Data from: Phenotypic variation of lake charr (Salvelinus namaycush) in North America
Data files
May 06, 2026 version files 1.42 MB
-
README.md
9 KB
-
Table_S1.csv
1.42 MB
Abstract
Lake charr (Salvelinus namaycush) head and body variation were compared among lakes across the species range in North America, to determine if variation was organized more strongly among morphs or lakes. For 3,473 lake charr, head-shape variation was >17-times greater among morphs than lakes, and body-shape variation was >16-times greater among morphs than lakes. Body depth, head depth, caudal peduncle length and depth, pectoral and pelvic fins lengths, and pre-orbital length varied more among morphs than lakes. Four morphs were delineated based on cluster analysis of combined head and body shape. The fat-fat (head-body) morph had the shortest pre-orbital head length, deepest head and body, shortest and deepest caudal peduncle, and longest paired fins, whereas the lean-lean morph had the longest pre-orbital length, shallowest head and body, longest and narrowest caudal peduncle, and shortest paired fins. The fat-lean morph was similar to the fat-fat morph in pre-orbital length, body depth, caudal peduncle length, and paired fin lengths, but not in head depth or caudal peduncle depth. The lean-fat morph was similar to the fat-fat morph in body depth, head depth at the eye, caudal peduncle length, and paired fin lengths, but was similar to the lean-lean morph in maximum head depth and pre-orbital length. The most extreme fat-fat morphs were from widely separated Great Bear Lake and Lake Superior, whereas the most extreme lean-lean morphs were from Skilak Lake (Alaska), Lake Superior, and Great Bear Lake. The lean-lean morph, present in every lake, was the common morph. A fat-fat morph was present in multiple lakes, but differed greatly in ecological function among lakes, which was more consistent with a hypothesis that multiple lake charr morphs differentiated in multiple lakes across the species’ range, rather than were derived from a single already-specialized morph that colonized multiple lakes from common glacial refugia.
Dataset DOI: 10.5061/dryad.f1vhhmhb8
Description of the data and file structure
Files and variables
File: Table_S1.csv
Description: Lake charr (Salvelinus namaycush) head and body variations from
Variables
- Sort: A continuous numeric variable used for sorting entries
- Lake: Great Bear Lake, NT, Canada; Great Slave Lake, NT, Canada; Lake Mistassini, QB, Canada; Naknek Lake, AK, USA; Rush Lake, MI, USA; Skilak Lake, AK, USA; Lake Superior, USA-Canada
- Location: Dease Arm and McTavish Arm in Great Bear Lake, NT, Canada; Big Reef, Grand Marais, Isle Royale, Klondike Reef, Stannard Rock, and Superior Shoal in Lake Superior, USA-Canada; Great Slave Lake, NT, Canada; Lake Mistassini, QB, Canada; Naknek Lake, AK, USA; Rush Lake, MI, USA; Skilak Lake, AK, USA
- Survey_ID: Unique alpha-numeric code to identify surveys in individual lakes and locations in Superior (ASL = Skilak; GBL = Great Bear, GSL = Great Slave, LSBR = Big Reef in Superior; LSGM = Grand Marais in Superior; LSIR = Isle Royale in Superior; LSKR = Klondike Reef in Superior; LSSR = Stannard Rock in Superior; LSSS = Superior Shoal in Superior; M = Mistassini; N = Naknek; and RL = Rush) and years (numeric code for last two digits of years between 01 = 2001 and 18 = 2018)
- Effort_ID: Unique alpha-numeric code to identify gill nets set and lifted within each survey, sequentially numbered for each Survey_ID from the first lift (1) to the last lift
- Fish_ID: Unique alpha-numeric code to identify each fish caught in each gill net lifted, sequentially numbered for each Effort_ID from the first fish (1) to the last fish
- Shape: Shape group defined by the joint head-body shape: lean-lean morph = long head and slender body; lean-fat morph = long head and plump body; fat-lean morph = short head and slender body; fat-fat morph = short head and plump body
- Body_PC1: First principal component computed from 18 body-shape warp scores that were derived from 16 homologous landmarks and 4 semi-landmarks along the belly curvature at 20, 30, 40, and 50% of standard length
- Body_PC2: Second principal component computed from 18 body-shape warp scores that were derived from 16 homologous landmarks and 4 semi-landmarks along the belly curvature at 20, 30, 40, and 50% of standard length
- Body_PC3: Third principal component computed from 18 body-shape warp scores that were derived from 16 homologous landmarks and 4 semi-landmarks along the belly curvature at 20, 30, 40, and 50% of standard length
- Head_PC1: First principal component computed from 26 head-shape warp scores that were derived from 8 homologous landmarks and 20 semi-landmarks at 10 equally-spaced horizontal positions along the top and bottom margins of the head
- Head_PC2: Second principal component computed from 26 head-shape warp scores that were derived from 8 homologous landmarks and 20 semi-landmarks at 10 equally-spaced horizontal positions along the top and bottom margins of the head
- Head_PC3: Third principal component computed from 26 head-shape warp scores that were derived from 8 homologous landmarks and 20 semi-landmarks at 10 equally-spaced horizontal positions along the top and bottom margins of the head
- TL (mm): Horizontal distance from the tip of the premaxilla to the end of the longest lobe of the caudal fin
- SL (mm): Horizontal distance from the tip of the premaxilla to the midpoint of the hypural plate (sum of head, trunk, dorsal, lumbar, and anal lengths)
- MXL (mm): Distance from the anterior point of the premaxillae to the posterior end of the maxilla (corrected for body size using the allometric relation between MXL and SL, and adjusted to the value for MXL at the same overall mean SL = 424 mm)
- HLL: Horizontal distance from the tip of the premaxilla to the posterior bony margin of the operculum (sum of preorbital, orbital, and post orbital lengths; corrected for body size using the allometric relation between HLL and SL, and adjusted to the value for HLL at the same overall mean SL = 424 mm)
- POL: Horizontal distance from the tip of the premaxilla to the anterior fleshy margin of the orbit (corrected for body size using the allometric relation between POL and SL, and adjusted to the value for POL at the same overall mean SL = 424 mm)
- OOL: Horizontal distance between anterior and posterior fleshy margins of the orbit (corrected for body size using the allometric relation between OOL and SL, and adjusted to the value for OOL at the same overall mean SL = 424 mm)
- PSL: Horizontal distance from the posterior fleshy margin of the orbit to posterior bony margin of the operculum (corrected for body size using the allometric relation between PSL and SL, and adjusted to the value for PSL at the same overall mean SL = 424 mm)
- HDE: Vertical distance through the pupil of the eye from the dorsal surface of the cranium to the ventral edge of the gular region (corrected for body size using the allometric relation between HDE and SL, and adjusted to the value for HDE at the same overall mean SL = 424 mm)
- HDM: Vertical distance from the dorsal to ventral surface of the cranium at the posterior bony margin of the operculum (corrected for body size using the allometric relation between HDM and SL, and adjusted to the value for HDM at the same overall mean SL = 424 mm)
- BDD: Vertical distance from the dorsal-fin origin to the ventral surface of the body (corrected for body size using the allometric relation between BDD and SL, and adjusted to the value for BDD at the same overall mean SL = 424 mm)
- CPD1: Distance from the anterior insertion of the adipose fin to the posterior insertion of the anal fin (corrected for body size using the allometric relation between CPD1 and SL, and adjusted to the value for CPD1 at the same overall mean SL = 424 mm)
- CPD2: Distance between the dorsal and ventral insertions of the upper and lower lobes of the caudal fin (corrected for body size using the allometric relation between CPD2 and SL, and adjusted to the value for CPD2 at the same overall mean SL = 424 mm)
- TTL: Horizontal distance from the posterior bony margin of the operculum to the anterior insertion of the dorsal fin (corrected for body size using the allometric relation between TTL and SL, and adjusted to the value for TTL at the same overall mean SL = 424 mm)
- DOL: Horizontal distance between the anterior and posterior insertions of the dorsal fin (corrected for body size using the allometric relation between DOL and SL, and adjusted to the value for DOL at the same overall mean SL = 424 mm)
- LUL: Horizontal distance from the posterior insertion of the dorsal fin to the anterior insertion of the anal fin (corrected for body size using the allometric relation between LUL and SL, and adjusted to the value for LUL at the same overall mean SL = 424 mm)
- ANL: Horizontal distance between the anterior and posterior insertions of the anal fin (corrected for body size using the allometric relation between ANL and SL, and adjusted to the value for ANL at the same overall mean SL = 424 mm)
- CPL: Horizontal distance from the posterior insertion of the anal fin to the midpoint of the hypural plate (corrected for body size using the allometric relation between CPL and SL, and adjusted to the value for CPL at the same overall mean SL = 424 mm)
- DOH: Distance from the anterior insertion of the dorsal fin to the tip of the longest ray (corrected for body size using the allometric relation between DOH and SL, and adjusted to the value for DOH at the same overall mean SL = 424 mm)
- CFU: Distance from the dorsal insertion of the caudal fin to the tip of the upper caudal fin lobe (corrected for body size using the allometric relation between CFU and SL, and adjusted to the value for CFU at the same overall mean SL = 424 mm)
- CFL: Distance from the ventral insertion of the caudal fin to the tip of the lower caudal fin lobe (corrected for body size using the allometric relation between CFL and SL, and adjusted to the value for CFL at the same overall mean SL = 424 mm)
- ANH: Distance from the anterior insertion of the anal fin to the tip of the longest ray (corrected for body size using the allometric relation between ANH and SL, and adjusted to the value for ANH at the same overall mean SL = 424 mm)
- PVL: Distance from the insertion of the pelvic fin to the tip of the longest ray (corrected for body size using the allometric relation between PVL and SL, and adjusted to the value for PVL at the same overall mean SL = 424 mm)
- PCL: Distance from the insertion of the pectoral fin to the tip of the longest ray
Code/software
Data were organized and curated using MS Excel and saved as a CSV file.
Access information
Other publicly accessible locations of the data:
- none
Data was derived from the following sources:
- none
To control for possible investigator bias, all body, head, and fin coordinates were mapped by a single individual (MJH) (Arnqvist and Mårtensson 1998, Fruciano 2016, Robinson and Terhune 2017, Fox et al., 2020). To distinguish locomotion-based body shape from feeding-based head shape, body and head shape were mapped separately as x-y pixel coordinates using tpsDig2w32 software (https://www.sbmorphometrics.org/index.html). To map body shape, a rectangular grid was overlaid on the full-body image, anchored at the tip of the snout and the mid-point of the hypural plate, to map coordinates of 16 homologous landmarks and 4 semi-landmarks along the belly curvature at 20, 30, 40, and 50% of standard length (Figure S1). To map head shape, a squared grid was overlaid on the head image, anchored at the tip of the snout and the posterior edge of the operculum, to map coordinates of 8 homologous landmarks and 20 semi-landmarks at 10 equally-spaced positions (Figure S1). To measure fin lengths, x-y coordinates were mapped at the tips of the dorsal, caudal (top and bottom lobes), ventral, pelvic, and pectoral fins on the full-body grid-image overlay. Distortions caused by horizontal orientation, location, and size were removed by converting all landmarks into size-free partial warp scores using tpsRelw32 software (https://www.sbmorphometrics.org/index.html).
Shape was analyzed separately for each sample to ensure shape-group assignments were independently developed within the head and body shape space of each lake or location (Perreault-Payette et al. 2017). First, principal component analysis (PCA) was used to reduce the number of 26 head-shape and 18 body-shape warp scores. The first principal component (PC) described overall size, from slender to plump body shape and from shallow (long pointed face) to deep (short steep face) head shape. Second, increasing numbers of principal components were used until >60% of total variance was explained in a Bayesian clustering analysis, MCLUST v.4, implemented in R (Fraley and Raftery 2006). MCLUST is a mixture model-based cluster analysis that selects the model with the best parameterization of the covariance matrix (EII = spherical distribution, equal volume, or equal shape or VII = spherical distribution, variable volume, equal shape) to describe the most likely number of groups based on the Bayesian information criterion (BIC), assigns individuals to groups, and computes the uncertainty of individual membership in a given group. The best model with highest BIC described the greatest separation between the two most morphologically distinct groups based on head shape (from lean = long pointed face to fat = short steep face) and body shape (from lean = slender body to fat = plump body) (Fraley and Raftery 2006, Muir et al. 2014). This method provided separate group assignments based on head shape and body shape.
Geometric shape analysis quantified head and body shape but provided little information about specific functional adaptations in morphology related to feeding or locomotion. Therefore, linear dimensions were computed from head, body, and fin landmarks (Table 2; Figure 2). Horizontal and vertical dimensions were computed by multiplying the image scale times pixel differences in either x- or y-coordinates. Angular dimensions were computed using Pythagoras’ theorem, by multiplying the image scale times pixel differences in both x- and y-coordinates. Linear dimensions were corrected for body size using allometric relations between each linear dimension and standard length (SL), adjusted to values they would assume if specimens were all of the same overall mean SL of 424 mm (Thorpe 1975, Reist 1985, Lleonart et al. 2005). Size-corrected linear dimensions were then compared among morphs and lakes in separate general linear models, with each linear measure as the dependent variable and morphs, lakes (Table 1; locations within Great Bear Lake and Lake Superior were pooled, as a single representative sample of phenotypic diversity in each lake), and the morph × lake interaction as independent variables. Pairwise multiple comparisons used Tukey’s HSD method.