Fig 1 - uploaded by Charles Krueger
Content may be subject to copyright.
Landmarks (LM) used to measure lake trout ( Salvelinus namaycush ) shape. Body LM ( a ) were anterior tip of the snout (1), posterior tip of the maxilla (2), center of the eye (3), top of the cranium (4), posterior of neurocranium above top of opercle (5), anterior insertion of dorsal fin (6), posterior insertion of dorsal fin (7), anterior insertion of adipose fin (8), dorsal insertion of caudal fin (9), midpoint of hypural plate (10), ventral insertion of caudal fin (11), posterior insertion of anal fin (12), anterior insertion of anal fin (13), insertion point of pelvic fin (14), insertion point of pectoral fin (19), ventral surface of head below maxilla tip (20), and belly curvature at 20%, 30%, 40%, and 50% of lake trout standard length (18–15, respectively). Head LM ( b ) were the center of the eye (22), the posterior tip of the maxilla (23), and a division of the head profile into 10 evenly spaced partitions (1–21).
Source publication
Resource polymorphisms are widely observed in fishes; however, ontogenetic contributions to morphological and ecological differences are poorly understood. This study examined whether ontogenetic changes in niche partitioning could explain morphological and buoyancy differences between lake trout ( Salvelinus namaycush) morphotypes in Great Slave L...
Contexts in source publication
Context 1
... and ontogeny is a rarely examined aspect of resource polymorphisms, though an understanding of this relationship is likely to ad- vance current understanding of morphotypic diversity (e.g., Parsons and Robinson 2007; Robinson et al. 2008). Depth partitioning is a broadly accepted difference among lake trout morphotypes; however, these results are primarily based on collections of large trout (Moore and Bronte 2001; Bronte et al. 2003; Zimmerman et al. 2006), and conclusions are biased toward the ecology of later life history stages. Small trout in monotypic lakes occupy deeper waters than large trout (Martin 1952; Elrod and Schneider 1987), suggesting that depth distributions also have an important ontogenetic component. Body form is correlated with ontogeny as well. Ontogenetic changes in siscowet morphology and buoyancy are as large as the difference between large siscowet and lean trout (Eschmeyer and Phillips 1965; Zimmerman et al. 2006). Therefore, lake trout ecology and morphology have the potential to be influenced by changes with size and age (i.e., ontogenic niche shifts) as well as interactions among morphotypes (i.e., resource partitioning). A combined morphological and ecological comparison of small and large trout should disentangle within-morphotype niche shifts from among-morphotype resource partitioning. This type of comparison has been hampered by limited information on small trout, which are difficult to identify. In this study, we identify morphological groupings of small and large trout using geometric morphometrics, a statistically powerful tool for differentiating morphotypes based on shape differences (Bookstein 1991; Zelditch et al. 2004). Our morphometric analysis is complemented by analysis of habitat depth, buoyancy, diet, and stable isotopes. By combining these multiple sources of information, we test the hypothesis that ontogenetic shifts in niche partitioning explain differences in morphology and buoyancy observed among lake trout morphotypes. The study was conducted in the east arm of Great Slave Lake, where lean and siscowet-like morphotypes were previously identified based on morphological measures of large (>43 cm standard length (SL)) trout (Zimmerman et al. 2006). Recent sampling in Great Slave Lake included 15 putative humpers (Eshenroder 2008), whose morphology and ecology are quantitatively examined in this paper. The objectives were to ( i ) determine whether morphological and buoyancy differences can be de- tected in small, as well as large, lake trout, ( ii ) investigate whether ontogenetic changes in lake trout morphology and buoyancy correlate with a shift in habitat depth or diet, and ( iii ) establish whether a subset of lake trout, identified as humper trout, were morphologically or ecologically distinct from the lean and siscowet-like morphotypes in Great Slave Lake. Great Slave Lake, Northwest Territories, is a deep, oligotrophic lake. The east arm of the lake lies in the Canadian Shield and is characterized by abundant deepwater habitat (maximum depth = 614 m) and rocky substrate (Evans 2000). Lake trout were caught from the eastern arm (Christie Bay) during August of 2001, 2002, and 2005 with graded mesh gill nets (64–114 mm stretch measure) set overnight for approximately 18 h and by angling with artifi- cial lures (latitude and longitude coordinates for each set are available by request). Of 30 sets, 4 were in depths of 0– 50 m, 22 were in 50–100 m, and 4 were in 100–150 m. All fish caught by angling were from water less than 30 m in depth. The shoreline has a steep grade; deepwater (>50 m) net sets were often within 100 m of shore. While among the better studied of Canada’s Great Lakes (Evans 2000), limited information has been published on Great Slave Lake’s fish community. Similar to Lake Superior, the cold-water fish community of Great Slave Lake is composed of sculpins, ninespine stickleback ( Pungitius pun- gitius ), coregonines, burbot ( Lota lota ), and lake trout (Rawson 1951). Sculpin species include slimy sculpin ( Cottus cognatus ), spoonhead sculpin ( Cottus ricei ), and deepwater sculpin ( Myoxocephalus thompsonii ) (Rawson 1951; Keleher 1972; Stewart 1997). Both lakes have a shallow-water (cisco, Coregonus artedi ) and deepwater pelagic coregonines (shortjaw cisco, Coregonus zenithicus ) as well as lake whitefish ( Coregonus clupeaformis ). Two invertebrates, opossum shrimp ( Mysis relicta ) and a benthic am- phipod ( Diporeia affinis ), are also abundant in the lake (Rawson 1953). The lateral image of each lake trout was captured in a full-body photograph (Zimmerman et al. 2006). Digital image files were used to quantify body shape and head shape with a series of digitized points classified as either landmarks or semilandmarks. Semilandmarks, though not homol- ogous points, were useful to describe belly and head curvatures (Sampson et al. 1996; Langerhans et al. 2003). Body shape was measured using 16 landmarks and 4 semilandmarks (Fig. 1 a ) (Zimmerman et al. 2006, 2007). Head shape was measured with 3 landmarks and 20 semilandmarks; positioning of these landmarks was based on a grid that subdivided the area between the snout and the opercle into 10 equally spaced regions (Fig. 1 b ). A landmark-based geometric morphometric method was used to quantify body shape (Zelditch et al. 2004). This method represents shape as partial warps that are calculated from digitized coordinates. Partial warps describe a set of shape deformations (2 k – 4, where k is the number of landmarks); a biologically meaningful interpretation of partial warps is based on the composite shape difference between groups. Semilandmarks were adjusted using a ‘‘sliding’’ calculation (Zelditch et al. 2004; Zimmerman et al. 2007). Shape variables were standardized to mean size prior to analysis (Zimmerman et al. 2006). The x , y coordinates were captured using tpsDIG software (life.bio.sunysb.edu/morph). Shape variables were calculated using CoordGen software, semilandmarks were adjusted using SemiLand software, and shape was size- standardized using Standard6 software. CoordGen, SemiLand, and Standard6 are part of a series of Integrated Morphometric Programs (IMP) produced in MATLAB6 (MathWorks 2000) and used for morphometric analysis (www2.canisius.edu/~sheets/morphsoft.html). Morphological groupings were identified from a model- based cluster analysis (Fraley and Raftery 2006) using MCLUST v.3 software (www.stat.washington.edu/mclust). The models assumed an underlying Gaussian distribution of the data set. This method uses Bayesian information crite- rion to select the number of groups most likely to exist, assigns individuals to groups, and calculates the uncertainty of individual membership in a given group. Cluster analysis was performed separately for head and body shape data and was based on the first two principal components for each data set. Small (<43 cm SL) and large (>43 cm SL) size classes were analyzed separately. These size groupings were chosen a priori based on earlier results demonstrating that size standardization, necessary for further analyses, could not be performed if data from small and large lake trout were combined (Zimmerman et al. 2006). Group assignments used in buoyancy and ecological comparisons were based on the model-based cluster results, with the exception of 15 individuals identified at the time of collection as humper trout. Identifying characteristics of humper trout were thin abdominal walls, eyes placed high on the head, and deep mid-body profiles. A priori assignments were used to evaluate the recently published assertion that a humper morphotype, distinct from the siscowet-like morphotype, exists in Great Slave Lake (Eshenroder 2008). Percent buoyancy was measured as weight in water div- ided by weight in air (Zimmerman et al. 2006). Fish were weighed to the nearest gram using a Pesola spring scale (Jennings 1989). Weights were taken in air and then in water. Prior to weighing in water, any remaining swim bladder gas was expelled via a longitudinal excision. The buoyancy measure accounts for all differences in soft and hard tissue that affect the specific gravity of lake trout tissue (Alexander 1972). As fat has lower specific gravity than water, tissue with high fat content will be lower in percent buoyancy. An analysis of covariance (ANCOVA) tested whether buoyancy differed between morphotypes. Buoyancy was the response variable, morphotype the explanatory variable, and SL the covariate. Post hoc pairwise tests were conducted on the estimated marginal means. Estimated marginal mean values were adjusted for the effect of the covariate on buoyancy. Small and large size classes, defined by a threshold of 43 cm SL, were analyzed separately, as the buoyancy – standard length relationship could not be linearized across all sizes. The same size threshold was used to assign small and large sizes classes in subsequent analyses. Mean depth for each net was calculated from nine evenly spaced bottom depth measures. Depth was recorded while nets were being set using a Lowrance X98DF sonar (50 kHz in deep water, 200 kHz in shallow water). A two- way analysis of variance (ANOVA) tested whether depth of capture differed between small and large size classes of each morphotype. Interactions were examined with simple effect tests that compared the estimated marginal means of size classes within morphotype and morphotypes within size class. Capture depths of the humper group, which consisted solely of the small size class, were compared with small lean and siscowet trout using a one-way ...
Context 2
... partitioning is a broadly accepted difference among lake trout morphotypes; however, these results are primarily based on collections of large trout (Moore and Bronte 2001; Bronte et al. 2003; Zimmerman et al. 2006), and conclusions are biased toward the ecology of later life history stages. Small trout in monotypic lakes occupy deeper waters than large trout (Martin 1952; Elrod and Schneider 1987), suggesting that depth distributions also have an important ontogenetic component. Body form is correlated with ontogeny as well. Ontogenetic changes in siscowet morphology and buoyancy are as large as the difference between large siscowet and lean trout (Eschmeyer and Phillips 1965; Zimmerman et al. 2006). Therefore, lake trout ecology and morphology have the potential to be influenced by changes with size and age (i.e., ontogenic niche shifts) as well as interactions among morphotypes (i.e., resource partitioning). A combined morphological and ecological comparison of small and large trout should disentangle within-morphotype niche shifts from among-morphotype resource partitioning. This type of comparison has been hampered by limited information on small trout, which are difficult to identify. In this study, we identify morphological groupings of small and large trout using geometric morphometrics, a statistically powerful tool for differentiating morphotypes based on shape differences (Bookstein 1991; Zelditch et al. 2004). Our morphometric analysis is complemented by analysis of habitat depth, buoyancy, diet, and stable isotopes. By combining these multiple sources of information, we test the hypothesis that ontogenetic shifts in niche partitioning explain differences in morphology and buoyancy observed among lake trout morphotypes. The study was conducted in the east arm of Great Slave Lake, where lean and siscowet-like morphotypes were previously identified based on morphological measures of large (>43 cm standard length (SL)) trout (Zimmerman et al. 2006). Recent sampling in Great Slave Lake included 15 putative humpers (Eshenroder 2008), whose morphology and ecology are quantitatively examined in this paper. The objectives were to ( i ) determine whether morphological and buoyancy differences can be de- tected in small, as well as large, lake trout, ( ii ) investigate whether ontogenetic changes in lake trout morphology and buoyancy correlate with a shift in habitat depth or diet, and ( iii ) establish whether a subset of lake trout, identified as humper trout, were morphologically or ecologically distinct from the lean and siscowet-like morphotypes in Great Slave Lake. Great Slave Lake, Northwest Territories, is a deep, oligotrophic lake. The east arm of the lake lies in the Canadian Shield and is characterized by abundant deepwater habitat (maximum depth = 614 m) and rocky substrate (Evans 2000). Lake trout were caught from the eastern arm (Christie Bay) during August of 2001, 2002, and 2005 with graded mesh gill nets (64–114 mm stretch measure) set overnight for approximately 18 h and by angling with artifi- cial lures (latitude and longitude coordinates for each set are available by request). Of 30 sets, 4 were in depths of 0– 50 m, 22 were in 50–100 m, and 4 were in 100–150 m. All fish caught by angling were from water less than 30 m in depth. The shoreline has a steep grade; deepwater (>50 m) net sets were often within 100 m of shore. While among the better studied of Canada’s Great Lakes (Evans 2000), limited information has been published on Great Slave Lake’s fish community. Similar to Lake Superior, the cold-water fish community of Great Slave Lake is composed of sculpins, ninespine stickleback ( Pungitius pun- gitius ), coregonines, burbot ( Lota lota ), and lake trout (Rawson 1951). Sculpin species include slimy sculpin ( Cottus cognatus ), spoonhead sculpin ( Cottus ricei ), and deepwater sculpin ( Myoxocephalus thompsonii ) (Rawson 1951; Keleher 1972; Stewart 1997). Both lakes have a shallow-water (cisco, Coregonus artedi ) and deepwater pelagic coregonines (shortjaw cisco, Coregonus zenithicus ) as well as lake whitefish ( Coregonus clupeaformis ). Two invertebrates, opossum shrimp ( Mysis relicta ) and a benthic am- phipod ( Diporeia affinis ), are also abundant in the lake (Rawson 1953). The lateral image of each lake trout was captured in a full-body photograph (Zimmerman et al. 2006). Digital image files were used to quantify body shape and head shape with a series of digitized points classified as either landmarks or semilandmarks. Semilandmarks, though not homol- ogous points, were useful to describe belly and head curvatures (Sampson et al. 1996; Langerhans et al. 2003). Body shape was measured using 16 landmarks and 4 semilandmarks (Fig. 1 a ) (Zimmerman et al. 2006, 2007). Head shape was measured with 3 landmarks and 20 semilandmarks; positioning of these landmarks was based on a grid that subdivided the area between the snout and the opercle into 10 equally spaced regions (Fig. 1 b ). A landmark-based geometric morphometric method was used to quantify body shape (Zelditch et al. 2004). This method represents shape as partial warps that are calculated from digitized coordinates. Partial warps describe a set of shape deformations (2 k – 4, where k is the number of landmarks); a biologically meaningful interpretation of partial warps is based on the composite shape difference between groups. Semilandmarks were adjusted using a ‘‘sliding’’ calculation (Zelditch et al. 2004; Zimmerman et al. 2007). Shape variables were standardized to mean size prior to analysis (Zimmerman et al. 2006). The x , y coordinates were captured using tpsDIG software (life.bio.sunysb.edu/morph). Shape variables were calculated using CoordGen software, semilandmarks were adjusted using SemiLand software, and shape was size- standardized using Standard6 software. CoordGen, SemiLand, and Standard6 are part of a series of Integrated Morphometric Programs (IMP) produced in MATLAB6 (MathWorks 2000) and used for morphometric analysis (www2.canisius.edu/~sheets/morphsoft.html). Morphological groupings were identified from a model- based cluster analysis (Fraley and Raftery 2006) using MCLUST v.3 software (www.stat.washington.edu/mclust). The models assumed an underlying Gaussian distribution of the data set. This method uses Bayesian information crite- rion to select the number of groups most likely to exist, assigns individuals to groups, and calculates the uncertainty of individual membership in a given group. Cluster analysis was performed separately for head and body shape data and was based on the first two principal components for each data set. Small (<43 cm SL) and large (>43 cm SL) size classes were analyzed separately. These size groupings were chosen a priori based on earlier results demonstrating that size standardization, necessary for further analyses, could not be performed if data from small and large lake trout were combined (Zimmerman et al. 2006). Group assignments used in buoyancy and ecological comparisons were based on the model-based cluster results, with the exception of 15 individuals identified at the time of collection as humper trout. Identifying characteristics of humper trout were thin abdominal walls, eyes placed high on the head, and deep mid-body profiles. A priori assignments were used to evaluate the recently published assertion that a humper morphotype, distinct from the siscowet-like morphotype, exists in Great Slave Lake (Eshenroder 2008). Percent buoyancy was measured as weight in water div- ided by weight in air (Zimmerman et al. 2006). Fish were weighed to the nearest gram using a Pesola spring scale (Jennings 1989). Weights were taken in air and then in water. Prior to weighing in water, any remaining swim bladder gas was expelled via a longitudinal excision. The buoyancy measure accounts for all differences in soft and hard tissue that affect the specific gravity of lake trout tissue (Alexander 1972). As fat has lower specific gravity than water, tissue with high fat content will be lower in percent buoyancy. An analysis of covariance (ANCOVA) tested whether buoyancy differed between morphotypes. Buoyancy was the response variable, morphotype the explanatory variable, and SL the covariate. Post hoc pairwise tests were conducted on the estimated marginal means. Estimated marginal mean values were adjusted for the effect of the covariate on buoyancy. Small and large size classes, defined by a threshold of 43 cm SL, were analyzed separately, as the buoyancy – standard length relationship could not be linearized across all sizes. The same size threshold was used to assign small and large sizes classes in subsequent analyses. Mean depth for each net was calculated from nine evenly spaced bottom depth measures. Depth was recorded while nets were being set using a Lowrance X98DF sonar (50 kHz in deep water, 200 kHz in shallow water). A two- way analysis of variance (ANOVA) tested whether depth of capture differed between small and large size classes of each morphotype. Interactions were examined with simple effect tests that compared the estimated marginal means of size classes within morphotype and morphotypes within size class. Capture depths of the humper group, which consisted solely of the small size class, were compared with small lean and siscowet trout using a one-way ...
Citations
... Digital full-body images of each specimen were photographed in the field (described by Muir et al., 2012), for use in assigning each specimen to one of four known morphs (Fig. 2) using both statistical and visual assignment methods (described by Muir et al., 2014). Statistical shape groups were identified using geometric morphometric analysis of head and body shape (Zelditch et al., 2012), as follows: (1) head and body shape were mapped separately using 2-dimensional coordinates (see Fig. 2 in Muir et al., 2014) by a single individual (MJH) to control for possible investigator bias (Arnqvist and Mårtensson, 1998;Fruciano, 2016;Robinson and Terhune, 2017;Fox et al., 2020); (2) head and body shape maps were converted into partial warp scores free of distortions from horizontal orientation, location, and size (Zelditch et al., 2012); (3) head and body warp scores were converted into principal components to reduce the number of dimensions (Zimmerman et al., 2009;Muir et al., 2014); and (4) individuals were assigned to head and body shape groups using a mixture model-based cluster analysis, with uncertainty of each individual belonging to a group (MCLUST v.4, implemented in R, Fraley and Raftery, 2009). Independent head and body shape groups were then used to assign individuals to morphs based on correspondence between head and body shape groups (Zimmerman et al., 2009): lean morph = long head + slender body; humper morph = short head + fat body; redfin morph = long head + fat body; and siscowet morph = short head + fat body. ...
... Statistical shape groups were identified using geometric morphometric analysis of head and body shape (Zelditch et al., 2012), as follows: (1) head and body shape were mapped separately using 2-dimensional coordinates (see Fig. 2 in Muir et al., 2014) by a single individual (MJH) to control for possible investigator bias (Arnqvist and Mårtensson, 1998;Fruciano, 2016;Robinson and Terhune, 2017;Fox et al., 2020); (2) head and body shape maps were converted into partial warp scores free of distortions from horizontal orientation, location, and size (Zelditch et al., 2012); (3) head and body warp scores were converted into principal components to reduce the number of dimensions (Zimmerman et al., 2009;Muir et al., 2014); and (4) individuals were assigned to head and body shape groups using a mixture model-based cluster analysis, with uncertainty of each individual belonging to a group (MCLUST v.4, implemented in R, Fraley and Raftery, 2009). Independent head and body shape groups were then used to assign individuals to morphs based on correspondence between head and body shape groups (Zimmerman et al., 2009): lean morph = long head + slender body; humper morph = short head + fat body; redfin morph = long head + fat body; and siscowet morph = short head + fat body. Three experts (CRB, CCK, AMM) visually assigned each individual specimen to a morph, as follows: (1) independent assignments by each expert based on overall shape, snout angle and length, eye size and position, paired fin lengths, and caudal peduncle depth and length ( Fig. 2; Burnham-Curtis, 1993;Burnham-Curtis and Smith, 1994); (2) consultation among the three experts to achieve consensus assignments for all specimens that disagreed in independent visual assignments; and (3) percent agreement was the proportion of same independent visual assignments among the three experts (Muir et al., 2014). ...
Life-history variation among four lake trout Salvelinus namaycush morphs was quantified at six geographically distant locations in Lake Superior (~30 to 250 km apart), one of the largest freshwater lakes in the world (82,100 km2). Lake trout were sampled using standardized multi-mesh gillnets in three depth strata at six locations in Lake Superior that were known or thought to have multiple morphs. Life-history traits were estimated using length-age analysis of back-calculated growth from sagittal otolith increments. Morphs, assigned using statistical and visual assignment rules, included 122 humpers, 646 leans, 86 redfins, and 1154 siscowets. Density (CPUE) varied 11-fold among morphs, 7-fold among locations, and 3-fold among depths. Morphs seemed to fill the same ecological niche at all locations, because life-history traits related to weight (body condition, buoyancy,
mean weight), age, and growth rate varied more among morphs than locations. However, abiotic and biotic variation among locations also seemed to exert control over life-history variation, because life-history traits related to length, maturity, and early life history varied more among locations than morphs. We conclude that lake trout morphs appeared to have a genetic component to their life history that was differentially expressed along environmental gradients.
... Previous literature indicates limited niche overlap between adult lake charr ecotypes as a function of habitat and resource partitioning (Conner et al. 1993;Harvey and Kitchell 2000;Ray et al. 2007;Zimmerman et al. 2009;Hoffman 2017). Adult lean and siscowet lake charr populations in Lake Superior are parapatric. ...
... Research from lakes in the western US confirmed that the introduction of Mysis alleviated a recruitment bottleneck allowing lake charr populations to increase (Ellis et al. 2011). Although there may be some dietary similarity between adult lean and siscowet lake charr, their ontogenetic shifts from early life occur with increasing spatial segregation, which minimizes overlap and competition (Zimmerman et al. 2009). ...
... Given that the majority of siscowet occupy waters deeper than that of leans (Pratt et al. 2016;Jasonowicz et al. 2022), the ecotypes partition prey resources very well. Our findings are consistent with Zimmerman et al. (2009), who reported ontogenetic shift in lake charr feeding around 430 mm in Great Slave Lake and Sitar et al. (2020), who measured a similar ontogenetic shift in prey fish energy density at 400 mm in Lake Superior. ...
We investigated the spatial overlap, diet, isotopic niche, and growth of juvenile lean and siscowet lake charr (Salvelinus namaycush) in Lake Superior to address concerns of potential competition with implications to the study of resource polymorphism. Catch data revealed the greatest levels of sympatry in waters from 40-60 m. Juvenile lean and siscowet diet changed ontogenetically with Mysis dominant prey item for the smallest lake charr but differentiating with onset of piscivory. As ecotypes increased in size, lean diets became dominated by pelagic prey whereas siscowets had equal proportions of benthic and pelagic prey. Isotopic niche overlap declined between ecotypes coincident with siscowet lake charr shifting to deeper habitats around 400 mm. Lean and siscowet exhibited different growth trajectories. However, length at age-4 declined in parallel for both ecotypes with no trend in condition suggesting that lake charr growth is sensitive to prey biomass and unlikely related to competition. Our findings indicate minimal evidence of competition and support the concept that multiple sympatric ecotypes of lake charr in Lake Superior are maintained by resource polymorphism.
... For example, Lake Superior Klondike Reef strain fish are derived from the native, deeper spawning humper ecotype that exists in Lake Superior (Lantry et al., 2020). This ecotype is also higher in fat content relative to the lean morphotype that is typified by the other strains used in the hatchery stocking programs, but also tends to selectively feed in deeper habitats relative to other lean lake trout strains (Zimmerman et al., 2009). Humper diets generally include a diversity of invertebrates relative to the piscivorous lean ecotype whose diets in Lake Ontario primarily include alewife (Alosa pseudoharengus), rainbow smelt (Osmerus mordax) and round goby (Neogobius melanostomus; Sitar et al., 2020, Futia et al., 2021. ...
... In Great Slave Lake, NWT, Lake Trout exist in three different morphs but their preferred feeding habitat appears to be more related to size rather than morph (Zimmerman et al. 2009). There are shallow and deep-water piscivorous morphs, and a zooplanktivorous morph, but all three tend to forage benthically, and shift to pelagic feeding at approximately 430 mm standard length (Zimmerman et al. 2009). ...
... In Great Slave Lake, NWT, Lake Trout exist in three different morphs but their preferred feeding habitat appears to be more related to size rather than morph (Zimmerman et al. 2009). There are shallow and deep-water piscivorous morphs, and a zooplanktivorous morph, but all three tend to forage benthically, and shift to pelagic feeding at approximately 430 mm standard length (Zimmerman et al. 2009). Smaller individuals of piscivorous morphs prey on sculpins during their benthic phase, but switch to coregonids as they grow larger and shift to pelagic feeding (Zimmerman et al. 2009). ...
... There are shallow and deep-water piscivorous morphs, and a zooplanktivorous morph, but all three tend to forage benthically, and shift to pelagic feeding at approximately 430 mm standard length (Zimmerman et al. 2009). Smaller individuals of piscivorous morphs prey on sculpins during their benthic phase, but switch to coregonids as they grow larger and shift to pelagic feeding (Zimmerman et al. 2009). All morphs in Great Slave Lake also consume terrestrial insects opportunistically (13% occurrence averaged across morphs) (Moshenko and Gilman 1983;Zimmerman et al. 2009). ...
Rapid climate change occurring in the Arctic may affect the diet of ecologically and culturally important northern fish species. Here, a systematic literature review was completed for eight fish species found across the North American Arctic, with a focus on Inuit Nunangat, to identify major prey items, summarize feeding strategies, and highlight data gaps. Arctic Char (Salvelinus alpinus), Dolly Varden Char (Salvelinus malma), Lake Trout (Salvelinus namaycush), Bull Trout (Salvelinus confluentus), Inconnu (Stenodus leucichthys), Lake Whitefish (Coregonus clupeaformis), Broad Whitefish (Coregonus nasus), and Burbot (Lota lota), were selected as species of interest due to their ecological and cultural importance. The 74 studies reviewed indicate that these species are generalist feeders that demonstrate wide dietary niches, as well as the tendency to avoid agonistic interactions by partitioning resources when they co-occur with an overlapping species. Across coastal, lacustrine, and riverine systems, the most commonly consumed prey items are insects (Diptera spp.), as well as benthic forage fishes such as sculpins (Family: Cottidae). Insects are major prey items in riverine systems, where diets appear to be more generalized, compared to lakes. Anadromous species in coastal waters most commonly feed on various crustaceans and forage fishes. Benthic forage fishes, insects, zooplankton, and mollusks are widely consumed prey items in Arctic lakes. Burbot, Inconnu, and resident Dolly Varden had the most specialized feeding strategies, due in part to their habitat requirements and morphology, while Lake Trout and resident Arctic Char often form multiple ecotypes in lakes, some with different feeding behaviors. Knowledge gaps regarding northern fish trophic ecology are highlighted, and in particular include riverine systems and winter foraging behavior. This review is intended to inform predictions regarding the impacts of climate change on fish tropic ecology and to guide future research.
... Diversity in body shape is one of the most overt responses members of the same species exhibit in accordance with such differences, and has been found to reflect factors including the presence of predators, habitat, food resources, and life history (6) (7) (8) (9) (10) (11). In fact the lake trout, a cousin species of brook trout, exhibits multiple ecomorphs in response to these pressures within a particular lake environment (12). ...
... Using the program Morphologica (13), brook trout images were marked at each of 17 identified physical landmarks in order to approximate the brook trout's morphology as a net. These points include (1) snout, (2) corner of mouth, (3) anterior eye, (4) closest point of tangency between outline and eye, (5) closest point D R A F T of tangency between gills and outline, (6) anterior dorsal, (7) posterior dorsal, (8) anterior adipose, (9) top of tail fin, (10) where the lateral line terminates, (11) bottom of tail fin, (12) posterior anal fin, (13) anterior anal fin, (14) origin of pelvic fin, (15) lowest point of body curve, (16) origin of pectoral fin, (17) posterior jaw. Morphologica converted these points into an ordered list of x,y coordinates, stored in a tab separated values sheet. ...
... Fig. 2 Using the program MorphoJ (14), 17 anatomical features were labeled for each fish image (Fig. 3). These points were chosen because they include fixed, non-mobile anatomical features that are similar to ones used in previous bodyshape studies of lake trout (12) and brook trout (16). Using the 17 anatomical features, we created geometric nets for each fish, reducing each individual's morphology to a standardized polygon. ...
Brook trout (Salvelinus fontinalis) is an invasive species in the Desolation Wilderness of California. But it is unknown to what extent this species is evolving to adapt to isolated high altitude lakes. We quantified morphological differences between three brook trout populations in Desolation Wilderness that are in isolation and of common origin. We took standardized photos of fish, created geometric nets of each specimen using points located at known morphological features, and performed a Procrustes superimposition and principal component analysis to examine and cluster morphological variation between individuals. Together, our results show morphological differences between three Salvelinus fontinalis populations in independent environments. Our results suggest that invasive species introduced from one source can show physical variation generations after introduction, and thus deserve attention for adapting to and perhaps becoming an increasingly complex part of their ecosystem.
... Landmarks used in this study were chosen based on prior research on cutthroat trout specifically (Seiler and Keeley, 2009) For each image, TPSDig2 software (Rohlf, 2017) was used to manually record X and Y coordinates of each of the 28 landmarks. The 18 landmarks used for linear measurements were selected based on demonstrated direct and indirect relationships to foraging and swimming (Nakano et al., 2020;Zimmerman et al. 2006Zimmerman et al. , 2007Zimmerman et al. , 2009Chavarie et al., 2013), and sensitivity to selection pressures in fishes (Webb, 1984;Kristjánsson et al., 2002;Kahilainen et al., 2005). ...
To mitigate westslope cutthroat trout (WCT; Oncorhynchus clarkii lewisi) declines, Montana Fish, Wildlife, & Parks carries out large scale restorations, including wild-origin stocking efforts from conservation hatcheries. Although hatcheries have made progress in limiting the effects of artificial selection on stocked populations, concerns remain that rearing practices inadvertently reduce the diversity propagated into the wild.
The objective of this research was to identify traits of WCT driving poor survival or reproduction in a hatchery, allowing managers to reduce artificial selection by screening for fish requiring alternative rearing. In Chapter 1, I repeatedly measured 18 behavior, morphology, and health traits from hatchery intake (July 2019) to spawn (June 2021). I identified traits with low within- relative to between-individual variation as traits likely to be indicative of specialization. As specialists tend to maximize performance under a narrow range of conditions, they may be vulnerable to artificial selection within hatcheries. In Chapter 2, I tested whether the specialized traits identified in Chapter 1, growth rate, or age at hatchery intake of individual WCT could predict survival or reproduction.
In Chapter 1, I identified nine specialized (relative condition, weighted health index, health index, body shape, energetic reserves, latency, and upper jaw, lower jaw, and anal fin residual length) traits. I hypothesized these traits would predict later survival or reproductive performance. In Chapter 2, I identified October 2019 daily growth rate in weight and every replicate length measurement after October 2019 to strongly predict total ovulated eggs and total number of hatch embryos produced by females. However, among individual variation in the median percent hatch embryos was not explained by maternal size. Male gamete quality and fertilization success was consistently high, and I found no biologically significant predictors of reproductive performance for males. I also found no predictors of survival for males or females.
Despite high total ovulated eggs and total hatch embryo success for females, variable female median percent hatch embryos suggests that quality of ovulated eggs is driving current limitations to WCT hatchery reproduction. I recommend further investigation into impacts of rearing stressors and post-ovulatory aging on female WCT and their reproductive performance.
... 11 Stable isotope values of nitrogen (δ 15 N) undergo predictable trophic enrichment from prey to predators, allowing us to estimate trophic position. 12,13 While information regarding MeHg bioaccumulation and biomagnification can be inferred from δ 13 C and δ 15 N analyses, the connection of Hg sources through dietary pathways is best assessed by including Hg stable isotope analyses. 1,12,14,15 Combining δ 13 C and δ 15 N with Hg stable isotope values facilitates determination of Hg sources and energy pathways to fish, but less commonly tested is the connection between fish habitat use and Hg sources in ecosystems where heterogeneity of Hg sources exists. ...
... 1,12,32 These multidimensional isotope values simultaneously allow us to track Hg reactionary processes 5−7,16,23 and sources 1,3,21,25,29,33−38 as well as fish habitat preference 11 and trophic placement. 12,13 The pristine and rugged landscape of southwest Alaska is characterized by a marked spectrum of topography, glaciation, wetland cover, and connectivity to the ocean, all of which potentially influence how Hg inputs that are both shared (e.g., global Hg sources) and regional (e.g., proximity to volcanoes and prevalence of salmon) are received by lakes and the resident fish therein. To assess the relative importance of these variables, we selected 13 lakes and performed limnological measurements, assessment of watershed features, Hg concentration analyses in water, and biological sample collections (seston [plankton], lake trout [Salvelinus namaycush], and sockeye salmon [Oncorhynchus nerka]) for Hg concentration and isotopic analyses. ...
... Using Hg stable isotope values to delineate Hg sources to fish can be complex when species undergo changes in life cycle 14 or dietary niche partitioning. 13 We propose that these lake trout niche-partition among preferred foraging habitats. In lakes where the largest observable differences in Hg and C isotope values exist between habitats, we can explore this partitioning. ...
Lake trout (Salvelinus namaycush), collected from 13 remote lakes located in southwestern Alaska, were analyzed for carbon, nitrogen, and mercury (Hg) stable isotope values to assess the importance of migrating oceanic salmon, volcanic activity, and atmospheric deposition to fish Hg burden. Methylmercury (MeHg) bioaccumulation in phytoplankton (5.0 - 6.9 kg L-1) was also measured to quantify the basal uptake of MeHg to these aquatic food webs. Hg isotope values in lake trout revealed that while the extent of precipitation-delivered Hg was similar across the entire study area, volcanic Hg is likely an important additional source to lake trout in proximate lakes. In contrast, migratory salmon (Oncorhynchus nerka) deliver little MeHg to lake trout directly, although indirect delivery processes via decay could exist. A high level of variability in carbon, nitrogen, and Hg isotope values indicate niche partitioning in lake trout populations within each lake and that a complex suite of ecological interactions is occurring, complicating the conceptually linear assessment of contaminant source to receiving organism. Without connecting energy and contaminant isotope axes, we would not have understood why lake trout from these pristine lakes have highly variable Hg burdens despite consistently low water Hg and comparable age-length dynamics.
... The characterization of lake charr habitat has been refined and expanded since the 1994 International Conference on Restoration of Lake Trout in the Laurentian Great Lakes (RESTORE) Marsden et al. 1995a) with advances in technology (e.g., high-resolution sonar, acoustic telemetry, modified remotely operated vehicles), dive surveys, and construction of research-based artificial reefs. Expansion of research to northern Canadian Great Lakes (Great Bear and Great Slave lakes; Zimmerman et al. 2006Zimmerman et al. , 2009Chavarie et al. 2016aChavarie et al. , b, 2019, other North American lakes (e.g., Lake Champlain, Otsego Lake, Alexie Lake, Mistassini Lake; Ellrott and Marsden 2004;Tibbits 2007;Blanchfield et al. 2009;Callaghan et al. 2016), and western lakes that lake charr have invaded has highlighted the commonalities that define basic habitat requirements and also revealed exceptions that demonstrate behavioral plasticity in the lake charr. ...
... Juveniles reside in deep water (>35 m), where they feed almost exclusively on benthic invertebrates during their first year (Martin and Olver 1980) or Mysis diluviana (Marshall et al. 1987;Marsden unpublished data), and then at age-1 or 2 begin to incorporate small fishes, such as sculpins Cottus spp., rainbow smelt Osmerus mordax, and alewife Alosa pseudoharengus. The diet gradually shifts to complete piscivory after lake charr reach a length between 300 and 490 mm (Zimmerman et al. 2009;Muir et al. 2016, Marsden unpublished data;Vinson et al. 2021). Pelagic fishes are the preferred prey for adult lake charr (Martin and Olver 1980), but in lakes with a low abundance of pelagic prey fish, adults forage in alternate habitats, including offshore benthic habitats (Rush et al. 2012;Colborne et al. 2016), nearshore littoral habitats (Morbey et al. 2006;Dolson et al. 2009;, and coastal brackish-water habitats (Swanson et al. 2010;Harris et al. 2014; Fig. 1). ...
Lake charr Salvelinus namaycush habitat is defined by the presence of cold (<15 °C), oligotrophic, oxygen-rich (>4 mg L–1) waters where rocky substrates suitable for spawning and forage is available. At low elevations and latitudes, lake charr habitat is confined to the hypolimnion of stratified lakes in summer, though river-spawning occurs at a few locations (e.g., Lake Superior). Description of spawning habitat has previously emphasized the importance of rocky shoals with steep bathymetric relief and deep, silt-free interstices. Recent work, including research in lakes invaded by lake charr, has broadened this view to include boulder and gravel habitat areas with little to no relief and emphasized the role of currents in spawning site suitability. Spawning habitat choice by lake charr is adaptable, such that charr readily spawn on new sites if previously used sites are degraded. Many nearshore structures such as breakwalls and water intake lines attract spawning lake charr. Advances in telemetry technology and work in invaded lakes have broadened understanding of the variability of habitat use by lake charr. Additional work is needed to describe the use of deepwater habitats in lakes and focus on habitat preferences of juvenile lake charr.
... Ontogenetic shifts in isotopic composition of wild fish have usually been attributed to changes in diet through life (de la Moriniere et al., 2003;Grey, 2001;Post, 2003;Zimmerman et al., 2009). However, we detected ontogenetic variation in both captive fish on a common diet and wild fish, suggesting that at least some of the ontogenetic variation seen in wild fish may arise from non-dietary factors. ...
Variation among individuals in stable isotope composition is increasingly being used as an ecological index of trophic niche size. This is based on the assumption that most of the observed variation arises from differences in diet. We tested this assumption by comparing variability in C and N isotopic compositions of adult female lake trout in five wild populations with those of their captive counterparts reared on a common diet, using mature ova as the tissue of comparison. Variation in δ13C and δ15N were related to female size and age, and to a lesser extent growth rate, in both wild and captive fish but relationships were generally stronger for wild fish. Variation among strains in mean δ13C and δ15N was much lower in captive than wild fish, but δ13C and δ15N still varied significantly among strains for the captive fish. As expected, the magnitude and orientation of individual dispersion in δ15N – δ13C space (ID) was much less variable among captive populations than among wild populations. The ID of captive populations expressed as a percentage of the ID of corresponding wild populations ranged from 3 to 16 % when δ13C and δ15N were adjusted for body size covariation, and from 3 to 28 % when data were not adjusted. Relationships between ID and growth rate at the population level were positive for captive lake trout and negative for wild lake trout but neither were significant. Our results suggest that non-dietary variation in δ13C and δ15N is usually a small component of the δ13C and δ15N variation seen in wild lake trout populations and current isotopic niche metrics primarily capture dietary variation.
... Morphological analysis of lake and streamdwelling rock bass and pumpkinseed populations suggests that smaller fins may be more common in stream-dwelling individuals (Brinsmead and Fox, 2002). Correlation of morphological characters and buoyancy were investigated in lake trout (Zimmerman et al., 2009), our results also showed that morphological markers can effectively distinguish species with large differences. ...
