Northeastern Naturalist Vol. 20, No. 4 K. Khidas, J. Duhaime, and H.M. Huynh 2013 587 Morphological Divergence of Continental and Island Populations of Canada Lynx Kamal Khidas1,*, Johannie Duhaime2, and Howard M. Huynh3,4 Abstract - Lynx canadensis (Canada Lynx) mostly occurs in the continental area of North America. Two populations in Atlantic Canada on Newfoundland and Cape Breton Island are geographically isolated. Past studies have revealed geographical and environmental barriers that have significantly impacted processes that ultimately influence the ecology, genetics, evolution, and conservation of the species’ populations. However, equivocal results were obtained as to the morphological and genetic characteristics of the species, and very little is known in this regard on the island populations. The aim of this study was to investigate skull morphometric variation between the species’ populations. We examined and measured 18 craniodental characters on 171 specimens spanning the species’ Canadian range, including most of its boreal forest range and the 2 island populations. Univariate and multivariate analyses provided evidence for significant morphological differentiation among the species’ populations. Factors pertaining to geographical isolation of populations accounted for the bulk of the craniometric variation for both males and females. Principal component analysis (PCA) identified 3 geographical groups: mainland Canada, Newfoundland, and Cape Breton Island. Consistent with the “island rule”, analysis of variance with the Sheffé significant difference post hoc test and PCA results indicated that continental individuals were significantly larger than those from Cape Breton Island, whereas those from Newfoundland, which is significantly larger than Cape Breton Island, exhibited intermediate size. Shape-related variations in the frontal bone among the 3 geographical groups were detected. Six out of the 18 craniodental characters, i.e., postorbital constriction, mastoid, mandible, upper tooth row, mandibular molar row, and upper canine to canine segment, can be used in discriminant function analysis to distinguish between these 3 groups and correctly classify more than 90% of the individuals from the insular populations. We showed for the first time differentiation between Canada Lynx populations on Cape Breton Island and those in the rest of Canada. Individuals from Cape Breton Island appeared the smallest. Contrary to expectations, the Rocky Mountains did not prove to be a significant geographical barrier, resulting in no morphological differentiation of the British Columbia populations. The morphological variations we reported in this study should benefit conservation and management programs in Atlantic Canada, where the levels of the Canada Lynx population are critically low. Conservation plans that strive to maintain the genetic variation documented in the present study should help ensure the preservation of sufficient variability for adaptation to changing environmental conditions required for the long-term viability of populations of Canada Lynx. 2013 NORTHEASTERN NATURALIST 20(4):587–608 1Canadian Museum of Nature, PO Box 3443 Station ‘D’, Ottawa, ON, Canada K1P 6P4. 2Biology Department, Ottawa University, 30 Marie Curie, Ottawa, ON, Canada K1N 6N5. 3Department of Natural Science, New Brunswick Museum, 277 Douglas Avenue, Saint John, NB, Canada E2K 1E7. 4Department of Biological Sciences, Texas Tech University, Texas, USA 79409. *Corresponding author - kkhidas@mus-nature.ca. 588 K. Khidas, J. Duhaime, and H.M. Huynh 2013 Northeastern Naturalist Vol. 20, No. 4 Introduction Lynx canadensis Kerr (Canada Lynx) occurs over a broad geographical range in North America. Most of its populations thrive in the boreal forests of northern North America, which extend from Alaska to the eastern parts of Canada (Aubry et al. 2000, McCord and Cardoza 1982, McKelvey et al. 2000, Mowat et al. 2000). These relatively stable populations occasionally serve as a reservoir for (re)colonization of the contiguous United States (McKelvey et al. 2000, Murray et al. 2008, Ruggiero et al. 2000b). Viable populations also occur marginally in the western mountains of the United States, Great Lakes region, and the Northeast US (Hall 1981, McCord and Cardoza 1982). Additionally, there are 2 geographically disjunct island populations in Atlantic Canada: Newfoundland and Cape Breton Island (Nova Scotia). Newfoundland, with a surface area of 111,390 km², and Cape Breton Island, with 10,311 km², are true islands that are isolated from the mainland by surrounding deep sea waters. Newfoundland, however, can be partially connected to the mainland when waters freeze and ice forms in winter in the Strait of Belle Isle, in the northern tip of the island. Cape Breton Island is artificially connected to mainland Nova Scotia by a bridge crossing the relatively narrow Strait of Canso. Isolation may have a significant impact on the ecology, evolution, and conservation of the insular populations (Whittaker and Fernández-Palacios 2007). The “island rule” stipulates that size tends to increase in smaller mammals and to decrease in larger mammals on islands (Van Valen 1973). According to Lomolino (2005), insular carnivores show a propensity towards dwarfism at the population level. However, Meiri et al. (2004, 2006) challenged Lomolino’s statement. Only two morphological studies addressed such an issue for the Canada Lynx, Saunders (1964) and van Zyll de Jong (1975), but they included only Newfoundland. Though van Zyll de Jong’s data (1975) revealed a trend towards smaller-bodied individuals in Newfoundland, this author concluded that the species remains morphologically homogeneous across its range, including British Columbia despite its relative isolation due to the Rocky Mountains. Cape Breton Island population remains unknown in this respect. Much attention has been paid in the past to genetic and morphological characterization of the Canada Lynx in relation to geographical and ecological variations, but studies have provided equivocal results. Owing to the high dispersal ability of the Canada Lynx coupled to the connectivity among its populations across North America (Houle and Côté 2005, McKelvey et al. 2000, Mowat et al. 2000), high levels of heterozygosity are observed in this species throughout its range, and high gene flow occurs between populations to the extent that a low level of differentiation or structuring is expected (Campbell and Strobeck 2006, Row et al. 2012, Rueness et al. 2003, Schwartz et al. 2002). However, Schwartz et al. (2003) observed decreased genetic variation at microsatellite loci in Canada Lynx occurring at the periphery of the species’ geographical distribution; though no spatial genetic structuring or highly significant deviation from Hardy-Weinberg equilibrium was detected. However, Rueness et al. (2003), using microsatellites and segments of the mitochondrial genome, revealed that present isolation of populations resulted in large-scale genetic Northeastern Naturalist Vol. 20, No. 4 K. Khidas, J. Duhaime, and H.M. Huynh 2013 589 structuring across Canada, with populations grouping into 5 geographical regions: Eastern, Prairie, Northern, Northwestern, and British Columbia. While genetic differentiation may be low within each of these climatic zones, Canada Lynx from British Columbia appeared to be genetically distinct from the other populations, probably as a result of allopatry and isolation produced by the Rocky Mountains. Interestingly, the Eastern population showed marked genetic differences from the other populations as a result of ecological isolation by an “invisible” environmental barrier that Stenseth et al. (2004b) suspected to be tightly linked to differential snow characteristics. Row et al. (2012) confirmed the differentiation of this Eastern population at microsatellite loci and showed a clear divergence of Newfoundland populations from those on mainland North America. In this study, we aimed to characterize skull morphology in relation to geography in the Canada Lynx. We specifically endeavored to answer the question of whether or not the species exhibits significant variations in skull morphometrics across its range. Significant variations should have implications on the ecology and conservation of the species. We investigated craniometric variations in specimens spanning the species’ Canadian geographical range, from British Columbia to Atlantic Canada, including the 2 geographically disjunct island populations, i.e., Newfoundland and Cape Breton Island. We hypothesized that the insular populations of the Canada Lynx differ morphologically from populations of mainland Canada. We also addressed the issue of isolation of the species’ populations from British Columbia, though not truly insular, by testing the hypothesis that they morphologically differ from populations of the rest of Canada. Methods Canada Lynx specimens that were examined in this study were collected from various regions of Canada (Fig. 1) and housed in museum collections. Vouchered specimens housed at the Canadian Museum of Nature (Ottawa, ON) constituted the majority of the sample. Other specimens were made available from the Royal Ontario Museum (Toronto, ON), the University of Laval (Quebec, QC), the Royal British Columbia Museum (Victoria, BC), and the Royal Saskatchewan Museum (Winnipeg, MB). Eighteen characters were examined and measured, 9 cranial (skull and mandible) and 9 dental (Fig. 2), following Pertoldi et al. (2006) and von den Driesch (1976): 1. CBL - condylobasal length, length from the anterior edge of the premaxillae to the posterior-most projections of the occipital condyles; 2. BAS - braincase base, length from the anterior edge of the basisphenoid to the anterior-most point on the lower border of the foramen magnum; 3. RL - rostral length, length from the anterior edge of the premaxillae to the posterior edge of the last tooth in maxilla; 4. ZB - breadth between zygomatics, maximum width across zygomatic arches; 5. POB - postorbital process breadth; 6. POC- postorbital constriction breadth, minimum width at the rear edge of the postorbital processes; 7. MB - mastoid breadth; 8. ML - mandible length; 9. RH - ramus height; 10. UTRL - upper tooth row length; 11. MMRL - mandibular molar row length; 12. LCL - lower canine length; 590 K. Khidas, J. Duhaime, and H.M. Huynh 2013 Northeastern Naturalist Vol. 20, No. 4 13. LCW - lower canine width; 14. CCB - maximum breadth between the 2 upper canines, distance from their exterior edges; 15. LPM4L - lower premolar 4 length; 16. LPM4W - lower premolar 4 width; 17. LM1L - lower molar 1 length; and 18. LM1W - lower molar 1width. Length and width measurements of teeth were taken at the edge of the alveolus. Some skulls were damaged so certain characters could not be reliably measured; these missing measurements were estimated by means of stepwise multiple linear regression equations (Sokal and Rohlf 1995). Additionally, 14 other craniodental characters were measured when present (see descriptions in Fig. 2). Measurements were recorded to the nearest 0.01 mm using digital callipers. To avoid undesirable variation due to possible asymmetry, only the left hand side was measured in paired characters. To minimize the effect of variation due to age and growth, only adult (fully grown) individuals were retained in the analyses. The age of the specimens was either known from museum records or assessed as adult as determined by complete cranial suture closure. Measurements were validated via re-examination and retaken in case outliers or odd values were detected. Normality of the data was tested using Kolmogorov-Smirnov statistic with a Lilliefors significance level. All measurements showed a normal distribution (P > 0.05). Statistical analyses on non-transformed measurements included univariate and multivariate tests. Skulls that had 3 or more missing characters were excluded from the multivariate analyses. Thus, a total of 171 specimens, 90 males and 81 females, Figure 1. Locations of the Canada Lynx specimens examined. Northeastern Naturalist Vol. 20, No. 4 K. Khidas, J. Duhaime, and H.M. Huynh 2013 591 from across the species’ Canadian range were used in these analyses. Sexual dimorphism was previously detected in Canada Lynx (McCord and Cardoza 1982, Quinn and Parker 1987, Saunders 1964, Sicuro and Oliveira 2011) and in other felids such as Felis silvestris (Wild Cat) (Dayan et al. 2002). Differences in means of measurements between males and females were tested using a t-test to reveal any presence of sexual dimorphism in the data used in the present study. Two-way Figure 2. Views of the skull and mandible of a Canada Lynx (Lynx canadensis subsolanus: CMNMA 42565, ♂, housed at the Canadian Museum of Nature, Ottawa, ON, Canada) showing the craniodental characters examined and measured in the study, including the 14 additional measurements (see descriptions below). Variables which most distinguish the 3 geographical groups are indicated with larger fonts. GLS = greatest length of the skull, length from the posteriormost point of the occipital bone to the anterior edge of the premaxillae; PM-FMD = premaxillae–foramen magnum distance, length from the anterior edge of the premaxillae to the posterior-most point on the upper border of the foramen magnum; BL = basal length, length from the anterior edge of the premaxillae to the anterior margin of the lower border of the foramen magnum; PL = palatal length, length from the posterior edge of incisors to posterior edge of palate; IF-PD = incisive foramen–palatal edge distance; IOB = interorbital breadth, least breadth; OL = orbital length; TBL = tympanic bone length; PSW = presphenoid width, distance between the 2 junctions of the presphenoid bone with the 2 other adjacent bones, the palatine and the pterygoid; MH = mandible height, height at the posterior edge of the lower molar 1; MCB = mandibular condyle breadth; UMRL = length of the maxillary molar row; LPM3L = lower premolar 3 length; and LPM3W = lower premolar 3 width. 592 K. Khidas, J. Duhaime, and H.M. Huynh 2013 Northeastern Naturalist Vol. 20, No. 4 multivariate analysis of variance (MANOVA) was used to assess the effects of sex (males and females) and geography (islands and mainland) simultaneously, and test the interaction between the two factors (Geography x Sex) and reveal any effect of one factor on the other. One-way analysis of variance (ANOVA) was used to test inter-group differences between the mean values of each single character. Scheffé tests for pairwise multiple comparisons of means were performed as post hoc tests. The differences in means between groups of individuals with respect to the 18 measured characters described above were tested by means of MANOVA with a Wilks’ λ. The critical α-level for these analyses was set at a P-value of less than 0.05. Ordination (principal components analysis [PCA]) and discrimination (discriminant function analysis [DFA]) statistical methods were used. PCA can be used to reveal non-a-priori clustering patterns in a set of observations. In the current study, PCA was performed on the covariance matrix derived from the morphometric data sets of the 18 craniodental measured characters in order to reveal structure in the relationships between them and for detecting groupings of the individuals. Kruskal-Wallis nonparametric ANOVA (NPANOVA) by ranks was used to test differences between the PC scores of groups of individuals. Scheffé tests for pairwise multiple comparisons of means of PCs were also performed as post hoc tests to these analyses. With respect to the Scheffé test, a large difference between means is required for significance. Discriminant function analysis allows discrimination between 2 or more mutually exclusive predefined groups with respect to a set of non-redundant quantitative variables (Brown and Wicker 2000, Sokal and Rohlf 1995). In this study, DFA stepwise procedure was performed to determine which of the 18 characters best describe each group defined by PCA, and the degree of difference between groups. The jackknife classification method, a leave-one-out cross-validation procedure, was performed to measure classification error. The equality of covariance matrices was tested using Box's M test. Multivariate normality was tested using Mardia’s test. Correlation matrices were examined to check for multicolinearity between variables. SYSTAT© 12 statistical program was used to compute the aforementioned analyses. Results The analyses confirmed the presence of sexual dimorphism in the data used in the present study. Except for postorbital constriction breadth (POC; P = 0.08), presphenoid bone (PSW; P = 0.20), and lower premolar 4 length (LPM4L; P = 0.50), males exhibited larger values than those of females, i.e., in 91% of the cases (results not shown in a table). A variance partition analysis indicated that relatively more variation occurred within sex than among sexes, i.e., 80.4% vs. 19.6% in average, respectively. The two-way MANOVA revealed effects for sex (Wilks’ λ = 0.64, F18, 148 = 4.63, P = 0.000), and for geography (Wilks’ λ = 0.13, F36, 296 = 14.81, P = 0.000), but no interaction between geography and sex (Wilks’ λ = 0.85, F36, 296 = 0.71, P = 0.89). In testing for differences for each character among mainland Canada (MLCAN) and the 2 islands groups, i.e., Newfoundland (NF) and Cape Breton Island (CBI), and using all 171 individuals, males and females combined, Northeastern Naturalist Vol. 20, No. 4 K. Khidas, J. Duhaime, and H.M. Huynh 2013 593 differences were detected in 27 (84.4%) characters (results not shown in a table). Relatively more variation occurred within the 3 geographical groups than among the groups, i.e., 77.5% vs. 22.5% on average, respectively. Consequently, to avoid that part of variation due to sexual dimorphism, sexes were analyzed separately in the subsequent analyses. In addition to the 171 individuals, the 32 aforementioned craniodental characters were measured when present in 27 other specimens, 13 males and 14 females, making a total number of 198 specimens examined and used in the following ANOVA. With sexes analyzed separately, the results of the ANOVA were significant for 28 (87.5%) characters in males and for 23 (72%) in females (Appendix 1). In both sexes, skulls of MLCAN specimens were longer than those of the specimens from the islands. Besides, they exhibited the largest maxillary and mandibular tooth rows (UTRL, UMRL, and MMRL). NF specimens tended to be wider at their frontal bone in comparison to those of the other 2 populations; in particular, they exhibited the broadest POC. CBI specimens generally tended to exhibit the smallest mean size of the skull traits in males and in females, with more marked differences detected in RL, interorbital segment (IOB), CCB, LCL, and PSW in comparison with the specimens from the other populations. Finally, skulls of the CBI specimens were narrower in ZB than those of MLCAN. The first 2 principal components encompassed the maximum amount of variation between the individuals in both sexes, and accounted for nearly 87% of the total variation in males and 81% in females (Table 1). Interpretations of the results of this PCA will be based on these 2 PCs, as the third PC accounted for only slightly more than 5% at the best, in females. For both sexes, the PC1-PC2 plane showed partitioning of the craniodental trait data into 3 clusters with partial overlapping: MLCAN, NF, and CBI (Fig. 3A, B). MLCAN specimens occupied most of the PC1 x PC2 plan space, suggesting that variations are much greater on mainland than on the islands. Table 1. Eigenvalues of the principal components (PC1, PC2) and percent of variance explained by each principal component, and results of nonparametric ANOVA for the significance of differences between group means of the principal components for 18 craniodental characters in male and female Canada Lynx. MLCAN = mainland Canada, CBI = Cape Breton Island, and NF = Newfoundland; ***P < 0.001, **P < 0.01, *P < 0.05. Principal % of variance Nonparametric component Eigenvalue explained ANOVA (F-value) P Scheffé F test Males 1 69.37 78.14 16.56 0.000 MLCAN > CBI *** MLCAN > NF* 2 7.74 8.71 27.38 0.000 MLCAN > NF *** CBI > NF *** Females 1 39.836 69.31 13.56 0.001 MLCAN > CBI *** 2 6.514 11.33 25.86 0.000 CBI > NF *** MLCAN > NF *** 594 K. Khidas, J. Duhaime, and H.M. Huynh 2013 Northeastern Naturalist Vol. 20, No. 4 The first principal component (PC1) explained most of the observed variation in both sexes (slightly more than 78% and 69% in males and females, respectively). Loadings of variables on this PC showed a wide range of variation in both sexes (Table 2). The highest loadings were observed with CBL, ZB, and ML in males Figure 3. Plot of the Canada Lynx specimens on the PC1 x PC2 plan. A: males, B: females. Open squares = mainland Canada (MLCAN) population, stars = Newfoundland Island population (NF), and open circles = Cape Breton Island population (CBI). Table 2. Factor loadings of principal component analysis for 18 craniodental characters in male and female Canada Lynx. *values indicate variables with high loadings. Males Females Craniodental charactersA PC 1 PC 2 PC 1 PC 2 CBL 4.796* 0.784 3.610* 0.430 BAS 1.820 0.162 1.374 0.033 RL 1.472 0.107 1.339 -0.165 ZB 3.414* -0.358 2.542* 0.118 POB 2.182 -2.289* 1.428 -1.938* POC -0.567 -1.201* -0.172 -1.483* MB 1.641 0.226 1.293 0.221 ML 3.426* 0.111 2.641* 0.083 RH 1.948 0.130 1.326 0.469 UTRL 1.704 0.158 1.507 -0.092 MMRL 0.975 0.308 0.753 0.059 LCL 0.258 -0.013 0.269 -0.056 LCW 0.232 -0.016 0.255 0.004 CCB 1.338 -0.228 1.010 -0.201 LPM4L 0.246 0.082 0.331 0.020 LPM4W 0.113 0.011 0.120 0.011 LM1L 0.284 0.110 0.251 0.030 LM1W 0.124 0.058 0.082 0.018 ASee full description of craniodental characters in the Methods section for acronyms. Northeastern Naturalist Vol. 20, No. 4 K. Khidas, J. Duhaime, and H.M. Huynh 2013 595 and females. In both sexes, MLCAN specimens clustered mainly on the positive side of PC1; the specimens from CBI positioned mainly on its negative side, with PC1 values of MLCAN being larger than those of CBI as confirmed by Sheffé’s test (Table 1). The NF specimens showed intermediate PC1 values. The NF male specimens were different from MLCAN. However, the NF and CBI specimens did not differ in PC1 for both males (P = 0.662) and females (P = 0.230) as determined by Scheffe’s test. The coefficients of PC1 were all positive, except for POC; these results in combination with the wide range of variation of the loadings indicate that PC1 contains information about shape as well. The subsequent principal component, PC2 (highest part of variation explained by other than PC1; see Table 1), bears shape-related morphological variations. PC2 was defined mainly by the width of the frontal bone of the skull (POC and POB) in both males and females. In both sexes, NF specimens showed smaller PC2 scores in comparison with those of the other 2 populations (Table 1). In females, NF and CBI were neatly split into 2 distinct groups by PC2 (Fig. 3B). A second PCA (results not shown) similar to the previous one was performed using individuals (57 males, and 55 females) from mainland Canada only. No evident geographic differentiation could be detected. Of particular interest, the individuals from British Columbia (15 males and 16 females) did not cluster; by contrast, they were scattered uniformly all across the main PC plane in males and females (82.6% and 76.8% of the total variation in males and females, respecti vely). A partitioning similar to that reported in the abovementioned PCAs was observed on the 2 discriminant functions space, with more or less overlapping between the 3 populations (CBI, MLCAN, and NF) (Fig. 4A, B; Table 3). Six variables were the most useful in distinguishing the 3 populations (Table 4). These 6 variables accounted for the width of the skull (POC and MB), the size of mandibular and maxillary dental characters (MMRL, UTRL, CCB), and ML. An average of 84% (jacknife: Figure 4. Plot of DF1 and DF2 scores from DFA in Canada Lynx specimens. A: males, B: females. Open squares = mainland Canada (MLCAN) population, stars = Newfoundland Island population (NF), and open circles = Cape Breton Island population (CBI). 596 K. Khidas, J. Duhaime, and H.M. Huynh 2013 Northeastern Naturalist Vol. 20, No. 4 83%) of the specimens was correctly assigned in males, and 88% (jacknife: 86%) in females. A high percentage, more than 90%, of the specimens from the insular populations were correctly assigned to their respective groups in both sexes. Differences among group centroids were observed in males (Wilks’ λ = 0.29, F12, 164 = 11.65, P = 0.0000) and females (Wilks’ λ = 0.19, F12, 146 = 15.82, P = 0.0000). In males, DF1 contained 55.6% (eigenvalue = 0.95) of the variance and described mainly a pattern of increasing POC and decreasing MB that more clearly separated NF specimens from those of CBI. NF specimens were associated with POC, while those of MLCAN and CBI were associated with MB. DF2 accounted for 44.4% of the variance (eigenvalue = 0.76) and described a pattern of decreasing ML and size of dental characters (UTRL, MMRL, and CCB). MLCAN specimens were associated with these 4 characters, whereas insular populations tended to have smaller DF2 scores. In females, DF1 (63.9% of the variance, eigenvalue = 1.7) described a pattern of increasing POC and CCB. CBI was neatly separated from NF on this first axis, and partially from MLCAN. MLCAN and NF were associated with larger POC and CCB. DF2 (36.1% of the variance, eigenvalue = 0.96) described a pattern of decreasing MB and MMRL that yielded more or less overlapping between the 3 groups in females; MB and MMRL were associated with MLCAN. F-statistics computed from Mahalanobis squared distance, D², showed that the centroids for MLCAN and CBI male specimens were closest (10.77), those for NF and CBI were farthest apart (12.86), while MLCAN and NF showed intermediate affinity (12.09). The centroids for NF and MLCAN female specimens were closest (12.44), those for CBI and NF (18.82) and CBI and MLCAN (18.74) were farthest apart. Table 3. Stepwise discriminant function analysis classification matrix for comparing the 3 Canada Lynx populations in males and females. Actual populations are shown in rows, predicted populations in columns. CBI = Cape Breton Island, MLCAN = mainland Canada, and NF = Newfoundland. Males Females Populations CBI MLCAN NF % correct CBI MLCAN NF % correct Cape Breton Island 17 1 0 94 14 1 0 93 Mainland Canada 7 45 5 79 4 47 3 87 Newfoundland 1 0 14 93 0 1 10 91 Table 4. Function loadings of stepwise discriminant function analysis for comparing the 3 Canada Lynx populations in males and females. * indicate variables with high loadings. Males Females Craniodental charactersA Function 1 Function 2 Function 1 Function 2 POC 0.938* 0.352 0.808* 0.750 MB -0.729* -0.688 -0.482 -0.959* ML -0.007 -1.000* 0.770 -0.457 UTRL -0.051 -0.999* 0.674 -0.577 MMRL -0.517 -0.859* 0.092 -0.953* CCB 0.329 -0.943* 0.998* 0.153 ASee full description of craniodental characters in the Methods section for acronyms. Northeastern Naturalist Vol. 20, No. 4 K. Khidas, J. Duhaime, and H.M. Huynh 2013 597 Discussion Werdelin (1981) reported that the Canada Lynx remained morphologically homogeneous all across mainland North America since its first appearance in the continent during the Wisconsinan. van Zyll de Jong (1975) supported the absence of significant morphological variation in the species. In his work, van Zyll de Jong (1975) ascribed this morphological uniformity to the similar selective pressures imposed by a supposed relative uniformity in ecological conditions throughout the species’ range. With significant variation in Canada Lynx craniodental characters detected, our study supports the phenotypic heterogeneity of the species across its geographical distribution in Canada. The patterns of differentiation revealed here were very complex. Most of the craniometric variation could be attributed to factors pertaining to geographical isolation of populations. Morphological differentiation was observed between mainland, Newfoundland and Cape Breton Island populations of the species. However, the presence of the Rocky Mountains did not appear in our study to have a significant impact on the isolation of the British Columbia population as suggested by Rueness et al. (2003) with regards to genetic characteristics. This finding on morphology is in agreement with the results of Row et al. (2012), who also did not reveal any significant impact of these mountains on the genetic structuring in the species. As in our results, Houle and Côté (2005) did not observe any difference in the morphology of the skulls of Canada Lynx from different regions of the province of Quebec, indicating that the Saint-Laurent River also did not prove to be a sufficient geographical barrier to result in skull morphological variation. Additionally, Stenseth et al. (1999) and Stenseth et al. (2004a) revealed spatial population dynamics structuring governed by differing large-scale climatic factors influenced by the North Atlantic Oscillation, with Canada Lynx grouping into 3 climatic zones: Pacific-Maritime, Continental, and Atlantic-Maritime. No evident grouping according to these 3 zones was observed in our analyses; as observed for British Columbia, individuals from these zones were scattered across the main PCA plane of the 2 PCAs performed (all Canada, and only mainland Canada). Our results showed that skull size tends to be smaller in the individuals from the insular populations. From our results, PC1 mainly summarized variation in skull size with the linear combination of several characters, out of which 3, i.e., CBL, ML, and ZB, contributed the most. These 3 traits altogether reflect the overall size (length and width) of the skull in cats (Mazák 2010). Individuals from populations of mainland Canada have significantly larger skulls in comparison with those from Cape Breton Island. Those from Newfoundland show intermediate overall size of the skulls that did not differ significantly from that of Cape Breton Island individuals in both sexes, but differed from mainland Canada individuals in males. Yom-Tov et al. (2007) used greatest length of skull, zygomatic breadth and inter-orbital constriction as surrogates for body size. By inference, our results would suggest that smallerbodied Canada Lynx are found on islands, with individuals from Cape Breton Island being the smallest. Consistent with Lomolino’s conclusions (2005), these findings conform to the predicted morphological patterns as postulated in the “island rule” (Van Valen 1973). Results from Mahalanobis D² reveal much dissimilarity in this 598 K. Khidas, J. Duhaime, and H.M. Huynh 2013 Northeastern Naturalist Vol. 20, No. 4 respect between CBI and NF populations, while MLCAN and NF populations share similarities. When considering only single characters, our results showed that individuals from mainland populations have significantly longer skulls relative to those from the 2 insular populations. Individuals from mainland Canada also have the widest skull in comparison with those from Cape Breton Island. Newfoundland behaves somewhat like a continent in comparison to Cape Breton Island. This study reports for the first time a morphological distinction between mainland populations and the Cape Breton Island population. Variation in size in insular Canada Lynx was previously reported in past studies, but concerned mainly females from Newfoundland. For example, Saunders (1964) observed smaller basilar length (BL in our study) in female skulls from Newfoundland compared to those from Alaska. Similarly, van Zyll de Jong (1975) reported that females from Newfoundland showed significantly smaller condylobasal length in comparison with those from mainland Canada. The 3 populations, MLCAN, NF, and CBI, differed in shape of craniodental characters as confirmed by PC2 and PC1. As indicated by the loadings (see Bookstein 1989, Humphries et al. 1981), PC1 also contained some shape-related variation as well. Results revealed much affinity between MLCAN and CBI populations, while NF population showed unique features. Separation of NF population from the other 2 populations was conspicuous when considering the shape of the frontal bone. With the recent development of genetic analyses, several authors reported a relationship between measures (size and shape) of skeletal traits and genes (Kenney- Hunt et al. 2008, Klingenberg and Leamy 2001, Klingenberg et al. 2001, Leamy et al. 1999, Simonsen et al. 2003, Workman et al. 2002). The variation in shape of the frontal bone that we report here, therefore, would indicate possible genetic divergence between the Newfoundland population and the other 2 populations, and between the Cape Breton and mainland Canada populations to some extent as well. Our results regarding the shape-related differentiation between MLCAN and NF are in agreement with those of the genetic analyses of Row et al. (2012). The smaller-sized Canada Lynx skulls in insular populations we report here, and thus possibly a smaller body size, may be a result of higher population densities on Newfoundland and Cape Breton Island (i.e., smaller size as a result of density-dependent effects). Yom-Tov et al. (2007) previously showed that Canada Lynx body size in Alaska was negatively correlated to population density. Population densities of the Canada Lynx are relatively high on Cape Breton Island (Parker et al. 1983) when compared to reported population densities in northern latitude boreal forest habitats during peak years of abundance (Aubry et al. 2000). The POC, along with POB, showed the highest loading on PC2. POC was identified in DFA also as the variable that best describes each population. Similarly, in felid species, including the Canada Lynx, Werdelin (1981) and Sicuro and Oliveira (2011) found that POC was the crucial skull character in the morphological aspect imparted by PC2 or the subsequent PCs. Besides, our results showed that the frontal bone is wider in Newfoundland individuals, with marked differences detected at the postorbital constriction in males and at the postorbital constriction and the postorbital processes in females (Appendix 1). A wider frontal bone Northeastern Naturalist Vol. 20, No. 4 K. Khidas, J. Duhaime, and H.M. Huynh 2013 599 would entail narrower available space in the temporal fossa in Newfoundland individuals. In addition, skulls of male individuals from Newfoundland have a significantly shorter ramus than those of individuals from mainland Canada (Appendix 1). A generally similar trend was observed in females. Results of PCA showed that the ramus is a character that varies much among the 3 populations. The cranial variation we report in this study suggests that the Newfoundland population exhibits different masses of the mandibular muscles, i.e., the masseter- pterygoid and temporalis muscles. Bulky mandibular muscles require more space in the temporal fossa to operate effectively and efficiently. Variations in the overall size of teeth detected among Canada Lynx populations in this study may indirectly support dietary specialization as suggested by Meiri et al. (2005a). In DFA, upper canine-canine segment, maxillary tooth row, and mandibular molar row proved to be important traits that best characterize each of the 3 populations in males and females. The ANOVA analyses we conducted on dental characters (Appendix 1) reveal significant differences in maxillary tooth and mandibular molar rows between the continental and insular populations. Similarly, Saunders (1964) reported that maxillary tooth row in skulls from Newfoundland were shorter in comparison to those from other skulls from Labrador and Maine. The Canada Lynx is a specialist carnivore adapted to preying mainly on Lepus americanus Erxleben (Snowshoe Hare). Lynx population levels are known to dramatically change during 9- to11-year cycles across their range in response to fluctuations in populations of this hare, the population densities of which are in turn closely associated with changes in vegetation dynamics in terms of abundance and quality (for review, see Murray et al. 2008, Ruggiero et al. 2000a). However, Canad Lynx are also known to feed upon a variety of alternative prey, especially during periods of low hare abundance (Aubry et al. 2000, Mowat et al. 2000, Murray et al. 2008, Roth et al. 2007). Saunders (1963) reported that Newfoundland Canada Lynx, though feeding mostly on Snowshoe Hare, showed seasonal variations in diet. Carrion formed a large proportion of its diet, and other prey species (e.g., meadow voles and birds) were also taken. The possible slimmer mandibular muscles suggested above for Newfoundland individuals could be related to this peculiar feeding habit. Christiansen (2008) and Slater and Van Valkenburgh (2009) reported that variation in the felid skull can be related to the species’ feeding habits. The morphological divergence in the skull and the teeth characters among the 3 populations may consequently indicate adjustments to prey types upon which the Canada Lynx feed. Environmental conditions vary within the range of the Canada Lynx. Agee (2000) and Buskirk et al. (2000) reported divergent ecological conditions throughout the species’ distributional range, with variability in climate, habitat (vegetation, disturbance regime, and successional dynamics), and predatorprey interactions. Significant north–south and east–west environmental gradients are observed for the species’ habitat quality and availability (Agee 2000). Hence, different selective forces along with different genetic characteristics and developmental patterns could, therefore, have resulted in the morphological divergence of these 3 forms of Canada Lynx in Canada. 600 K. Khidas, J. Duhaime, and H.M. Huynh 2013 Northeastern Naturalist Vol. 20, No. 4 Our findings can have implications on the conservation of the Canada Lynx in Atlantic Canada. Two subspecies of Canada Lynx are currently recognized in Canada (Hall 1981, Wilson and Reeder 2005) on the basis of fur coloration and size: the nominate subspecies L. c. canadensis in mainland Canada, and L. c. subsolanus in Newfoundland. Genetic distinction of the latter subspecies was recently confirmed by Row et al. (2012). In this eastern part of the country, the Canada Lynx populations reached critically low numbers owing to shortages in food availability (Nowak 1999) and because of anthropogenic activity that dramatically altered the habitat (Buskirk et al. 2000). Moreover, Cape Breton Island is very small, and Canada Lynx occur only on the western plateau of Victoria and Inverness counties in the highlands. Open and barren lands between this small area and mainland Nova Scotia isolate this population from the rest of Canada, with mainland Nova Scotia connected to mainland Canada by a land bridge, the Chignecto Isthmus. The Canada Lynx is considered endangered and is legally protected in Nova Scotia, but not in Newfoundland. Here, we emphasize the need to preserve the integrity of the island gene pools. Kyle and Strobeck (2003) and Paetkau and Strobeck (1994) reported that Newfoundland populations of Martes americana Turton (American Marten) and Ursus americanus Pallas (Black Bear) show less genetic variation than their mainland counterparts. Lomolino (2005) stated that long after populations have settled on islands changes in size within insular ecological settings reflect combined effects of several factors and are driven by selective forces (e.g., resource limitation and ecological/predatory release) that would favor an optimal size (but see Meiri et al. 2005b). Limited or absence of gene flow from other conspecific populations would further reinforce directional selection for optimal body size on islands. The question of when the Canada Lynx first appeared on the 2 islands, and when the 2 populations became isolated is still a matter of debate. The species as currently identified occurred in North America either during the Sangamon interglacial or the Wisconsin glaciation (Kurtén and Anderson 1980, Kurtén and Rausch 1959, Werdelin 1981). It could have appeared on Newfoundland either sometime soon after the last glaciation or as recently as the early 19th century (van Zyll de Jong 1975). In fact, Millien (2006) demonstrated that mammals evolve faster on islands than on the mainland and significant morphological changes can occur within a relatively short period of time. Schwartz et al. (2002) reported that the Canada Lynx population from the Kenai Peninsula, though not completely isolated, was genetically divergent from other populations, a divergence that was produced only after a few generations. Pertoldi et al. (2006) showed that a recent human-induced isolation of populations of Lynx pardinus Temminck (Iberian Lynx) produced significant morphometric differentiation within the isolated population in less than 100 years. Ruggiero et al. (2000b) argued that conservationists and wildlife managers should exercise caution in supplementing depleted populations with translocated individuals from other regions as any such introduction could potentially undermine and threaten the integrity of the indigenous gene pool, and could result in damaging effects on the local population (e.g., introduction of diseases; Northeastern Naturalist Vol. 20, No. 4 K. Khidas, J. Duhaime, and H.M. Huynh 2013 601 eventually swamping out unique alleles found in isolated populations). Similarly, in our case, the influx of lynx from mainland Canada could disrupt co-adapted gene complexes that have evolved in situ. Maintaining and enhancing the natural habitat requirements of the Canada Lynx on these 2 islands should be a priority in conservation programs. Unique ecological conditions that prevail on the islands of Newfoundland and Cape Breton, coupled with geographical isolation, could have contributed to these 2 different phenotypes. Considerable attention has been paid to the Canada Lynx in western Canada and in the United States; in contrast, studies on this species in Newfoundland and Cape Breton islands are relatively few. Continental populations have been the subject of several genetic studies (see Introduction); yet, as far as we are aware, no DNA analyses were carried out on the Cape Breton Island population. Future DNA analyses on this insular population should reveal interesting results and should confirm genetic differentiation between Island and mainland populations. An extensive program of studies on the biology and ecology of this species would be an invaluable component to the success of any protection or rehabilitation plan. Acknowledgments We are grateful to our colleagues who helped in completing our sample of museum specimens. Our special thanks go in particular to Judith Eger and Susan Woodward (Royal Ontario Museum, Toronto, ON), Cyrille Barrette (University of Laval, Quebec, QC), Kelly Sendall (Royal British Columbia Museum, Victoria, BC), and Ray Poulin (Royal Saskatchewan Museum, Regina, SK) who eagerly offered their assistance and made easily accessible the collections they manage. François Chapleau offered valuable support at the University of Ottawa. Natalia Rybczynski (Canadian Museum of Nature, Ottawa, ON) and Donna Naughton (Canadian Museum of Nature) willingly reviewed a former draft. Virginie Millien (McGill University, Montreal, QC) made invaluable comments and suggestions on our results. The editor and two anonymous referees made comments that improved the manuscript. Micheline Beaulieu-Bouchard (Canadian Museum of Nature) provided the drawings. Literature Cited Agee, J.K. 2000. Disturbance ecology of North American boreal forests and associated northern mixed/subalpine forests. Pp. 39–82, In L.F. Ruggiero, K.B. Aubry, S.W. Buskirk, G.M. Koehler, C.J. Krebs, K.S. McKelvey, and J.R. Squires (Eds.). Ecology and Conservation of Lynx in the United States. University Press of Colorado, Boulder, CO. 350 pp. Aubry, K.B., G.M. Koehler, and J.R. Squires. 2000. Ecology of Canada Lynx in southern boreal forests. Pp. 373–396, In L.F. Ruggiero, K.B. Aubry, S.W. Buskirk, G.M. Koehler, C.J. Krebs, K.S. McKelvey, and J.R. Squires (Eds.). Ecology and Conservation of Lynx in the United States. University Press of Colorado, Boulder , CO. 350 pp. Bookstein, F.L. 1989. Size and Shape: A comment on semantics. Systematic Zoology 38:173–180. Brown, M.T., and L.R. Wicker. 2000. Discriminant analysis. Pp. 209–235, In H.E.A. Tinsley and S.D. Brown (Eds.). Handbook of Applied Multivariate Statistics and Mathematical Modeling. Academic Press, San Diego, CA. 721 pp. 602 K. Khidas, J. Duhaime, and H.M. Huynh 2013 Northeastern Naturalist Vol. 20, No. 4 Buskirk, S.W., L.F. Ruggiero, K.B. Aubry, D.E. Pearson, J.R. Squires, and K.S. McKelvey. 2000. Comparative ecology of lynx in North America. Pp. 397–417, In L.F. Ruggiero, K.B. Aubry, S.W. Buskirk, G.M. Koehler, C.J. Krebs, K.S. McKelvey, and J.R. Squires (Eds.). Ecology and Conservation of Lynx in the United States. University Press of Colorado, Boulder, CO. 350 pp. Campbell, V., and C. Strobeck. 2006. Fine-scale genetic structure and dispersal in Canada Lynx (Lynx canadensis) within Alberta, Canada. Canadian Journal of Zoology 84:1112–1119. Christiansen, P. 2008. Evolution of skull and mandible shape in cats (Carnivora: Felidae). PLoS ONE 3:e2807. doi:10.1371/journal.pone.0002807 Dayan, T., D. Wool, and D. Simberloff. 2002. Variation and covariation of skulls and teeth: Modern carnivores and the interpretation of fossil mammals. Pal eobiology 28:508–526 Hall, E.R. 1981. The Mammals of North America. 2nd Edition. John Wiley and Sons, New York, NY.1181 pp. Houle, M., and S.D. Côté. 2005. Croissance, dimorphisme sexuel et variabilité morphométrique du crâne entre différentes populations de lynx du Canada (Lynx canadensis) au Québec. Université Laval, Laval, QC, Canada. 24 pp. Humphries, J.M., F.L. Bookstein, B. Chernoff, G.R. Smith, R.L. Elder, and S.G. Poss. 1981. Multivariate discrimination by shape in relation to size. Systematic Zoology 30:291–308. Kenney-Hunt, J.P., B. Wang, E.A. Norgard, G. Fawcett, D. Falk, L.S. Pletscher, J.P. Jarvis, C. Roseman, J. Wolf, and J.M. Cheverud. 2008. Pleiotropic patterns of quantitative trait loci for 70 murine skeletal traits. Genetics 178:2275–2288. Klingenberg, C.P., and L.J. Leamy. 2001. Quantitative genetics of geometric shape in the mouse mandible. Evolution 55:2342–2352. Klingenberg, C.P., L.J. Leamy, E.J. Routman, and J.M. Cheverud. 2001. Genetic architecture of mandible shape in mice: Effects of quantitative trait loci analyzed by geometric morphometrics. Genetics 157:785–802. Kurtén, B., and R.L. Rausch. 1959. Biometric comparisons between North American and European mammals. II. A comparison between the northern lynxes of Fennoscandia and Alaska. Acta Artica 11:21–44. Kurtén, B., and E. Anderson. 1980. Pleistocene mammals of North America. Columbia University Press, New York, NY. 442 pp. Kyle, C.J., and C. Strobeck. 2003. Genetic homogeneity of Canadian mainland marten populations underscores the distinctiveness of Newfoundland Pine Martens (Martes americana atrata). Canadian Journal of Zoology 81:57–66. Leamy, L.J., E.J. Routman, and J.M. Cheverud. 1999. Quantitative trait loci for early- and late-developing skull characters in mice: A test of the genetic independence model of morphological integration. American Naturalist 153:201–214. Lomolino, M.V. 2005. Body size evolution in insular vertebrates: Generality of the island rule. Journal of Biogeography 32:1683–1699. Mazák, J.H. 2010. Craniometric variation in the Tiger (Panthera tigris): Implications for patterns of diversity, taxonomy, and conservation. Mammalian Biology 75:45–68. McCord, C.H., and J.E. Cardoza. 1982. Bobcat and Lynx (Felis rufus and F. lynx). Pp. 728–766, In J.A. Chapman and G.A. Feldhame (Eds.). Wild Mammals of North America: Biology, Management, and Economics. The Johns Hopkins University Press, Baltimore, MD. 1123 pp. Northeastern Naturalist Vol. 20, No. 4 K. Khidas, J. Duhaime, and H.M. Huynh 2013 603 McKelvey, K.S., K.B. Aubry, and Y.K. Ortega. 2000. History and distribution of lynx in the contiguous United States. Pp. 207–264, In L.F. Ruggiero, K.B. Aubry, S.W. Buskirk, G.M. Koehler, C.J. Krebs, K.S. McKelvey, and J.R. Squires (Eds.). Ecology and Conservation of Lynx in the United States. University Press of Colorado, Boulder, CO. 350 pp. Meiri, S., T. Dayan, and D. Simberloff. 2004. Body size of insular carnivores: Little support for the island rule. The American Naturalist 163:469–479. Meiri, S., T. Dayan, and D. Simberloff. 2005a. Variability and correlations in carnivore crania and dentition. Functional Ecology 19:337–343. Meiri, S., D. Simberloff, and T. Dayan. 2005b. Insular carnivore biogeography: Island area and mammalian optimal body size. The American Naturalist 165:505–514. Meiri, S., T. Dayan, and D. Simberloff. 2006. The generality of the island rule re-examined. Journal of Biogeography 33:1571–1577. Millien, V. 2006. Morphological evolution is accelerated among island mammals. PLoS Biology 4:e321. doi:10.1371/journal.pbio.0040321 Mowat, G., K.G. Poole, and M. O’Donoghue. 2000. Ecology of lynx in Northern Canada and Alaska. Pp. 265–306, In L.F. Ruggiero, K.B. Aubry, S.W. Buskirk, G.M. Koehler, C.J. Krebs, K.S. McKelvey, and J.R. Squires (Eds.). Ecology and Conservation of Lynx in the United States. University Press of Colorado, Boulder , CO. 350 pp. Murray, D.L., T.D. Steury, and J.D. Roth. 2008. Assessment of Canada Lynx research and conservation needs in the Southern range: Another kick at the cat. Journal of Wildlife Management 72:1463–1472. Nowak, R.M. 1999. Walker’s Mammals of the World. 6th Edition. The Johns Hopkins University Press, Baltimore, MD. 1936 pp. Paetkau, D., and C. Strobeck. 1994. Microsatellite analysis of genetic variation in Black Bear populations. Molecular Ecology 3:489–495. Parker, G.R., J.W. Maxwell, L.D. Morton, and G.E.J. Smith. 1983. The ecology of the lynx (Lynx canadensis) on Cape Breton Island. Canadian Journal of Zoology 61:770–786 . Pertoldi, C., R. García-Perea, J.A. Godoy, M. Delibes, and V. Loeschcke. 2006. Morphological consequences of range fragmentation and population decline on the endangered Iberian Lynx (Lynx pardinus). Journal of Zoology 268:73–86. Quinn, N.W.S., and G. Parker. 1987. Lynx. Pp. 683–694, In M. Novak, J.A. Baker, M.E. Obbard, and B. Malloch (Eds.). Wild Furbearer Management and Conservation in North America. Ontario Trappers Association, North Bay, ON, Canada. 1150 pp. Roth, J.D., J.D. Marshall, D.L. Murray, D.M. Nickerson, and T.D. Steury. 2007. Geographical gradients in diet affect population dynamics of Canada Lynx. Ecology 88:2736–2743. Row, J.R., C. Gomez, E.L. Koen, J. Bowman, D.L. Murray, and P.J. Wilson. 2012. Dispersal promotes high gene flow among Canada Lynx populations across mainland North America. Conservation Genetics 13:1259–1268. Rueness, E.K., N.C. Stenseth, M. O’Donoghue, S. Boutin, H. Ellegren, and K.S. Jakobsen. 2003. Ecological and genetic spatial structuring in the Canadian Lynx. Nature 425:69–72. Ruggiero, L.F., K.B. Aubry, S.W. Buskirk, G.M. Koehler, C.J. Krebs, K.S. McKelvey, and J.R. Squires. 2000a. Ecology and Conservation of Lynx in the United States. University Press of Colorado, Boulder, CO. 350 pp. Ruggiero, L.F., M.K. Schwartz, K.B. Aubry, C.J. Krebs, A. Stanley, and S.W. Buskirk. 2000b. Species conservation and natural variation among populations. Pp. 101–116, In L.F. Ruggiero, K.B. Aubry, S.W. Buskirk, G.M. Koehler, C.J. Krebs, K.S. McKelvey, and J.R. Squires (Eds.). Ecology and Conservation of Lynx in the United States. University Press of Colorado, Boulder, CO. 350 pp. 604 K. Khidas, J. Duhaime, and H.M. Huynh 2013 Northeastern Naturalist Vol. 20, No. 4 Saunders, J.K. 1963. Food habits of the Lynx in Newfoundland. Journal of Wildlife Management 27:384–390. Saunders, J.K. 1964. Physical characteristics of the Newfoundland lynx. Journal of Mammalogy 45:36–47. Schwartz, M.K., L.S. Mills, K.S. McKelvey, L.F. Ruggiero, and F.W. Allendorf. 2002. DNA reveals high dispersal synchronizing the population dynamics of Canada Lynx. Nature 415:520–522. Schwartz, M.K., L.S. Mills, Y. Ortega, L.F. Ruggiero, and F.W. Allendorf. 2003. Landscape location affects genetic variation of Canada Lynx (Lynx canadensis). Molecular Ecology 12:1807–1816. Sicuro, F.L., and L.F.B. Oliveira. 2011. Skull morphology and functionality of extant Felidae (Mammalia: Carnivora): A phylogenetic and evolutionary perspective. Zoological Journal of the Linnean Society 161:414–462. Simonsen, V., C. Pertoldi, A.B. Madsenc, and V. Loeschcke. 2003. Genetic differentiation of foxes (Vulpes vulpes) analysed by means of craniometry and isozymes. Journal for Nature Conservation 11:109–116. Slater, G.J., and B. Van Valkenburgh. 2009. Allometry and performance: The evolution of skull form and function in felids. Journal of Evolutionary Biol ogy 22:2278–2287. Sokal, R.R., and F.J. Rohlf. 1995. Biometry: The Principles and Practice of Statistics in Biological Research.3rd Edition. Freeman and Company, New York, NY. 887 pp. Stenseth, N.C., K.S. Chan, H. Tong, R. Boonstra, S. Boutin, C.J. Krebs, E. Post, M. O’Donoghue, N.G. Yoccoz, M.C. Forchhammer, and J.W. Hurrell. 1999. Common dynamic structure of Canada Lynx populations within three climatic regions. Science 285:1071–1073. Stenseth, N.C., D. Ehrich, E.K. Rueness, O.C. Lingjaerde, K.S. Chan, S. Boutin, M. O’Donoghue, D.A. Robinson, H. Viljugrein, and K.S. Jakobsen. 2004a. The effect of climatic forcing on population synchrony and genetic structuring of the Canadian Lynx. Proceedings of the National Academy of Sciences of the United States of America 101:6056–6061. Stenseth, N.C., A. Shabbar, K.S. Chan, S. Boutin, E.K. Rueness, D. Ehrich, J.W. Hurrell, O.C. Lingjaerde, and K.S. Jakobsen. 2004b. Snow conditions may create an invisible barrier for lynx. Proceedings of the National Academy of Sciences of the United States of America 101:10632–10634 Yom-Tov, Y., S. Yom-Tov, D. MacDonald, and E. Yom-Tov. 2007. Population cycles and changes in body size of the lynx in Alaska. Oecologia 152:239–244. Van Valen, L. 1973. Pattern and the balance of nature. Evolutionary Theory 1:31–49. van Zyll de Jong, C.G. 1975. Differentiation of the Canada Lynx, Felis (Lynx) canadensis subsolana, in Newfoundland. Canadian Journal of Zoology 53:699–705. von den Driesch, A. 1976. A guide to the measurement of animal bones from archaeological sites. Peabody Museum Bulletin 1:1–137. Werdelin, L. 1981. The evolution of lynxes. Annales Zoologici Fennici 18:37–71. Whittaker, R.J., and J.M. Fernández-Palacios. 2007. Island Biogeography: Ecology, Evolution, and Conservation. 2nd Edition. Oxford University Press, Inc., New York, NY. 401 pp. Wilson, D.E., and D.M. Reeder. 2005. Mammal Species of the World. A Taxonomic and Geographic Reference. 3rd Edition. The Johns Hopkins University Press, Baltimore, MD. 2142 pp. Workman, M.S., L.J. Leamy, E.J. Routman, and J.M. Cheverud. 2002. Analysis of quantitative trait locus effects on the size and shape of mandibular molars in mice. Genetics 160:1573–1586. Northeastern Naturalist Vol. 20, No. 4 K. Khidas, J. Duhaime, and H.M. Huynh 2013 605 Appendix 1. Means and standard errors of craniodental characters in 198 male and female Canada lynx and according to geographical areas (MLCAN = mainland Canada, CBI = Cape Breton Island, and NF = Newfoundland), and results of ANOVA (ns = not significant, *** = P < 0.001, ** = P < 0.01, * = P < 0.05). See full description of craniodental characters in th e Methods section and figure 2 for acronyms. Males Females Craniodental ANOVA ANOVA characters Population n Mean SE P Scheffé F test n Mean SE P Scheffé F test Skull and mandible GLS MLCAN 71 128.12 0.55 0.000 MLCAN > NF** 68 124.16 0.47 0.003 MLCAN > CBI* CBI 17 124.85 1.13 MLCAN > CBI* 16 120.96 0.98 MLCAN > NF* NF 15 123.39 1.20 11 120.93 1.19 CBL MLCAN 71 117.31 0.51 0.000 MLCAN > CBI** 68 113.62 0.43 0.000 MLCAN > CBI*** CBI 17 112.61 1.04 MLCAN > NF** 16 109.30 0.90 MLCAN > NF* NF 15 113.04 1.11 11 110.45 1.08 PM-FMD MLCAN 71 120.46 0.50 0.000 MLCAN > CBI*** 68 116.85 0.44 0.000 MLCAN > CBI** CBI 17 115.70 1.02 MLCAN > NF ** 16 113.00 0.91 NF 15 116.52 1.09 11 114.05 1.10 BL MLCAN 71 108.13 0.51 0.000 MLCAN > CBI** 67 104.52 0.48 0.000 MLCAN > CBI*** CBI 17 102.83 1.05 MLCAN > NF** 16 99.75 1.00 MLCAN > NF* NF 15 103.31 1.11 11 100.65 1.20 BAS MLCAN 70 38.26 0.22 0.001 MLCAN > CBI** 68 36.34 0.19 0.019 MLCAN > CBI * CBI 17 36.31 0.45 16 35.05 0.40 NF 15 37.47 0.48 11 35.98 0.48 PL MLCAN 71 49.24 0.21 0.000 MLCAN > CBI***, 68 47.84 0.19 0.000 MLCAN > CBI*** CBI 17 46.87 0.44 MLCAN > NF*** 16 45.09 0.40 MLCAN > NF** NF 15 46.99 0.47 11 46.16 0.49 RL MLCAN 71 45.85 0.18 0.000 MLCAN > CBI*** 68 44.49 0.17 0.000 MLCAN > CBI*** CBI 17 43.80 0.37 NF > CBI * 16 42.17 0.35 NF > CBI** NF 15 45.19 0.39 11 43.98 0.42 606 K. Khidas, J. Duhaime, and H.M. Huynh 2013 Northeastern Naturalist Vol. 20, No. 4 Males Females Craniodental ANOVA ANOVA characters Population n Mean SE P Scheffé F test n Mean SE P Scheffé F test IF-PD MLCAN 71 36.41 0.17 0.000 MLCAN > NF *** 68 35.50 0.16 0.000 MLCAN > CBI *** CBI 17 34.61 0.35 MLCAN > CBI *** 16 33.23 0.34 MLCAN > NF ** NF 15 34.23 0.37 11 33.76 0.41 ZB MLCAN 71 90.55 0.39 0.001 MLCAN > CBI ** 68 88.34 0.33 0.000 MLCAN > CBI ** CBI 17 87.34 0.8 16 85.3 0.68 NF 15 88.38 0.85 11 86.81 0.82 POB MLCAN 57 57.64 0.40 0.012 NF > CBI * 55 56.48 0.31 0.000 NF > CBI *** CBI 17 56.21 0.74 16 54.30 0.57 NF > MLCAN * NF 15 59.49 0.78 11 58.65 0.69 MLCAN > CBI ** POC MLCAN 57 39.99 0.20 0.000 NF > MLCAN *** 55 39.74 0.24 0.001 NF > CBI ** CBI 17 40.15 0.36 NF > CBI ** 16 39.11 0.44 NF > MLCAN ** NF 15 42.18 0.39 11 41.81 0.54 IOB MLCAN 57 28.99 0.22 0.000 NF > CBI ** 55 27.64 0.18 0.000 NF > CBI *** CBI 17 27.19 0.41 MLCAN > CBI ** 16 26.38 0.34 MLCAN > CBI ** NF 15 29.61 0.44 11 28.64 0.41 MB MLCAN 57 55.43 0.25 0.003 MLCAN > NF ** 68 52.8 0.23 0.634 CBI 17 54.68 0.46 16 53.19 0.48 (ns) NF 15 53.51 0.49 11 52.48 0.59 OL MLCAN 71 31.99 0.14 0.057 68 31.44 0.11 0.008 MLCAN > CBI ** CBI 17 31.21 0.29 (ns) 16 30.62 0.23 NF 15 32.04 0.31 11 31.32 0.27 TBL MLCAN 71 26.46 0.13 0.002 MLCAN > NF ** 68 25.67 0.13 0.058 CBI 17 25.96 0.28 16 25.39 0.27 (ns) NF 15 25.28 0.3 11 24.83 0.33 PSW MLCAN 57 8.61 0.08 0.000 MLCAN > CBI *** 55 8.43 0.08 0.000 MLCAN > CBI *** CBI 17 7.80 0.16 NF > CBI * 16 7.60 0.15 NF > CBI ** NF 15 8.46 0.17 11 8.46 0.19 Northeastern Naturalist Vol. 20, No. 4 K. Khidas, J. Duhaime, and H.M. Huynh 2013 607 Males Females Craniodental ANOVA ANOVA characters Population n Mean SE P Scheffé F test n Mean SE P Scheffé F test ML MLCAN 71 83.78 0.37 0.001 MLCAN > CBI ** 68 81.28 0.31 0.000 MLCAN > CBI *** CBI 17 80.67 0.76 16 77.74 0.64 NF 15 81.70 0.81 11 79.86 0.77 MH MLCAN 57 17.21 0.11 0.000 MLCAN > CBI *** 55 16.55 0.11 0.045 (ns) CBI 17 16.17 0.21 NF > CBI * 16 16.04 0.20 NF 15 17.06 0.22 11 16.78 0.24 RH MLCAN 71 35.94 0.26 0.026 MLCAN > NF * 68 34.18 0.23 0.450 CBI 17 35.14 0.53 16 34.59 0.48 (ns) NF 15 34.31 0.57 11 33.63 0.58 MCB MLCAN 57 17.44 0.13 0.266 55 16.95 0.10 0.003 MLCAN > CBI ** CBI 17 16.97 0.25 (ns) 16 16.17 0.19 NF 15 17.37 0.26 11 16.75 0.23 Dental UTRL MLCAN 71 50.68 0.18 0.000 MLCAN > CBI *** 68 49.11 0.17 0.000 MLCAN > CBI *** CBI 17 48.25 0.37 MLCAN > NF ** 16 46.37 0.36 MLCAN > NF * NF 15 48.96 0.40 11 47.67 0.44 UMRL MLCAN 57 26.35 0.12 0.000 MLCAN > CBI *** 55 25.49 0.14 0.000 MLCAN > CBI *** CBI 17 24.97 0.23 MLCAN > NF * 16 24.10 0.69 NF > CBI * NF 15 25.60 0.24 11 25.17 0.84 MMRL MLCAN 71 31.54 0.27 0.000 MLCAN > NF *** 67 30.67 0.12 0.000 MLCAN > NF ** CBI 17 30.54 0.13 MLCAN > CBI ** 16 29.75 0.25 MLCAN > CBI** NF 15 29.94 0.29 11 29.51 0.30 LCL MLCAN 71 7.14 0.05 0.003 NF > CBI * 68 6.68 0.04 0.000 NF > CBI *** CBI 17 6.75 0.10 MLCAN > CBI ** 16 6.21 0.10 MLCAN > CBI *** NF 15 7.19 0.11 11 6.87 0.12 608 K. Khidas, J. Duhaime, and H.M. Huynh 2013 Northeastern Naturalist Vol. 20, No. 4 Males Females Craniodental ANOVA ANOVA characters Population n Mean SE P Scheffé F test n Mean SE P Scheffé F test LCW MLCAN 71 5.13 0.04 0.003 MLCAN > CBI * 68 4.86 0.04 0.043 MLCAN > CBI * CBI 17 4.86 0.08 16 4.58 0.09 NF 15 4.92 0.08 11 4.81 0.11 CCB MLCAN 57 32.73 0.18 0.000 MLCAN > CBI *** 55 31.20 0.14 0.00 NF > CBI ** CBI 17 30.89 0.33 16 30.02 0.27 MLCAN > CBI ** NF 15 32.08 0.35 11 31.53 0.32 LPM3L MLCAN 57 8.03 0.06 0.004 MLCAN > NF ** 55 7.75 0.07 0.598 CBI 17 7.84 0.11 16 7.61 0.12 (ns) NF 15 7.59 0.11 11 7.66 0.15 LPM3W MLCAN 70 4.14 0.02 0.049 66 4.08 0.03 0.368 CBI 17 4.27 0.04 16 3.99 0.06 (ns) NF 15 4.19 0.05 11 4.10 0.07 LPM4L MLCAN 57 9.97 0.09 0.013 MLCAN > CBI * 55 9.84 0.09 0.068 CBI 17 9.45 0.16 16 9.43 0.17 (ns) NF 15 9.57 0.18 11 9.51 0.20 LPM4W MLCAN 71 4.60 0.02 0.369 68 4.48 0.03 0.531 CBI 17 4.64 0.05 (ns) 16 4.54 0.06 (ns) NF 15 4.52 0.06 11 4.44 0.07 LM1L MLCAN 57 12.59 0.10 0.273 55 12.19 0.12 0.610 CBI 17 12.56 0.19 (ns) 16 12.25 0.22 (ns) NF 15 12.22 0.20 11 11.92 0.26 LM1W MLCAN 71 5.25 0.02 0.018 (ns) 67 5.14 0.02 0.325 CBI 17 5.11 0.05 16 5.11 0.05 (ns) NF 15 5.09 0.06 11 5.03 0.06