbot Proporon ofProporon with Catch Panels 0.15 0.0 1953 1959 1961-1970 1971-1978 1981-1986 1994-2000 Period 0.1 t 12.0 h g Proporon i

n 0.05 659 /

' 10.0 0 0 0

1 0 / 8.0 h s 480 520 600 640 680 400 160 200 240 280 360 320 560 440 i f C #

( 6.0 0.2 t

o 1974-1984 b r

u 4.0

B y = -1.0799x + 9.3884 0.15 f

o R² = 0.9278

y 2.0 t i

s 0.1 n e

D 0.0 Proporon 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 0.05 2375 Density of Lake Trout (# fish /1000' /night) 0 FIGURE 2. Trends in the mean density of Burbot and wild Lake Trout from Downloaded by [Department Of Fisheries] at 22:58 25 November 2013 200 680 280 320 360 440 gill-net surveys conducted in Michigan waters of Lake Superior, 1953–2000. 160 240 400 480 520 560 600 640 0.2 Gill nets with randomized multimesh (2.0–4.5 in [51–114 mm]) panels were 1989-2001 deployed in waters 20–188 m deep. Panel (A) shows the trends in catch per unit effort (CPUE) for Burbot and wild Lake Trout; the error bars represent 0.15 SEs. Panel (B) shows the trends in the proportion of gill-net panels in which there were Burbot or Lake Trout. Panel (C) portrays the regression of Burbot 0.1 densities on Lake Trout densities from panel (A), omitting the 1959 values.

0.05Proporon 147

The length-frequency distributions of Burbot derived from 0 catch data from gill nets set in Michigan waters showed chang- 360 160 200 240 280 400 560 600 680 520 320 640 440 480 ing population size structure from 1953 to 2001 (Figure 3). Size bin, mm The 1953 distribution was composed of 14% juveniles, 71% small adults, and 15% large adults. The 1958–1965 distribu- FIGURE 3. Length-frequency distributions of Burbot caught in gill-net sur- tion showed truncations of juveniles (to 10%) and large adults veys conducted in Michigan waters of Lake Superior for various periods. The data shown for 1966–1969 are from Bailey (1972) and contain records from (to 3%), with most of the population being composed of small Michigan and Wisconsin waters of Lake Superior. To account for the differ- adults (87%). During 1966–1969, the proportions of juveniles ences in sample size (reported in the lower right of each panel) and improve and large adults increased to 13% and 14%, respectively, and clarity, the distributions are shown as proportional abundances by size. 1762 GORMAN AND SITAR

TABLE 2. Comparisons of length-frequency distributions of Burbot, 1953- 2001 based on counts (Figure 3) with Kolmogorov-Smirnov (KS) test statistic D (values above the diagonal) and probability of being the same (italicized values below the diagonal). Significant values of D and corresponding probabilities are showninbold(P < 0.05).

Period 1953 1958–1965 1966–1969 1974–1984 1985–1987 1989–2001 1953 0.4640 0.7860 0.7140 0.6070 0.2500 1958–1965 0.0029 0.4290 0.3210 0.2140 0.4290 1966–1969 0.0000 0.0077 0.3570 0.2500 0.6790 1974–1984 0.0000 0.0870 0.0420 0.2140 0.5710 1985–1987 0.0000 0.4900 0.3003 0.4900 0.4640 1989–2001 0.3003 0.0077 0.0000 0.0001 0.0029

that of small adults decreased to 73%. In the period 1974–1984, (0.36–0.67 fish/ha) in 1958–1969 to peak levels in 1972–1974 the proportion of juveniles increased again (to 19%), while the (0.96 fish/ha; significantly positive slope; P < 0.001; Figure 5). proportion of small adults increased slightly (to 76%) and that Afterwards, the density of all Burbot declined gradually with of large adults decreased to 5%. In the final period, 1989–2001, fluctuation to 0.07 fish/ha in 2011 (significantly negative slope; the proportion of juveniles in the population contracted to 8%, P < 0.001; Figure 5). Inspection of the trends in density by that of small adults increased to 87%, and that of large adults re- size-class starting in 1975 shows that most of the decline after mained at 5%. Comparison of the length-frequency distributions 1972–1974 was driven by the declining densities of juveniles based on counts by means of the KS test statistic showed that and small adults (significantly negative slopes; P < 0.001; Fig- the 1953 and 1989–2001 distributions were similar (D = 0.25, ure 5). The density of small adults showed the greatest overall P > 0.30; Table 2) and distinct from those in other periods (D > decline, fluctuating about 0.3 fish/ha in the period 1975–1993 0.46, P < 0.01; Table 2). Comparisons of the distributions for and then declining more than 10-fold to 0.02 fish/ha in 2011. 1958–1965, 1966–1969, 1974–1984, and 1985–1987 revealed The density of juveniles was characteristically much lower than only two significant differences: that between 1958–1965 and the density of small adults, fluctuating about 0.1 fish/ha in the 1966–1969 and that between 1966–1969 and 1974–1984 (D > period 1975–1995 and then declining to <0.02 fish/ha after 0.35, P < 0.05; Table 2). 1998 and <0.01 fish/ha after 2005. The density of large adults did not show a discernible trend (slope not significantly differ- Bottom Trawl Surveys ent from 0; P > 0.17; Figure 5), fluctuating between 0.00 and Between 1953 and 2011, the long-term trends in wild adult 0.11 fish/ha over the time series. During 2005–2011, the density Lake Trout and Burbot of various size- and age-classes in of large adults exceeded that of juveniles and small adults. nearshore waters were characterized by 2,573 trawl tows cover- Superimposed on the downward trending plots of Burbot ing 2,727 ha in Lake Superior (Figure 1; Table 1). Over the densities are waves of increasing and decreasing densities of ju- period 1978–2011, bottom trawl surveys in U.S. waters re- venile and small adult Burbot in both U.S. waters (1978–2011; vealed changing size and age structure in Burbot populations; Figure 4) and the Apostle Islands region (1975–2011; Figure 5). overall density declined, driven by declining densities of juve- The peaks and troughs along the trend plots may reflect cycles niles and small adults (significant negative slopes; P < 0.001; of stronger and weaker year-classes. This hypothesis was tested Figure 4). Adult densities showed no trend (slope not signif- by comparing the estimated strong year-classes for juveniles

Downloaded by [Department Of Fisheries] at 22:58 25 November 2013 icantly different from 0; P > 0.88; Figure 4). The density of with those for small adults to see whether there was agreement juveniles increased from 0.08–0.14 fish/ha in 1978–1984 to a between them. Strong year-classes were estimated by subtract- peak of 0.24 fish/ha in 1987 and then abruptly dropped to 0.04 ing the estimated ages of juveniles and small adults from the fish/ha in 1990 and fluctuated with a downward trend to a low of years of peak densities. For juveniles, we used the estimated 0.01 fish/ha in 2011. Though fluctuating, the density of small median age (5 years) of larger juveniles (201–300 mm TL), adults declined more evenly over the time series, from peak as they dominated that size-group. For small adults, we found densities of 0.25 fish/ha in 1978–1981 to 0.01–0.03 fish/ha in that the minimum estimated age (7 years) provided the great- 2007–2011. The density of large adults fluctuated without a est agreement with year-class estimates from larger juveniles discernible trend between 0.01 and 0.10 fish/ha. During 2007– for both U.S. waters and the Apostle Islands region (Cohen’s 2011, the density of large adults exceeded that of juveniles and K > 0.94) and were not significantly different (Wilcoxon’s W ≥ small adults. 0.50). The consensus years for estimated stronger year-classes of The longer time series of bottom trawl samples from the Burbot in U.S. waters and the Apostle Islands region were 1970, Apostle Islands region of Lake Superior (1958–2011) allowed a 1974, 1979, 1985, 1990, 1997, and 2003. This array of stronger more extensive evaluation of the trends in Burbot density (Fig- Burbot year-classes is surprisingly similar to that reported for ure 5). The density of all Burbot increased from moderate levels Cisco Coregonus artedi and Bloater C. hoyi in Lake Superior BURBOT AND LAKE TROUT IN LAKE SUPERIOR 1763

0.50 All sizes 0.40

0.30

0.20

0.10 y = -0.0108x + 0.444 R² = 0.7642 0.00

0.20 < 301 mm

0.15

0.10

0.05 y = -0.0047x + 0.161 R² = 0.6138 0.00

0.20 301-500 mm

Mean Density, No. / ha Density, Mean 0.15

0.10

0.05 y = -0.006x + 0.2373 R² = 0.7195 0.00 Downloaded by [Department Of Fisheries] at 22:58 25 November 2013 0.20 > 500 mm y = -7E-05x + 0.0457 0.15 R² = 0.0007

0.10

0.05

0.00 1978 1982 1984 1986 1988 1990 1994 1996 1998 2000 2002 2006 2008 2010 1980 2004 1992 Year

FIGURE 4. Trends in the density of Burbot from lakewide bottom trawl surveys in the nearshore areas of U.S. waters of Lake Superior, 1978–2011, by size-class. Burbot <301 mm TL represent fish 6 years old and less, those 301–500 mm represent fish 7–9 years old, and those >500 mm represent fish ≥10 years old. The data shown are 3-year moving averages of the present and prior 2 years. Regression lines (dashes) are superimposed on the time series plots. 1764 GORMAN AND SITAR

1.0

0.9 0.8 All sizes 0.7

0.6

0.5

0.4

0.3

0.2 y = 0.0308x + 0.2984 y = -0.0131x + 0.6051 0.1 R² = 0.5591 R² = 0.7734 0.0 2008 1986 1990 1992 1994 1996 2000 2004 2006 2010 1978 1980 1982 1984 1964 1966 1974 1988 1998 2002 1958 1960 1962 1968 1970 1972 1976 Year y = -0.0033x + 0.1291 R² = 0.7479 < 301 mm 0.1

0.0 Mean Density, No. / Meanha Density,

0.4 301-500 mm

0.3

0.2 y = -0.0091x + 0.4071 R² = 0.721 0.1 Downloaded by [Department Of Fisheries] at 22:58 25 November 2013

0.0

y = 0.0005x + 0.042 > 500 mm R² = 0.0508 0.1

0.0 1993 1979 1981 1999 2003 1975 1977 1983 1985 1987 1989 1991 1995 1997 2001 2005 2007 2009 2011 Year

FIGURE 5. Trends in the density of Burbot from bottom trawl surveys in nearshore waters of the Apostle Islands region of Lake Superior, 1958–2011, by size-class. The upper panel (for all sizes of Burbot) is divided into two series, 1958–1974 and 1975–2011; the 1975–2011 series reflects sampling conducted in May–June from a fixed set of locations (Figure 1), whereas the 1958–1974 series reflects sampling conducted in May–September at varying locations. See Figure 4 for more details. BURBOT AND LAKE TROUT IN LAKE SUPERIOR 1765

TABLE 3. Results of regressions of the density of adult Lake Trout on that of Burbot, by size-class (all Burbot, juveniles [<301 mm TL], small adults [301– 500 mm TL], and large adults [>500 mm TL]) with time lags of 0–6 years in the densitie of Lake Trout. The regressions for U.S. waters are based on annual density estimates for 1978–2007, those for the Apostle Islands region on annual density estimates for 1975–2007. P-values ≤0.05 and R2 values ≥0.50 are indicated by bold italics.

All Burbot Juveniles Small adults Large adults Lag (years) R2 rPR2 rPR2 rPR2 rP U.S. waters 0 0.03 0.18 0.392 0.02 0.13 0.541 0.03 0.16 0.461 0.01 0.11 0.616 1 0.03 0.16 0.449 0.06 0.24 0.245 0.00 0.02 0.909 0.00 0.06 0.798 2 0.10 0.31 0.133 0.23 0.48 0.480 0.02 0.15 0.487 0.01 0.08 0.703 3 0.25 0.50 0.013 0.50 0.71 0.000 0.07 0.27 0.199 0.01 0.09 0.682 4 0.37 0.61 0.002 0.61 0.78 0.000 0.11 0.34 0.109 0.00 0.03 0.881 5 0.34 0.59 0.003 0.55 0.74 0.000 0.08 0.28 0.183 0.00 0.03 0.889 6 0.28 0.53 0.008 0.44 0.66 0.000 0.09 0.31 0.148 0.00 0.05 0.801 Apostle Islands 0 0.20 0.44 0.026 0.11 0.34 0.085 0.22 0.47 0.013 0.03 0.17 0.395 1 0.15 0.39 0.045 0.19 0.44 0.023 0.21 0.45 0.018 0.05 0.22 0.267 2 0.16 0.40 0.040 0.37 0.61 0.001 0.17 0.41 0.035 0.01 0.11 0.587 3 0.19 0.44 0.023 0.41 0.64 0.000 0.18 0.43 0.025 0.00 0.02 0.901 4 0.24 0.49 0.009 0.51 0.71 0.000 0.17 0.41 0.034 0.04 0.19 0.342 5 0.28 0.53 0.004 0.43 0.65 0.000 0.20 0.45 0.019 0.14 0.38 0.053 6 0.41 0.64 0.000 0.39 0.62 0.000 0.30 0.55 0.003 0.17 0.42 0.031

from 1978, namely, 1978, 1984, 1988–1990, 1998, and 2003 largely by the significant negative relationship between the den- (Bronte et al. 2003; Gorman and Hoff 2009; Gorman et al. sities of adult Lake Trout and juvenile Burbot, again with a 2012a); the agreement of strong year-classes of Burbot, Cisco, 4-year lag giving the highest R2 (0.61) and level of significance and Bloater between 1978 and 2011 was high (Cohen’s K = (P ≤ 0.0001) (Table 3; Figure 7A). There were no significant 0.94), and the differences were not significant (Wilcoxon’s W regressions between the densities of adult Lake Trout and small ≥ 0.28). and large adult Burbot (Table 3; Figure 7). Comparison of the The density of adult wild Lake Trout in nearshore U.S. waters density of adult Lake Trout lagged 4 years and the density of of Lake Superior over the period 1978–2011 showed a sharp small adult Burbot suggests a negative relationship, but it was increase after 1984, fluctuated about 0.25 fish/ha from 1985 not significant (P = 0.109) and the R2 was a low 0.11 (Table 3; to 1997, declined to levels fluctuating about 0.20 fish/ha from Figure 7A). 1998 to 2006, and declined thereafter (Figure 6A). The more The regressions between the densities of adult Lake Trout extensive data series from the Apostle Islands region permitted and all Burbot in the Apostle Islands region over the period inspection of the trend in the density of adult Lake Trout back to 1975–2007 were significantly negative for all time lags (P < 1958 (Figure 6B). Starting from near zero in 1958, the density 0.05), and the R2 values ranged from weak to moderate (0.15– Downloaded by [Department Of Fisheries] at 22:58 25 November 2013 of adult Lake Trout increased with fluctuating waves to a peak 0.41) (Table 3; Figure 7B). The interaction between the densities of 0.59 fish/ha in 1997 and then declined steadily to 0.06– of all Burbot and adult Lake Trout was driven by the significant 0.10 fish/ha in 2009–2011. For comparison, the estimated negative relationship between the densities of adult Lake Trout lakewide density of adult Lake Trout in 1953 was 0.31 fish/ha. and juvenile and small adult Burbot (Table 3). Significant nega- The interactions between adult Lake Trout and Burbot pop- tive regressions were obtained between the densities of juvenile ulation trends derived from nearshore bottom trawl catch data Burbot and adult Lake Trout with time lags of 1–6 years (P < in U.S. waters were evaluated by regression of the densities of 0.03); the regression with a 4-year lag was the most significant the two species over the period 1978–2007. In U.S. waters, the (P ≤ 0.0001) and had the highest R2 (0.51) (Table 3; Figure 7B). density of Burbot (all sizes) was negatively related to the density The regressions of the densities of small adult Burbot and adult of adult Lake Trout with 3–6-year time lags. In this compari- Lake Trout were significantly negative for time lags of 0–5 years son, the 4-year time lag was the most significant (P = 0.002), (P < 0.04), but the correlations were weak (R2 = 0.17–0.22) though in all cases the R2 values were <0.38, indicating only (Table 3; Figure 7B). Notably, the regression with a 6-year lag a moderate correlation (Table 3; Figure 7A). The interactions was the most significant (P = 0.003) and had a moderate cor- between the densities of all Burbot and Lake Trout were driven relation (R2 = 0.30). The regressions of the densities of large 1766 GORMAN AND SITAR

0.40

0.35 A

0.30

0.25

0.20

0.15

0.10

0.05

0.00 2002 2004 2006 2008 1996 1998 2000 2010 1984 1986 1988 1992 1994 1978 1980 1982 1990

0.60 Density, Fish / ha Density, 0.50 B

0.40

0.30

0.20

Downloaded by [Department Of Fisheries] at 22:58 25 November 2013 0.10

0.00 1985 1961 1964 1958 1973 1976 1979 1994 1997 2000 2009 1988 1970 1982 1967 2003 2006 1991 Year

FIGURE 6. Trends in the density of wild adult Lake Trout from bottom trawl surveys in nearshore waters of Lake Superior. Panel (A) shows the trends in U.S. waters over the period 1978–2011, panel (B) the trends in the Apostle Islands region over the period 1958–2011. The horizontal dashed lines indicate the density of adult Lake Trout in U.S. waters in 1953. The dashed vertical line in panel (B) is a reference for the start of the lakewide bottom trawl surveys by the U.S. Geological Survey in 1978. The data shown are 3-year moving averages of the present and prior 2 years. BURBOT AND LAKE TROUT IN LAKE SUPERIOR 1767

0.7 1 2 3 3 454 5 6 6 7 7 8 899 Bbt ALL A 25 0.6 Bbt <301 A Bbt 301-500 20 0.5 Bbt >500 15 0.4 2

R 10

0.3 Frequency 5 0.2

0.1 0 60 90 30 330 480 120 150 180 210 240 270 390 300 0.0 360 420 450 Size bin, mm TL 0123456 0.6 B 1.0 Bbt ALL B 0.5 Bbt <301 0.8 Bbt 301-500 0.4 Bbt >500 0.6 2 R 0.3 0.4

0.2 0.2 Cumulave proporon 0.1 0.0 30 90 60 120 210 270 330 420 450 150 180 240 300 480 0.0 360 390 0123456 Size bin, mm TL

Lag period, years C 500 y = 0.5456x -149.36 2 R² = 0.5099 FIGURE 7. R values from linear regression models comparing density esti- 400 mates of Burbot (Bbt) by size-class with density estimates of wild adult Lake Trout. Shown are the results of regressions reflecting 0–6-year time lags in Lake 300 Trout densities. Panel (A) shows the results for regression models for U.S. wa- ters over the period 1978–2007, panel (B) the results for regression models for 200 the Apostle Islands region over the period 1975–2007. 100 Burbot length length (mm) Burbot adult Burbot and adult Lake Trout were not significant with time > 0 lags of 0–5 years (P 0.05), and no correlations were found 300 400 500 600 700 800 900 1000 2 < (R 0.15) (Table 3; Figure 7B). The regression with a 6-year Lean lake trout length (mm) lag was significant (P = 0.03), but the correlation was weak (R2 = 0.17). FIGURE 8. Analyses of the sizes of wild lean Lake Trout and Burbot con- sumed in Michigan waters of Lake Superior, 2001–2012. Panel (A) shows the Downloaded by [Department Of Fisheries] at 22:58 25 November 2013 Lake Trout Consumption of Burbot length-frequency distribution of 170 Burbot removed from lean Lake Trout stomachs; the estimated ages of the Burbot (years) are indicated above the bars. The length-frequency distributions of 170 Burbot taken from Panel (B) shows the cumulative distribution of the Burbot in Lake Trout stom- lean Lake Trout stomachs showed that 84% were juveniles and achs, by size. Panel (C) portrays the regression of the TL of Burbot consumed 61% were <210 mm TL (Figure 8). Comparison of the size on the TL of lean Lake Trout. The dashed horizontal line indicates the estimated of the lean Lake Trout with the size of the Burbot consumed prey size limitation for lean Lake Trout consumption of Burbot (310 mm). yielded a significant correlation (R2 = 0.51, P < 0.0001) and regression (Wilk’s lambda = 0.49, P < 0.0001), indicating that correlation (R2 = 0.49, P < 0.0001) and regression (Wilk’s larger Lake Trout consumed larger Burbot (Figure 8). In this lambda = 0.51, P < 0.0001), indicating that larger siscowets sample, 86% of the Burbot consumed by lean Lake Trout were consumed larger Burbot (Figure 9). The slope of the regression smaller than the estimated asymptotic prey size of 310.7 mm. (0.805) was steeper than that of the regression for lean Lake Similarly, the length-frequency distributions of 196 Burbot Trout (0.545), indicating that for any given size siscowets con- taken from siscowet stomachs showed that 74% were juveniles sumed larger Burbot than lean Lake Trout (Figures 8, 9). In this and 48% were <210 mm TL (Figure 9). Comparison of siscowet sample, 87% of the Burbot consumed by siscowets were smaller size with the size of the Burbot consumed yielded a significant than the estimated asymptotic prey size of 344.0 mm. 1768 GORMAN AND SITAR

1 2 3 3 454 5 6 6 7 7 8 899 populations have been tied to those in Sea Lampreys (a predator 30 A on adult Burbot; Smith 1968, 1972; Wells and McLain 1973; 25 Christie 1974; Madenjian et al. 2002) and Alewives Alosa pseu- 20 doharengus (a predator on and potential competitor with larval Burbot; Wells and McLain 1973; Eshenroder and Burnham- 15 Curtis 1999; Madenjian et al. 2002). The institution of Sea Lam- prey control measures across the Great Lakes by the 1960s re- Frequency 10 duced the mortality of adult Burbot in Lakes Michigan (Maden- 5 jian et al. 2002), Huron (Collins et al. 1988), Erie (Stapanian 0 et al. 2006), and Ontario (Elrod et al. 1995; Mills et al. 2003, 2005; Owens et al. 2003) and set the stage for a recovery when 30 60 90 450 480 150 180 210 240 270 300 330 360 390 420 120 Alewife populations declined. That recovery became evident in Size bin, mm TL Lakes Huron, Michigan, and Ontario in the 1980s and 1990s following decreases in Alewife abundance (Elrod et al. 1995; B 1.0 Madenjian et al. 2002; Mills et al. 2003, 2005; Owens et al. 0.8 2003; Stapanian et al. 2008). In Lake Erie, where Alewives have never been abundant, the recovery of Burbot populations 0.6 in the 1980s and 1990s appeared to be the result of increased Sea Lamprey control and improved water quality (Stapanian 2006). 0.4 However, the recovery of Burbot populations in the lower Great Lakes was short-lived, as declines were evident by the 2000s. 0.2 These have been attributed to a mix of factors, including reduced Cumulave proporon 0.0 recruitment, increased Sea Lamprey depredation, and predation on larval fish (Stapanian et al. 2008, 2010), though food web 90 60 30 270 300 330 480 120 150 180 210 240 360 390 420 450 changes driven by the invasion of dreissenid mussels in these Size bin, mm TL Great Lakes have yet to be implicated. In Lake Superior, the abundance of Burbot has been declin- C 500 y = 0.8051x -285.29 ing since the early 1980s, when the recovery of Lake Trout 400 R² = 0.4909 populations became evident (Sitar and He 2006), thus implicat- ing Lake Trout predation as a principal cause of that decline 300 (Schram et al. 2006). Stapanian et al. (2008) hypothesized that prior to the 1990s, the Burbot populations in Lake Superior 200 were maintained at a relatively low abundance by predation by 100 residual siscowet stocks. Moreover, Stapanian et al. (2008) hy- Burbot length length (mm) Burbot pothesized that Sea Lamprey predation on Burbot populations 0 prior to the institution of control measures in the 1960s was 300 400 500 600 700 800 900 1000 buffered by Lake Trout, suggesting that Burbot abundance was Siscowet lake trout length (mm) relatively stable before and after the expansion of Sea Lam- prey populations. Our analysis of newly available and extensive

Downloaded by [Department Of Fisheries] at 22:58 25 November 2013 FIGURE 9. Analyses of the sizes of siscowets and Burbot consumed in data has led to a very different interpretation. The abundance Michigan waters of Lake Superior, 2001–2012. Panel (A) shows the length- of Burbot in Lake Superior declined following the expansion of frequency distribution of 196 Burbot removed from siscowet stomachs; the estimated ages of the Burbot (years) are indicated above the bars. Panel (B) Sea Lamprey populations in the 1950s, increased rapidly fol- shows the cumulative distribution of the Burbot in siscowet stomachs, by size. lowing the implementation of Sea Lamprey control measures in Panel (C) portrays the regression of the TL of Burbot consumed on the TL of the 1960s, and then began to decline following the recovery of siscowets. The dashed horizontal line indicates the estimated prey size limitation Lake Trout populations in the 1980s. These results support our for siscowet consumption of Burbot (344 mm). hypothesis that predation by both Sea Lampreys and Lake Trout has contributed to the decline of the Burbot populations in Lake DISCUSSION Superior over the past 59 years. Although the decline of Burbot populations following the Fortunately, an extensive body of data describing the state of expansion of Sea Lampreys throughout the Great Lakes in the the Lake Superior fish community back to 1953 was available 1940s and 1950s has been amply described in the literature for this study. Survey data from Lake Superior in 1953 serve as a (Smith 1968, 1972, 1973; Lawrie and Rahrer 1973; Christie snapshot of the state of the lake’s fish community at a time before 1974; Lawrie 1978; Smith and Tibbles 1980), hard data demon- Sea Lamprey depredation began to have a noticeable impact; strating this decline are lacking. In the lower Great Lakes, where after 1953, Sea Lamprey numbers began to increase sharply Lake Trout populations have not recovered, the trends in Burbot at electrical barriers and weirs in tributary streams (McLain BURBOT AND LAKE TROUT IN LAKE SUPERIOR 1769

et al. 1965; Lawrie 1970). Thus, the 1953 data set describes tion recovery) and the 1974–1984 distribution (which reflects the Lake Superior fish community in a relatively natural state increasing predation from a recovering Lake Trout population). of equilibrium in the middle of the 20th century. A follow-up The gill-net survey time series were not evaluated with time survey of Michigan waters of Lake Superior in 1959 provided lags to further elucidate the interaction between Burbot and Lake evidence of the dramatic changes in the fish community after Trout abundance for several reasons. The 1953 survey results 6 years of Sea Lamprey depredation. A nearly continuous stream represent a snapshot of the Lake Superior fish community fol- of survey data from 1958 to the present chronicles the changes lowing a protracted state of relative equilibrium; up to that point, in the fish community and provides an opportunity to discern there had been no major perturbations in the relationship be- the factors driving these changes and to make predictions about tween the Burbot and Lake Trout populations. The 1959 survey future changes. results represent a snapshot of the Lake Superior fish community Gill-net survey data provide a record of the changes in at its maximum disequilibrium, driven by the rapid expansion Burbot and Lake Trout populations dating from the 1953 bench- of invasive Sea Lamprey populations and heavy depredation mark year. By 1959, the density of adult Lake Trout had dropped of large-bodied fishes, notably Lake Trout and Burbot. Finally, precipitously, and it recovered slowly over the next 25 years the gill-net survey time series for the period 1961–2000 was (Figure 2). In contrast, the density of Burbot declined after 1953 grouped into periods of 6–10 years and thus incorporated time (though much less dramatically than that of Lake Trout), and lags into the interactions between the densities of Burbot and after its nadir in the early 1960s rose sharply in the late 1960s Lake Trout. and early 1970s, though by the turn of the 21st century it had de- The interpretation of trends based on gill-net surveys was clined to levels at or lower than the nadir of the early 1960s. The reinforced by parallel trends independently derived from bot- significant negative relationship between Lake Trout density and tom trawl data. Because the bottom trawl data series represents Burbot density is strongly suggestive of a predator–prey rela- standardized surveys that sampled extensive regions of Lake tionship, and changes in the population age–size structure pro- Superior and serves as annual snapshots of the fish community, vide further evidence that predation-mediated factors are driving they were amenable to investigation of the effects of time lags on population trends. The mix of juveniles, small adults, and large Burbot–Lake Trout interactions. Partitioning the bottom trawl adults in the 1953 Burbot population serves as a benchmark for density estimates by size-classes showed that the density of ju- a natural size–age distribution at near-equilibrium conditions. venile Burbot began to decline rapidly following the recovery The 1958–1965 population size structure showed a truncation of Lake Trout in the 1980s and declined further after 1998 (Fig- of large adults, a predicted outcome of nearly a decade of Sea ures 4–6) as continued Lake Trout predation began to reduce Lamprey depredation. The increase in the proportions of juve- the biomass of all prey species (Gorman 2007, 2012a; Gorman niles and large adults observed in 1966–1969 was the likely and Hoff 2009). The declining abundance of juvenile Burbot outcome of successful Sea Lamprey controls instituted by 1962 translated into a declining abundance of small adults and should (Smith and Tibbles 1980). The rapid expansion of the Burbot be reflected in a declining density of large adults, though the population at a time of low levels of depredation on adults by trawl data showed no discernible trend in the density of large Sea Lampreys and low levels of predation on juveniles due to adults, perhaps because bottom trawls do not accurately reflect a paucity of adult Lake Trout favored expansion of the juvenile the abundance of large Burbot. By 2007, the densities of juvenile and large adult segments of the population. The truncation of and small adult Burbot had dropped below that of large adults; small Burbot in the population size distribution after 1989 was this trend serves as evidence of strong, protracted predation by a predictable outcome of Lake Trout recovery in the mid-1980s Lake Trout. Paralleling this trend was a reduction in the abun- and was concurrent with the increased numbers of adult Lake dance of prey species to the lowest levels since lakewide bottom

Downloaded by [Department Of Fisheries] at 22:58 25 November 2013 Trout detected in gill-net surveys following recovery (Sitar and trawl surveys began in 1978 (Gorman et al. 2008, 2012a, 2013). He 2006; Figure 2). Comparison of the length-frequency dis- As with Burbot, the changes in the abundance of adult Lake tributions by means of the KS test statistic further supports Trout derived from the bottom trawl surveys corroborated the these interpretations (Table 2; Figure 3). The two distributions trends obtained from the gill-net surveys: the density of adult that reflect periods in which viable Lake Trout populations ex- Lake Trout increased sharply after 1984, signaling the presence erted strong, top-down control of the fish community (1953 and of a formidable predator population in Lake Superior that was the most recent period examined, 1989–2001) were not signifi- capable of reducing the abundance of prey species, including cantly different but differed significantly from the distributions smaller Burbot. The extended time series of bottom trawl data for the other periods. The 1966–1969 distribution in particular from the Apostle Islands region shows that the density of Bur- is unique, as it is positioned after Sea Lamprey controls were in- bot increased most rapidly when adult Lake Trout densities were stituted but before the recovery of native Lake Trout stocks and low (1962–1975), began to decline with the recovery of Lake reflects an expanded Burbot population following the suppres- Trout after 1975, and continued to decline up to the present time sion of its principal predator. Not surprisingly, the 1966–1969 (Figures 5, 6). distribution differs significantly from the adjacent distributions, These interpretations are further supported by models that the 1958–1965 distribution (which reflects the effects of Sea investigated the interaction between the density of Burbot Lamprey depredation and the early stages of Burbot popula- and that of adult Lake Trout based on bottom trawl catches. 1770 GORMAN AND SITAR

Significant negative regressions were obtained between the den- long. Similarly, the proportional biomass of Burbot in siscowet sities of adult Lake Trout and juvenile Burbot (Table 3; Figure 7), diets increased from 1% in fish 200–399 mm long to 64.7% in implicating predation on younger Burbot life stages as a princi- fish ≥800 mm long. pal means by which Lake Trout may control Burbot populations Oddly, the density of juvenile Burbot was substantially lower in Lake Superior. Moreover, these models indicated a time lag than that of small adults until the last few years of the time series, of 3–4 years for the effect of Lake Trout predation on the density when they were nearly equal (with the exception of U.S. waters of juvenile Burbot to be realized. Importantly, these time lags in the early 1980s; Figure 4). This suggests that Burbot, which were found in lakewide surveys in U.S. waters and the Apostle in their early life stages are found in shallow, inshore waters Islands. Such time lags are typical of predator–prey interactions (McPhail and Paragamian 2000) and are a conspicuous element and not unexpected for our broadly defined size-group encom- of shoreline habitat in Lake Superior embayments (Gorman and passing smaller Burbot ranging from 1 to 6 years of age. Closer Moore 2009), begin to move into deeper waters as juveniles and examination of the size composition of our <301-mm size- small adults. The results of this study suggest that the migra- class revealed that very few Burbot <201 mm were captured tion of juvenile Burbot <210 mm long into nearshore waters of in our bottom trawls. These smaller Burbot were estimated to Lake Superior exposes them to considerable risk of predation be 1–3 years of age, and the larger juveniles (201–300 mm), by a recovered Lake Trout population (Figures 8, 9). We also which represent the bulk of the size-class, were estimated to be note that while the growth rates of Burbot in Lake Superior 4–6 years of age. If predation is concentrated on smaller juve- are substantially lower than those of Burbot in other regions niles, the 3–4-year time lag that we detected in the response of of North America (Bonar et al. 2000), slower growth is most the <301-mm size-class can be explained as an artifact of the noticeable after fish attain age 2 and 200 mm in length. This sug- breadth of the size-class. That is, increased predation on Burbot gests that Burbot grow more rapidly while inhabiting warmer, of ages 1–3 would result in decreased abundance at ages 4–6 and shallow, inshore waters as juveniles and that their growth slows thus a time lag of 3–4 years for the response of this size-class to considerably when they move into the deep, cold waters of Lake greater predation. Superior. Further study of Burbot ecology in the inshore waters Examination of the Burbot consumed by Lake Trout further of Lake Superior <15 m deep would provide insight into this supports the differential predation on younger life stages. Of all aspect of Burbot early life history and the species’ ontogenetic the Burbot eaten by lean Lake Trout and siscowets, 61% and shifts in habitat use. 48%, respectively, were <210 mm (Figures 8, 9). This obser- The concordance in the occurrence of stronger year-classes vation, along with the paucity of small juvenile Burbot in our of Burbot with those of ciscoes suggests a shared vulnerability to bottom trawl samples, underscores the high level of predation factors that affect the success of recruitment in early life history on small Burbot. The close agreement between the size compo- stages. The larvae of both Burbot and ciscoes hatch in the spring, sition of Burbot consumed by Lake Trout (84% of the Burbot are pelagic, and depend on transport to shallow, nearshore nurs- consumed by lean Lake Trout and 74% of those consumed by ery habitat for survival (Anderson and Smith 1971; Scott and siscowets were ≤300 mm) and the estimated asymptotic prey Crossman 1973; Becker 1983; Selgeby et al. 1994; McPhail size of Burbot consumed by Lake Trout (310 mm for lean Lake and Paragamian 2000; Oyadomari and Auer 2004). Synchrony Trout, 344 mm for siscowets), along with the significant neg- in the appearance of weak and strong year-classes across large ative relationship between the density of adult Lake Trout and areas and among multiple species is likely driven by interannual that of Burbot ≤300 mm, provide strong evidence that Lake variation in climatic conditions on a global scale (Rook et al. Trout influence Burbot populations by differentially preying on 2012). For the Great Lakes, variation in large-scale climatic younger life stages. Further evidence that Lake Trout predation conditions may have manifold effects on their ecosystems and

Downloaded by [Department Of Fisheries] at 22:58 25 November 2013 suppresses the density of adult Burbot is provided by the signif- communities. icant negative regression (with its moderate correlation) of the Most of the survey data analyzed in this study came from density of small adult Burbot (7–9 years of age) on that of adult nearshore waters 15–80 m deep, which represents approxi- Lake Trout (with a 6-year lag) in the Apostle Islands, a finding mately 20% of the surface area of Lake Superior (Gorman et al. that is not surprising for a discrete region where Burbot–Lake 2012c). Burbot occur throughout the lake, and most individuals Trout interactions are likely to be strong (Table 3; Figure 7). in deep (>80 m), offshore waters are large adults (U.S. Geo- Extensive diet analysis of hatchery and wild lean Lake Trout logical Survey, unpublished data). Approximately 53% of all and siscowets provides additional evidence of strong predation Burbot biomass occurs in offshore waters that account for 80% pressure on Burbot populations in Lake Superior. Sitar (in press) of the lake area (Gorman et al. 2012c), suggesting that Burbot found that overall Burbot made up only 2.9% the biomass of populations have a substantial number of reproductive adults to prey items consumed by hatchery Lake Trout but that the pro- buffer future declines in recruitment. portion was much higher for wild lean Lake Trout (12.3%) and As of 2011, the Burbot populations in Lake Superior had siscowets (26.1%). Moreover, the proportional biomass of Bur- reached a new nadir that was likely driven by Lake Trout preda- bot in the diet of wild lean Lake Trout increased with size, from tion. This low density is the result of sharply reduced numbers 3.0% in those 200–399 mm long to 33.5% in those ≥800 mm of juvenile and small adult Burbot; the density of large adults BURBOT AND LAKE TROUT IN LAKE SUPERIOR 1771

appears to be unchanged. Thus, the new population nadir is Taylor and C. P.Ferreri, editors. Great Lakes fisheries and policy management: different from that in the late 1950s–early 1960s, when the den- a binational perspective. Michigan State University Press, East Lansing. sity of large adults was greatly reduced due to Sea Lamprey Gorman, O. T. 2007. Changes in a population of exotic rainbow smelt in Lake Superior: boom to bust, 1974–2005. Journal of Great Lakes Research 33: depredation. Given that the abundance of large adult Burbot in 75–90. nearshore waters is relatively unchanged, that there is a large Gorman, O. T. 2012. Successional change in the Lake Superior fish community: reserve of adult Burbot in offshore waters, that Burbot mature population trends in ciscoes, Rainbow Smelt, and Lake Trout, 1958–2008. at a relatively young age and size (4–6 years and ∼300 mm), Fundamental and Applied Limnology, Advances in Limnology 63:337–362. and are highly fecund (Bailey 1972; McPhail and Paragamian Gorman, O. T., L. M. Evrard, G. A. Cholwek, and M. R. Vinson. 2013. Status and trends in the fish community of Lake Superior, 2012. U.S. Geological 2000), the present and future reproductive potential of Burbot Survey, Great Lakes Science Center, Ann Arbor, Michigan. Available: is considerable. This suggests that the future of Burbot in Lake http://www.glsc.usgs.gov/sites/default/files/product files/2011LakeSuperior Superior is not bleak, though no substantial increases in abun- Preyfish.pdf. (September 2013). dance are to be expected as long as the population of Lake Trout Gorman, O. T., L. M. Evrard, G. A. Cholwek, D. L. Yule, and J. D. remains stable. Stockwell. 2008. Status and trends of prey fish populations in Lake Su- perior, 2007. U.S. Geological Survey, Great Lakes Science Center, Ann Arbor, Michigan. Available: http://www.glsc.usgs.gov/sites/default/files/ product files/2008LakeSuperiorPreyfish.pdf. (September 2013). ACKNOWLEDGMENTS Gorman, O. T., L. M. Evrard, G. A. Cholwek, D. L. Yule, and M. R. This article would not have been possible without the help Vinson. 2012a. Status and trends of the nearshore fish community of of many people dedicated to revealing the wonders of Lake Lake Superior, 2011. U.S. Geological Survey, Great Lakes Science Cen- Superior, including Lori Evrard for providing trawl-based data ter, Ann Arbor, Michigan. Available: http://www.glsc.usgs.gov/sites/deflautt/ summaries, and numerous Northland College student interns for files/product files/2011LakeSuperiorPreyfish.pdf. (September 2013). Gorman, O. T., and M. H. Hoff. 2009. Changes in the fish community of assisting with the entry of old field data records: Tyler Sikora, Lake Superior during 1978–2003: chronicling the recovery of a native fauna. Jill Falck, Steve Whitlock, Kari Kudick, Becky Walters, and Pages 401–437 in M. Munawar and I. F. Munawar, editors. The state of Lake John “Logan” Tucker. Thanks also to Laura Graf for providing Superior. Aquatic Ecosystem Health and Management Society, Burlington, editorial assistance in preparing the manuscript. This work was Ontario. supported by funds from the U.S. Geological Survey. This article Gorman, O. T., and S. A. Moore. 2009. Inventory of nearshore fish population densities and community structures at Apostle Islands National Lakeshore is contribution 1783 of the U.S. Geological Survey, Great Lakes and Isle Royale National Park. U.S. National Park Service, Natural Resource Science Center. Technical Report NPS/GLKN/NRTR—2009/163, Fort Collins, Colorado. Available: https://irma.nps.gov/App/Reference/Profile?Code=661080. (July 2013). REFERENCES Gorman, O. T., D. L. Yule, and J. D. Stockwell. 2012b. Habitat use by fishes of Anderson, E. D., and L. L. Smith Jr. 1971. Factors affecting the abundance of Lake Superior. I. Diel patterns of habitat use in nearshore and offshore waters Lake Herring (Coregonus artedii Lesueur) in western Lake Superior. Trans- of the Apostle Islands region. Aquatic Ecosystem Health and Management actions of the American Fisheries Society 100:691–707. 15:333–354. Bailey, M. M. 1972. Age, growth, reproduction, and food of the Burbot, Lota Gorman, O. T., D. L. Yule, and J. D. Stockwell. 2012c. Habitat use by fishes lota (Linnaeus), in southwestern Lake Superior. Transactions of the American of Lake Superior. II. Consequences of diel habitat use for habitat linkages, Fisheries Society 101:667–674. habitat coupling and energy and nutrient transfers in nearshore and offshore Becker, G. C. 1983. Fishes of Wisconsin. University of Wisconsin Press, waters. Aquatic Ecosystem Health and Management 15:355–368. Madison. Harvey, C. F., S. T. Schram, and J. F. Kitchell. 2003. Trophic relationships Bonar, S. A., L. G. Brown, P. E. Mongillo, and K. Williams. 2000. Biology, among lean and siscowet Lake Trout in Lake Superior. Transactions of the distribution and management of Burbot (Lota lota) in Washington State. American Fisheries Society 132:219–228. Northwest Science 74:87–96. Lawrie, A. H. 1970. The Sea Lamprey in the Great Lakes. Transactions of the Bronte, C. R., M. P. Ebener, D. R. Schreiner, D. S. DeVault, M. M. Petzold, American Fisheries Society 99:66–775. Downloaded by [Department Of Fisheries] at 22:58 25 November 2013 D. A. Jensen, C. Richards, and S. J. Lozano. 2003. Fish community change Lawrie, A. H. 1978. The fish community of Lake Superior. Journal of Great in Lake Superior, 1970–2000. Canadian Journal of Fisheries and Aquatic Lakes Research 4:513–549. Sciences 60:1552–1574. Lawrie, A. H., and J. F. Rahrer. 1972. Lake Superior: effects of exploitation and Christie, W. J. 1974. Changes in fish species composition of the Great Lakes. introductions on the salmonid community. Journal of the Fisheries Research Journal of the Fisheries Research Board of Canada 31:827–854. Board of Canada 29:765–776. Collins, J. J. 1988. Changes in the north channel fish community, with em- Lawrie, A. H., and J. F. Rahrer. 1973. Lake Superior: a case history of the phasis on Pink Salmon (Oncorhyncos gorbuscha Walbaum). Hydrobiologia lake and its fisheries. Great Lakes Fishery Commission Technical Report 19. 163:195–213. Available: http://www.glfc.org/pubs/TechReports/Tr19.pdf. (July 2013). Conner, D. J., C. R. Bronte, J. H. Selgeby, and H. L. Collins. 1993. Madenjian, C. P., G. L. Fahnenstiel, T. H. Johengen, T. F. Nalepa, H. A. Van- Food of salmonine predators in Lake Superior, 1981–1987. Great Lakes derploeg, G. W. Fleischer, P. J. Schneeberger, D. M. Benjamin, E. B. Smith, Fishery Commission Technical Report 59. Available: http://www.glfc.org/ J. R. Bence, E. S. Rutherford, D. S. Lavis, D. M. Robertson, D. J. Jude, and pubs/TechReports/Tr59.pdf. (July 2013). M. P. Ebener. 2002. Dynamics of the Lake Michigan food web, 1970–2000. Elrod, J. H., R. O’Gorman, C. P. Schneider, T. H. Eckert, T. Schaner, J. N. Canadian Journal of Fisheries and Aquatic Sciences 59:736–753. Bowlby, and L. P. Schleen. 1995. Lake Trout rehabilitation in Lake Ontario. McLain, A. L., B. R. Smith, and H. H. Moore. 1965. Experimental control of Journal of Great Lakes Research 21(Supplement 1):83–107. Sea Lampreys with electricity on the south shore of Lake Superior, 1953– Eshenroder, R. L., and M. K. Burnham-Curtis. 1999. Species succession and 1960. Great Lakes Fishery Commission Technical Report 10. Available: sustainability of the Great Lakes fish community. Pages 145–184 in W. W. http://www.glfc.org/pubs/TechReports/Tr10.pdf. (July 2013). 1772 GORMAN AND SITAR

McPhail, J. D., and V. L. Paragamian. 2000. Burbot biology and life history. Selgeby, J. H., and M. H. Hoff. 1996. Seasonal bathymetric distributions of 16 Pages 11–23 in V. L. Paragamian and D. W. Willis, editors. Burbot: biology, fishes in Lake Superior, 1958–1975. National Biological Service, Biological ecology, and management. American Fisheries Society, Fisheries Manage- Science Report 7, Springfield, Virginia. ment Section Publication 1, Spokane, Washington. Selgeby, J. H., W. R. MacCallum, and M. H. Hoff. 1994. Rainbow Smelt— Mills, E. L., J. M. Casselman, R. Dermott, J. D. Fitzsimons, G. Gal, K. T. larval Lake Herring interactions: competitors or casual acquaintances? U.S. Holek, J. A. Hoyle, O. E. Johannsson, B. F. Lantry, J. C. Markarewicz, Fish and Wildlife Service Biological Report 25. E. S. Millard, I. F. Munawar, M. Munawar, R. O’Gorman, R. W. Owens, L. Sitar, S. P. In press. Assessment of Lake Trout stocks in Michigan waters of G. Rudstam, T. Schaner, and T. J. Stewart. 2005. A synthesis of ecological and Lake Superior, 1998 to 2007. Michigan Department of Natural Resources, fish-community changes in Lake Ontario, 1970–2000. Great Lakes Fishery Fisheries Research Report, Lansing. Commission Technical Report Series 67. Sitar, S. P., and J. X. He. 2006. Growth and maturity of hatchery and wild lean Mills, E. L., J. M. Casselman, R. Dermott, J. D. Fitzsimons, G. Gal, J. A. Hoyle, Lake Trout during population recovery in Michigan waters of Lake Superior. O. Johannsson, B. F. Lantry, J. C. Makarewicz, S. Millard, M. Munawar, Transactions of the American Fisheries Society 135:915–923. I. F. Munawar, R. O’Gorman, R. W. Owens, L. G. Rudstam, T. Shaner, and T. Sitar, S. P., H. M. Morales, M. T. Mata, B. B. Bastar, D. M. Dupras, G. D. J. Stewart. 2003. Lake Ontario: food web dynamics in a changing ecosystem Kleaver, and K. D. Rathbun. 2008. Survey of siscowet Lake Trout at their (1970–2000). Canadian Journal of Fisheries and Aquatic Sciences 60:471– maximum depth in Lake Superior. Journal of Great Lakes Research 34:276– 490. 286. Owens, R. W., R. O’Gorman, T. H. Eckert, and B. F. Lantry. 2003. The offshore Smith, B. R., and J. J. Tibbles. 1980. Sea Lamprey (Petromyzon marinus)in fish community in southern Lake Ontario, 1972–1998. Pages 407–441 in M. Lakes Huron, Michigan, and Superior: history of invasion and control, 1936– Munawar, editor. State of Lake Ontario—past, present, and future. Goodword 78. Canadian Journal of Fisheries and Aquatic Sciences 37:1780–1808. Books, New Delhi. Smith, S. H. 1968. Species succession and fishery exploitation in the Great Oyadomari, J. K., and N. A. Auer. 2004. Inshore-offshore distribution of Lakes. Journal of the Fisheries Research Board of Canada 25:667–693. larval fishes of Lake Superior off the coast of the Keweenaw Penin- Smith, S. H. 1972. Factors of ecologic succession in oligotrophic fish commu- sula, Michigan. Journal of Great Lakes Research 30(Supplement 1):369– nities of the Laurentian Great Lakes. Journal of the Fisheries Research Board 384. of Canada 29:717–730. Peck, J. W., and S. P. Sitar. 2000. Lake Trout assessment and management Smith, S. H. 1973. Application of theory and research in fishery management of in Michigan Waters of Lake Superior, 1993–97. Michigan Department of the Laurentian Great Lakes. Transactions of the American Fisheries Society Natural Resources, Fisheries Research Report 2053, Lansing. 102:156–163. Ray, B. A., T. R. Hrabik, M. P. Ebener, O. T. Gorman, D. R. Schreiner, S. T. Stapanian, M. A., C. P. Madenjian, C. R. Bronte, M. P. Ebener, B. F. Lantry, and Schram, S. P. Sitar, W. P. Mattes, and C. R. Bronte. 2007. Diet and prey J. D. Stockwell. 2008. Status of Burbot populations in the Laurentian Great selection by Lake Superior Lake Trout during spring, 1986–2001. Journal of Lakes. Pages 111–130 in V.L. Paragamian and D. H. Bennett, editors. Burbot: Great Lakes Research 33:104–113. ecology, management, and culture. American Fisheries Society, Symposium Rook, B. J., M. J. Hansen, and O. T. Gorman. 2012. The spatial scale for cisco 59, Bethesda, Maryland. recruitment dynamics in Lake Superior during 1978–2007. North American Stapanian, M. A., C. P. Madenjian, and L. D. Witzel. 2006. Evidence that Journal of Fisheries Management 32:499–514. Sea Lamprey control led to recovery of the Burbot population in Lake Erie. Schram, S. T., T. B. Johnson, and M. J. Seider. 2006. Burbot consumption and Transactions of the American Fisheries Society 135:1033–1043. relative abundance in the Apostle Islands region of Lake Superior. Journal of Stapanian, M. A., L. D. Witzel, and A. Cook. 2010. Recruitment of Burbot (Lota Great Lakes Research 32:798–805. lota L.) in Lake Erie: an empirical modeling approach. Ecology of Freshwater Schreiner, D. R., and S. T. Schram. 2001. Lake Superior fish aging manual. Fishes 19:326–337. Great Lakes Fishery Commission, Ann Arbor, Michigan. Wells, L., and A. L. McLain. 1973. Lake Michigan—man’s effects on na- Scott, W. B., and E. J. Crossman. 1973. Freshwater fishes of Canada. Fisheries tive fish stocks and other biota. Great Lakes Fishery Commission Technical Research Board of Canada Bulletin 184. Report 20. Downloaded by [Department Of Fisheries] at 22:58 25 November 2013 This article was downloaded by: [Department Of Fisheries] On: 25 November 2013, At: 22:59 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Transactions of the American Fisheries Society Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/utaf20 A Review of “Biology of Sharks and Their Relatives” David Shiffman a a Abess Center for Ecosystem Science and Policy , R. J. Dunlap Marine Conservation Program University of Miami , Miami , Florida , USA Published online: 20 Nov 2013.

To cite this article: David Shiffman (2013) A Review of “Biology of Sharks and Their Relatives”, Transactions of the American Fisheries Society, 142:6, 1773-1773, DOI: 10.1080/00028487.2013.849493 To link to this article: http://dx.doi.org/10.1080/00028487.2013.849493

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions Transactions of the American Fisheries Society 142:1773, 2013 C American Fisheries Society 2013 ISSN: 0002-8487 print / 1548-8659 online DOI: 10.1080/00028487.2013.849493

BOOK REVIEW

Biology of Sharks and Their Relatives, second edition. Edited by molecular techniques, tagging studies are the aspect of elasmo- Jeffrey C. Carrier, John A. Musick, and Michael R. Heithaus. CRC branch research that has benefitted the most from revolutionary Press, Boca Raton, Florida. 2012. 666 pages. $99.95 new technologies in recent years. Though it is undoubtedly a consequence of the focus of re- The first edition of Biology of Sharks and Their Relatives searchers worldwide and not an editorial decision, the entire is rightly considered an indispensable resource among elasmo- Sharks and Their Relatives series concentrates much more on branch researchers and specialists in related marine science dis- sharks than on rays, skates, and chimeras. Increased focus on ciplines. The much-anticipated second edition contains updated these “relatives” in future editions would strengthen the series information on every topic covered in the first edition as well as even more. two entirely new chapters. Unlike the previous edition, this one This series would also benefit from greater inclusion of cur- features color images in the form of 29 plates, which greatly rent social science research. The knowledge and attitudes of enhance these visual aids. Future editions should consider ex- different stakeholder groups (fishers, policymakers, etc.) and panding this section even further. studies of how different stakeholders interact with elasmo- Like the first edition, the new one is divided into three branchs (SCUBA ecotourism, catch-and-release fishing, etc.) broad sections that progress logically from how elasmobranchs have been identified as research priorities in recent years but evolved to how individuals function on a physiological level to are absent from the Sharks and Their Relatives series. A thor- how multiple individuals interact with their ecosystem. Topics ough summary of the conservation and management policy tools related to biodiversity, behavior, and conservation are covered in available might make an interesting chapter, as this is a much- a separate volume, Sharks and Their Relatives II. Each section discussed topic among researchers, managers, and the interested features a series of chapters written by a team of established public. These topics may be a better fit for the updated edition experts and rising stars from that specialty. Each chapter is of Sharks and Their Relatives II. thoroughly researched and clearly written, making both founda- The diversity of topics covered in the updated edition, the tional research and the latest discoveries accessible to scientists exhaustive reference lists at the end of each chapter, and the from other specialties. Nonscientists interested in these topics new chapter focusing on cutting-edge technology solidify this (of which there are many) will find most chapters accessible and updated edition’s place at the top of the long list of elasmobranch informative. reference guides. Perhaps the most exciting contribution of this new edition is a chapter on emerging technologies and how they can be used DAV I D SHIFFMAN to study many aspects of elasmobranch biology and ecology. Multisensor tags are an increasingly important tool in such Abess Center for Ecosystem Science and Policy, studies, and the new chapter reviews recent technological R. J. Dunlap Marine Conservation Program developments that allow researchers to study everything University of Miami, from migration patterns to digestive physiology. After certain Miami, Florida, USA Downloaded by [Department Of Fisheries] at 22:59 25 November 2013

1773