Skip to main content
Download PDF

The natural history and ecology of melanism in red wolf and coyote populations of the southeastern United States – evidence for Gloger’s rule

BMC Zoology volume 7, Article number: 33 (2022) Cite this article

Abstract

Background

Gloger’s rule postulates that animals should be darker colored in warm and humid regions where dense vegetation and dark environments are common. Although rare in Canis populations, melanism in wolves is more common in North America than other regions globally and is believed to follow Gloger’s rule. In the temperate forests of the southeastern United States, historical records of red wolf (Canis rufus) and coyote (Canis latrans) populations document a consistent presence of melanism. Today, the melanistic phenotype is extinct in red wolves while occurring in coyotes and red wolf-coyote hybrids who occupy the red wolf's historical range. To assess if Gloger’s rule could explain the occurrence and maintenance of melanistic phenotypes in Canis taxa, we investigated differences in morphology, habitat selection, and survival associated with pelage color using body measurements, GPS tracking data, and long-term capture-mark-recapture and radio-telemetry data collected on coyotes and hybrids across the southeastern United States.

Results

We found no correlation between morphometrics and pelage color for Canis taxa. However, we observed that melanistic coyotes and hybrids experienced greater annual survival than did their gray conspecifics. Furthermore, we observed that melanistic coyotes maintained larger home ranges and exhibited greater selection for areas with dense canopy cover and wetlands than did gray coyotes.

Conclusions

In the southeastern United States, pelage color influenced habitat selection by coyotes and annual survival of coyotes and hybrids providing evidence that Gloger’s rule is applicable to canids inhabiting regions with dense canopy cover and wetlands. Greater annual survival rates observed in melanistic Canis may be attributed to better concealment in areas with dense canopy cover such as coastal bottomland forests. We suggest that the larger home range sizes of melanistic coyotes may reflect the trade-off of reduced foraging efficiency in lower quality wetland habitat for improved survival. Larger home ranges and differential use of land cover by melanistic coyotes may facilitate weak assortative mating in eastern coyote populations, in which melanistic animals may have lower success of finding compatible mates in comparison to gray conspecifics. We offer that our observations provide a partial explanation for why melanism is relatively low (< 10%) but consistent within coyote populations throughout southeastern parts of their range.

Peer Review reports

Background

In mammals, pelage color has been linked to fitness-relevant traits such as crypsis, sexual behavior, fecundity, aggressiveness, and immunity [1,2,3,4,5]. Gloger’s rule, the primary ecogeographical rule on animal coloration, postulates that animals should be darker colored in warm and humid regions because these climatic conditions promote dense vegetation and darker environments [6, 7]. To improve camouflage, animals inhabiting humid environments with dense canopy may acquire darker colors than conspecifics inhabiting drier, non-forest habitats. For example, melanism (black pelage color) is observed in felids living in tropical forests, such as melanistic jaguars (Panthera onca) and leopards (Panthera pardus). However, melanism is rare in other carnivores such as Canis, in which black pelage color occurs most commonly in North American gray wolves (Canis lupus) [8, 9] and is believed to follow Gloger’s rule [10]. Melanistic gray wolves are rare in Eurasia, but some isolated occurrences of melanism have been documented [11,12,13]. Collectively, most research aimed at elucidating melanistic traits in Canis has focused on the origin and function of melanism in North American gray wolves [9, 14,15,16] and to a lesser extent, Eurasian wolves [17, 18].

Historical and current records demonstrate a consistent presence of melanism in Canis populations of temperate forests of the southeastern United States (hereafter Southeast) [19,20,21,22,23,24,25] (Figs. 1 and 2). Historically, melanism was common in the red wolf (Canis rufus), the only wild Canis species that occurred in the Southeast from the terminal Pleistocene until the early twentieth century [23, 26]. Indeed, melanistic wolves were so distinguishable in the Southeast that Goldman [21] combined all red wolf forms into a single species, Canis niger. However, the red wolf suffered a severe population bottleneck because of government-sponsored eradication campaigns that resulted in the eventual extirpation of the species from the wild and the use of 14 individuals to establish a captive-breeding program to prevent extinction [27, 28]. Currently, the melanistic phenotype is extinct within the extant red wolf population because only russet-colored individuals were included as founders for the captive population (USFWS, unpublished data). However, melanism occurs in contemporary coyote (Canis latrans) populations that replaced red wolves throughout the Southeast [22, 24, 25].

Fig. 1
figure 1

A “Black American Wolf”, a nineteenth-century hand-colored lithograph drawn from a Florida specimen [20]. B A melanistic red wolf killed in Oklahoma, May–June 1919. Reverend James O. Arthur photograph collection from the National Museum of the American Indian repository at the Smithsonian Institute. C Melanistic red wolf, taken from Evangeline Parish and exhibited in the Audubon Park Zoo in New Orleans, LA in the late 1920s. Photo featured in The Fur Animals of Louisiana, 1931 by Stanley Clisby Arthur. D A melanistic red wolf in Tensas Parish, Louisiana, 1934. Copyright Tappan Gregory (1886–1961). E Melanistic wolf that weighed 32.2 kg. Shown in the trap in Winn Parish, Louisiana, 1948. Courtesy of the T.E. “Doc” Harris family. F Photo of a large melanistic wolf killed in Caldwell Parish, Louisiana, 1949 by R.E. Walters. Photo appeared in The Richland Beacon News, August 20, 1949

Full size image
Fig. 2
figure 2

A A melanistic coyote (458 M) captured and radio-collared in Tyrrell County, NC, 2009. Photo by Joseph W. Hinton. B A melanistic coyote (586 M) captured and radio-collared in Tyrrell County, NC, 2010. Photo by Joseph W. Hinton. C A melanistic coyote (634 M) captured in Washington County, NC, 2011 and euthanized at the behest of the landowner. Photo by Joseph W. Hinton. D A melanistic coyote (SC34M) captured and radio-collared in Saluda County, SC, 2016. Photo by Joseph W. Hinton. E Male melanistic coyote shot by a landowner in Livingston Parish, Louisiana, 2018. Photo courtesy Amy C. Shutt. F A melanistic coyote (LA51M) captured and radio-collared in Cameron Parish, Louisiana, 2022. Photo by Amy C. Shutt

Full size image

Following their colonization of eastern North America, coyotes experienced increased geographic variation in phenotypic and genetic traits [29]. Melanistic pelage, noticeably absent from western coyote populations, is unique to eastern coyote populations and more commonly found in populations of the Southeast [23,24,25]. The origin of melanism in eastern coyotes is linked with hybridization with other Canis taxa including dogs (Canis lupus familiaris) [9], eastern wolves (Canis lycaon) [30], and red wolves [22, 23]. Despite the coyote's widespread co-occurrence with gray wolves and dogs in central and western North America, only one case of melanism was reported in Colorado during the 1920s via Young and Jackson’s [31] extensive survey of coyote populations prior to the coyote’s colonization of the Southeast. However, reports of melanistic coyotes became more common in areas of the Southeast where they colonized areas inhabited by melanistic red wolves during 1940–1975 [22, 23, 31,32,33]. Several studies conducted between 1976 and 2015 indicated that melanistic coyotes were uncommon in the Southeast and comprised 2–9% of the surveyed populations [22, 24, 25]. Recently, Caudill and Caudill [25] concluded that Gloger’s rule did not hold true for coyotes in Florida because they could not conclude that melanistic traits were under selective pressure in the temperate forests of the state.

Given their broad geographic range across North America, coyotes are an ideal species to explore and test ecogeographical rules believed to influence geographic variation in phenotypic and genotypic traits of animal populations [29]. For example, geographic distribution of body mass in coyote populations does not follow Bergmann’s rule because of coyote hybridization with wolves in eastern North America facilitated longitudinal variation in mass rather than latitudinal variation as predicted by Bergmann’s rule [29, 34]. However, longitudinal variation of melanism in coyote populations does not violate Gloger’s rule, because, as observed in other large carnivores such as leopards that exhibit similar west-to-east gradients in melanism [35], factors influencing melanism are associated with canopy cover and humidity rather than colder climates associated with increasing latitude [7].

It is commonly argued that, following hybridization events during colonization of eastern North America, the eastern coyote’s phenotype reflected an adaptive response to larger prey, specifically white-tailed deer (Odocoileus virginianus) [34, 36,37,38,39]. However, given the near extirpation of white-tailed deer from eastern North America during the early stages of coyote colonization and the coyote’s regular use of ungulates in western North America, Hinton et al. [29] suggested that increased body size improved coyote dispersal capabilities necessary for improving connectivity among coyote metapopulations on the colonization front. If hybridization influenced coyote movement behaviors, we should expect melanism, a conspicuous trait associated with hybridization, to be associated with space use behaviors such as larger home range sizes and increased selection for canopy and wetland cover. Therefore, our objectives were to describe the occurrence, morphometrics, and spatial ecology of melanistic individuals in extant wild populations of Canis in the Southeast. We believe examining melanism at a regional scale may provide interesting insight into differences in morphology and behavior associated with pelage color of Canis taxa and help explore hypotheses that best explain occurrence and maintenance of melanistic traits such as Gloger’s rule. We used morphometrics to test the hypothesis that melanistic coyotes and hybrids were larger than their gray conspecifics. We used global positioning system (GPS) radiotelemetry data to quantify habitat selection by southeastern coyotes to test the hypothesis that melanistic individuals would exhibit stronger selection for areas with greater canopy and wetland cover than would gray conspecifics. To test the hypothesis that melanistic individuals experience greater survival than gray individuals, we used long-term monitoring data on coyotes and red wolf-coyote hybrids collected by the United States Fish and Wildlife Service (USFWS) Red Wolf Recovery Program (hereafter Recovery Program). Finally, we discuss which mechanisms most likely influence patterns of variation in Canis coat color under the context of Gloger’s rule.

Results

Morphometrics

Overall, 460 coyotes, 532 red wolves, and 160 hybrids were captured and measured between 1987 and 2016. Morphometric measurements differed among coyotes, red wolves, and hybrids with hybrid measurements falling between the larger red wolf and smaller coyote (Tables 1 and 2). As expected, no red wolves were melanistic, whereas 5.7% of coyotes and 8.5% of hybrids consisted of melanistic individuals.

Table 1 Means (± SD) and ranges for body mass and measurements of red wolves, coyotes, and hybrids in the southeastern United States
Full size table
Table 2 Means (± SD) and ranges for head and appendage measurements of red wolves, coyotes, and hybrids in northeastern North Carolina, USA
Full size table

For our analysis, we used canids with > 4 of the 6 body measurements recorded. We assessed the measurements of 425 coyotes, 449 red wolves, and 153 hybrids using principal component analysis (PCA) (Fig. 3). Approximately 31.8, 37.6, and 30.1% of the data sets were complete for coyotes, red wolves, and hybrids, respectively. Width of head was the most commonly missing body measurement (46.5% of individuals) followed by ear (17.5%), tail (17.2%), body mass (4.1%), body length (1.4%), shoulder height (1.1%), and hind foot length (0.3%).

Fig. 3
figure 3

Scatter plots of component loadings for between-group principal components of the principal component analysis. Shaded circles represent 95% ellipses

Full size image

Only principal component (PC) 1, which explained 75.84% of the cumulative variation, had an eigenvalue > 1 (Table 3). The eigenvalues of PC1 consisted of strong positive loadings for all body measurements. Although the eigenvalues for PC2 and PC3 were < 1, both contributed 14.0% of the cumulative variance explained in our PCA (Table 3) and contained loadings of body measurements that were independent of body size. Collectively, these PC scores indicated that once PC1 accounted for body size, PC2 and PC3 accounted for variation in tail length and ear length, respectively. Mean PC1 (body size) scores for hybrids were intermediate to those for coyotes and red wolves (F4, 1021 = 449.300, P < 0.001; Fig. 4). We observed differences in mean PC2 (tail length) scores (F4, 1021 = 8.729, P < 0.001; Fig. 4) which indicated red wolves had shorter tails relative to their body size than did coyotes and hybrids, although mean tail length of red wolves was greater than mean tail length of coyotes and hybrids (Table 2). We detected no differences in mean PC3 (ear length) scores (F4, 1021 = 1.098, P = 0.356; Fig. 4), although mean ear length of red wolves was greater than mean ear length of coyotes and hybrids (Table 2). We detected no difference in mean PC scores of melanistic and gray coyotes and mean PC scores of melanistic and gray hybrids (Fig. 4). In other words, we found no evidence that melanism was correlated with differences in morphometrics observed within our Canis taxa.

Table 3 Factor loadings and percent contribution of body measurements for each of the top 3 principal components for body measurements recorded from red wolves, coyotes, and their hybrids across southeastern United States. Also included are eigenvalue, percent of total variance explained, and descriptions of principal components
Full size table
Fig. 4
figure 4

Mean principal components scores for (A) PC1 (body size), (B) PC2 (tail length), and (C) PC3 (ear length) of coyotes, red wolves, and hybrids in the southeastern United States

Full size image

Space use and habitat selection

To assess space use and habitat selection, we monitored 6 clusters of coyotes, each cluster consisting of 5 animals fitted with GPS radio-collars (n = 30 coyotes, Table 4). Our 6 clusters consisted of 20 resident (14 gray, 6 melanistic) and 10 transient (7 gray, 3 melanistic) coyotes. We included random intercepts for each coyote nested within clusters in our generalized linear mixed models (GLMMs) to account for unbalanced telemetry data and differences in land cover among 3 geographic regions (Fig. 5) when comparing space use and habitat selection by melanistic coyotes to selection by gray coyotes. Our GLMMs indicated that gray coyotes maintained smaller home ranges than did melanistic coyotes (β = − 17.432 ± 7.188 SE, P = 0.015), whereas no difference was observed between the size of core areas (β = − 1.518 ± 1.932 SE, P = 0.432), transient ranges (β = 191.9 ± 184.0 SE, P = 0.297), and biding areas (β = 20.86 ± 32.95 SE, P = 0.527) of gray and melanistic coyotes (Table 4). Mean home range sizes for gray and melanistic coyotes were 17.4 km2 (SD = 10.6) and 27.1 km2 (SD = 12.7), respectively. Mean core area size for coyotes was 4.0 km2 (SD = 2.3), whereas mean sizes for transient ranges and biding areas were 179.9 km2 (SD = 138.0) and 22.7 km2 (SD = 22.1), respectively.

Table 4 Summary of coyote collar deployments, geographical clusters, color, space use status, monitoring dates, and extent of space used (95 and 50% contour intervals representing home ranges and core areas, respectively for residents and transient ranges and biding areas, respectively for transients) in the southeastern United States. The “M” of “F” in coyote ID indicates sex
Full size table
Fig. 5
figure 5

Map of the three study areas (indicated by boxes) in the southeastern United States. The Savannah River Area (SRA) represented the Georgia and South Carolina study areas

Full size image

We detected differences in habitat selection by gray and melanistic coyotes (Tables 5 and 6) and our Spearman’s rank correlations from k-fold cross-validation indicated that our best models consistently predicted habitat selection patterns for gray (rs = 0.814, P < 0.001) and melanistic (rs = 0.813, P < 0.001) coyotes.

Table 5 Summary of top generalized linear mixed models (cumulative Akaike weight ≤ 1.00) for predicting 3rd-order habitat selection by melanistic and gray coyotes in the southeastern United States. Shown are Akaike’s Information Criteria for small sample sizes (AICc), differences among AICc (ΔAIC), AICc weights (ωi) and cumulative AICc weights
Full size table
Table 6 Model averaged parameter estimates for 3rd-order resource selection functions for radio-collared coyotes in the southeastern United States. Shown are β coefficients, standard error (SE), and 95% confidence intervals (CI)
Full size table

For gray coyotes, all land cover types except for forest were important predictors of habitat selection (Table 5). Gray coyotes exhibited strong selection for agriculture and strong avoidance of wetlands (Table 6 and Fig. 6). They also showed selection for areas in or proximate to shrubland and human development with increasing canopy cover. Gray coyotes avoided roads. For melanistic coyotes, all landcover types except for agriculture and shrubland were important predictors of habitat selection (Table 5). Melanistic coyotes exhibited strong selection for areas with increasing canopy cover (Table 6 and Fig. 6). They also showed selection for areas in or proximate to wetlands, human development, and roads and avoided areas with forest cover.

Fig. 6
figure 6

Model averaged parameter estimates for 3rd-order resource selection functions for radio-collared melanistic and gray coyotes in the southeastern United States during 2009–2017. The midpoint location and 95% confidence intervals of each distribution are represented by shapes and lines, respectively. Canopy = % canopy cover, Ag = distance to agriculture, Forest = distance to forest, Roads = distance to roads, Urban = distance to human development, Shrub = distance to shrubland, and Wetland = distance to wetland

Full size image

Survival

We documented 256 mortalities of radio-collared coyotes (64.1% of mortalities) and hybrids (35.9% of mortalities) from the Red Wolf Experimental Population Area (hereafter NC Recovery Area) in northeastern North Carolina during 1992–2018 in which melanistic animals accounted for 7 mortalities. Global tests suggested that our data did not violate the proportional hazards assumption (maximum χ2 = 0.322, P = 0.85) in which the relative hazard between coyotes and hybrids or gray and black phenotypes was constant over time and did not cause survival curves to diverge. Mean annual survival did not differ between coyotes (0.673 [0.629–0.720 95% CI]) and hybrids (0.689 [0.630–0.754 95% CI]; z = 0.357, P = 0.721, hazard = 1.062 [0.765, 1.474 95% CI]). However, melanistic animals exhibited greater annual survival than did gray conspecifics (0.827 [0.716–0.955 95% CI] vs 0.671 [0.639–0.705 95% CI]; z = − 1.892, P = 0.059, hazard = 0.478 [0.222, 1.027]).

Discussion

Although Anderson et al. [9] postulated that melanism in North American gray wolves and coyotes was introduced through hybridization with domestic dogs, Rutledge et al. [30] suggested that a more comprehensive examination of black canids from eastern North America was required before such conclusions of introgressive hybridization from dogs to gray wolves and coyotes were drawn. We agree with Rutledge et al. [30] because melanism was historically common in red wolves [21, 26, 32, 40, 41] and absent from coyotes [31]. Coyotes that colonized the Southeast during the twentieth century likely acquired melanism through hybridization with red wolves, given that melanistic coyotes were reported soon after they made inroads into the region following the decline of red wolves [21, 32] and occur predominantly in populations inhabiting the red wolf’s historical range.

Regarding dog introgression that co-occurs with melanism in southeastern coyote populations, the origin of that introgression may have occurred through human-facilitated hybridization in captive settings by which some hybrid escapees backcrossed with wild coyotes or hybridization with red wolves in the wild who may have carried dog alleles. Indeed, Mengel [42] reported that some of his F1 and F2 coyote-dog hybrids escaped from their pens and were not retrieved and Goldman [21] reported the existence of a red wolf-dog hybrid in Reynolds County, Missouri (Fig. 7c) indicating that humans crossed red wolves and dogs in captivity. In fact, there is considerable evidence of human-facilitated interbreeding of wolves, coyotes, and dogs in captive environments (Fig. 7) [21, 31, 42,43,44,45], whereas direct interbreeding (e.g., copulation; Fig. 7a) between coyote and dogs has not been documented in the wild (see review by vonHoldt and Aardema [46]) despite numerous research and monitoring programs and the common and widespread use of camera surveys in modern research. However, interbreeding between coyotes and red wolves in the wild has been repeatedly and comprehensively documented by ecological [47,48,49,50] and molecular [51,52,53] studies by which more realistic and parsimonious pathways (e.g., red wolf and coyote hybridization) can be formulated for how southeastern coyotes may have acquired melanistic traits and dog alleles.

Fig. 7
figure 7

A A captive western coyote copulating with a collie-like domestic dog [32]. B Two captive coyote-dog hybrids in Wyoming, USA [32]. C A captive red wolf-dog hybrid from Reynolds County, Missouri, USA [43]. D An F2 coyote-beagle hybrid who was part of a scientific study conducted in St. Louis, Missouri, USA [43]. E An F1 coyote-dog hybrid who was part of a scientific study from Lawrence, Kansas, USA [44]. F A captive female gray wolf nursing her litter of hybrid pups in Hall County, Texas, USA [44]. Her pups were the result of humans crossing her with a hound dog

Full size image

Melanistic coyotes and hybrids comprised 2.0–8.5% of individuals captured at our study areas indicating that the occurrence of melanism was relatively low but consistent in the region’s coyote populations. This observation corroborates previous studies reporting that melanistic individuals comprised < 10% of coyote populations surveyed in other parts of the Southeast [22, 25]. Few studies on the occurrence of melanism in red wolves exist, and the actual proportion of the historical wolf population that consisted of melanistic individuals is unknown. However, anecdotal accounts and population surveys suggest that melanistic individuals were more common in the historical red wolf population than they are in today’s coyote population. For example, the only three pictures of free roaming wild red wolves in Louisiana were photographs of melanistic wolves captured by remote cameras [40], and red wolves were often referred to as the “black timber wolf” by Louisiana zoologists [26]. During the same period of Gregory’s [40] photographic expedition of northeastern Louisiana, a survey of red wolves in Arkansas reported that 25% of the population consisted of melanistic wolves [41]. Elder and Hayden [54] reported that five of seven Missouri specimens considered to be red wolves or red wolf-coyote hybrids were melanistic. Therefore, we suggest that melanism in the historical red wolf population may have occurred at similar frequencies as those observed in gray wolves inhabiting boreal forests of North America [8,9,10, 16] and occurred more frequently in historical red wolves than in contemporary coyotes of the Southeast.

We observed no correlation between Canis morphometrics and pelage color, as melanistic coyotes and hybrids were similar in size as their gray conspecifics. Instead, body size of canids was driven by red wolf ancestry in which red wolves represented a uniquely large Canis phenotype that was not achieved by coyotes and hybrids [49]. Our findings offer evidence that the tails and ears of red wolves and coyotes are distinctive characteristics [49]. Our interpretation of PC2 (tail length) and PC3 (ear length) suggests that the distinctiveness was minor, only meaningful when related to their body sizes, and consisted of red wolves having shorter and less bushier tails than did coyotes. Additionally, the red wolf’s triangular facial appearance is accentuated by the angle at which the wolf carries its ears [47]. Despite efforts to quantify coyote niche expansion through hybridization in the northeastern United States and eastern Canada [36, 55,56,57], no field research has substantiated the assumption that intermediate body sizes facilitate unique niches for introgressed coyote populations in the Southeast. To our knowledge, coyote populations with the greatest levels of red wolf ancestry are largely restricted to the NC Recovery Area of northeastern North Carolina [58] and isolated pockets along the American Gulf Coast [59,60,61] indicating reduced niche dimensions when compared to other southeastern coyote populations with less wolf ancestry. Furthermore, we detected no differences in annual survival rates for coyotes and hybrids in the NC Recovery Area. We suggest that future research directly study the effects of red wolf ancestry in Canis populations of the region to confirm claims that hybridization can confer adaptive benefits to Canis taxa.

Mean home-range size of melanistic coyotes was 1.6 times larger than that observed for gray coyotes, and melanistic coyotes exhibited stronger selection for areas with canopy and wetland cover than did gray coyotes. Research on space use of coyotes and red wolves in the NC Recovery Area reported that body size influenced home range sizes and that coyote home range size was negatively correlated with agriculture [50, 62]. Given that we detected no difference in body sizes of melanistic and gray coyotes, it is likely that differences in land cover preferences by coyotes influenced their home range sizes such as gray coyotes exhibiting stronger selection for agriculture than did melanistic coyotes. Wetlands in our study areas were predominantly woody wetlands characterized by coastal bottomland forest and other types of woody riparian cover [62, 63]. Therefore, forest cover that melanistic coyotes avoided were drier deciduous, evergreen, and mixed forests that gray coyotes used according to their availability. Regardless of pelage color, coyotes exhibited selection for areas impacted by human development; however, gray coyotes showed avoidance of roads whereas melanistic coyotes exhibited selection for them. We speculate that the selection of roads by melanistic coyotes likely resulted from their need to use linear corridors to maneuver around inundated areas associated with woody wetland cover whereas gray coyotes could avoid roads as they selected for drier, open habitats that allowed for more diffuse movements. Nevertheless, our findings corroborate observations and insights spanning from 18th- and 19th-century naturalists to modern biologists, who linked the occurrence of melanism in wild Canis populations of the Southeast with dense canopy cover [19,20,21,22,23, 25] and support hypotheses such as Gloger’s rule that postulate an adaptive role for melanism in canopy dense environments [7, 35, 64].

Melanistic coyotes can be cryptic under dense canopy cover in which their black pelage may improve anti-predator (i.e., human hunters) behaviors through superior camouflage [35, 64]. For example, melanistic coyotes and hybrids in the NC Recovery Area exhibited greater survival than did their gray counterparts, which may be attributed to better concealment in coastal bottomland forests. We suggest that the large home range sizes of melanistic coyotes was caused by their preferences for areas with dense canopy, and their selection for lower quality wetland cover was likely a trade-off of reduced foraging efficiency for improved survival. Large home ranges and differential use of land cover by melanistic coyotes may facilitate weak assortative mating in eastern coyote populations, whereby melanistic animals have lower success of finding compatible mates in comparison to their gray conspecifics. Indeed, space use behaviors influenced assortative mating in red wolves and coyotes [50] and melanistic coyotes selected for similar land cover types (i.e., wetland cover) as did red wolves in eastern North Carolina [65]. Furthermore, Hinton et al. [29] suggested that red wolf ancestry improved coyote dispersal capabilities, rather than their ability to kill deer [36, 38, 66], improving connectivity among coyote metapopulations in a region dominated by forest cover. Given that melanism is a newly acquired trait in eastern coyote populations that may influence coyote space use behaviors, we believe some reproductive isolation between gray and melanistic individuals along the colonization front may have been important for engendering diversity and dispersal of newly acquired traits.

Methods

Study area

We compiled data from 3 regions in the Southeast representing 3 separate Canis populations: North Carolina’s Albemarle Peninsula, the Savannah River area along the Georgia and South Carolina border, and southeastern Alabama (Fig. 5). Since 1987, the Albemarle Peninsula of northeastern North Carolina served as the NC Recovery Area [28]. The peninsula included 5 counties (Beaufort, Dare, Hyde, Tyrrell, and Washington) and consisted of approximately 6000 km2 of federal, state, and private lands. The NC Recovery Area was predominantly an intensively farmed agricultural-hardwood bottomland mosaic in which approximately 30% of the landscape was croplands. More details of the NC Recovery Area can be found in Hinton et al. [67].

During 2015–2016, we captured and monitored coyotes across several broad areas in southeastern Alabama (Barbour, Macon, and Pike Counties), east-central Georgia (Columbia, Jefferson, Lincoln, McDuffie, and Warren Counties), and western South Carolina (Aiken, Edgefield, McCormick, and Saluda Counties) totaling approximately 16,200 km2 (Fig. 5). Because some coyotes captured at the Georgia and South Carolina study areas dispersed into each respective study area, we considered coyotes in both areas along the Savannah River to be one population and referred to the area as the Savannah River area (SRA). Land cover in Alabama and SRA contained a mix of early successional, agricultural, forested, and urban habitats. More details on these two study areas can be found in Ward et al. [63].

Animal captures

Red wolves, coyotes, and hybrids were captured with foothold traps with offset jaws (Victor #1.5 and #3 softcatch, Woodstream Corporation, Lititz, Pennsylvania, USA, and Minnesota Brand 550, Minnesota Trapline Products, Pennock, Minnesota, USA). From 1987 to 2011, the Recovery Program captured red wolves, coyotes, and hybrids in northeastern North Carolina. During 2015–2016, the University of Georgia captured coyotes in Alabama and SRA.

Once captured, animals were restrained with a catchpole, muzzle, and hobbles. However, we chemically immobilized some animals with an intramuscular injection of 1.3–1.8 mg/kg ketamine HCl and 0.2–0.4 mg/kg xylazine HCl to inspect inside their mouths for injuries. Sex, weight, body measurements, and pelage color were recorded for all animals, and ages of coyotes and hybrids were estimated by tooth wear [68, 69]. We acquired accurate estimates of red wolf ages through the Recovery Program’s detailed life-history data [49, 70]. Ancestry of all animals was confirmed with microsatellite markers and genomic information [29, 49, 71, 72]. We categorized animals ≥2 years as adults, 1–2 years old as juveniles, and < 1 year old as pups. Our capture and handling of animals followed guidelines approved by the American Society of Mammalogists [73] and were approved by the Institutional Animal Care and Use Committees at the University of Georgia (A2014 08–025-R2) and Louisiana State University (AE2009–19).

Prior to release at their capture sites, all red wolves, coyotes, and hybrids were fitted with mortality-sensitive radio-collars. In the NC Recovery Area, animals were predominantly fitted with very high frequency (VHF) radio collars (Telonics, Inc., Mesa, AZ) for monitoring space use, breeding status, and survival [70]. During 2005–2011, some red wolves and coyotes were fitted with global positioning system (GPS) radio collars for research purposes [62, 65, 74, 75]. Red wolves equipped with GPS radio-collars (Lotek 4400S, Newmarket, Ontario, Canada) had their locations recorded every 5 hours on a 24-hour rotating schedule throughout the year. Coyotes in the NC Recovery Area equipped with smaller Lotek 3300 s GPS collars had their locations recorded every 4 hours (e.g., 0000, 0400, 0800, 1200, and so on). The Recovery Program monitored radio-collared red wolves, coyotes, and hybrids 2–3 times a week from aircraft to identify and monitor wolf territories in the NC Recovery Area. Coyotes in Alabama and SRA fitted with G2110E satellite GPS collars (Iridium; Advanced Telemetry Systems, Isanti, Minnesota, USA) had their locations recorded every 4 hours beginning at midnight.

Morphometrics

We recorded postcranial measurements including body length (anterior tip of the nose pad to the tail base), tail length (tip of the fleshy part of the tail to the tail base), hind foot length (hock to the tip of the digital pads), and shoulder height (tip of the scapula to tip of the digital pads). Cranial measurements included width of head (most widely separated points) and ear length (edge of the external auditory canal to the tip of the ear).

We used a PCA to extract the dominant, underlying gradients of variation (principal components) in our dataset [76]. The PCs are weighted linear combinations of the original variables ordered according to the amount of variation each PC explained. We logarithmically transformed our data, as body mass was measured on a different scale than linear body measurements.

Approximately 66% of our study animals were missing at least 1 body measurement which makes deletion of individuals or traits from the analysis impractical. In this context, we addressed the issue of missing values within our morphometrical dataset by using a joint modeling approach in the R package Amelia [77] to create a completed dataset to perform the PCA. The joint modeling approach obtains maximum likelihood estimates using an expectation-maximization algorithm and considers the relationships between variables to fill the gaps [78]. In doing so, the joint modeling approach allowed for missing value uncertainty to be incorporated into our PCA [78]. We used the relative percent variance criterion jointly with the latent root criterion (PCs with eigenvalues > 1) to determine the number of significant PCs to retain and interpret, because the latent root criterion is known to be overly conservative when the number of variables is < 20 [79]. We then based our interpretation of each PC on those variables with loadings ≥0.50 or ≤ − 0.50 and placed most emphasis on those with loadings ≥0.60 or ≤ − 0.60 [79]. We used variables with the strongest loadings to interpret the ecological meaning of each PC.

Space use and habitat selection

We used only GPS-collared coyotes for our space use and habitat selection analyses for two reasons. First, red wolves could not be included in this analysis because melanistic individuals were absent in the extant wolf population. Additionally, estimates of wolf space use and habitat selection have been reported [65, 75, 80]. Second, we excluded coyotes and hybrids in the NC Recovery Area who were fitted with VHF radio-collars because they were not monitored intensively to have achieved sufficient numbers of locations (e.g., ≥30) required for reliable estimates of home range size and habitat selection [81].

To identify resident coyotes, we used an animal’s spatial association with other animals and fidelity to an area for ≥4 months as our primary criteria [62, 63]. We confirmed the presence of mates and pack members through field inspection for sign (i.e., visual observation and tracks) of other individuals over the course of several weeks [62, 63]. Wide-ranging and unstable space use by animals was characteristic of transient behaviors, as these individuals were typically young dispersers that moved nomadically on the landscape in search of mates and territories and, therefore, did not maintain home ranges [62]. Accordingly, we did not refer to transient space use as home ranges and core areas, but instead refer to space used by transient coyotes as transient ranges and biding areas [62]. Transient ranges and biding areas are analogous to home ranges and core areas, but are assigned to non-breeding, solitary animals traversing the landscape seeking mates and territories whereas home ranges and core areas were assigned to animals belonging to breeding pairs and packs who defended territories.

To investigate space use and habitat selection by coyotes, we identified local clusters of GPS-collared coyotes that included a melanistic individual (Table 1). This approach allowed us to account for differences in land cover available to coyotes inhabiting the 3 geographic areas when comparing habitat selection by melanistic coyotes and gray coyotes. Using the locations of our GPS-collared melanistic coyotes, we identified 6 unique clusters of coyotes across our Alabama, SRA, and North Carolina study areas where melanistic individuals resided. We created clusters of 5 coyotes by identifying the social status of melanistic coyotes (resident vs. transient) and assigned the 3–4 closest gray coyotes of the same social status to the cluster. For example, in Alabama, we captured and GPS-collared 54 coyotes between 2015 and 2016. Of those 54 coyotes, only 1 coyote was melanistic. We used that coyote to create the Alabama cluster, and because that melanistic coyote was a resident animal, we selected the 4 closest resident gray coyotes to create the Alabama cluster (Table 1).

We estimated space use of resident and transient coyotes using dynamic Brownian bridge movement models (dBBMM) in Program R 3.6.3, using the package move [82, 83]. For full tracks of each animal, we measured variation in movements using a moving window size of 7 locations (equivalent to 14 hours) with a margin of 3 locations while assuming an error estimate of 20 m for all locations [62]. For resident coyotes, we defined 95 and 50% contour intervals as home ranges and core areas, respectively [62]. For transient coyotes, we defined 95 and 50% contour intervals as transient ranges and biding areas, respectively [62]. To account for effects of pelage color on coyote space use, we used generalized linear mixed models (GLMM) with a logit link in Program R [84]. Our response variables were home range and core area sizes for resident animals, and transient range and biding area sizes for transients. We modeled pelage color as a binary (0 = gray, 1 = melanistic) predictor variable and included random intercepts for each coyote nested within clusters to account for the influence of unbalanced telemetry data and unmeasured geographic-related factors.

To develop resource selection functions (RSF), we followed a 3rd-order resource selection design [85] to examine the relationship between land cover and coyote space use within 95% ranges. We used individual coyotes as sampling units and measured resource availability for each animal at the 95% contour intervals estimated via dBBMMs. To estimate RSFs, we used a binomial approach by comparing characteristics of known locations to 3 times the number of random locations within 95% contour intervals for each coyote [85]. We generated 3 times more random locations than GPS locations for each coyote to ensure accurate estimates of land cover availability for use with RSFs [86].

We overlaid GPS locations and random locations onto 30-m resolution digital maps of 7 measurements of major land cover types (percent canopy cover, distance to roads, distance to agriculture, distance to forest, distance to shrubland, distance to human development, and distance to wetlands) that likely influenced coyote habitat selection. Using the Euclidean Distance tool in the Spatial Analyst toolbox in ArcGIS 10.7 (Environmental Systems Research Institute Inc., Redlands, California), we calculated distances from every 30-m pixel to the closest landscape features (agriculture, forest, shrubland, human development, and wetlands) using the United States Geological Survey (USGS), National Land Cover Database from 2011 and 2016 [87, 88]. Our forest cover category was created by reclassifying the NLCD’s deciduous, evergreen, and mixed forest land covers as “forest” cover. We acquired roads layers from the Alabama, Georgia, and North Carolina Departments of Transportation to create a distance to roads layer. Finally, we used percent canopy cover derived from the 2011 and 2016 NLCD tree canopy cover layers.

We used GLMMs with a logit link in Program R to evaluate 3rd-order selection by coyotes [84, 85, 89]. We included random intercepts for each coyote nested within clusters in each model using the lme4 package in R [90]. Including random intercepts for individual coyotes and clusters accounted for the influence of unbalanced telemetry data and unmeasured geographic-related factors. We modeled resource selection with a binary (0 = random, 1 = known) response variable. Before modeling, we rescaled and centered values for distance-based variables and canopy cover by subtracting their mean and dividing by 1 standard deviation. To model the influence of land cover correlates on the relative probability of habitat selection by melanistic and gray coyotes, we constructed separate but identical GLMMs for each coyote type and compared the coefficients and their 95% confidence intervals [86]. As noted by Northrup et al. [86], non-spatial factors such as pelage color or sex cannot be modeled as covariates in spatial models because these models are typically approximating a Poisson point process model.

We then restricted models to first-order terms and explored all possible subsets of the 7 predictors including the null model as candidate models to investigate coyote habitat selection. We evaluated model sets using Akaike’s Information Criterion adjusted for small sample sizes (AICc) and used ΔAICc to select which models best supported factors influencing resource selection by coyotes [91]. We considered the model with the lowest AICc and the greatest model weight as the best approximating model. However, when model sets had ≥2 models that were within 2 ΔAICc of the top model, we performed model-averaging across the top model set to calculate effect sizes of explanatory variables. We only considered parameter estimates with 95% confidence intervals that excluded 0 to be informative.

We then validated our top models describing coyote habitat selection influenced by pelage color using k-fold (k = 5 repetitions) procedures to assess the predictive performance of RSF models by randomly dividing our data into 5 equally sized folds and using 4 folds to create our training data set (80% of the data) and the last fold (20% of the data) as our test data set. We classified RSF values into 5 quantile bins and calculated each bin from the test data set. We used Spearman’s rank correlation to compare expected and observed frequencies [92]. Models with good predictive abilities are expected to show a strong correlation with greater numbers of locations falling into higher probability bins.

Annual survival

We modeled annual survival rates between melanistic and gray Canis taxa using the Kaplan-Meier estimator. To provide robust hazard ratio estimates among melanistic and gray Canis, we used a Cox proportional hazard model [93] with coat color as a dummy variable (melanistic = 1, gray = 0). We used a right-censored design with time-at-risk based on time (days) since the animal’s first capture [94] and evaluated main effects only. We tested the proportional hazards assumption of Cox PH using the formal test recommended by [93] and found no significant violations of proportionality in any of the predictor variables included in models (all P > 0.05). We used a 365-day (recurrent) time scale to model the baseline hazard [94], standardized to a year beginning on 1 January and ending on 31 December.

Availability of data and materials

The datasets generated during the current study are available from the corresponding author on reasonable request.

Abbreviations

USFWS:

United States Fish and Wildlife Service

GPS:

Global Positioning System

PCA:

Principal component analysis

PC:

Principal component

GLMM:

Generalized linear mixed model

SD:

Standard deviation

AIC:

Akaike information criterion

SRA:

Savannah River area

VHF:

Very-high frequency

REML:

Restricted maximum-likelihood

dBBMM:

Dynamic Brownian bridge movement model

RSF:

Resource selection function

USGS:

United States Geological Survey

NLCD:

National Land Cover Database

References

  1. Caro T. The colours of extant mammals. Semin Cell Dev Biol. 2013;24:542–52.

    Article  PubMed  Google Scholar 

  2. Caro T, Mallarino J. Coloration in mammals. Trends Ecol Evol. 2020;35:357–66.

    Article  PubMed  Google Scholar 

  3. Cerezer FO, Ribeiro JR, Graipel M, Cáceres NC. The dark side of coloration: Ecogeographical evidence supports Gloger’s rule in American marsupials. Evolution. 2020;74:2046–58.

    Article  PubMed  Google Scholar 

  4. Majerus ME, Mundy NI. Mammalian melanism: natural selection in black and white. Trends Genet. 2003;19:585–8.

    Article  CAS  PubMed  Google Scholar 

  5. San-Jose LM, Roulin A. Toward understanding the repeated occurrence of associations between melanin-based coloration and multiple phenotypes. Am Nat. 2018;192:111–30.

    Article  PubMed  Google Scholar 

  6. Caro T. The adaptive significance of coloration in mammals. BioScience. 2005;55:125–36.

    Article  Google Scholar 

  7. Delhey K. A review of Gloger’s rule, an ecogeographical rule of colour: definitions, interpretations and evidence. Biol Rev. 2019;94:1294–316.

    PubMed  Google Scholar 

  8. Musiani M, Leonard JA, Cluff HD, Gates CC, Mariani S, Paquet PC, et al. Differentiation of tundra/taiga and boreal coniferous forest wolves: genetics, coat colour and association with migratory caribou. Mol Ecol. 2007;16:4149–70.

    Article  CAS  PubMed  Google Scholar 

  9. Anderson TM, Candille SI, Musiani M, Greco C, Stahler DR, Smith DW, et al. Molecular and evolutionary history of melanism in north American gray wolves. Science. 2009;323:1339–43.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  10. Gipson PS, Bangs EE, Bailey TN, Boyd DK, Cluff HD, Smith DW, et al. Color patterns among wolves in western North America. Wildl Soc Bullet. 2002;30:821–30.

    Google Scholar 

  11. Lokhande AS, Bajaru SB. First record of melanistic Indian wolf Canis lupus pallipes from the Indian subcontinent. J Bombay Nat Hist Soc. 2013;110:220–1.

    Google Scholar 

  12. Saleh M, Younes M, Sarhan M, Abdel-Hamid F. Melanism and coat colour polymorphism in the Egyptian wolf Canis lupaster Hemprich & Ehrenberg (Carnivora: Canidae) from Egypt. Zool Middle East. 2018;64:195–206.

    Article  Google Scholar 

  13. Zafar-ul Islam M, Boug A, Shehri A, da Silva LG. Geographic distribution patterns of melanistic Arabian wolves, Canis lupus arabs (Pocock), in Saudi Arabia (Mammalia: Carnivora). Zool Middle East. 2019;65:95–103.

    Article  Google Scholar 

  14. Coulson T, MacNulty DR, Stahler DR, vonHoldt B, Wayne RK, Smith DW. Modeling effects of environmental change on wolf population dynamics, trait evolution, and life history. Science. 2011;334:1275–8.

    Article  CAS  PubMed  Google Scholar 

  15. Hedrick PW, Smith DW, Stahler DR. Negative-assortative mating for color in wolves. Evolution. 2016;70:757–66.

    Article  PubMed  Google Scholar 

  16. Schweizer RM, Durvasula A, Smith J, Vohr SH, Stahler DR, Galaverni M, et al. Natural selection and origin of a melanistic allele in north American gray wolves. Mol Biol Evol. 2018;35:1190–209.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  17. Caniglia R, Fabbri E, Greco C, Galaverni M, Manghi L, Boitani L, et al. Black coats in an admixed wolf× dog pack is melanism an indicator of hybridization in wolves? European J Wildl Res. 2013;59:543–55.

    Article  Google Scholar 

  18. Khosravi R, Aghbolaghi MA, Rezaei HR, Nourani E, Kaboli M. Is black coat color in wolves of Iran an evidence of admixed ancestry with dogs? J Appl Genet. 2015;56:97–105.

    Article  PubMed  Google Scholar 

  19. Bartram W. Travels through North & South Carolina, Georgia, East & West Florida. New York: Dover Publications, Inc.; 1791.

  20. Audubon JJ, Bachman J. The quadrupeds of North America, vol. 2. New York: V. G. Audubon; 1851.

    Google Scholar 

  21. Goldman EA. Classification of wolves. In: Young SP, Goldman EA, editors. The wolves of North America. Part 2. Washington, D.C.: American Wildlife Institute; 1944.

    Google Scholar 

  22. Gipson P. Melanistic Canis in Arkansas. Southwest Nat. 1976;21:124–6.

    Article  Google Scholar 

  23. Nowak RM. North American quaternary Canis. Monograph Museum Nat Hist Univ Kans. 1979;6:1–154.

    Google Scholar 

  24. Mowry CB, Edge JL. Melanistic coyotes in Northwest Georgia. Southeast Nat. 2014;13:280–7.

    Article  Google Scholar 

  25. Caudill G, Caudill D. Melanism of coyotes (Canis latrans) in Florida. Am Midl Nat. 2015;174:335–42.

    Article  Google Scholar 

  26. Nowak RM. The red wolf in Louisiana. Defenders Wildl News. 1967;42:60–70.

    Google Scholar 

  27. United States Fish and Wildlife Service. Red wolf recovery plan. Atlanta: United States Fish and Wildlife Service; 1989.

    Google Scholar 

  28. Hinton JW, Chamberlain MJ, Rabon DR Jr. Red wolf (Canis rufus) recovery: a review with suggestions for future research. Animals. 2013;3:722–44.

    Article  PubMed  PubMed Central  Google Scholar 

  29. Hinton JW, Heppenheimer E, West KM, Caudill D, Karlin ML, Kilgo JC, et al. Geographic patterns in morphometric and genetic variation for coyote populations with emphasis on southeastern coyotes. Ecol Evol. 2019;9:3389–404.

    Article  PubMed  PubMed Central  Google Scholar 

  30. Rutledge LY, Wilson PJ, Kyle CJ, Wheeldon TJ, Patterson BR, White BN. How the gray wolf got its color. Science. 2009;325:33–4.

    Article  CAS  PubMed  Google Scholar 

  31. Young SP, Jackson HHT. The clever coyote. Harrisburg: Stackpole Company; 1951.

    Google Scholar 

  32. Halloran AF. Black red wolves. Oklahoma Wildl. 1958;14:6–8.

    Google Scholar 

  33. Halloran AF. A melanistic coyote from Oklahoma. Southwest Nat. 1963;8:48–9.

    Article  Google Scholar 

  34. Way JG. A comparison of body mass of Canis latrans (coyotes) between eastern and western North America. Northeast Nat. 2007;14:111–24.

    Article  Google Scholar 

  35. da Silva LG, Kawanishi K, Henschel P, Kittle A, Sanei A, Reebin A, et al. Mapping black panthers: macroecological modeling of melanism in leopards (Panthera pardus). PLoS One. 2017;12:e0170378.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  36. Kays R, Curtis A, Kirchman JJ. Rapid adaptive evolution of northeastern coyotes via hybridization with wolves. Biol Lett. 2010;6:89–93.

    Article  PubMed  Google Scholar 

  37. Power JW, LeBlanc N, Bondrup-Nielsen S, Boudreau MJ, O'Brien MS, Stewart DT. Spatial genetic and body-size trends in Atlantic Canada Canis latrans (coyote) populations. Northeast Nat. 2015;22:598–612.

    Article  Google Scholar 

  38. Way JG, Lynn WS. Northeastern coyote/coywolf taxonomy and admixture: a meta-analysis. Canid Biol Conserv. 2016;19:1–7.

    Google Scholar 

  39. vonHoldt BM, Kays R, Pollinger JP, Wayne RK. Admixture mapping identifies introgressed genomic regions in north American canids. Mol Ecol. 2016;25:2443–53.

    Article  PubMed  Google Scholar 

  40. Gregory T. The black wolf of the Tensas. Chicago Acad Sci. 1935;6:35–68.

    Google Scholar 

  41. Black JD. Mammals of northwestern Arkansas. J Mammal. 1936;17:29–35.

    Article  Google Scholar 

  42. Mengel RM. A study of dog-coyote hybrids and implications concerning hybridization in Canis. J Mammal. 1971;52:316–36.

    Article  CAS  PubMed  Google Scholar 

  43. Mech LD, Asa CS, Callahan M, Christensen BW, Smith F, Young JK. Studies of wolf x coyote hybridization via artificial insemination. PLoS One. 2017;12:e0184342.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  44. Mech LD, Christensen BW, Asa CS, Callahan M, Young JK. Production of hybrids between western gray wolves and western coyotes. PLoS One. 2014;9:e88861.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  45. Fox MW. Behavior genetics of F1 and F2 coyote-dog hybrids. Appl Anim Ethol. 1975;1:185–95.

    Article  Google Scholar 

  46. vonHoldt BM, Aardema ML. Updating the bibliography of interbreeding among Canis in North America. J Hered. 2020;111:249–62.

    Article  PubMed  Google Scholar 

  47. Riley GA, McBride RT. A survey of the red wolf (Canis rufus). In: Fox MW, editor. The wild canids: their systematics, behavioral ecology and evolution. New York: Van Nostrand Reinhold; 1975. p. 263–77.

    Google Scholar 

  48. Shaw JH. Ecology, behavior, and systematics of the red wolf (Canis rufus). In: Doctoral dissertation. New Haven: Yale University; 1975.

    Google Scholar 

  49. Hinton JW, Chamberlain MJ. Morphometrics of Canis taxa in eastern North Carolina. J Mammal. 2014;95:855–61.

    Article  Google Scholar 

  50. Hinton JW, Gittleman JL, van Manen FT, Chamberlain MJ. Size assortative choice and mate availability influences hybridization between red wolves (Canis rufus) and coyotes (Canis latrans). Ecol Evo. 2018;8:3927–40.

    Article  Google Scholar 

  51. Adams JR, Leonard JA, Waits LP. Widespread occurrence of a domestic dog mitochondrial DNA haplotype in southeastern US coyotes. Mol Ecol. 2003;12:541–6.

    Article  CAS  PubMed  Google Scholar 

  52. Bohling JH, Waits LP. Assessing the prevalence of hybridization between sympatric Canis species surrounding the red wolf (Canis rufus) recovery area in North Carolina. Mol Ecol. 2011;20:2142–56.

    Article  PubMed  Google Scholar 

  53. Bohling JH, Waits LP. Factors influencing red wolf–coyote hybridization in eastern North Carolina, USA. Biol Conserv. 2015;184:108–16.

    Article  Google Scholar 

  54. Elder WH, Hayden CM. Use of discriminant function in taxonomic determination of canids from Missouri. J Mammal. 1977;58:17–24.

    Article  Google Scholar 

  55. Thornton DH, Murray DL. Influence of hybridization on niche shifts in expanding coyote populations. Divers Distrib. 2014;20:1355–64.

    Article  Google Scholar 

  56. Ellington EH, Murray DL. Influence of hybridization on animal space use: a case study using coyote range expansion. Oikos. 2015;124:535–42.

    Article  Google Scholar 

  57. Otis JA, Thornton D, Rutledge L, Murray DL. Ecological niche differentiation across a wolf-coyote hybrid zone in eastern North America. Divers Distrib. 2017;23:529–39.

    Article  Google Scholar 

  58. Bohling JH, Dellinger J, McVey JM, Cobb DT, Moorman CE, Waits LP. Describing a developing hybrid zone between red wolves and coyotes in eastern North Carolina, USA. Evol Appl. 2016;9:791–804.

    Article  PubMed  PubMed Central  Google Scholar 

  59. Heppenheimer E, Brzeski KE, Hinton JW, Chamberlain MJ, Robinson J, Wayne RK, et al. A genome-wide perspective on the persistence of red wolf ancestry in southeastern canids. J Hered. 2020;111:277–86.

    Article  CAS  PubMed  Google Scholar 

  60. Murphy SM, Adams JR, Cox JJ, Waits LP. Substantial red wolf genetic ancestry persists in wild canids of southwestern Louisiana. Conserv Lett. 2019;12:e12621.

    Article  Google Scholar 

  61. vonHoldt BM, Hinton JW, Shutt AC, Murphy SM, Karlin ML, Adams JR, et al. Reviving ghost alleles: genetically admixed coyotes along the American Gulf Coast are critical for saving the endangered red wolf. Sci Adv. 2022;8:eabn7731.

  62. Hinton JW, van Manen FT, Chamberlain MJ. Space use and habitat selection by resident and transient coyotes (Canis latrans). PLoS One. 2015;10:e0132203.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  63. Ward JN, Hinton JW, Johannsen KL, Karlin ML, Miller KV, Chamberlain MJ. Home range size, vegetation density, and season influences prey use by coyotes (Canis latrans). PLoS One. 2018;13:e0203703.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  64. Mooring MS, Eppert AA, Botts RT. Natural selection of melanism in Costa Rican jaguar and oncilla: a test of Gloger’s rule and the temporal segregation hypothesis. Trop Conserv Sci. 2020;2020(13):194008290910364.

    Google Scholar 

  65. Hinton JW, Proctor C, Kelly MJ, van Manen FT, Vaughan MR, Chamberlain MJ. Space use and habitat selection by resident and transient red wolves (Canis rufus). PLoS One. 2016;11:e0167603.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  66. Power JWB, LeBlanc N, Bondrup-Nielsen S, Boudreau MJ, O’Brien MS, Stewart DT. Spatial genetic and body-size trends in Atlantic Canada Canis latran (coyote) populations. Northeast Nat. 2015;22:598–612.

    Article  Google Scholar 

  67. Hinton JW, Ashley AK, Dellinger JA, Gittleman JL, van Manen FT, Chamberlain MJ. Using diets of Canis breeding pairs to assess resource partitioning between sympatric red wolves and coyotes. J Mammal. 2017;98:475–88.

    Article  Google Scholar 

  68. Gier HT, Coyotes in Kansas. Kan. State Coll. Ag Exp Stat Bullet. 1968;939:1–118.

    Google Scholar 

  69. Gipson PS, Ballard WB, Nowak RM, Mech LD. Accuracy and precision of estimating age of gray wolves by tooth wear. J Wildl Manag. 2000;64:752–8.

    Article  Google Scholar 

  70. Hinton JW, White GC, Rabon DR Jr, Chamberlain MJ. Survival and population estimates of the red wolf. J Wildl Manag. 2017;81:417–28.

    Article  Google Scholar 

  71. Heppenheimer E, Cosio DS, Brzeski KE, Caudill D, Van Why K, Chamberlain MJ, et al. Demographic history influences spatial patterns of genetic diversity in recently expanded coyote (Canis latrans) populations. Heredity. 2018;120:183–95.

    Article  PubMed  Google Scholar 

  72. Heppenheimer E, Brzeski KE, Hinton JW, Patterson BR, Rutledge LY, DeCandia AL, et al. High genomic diversity and candidate genes under selection associated with range expansion in eastern coyote (Canis latrans) populations. Ecol Evol. 2018;8:12641–55.

    Article  PubMed  PubMed Central  Google Scholar 

  73. Sikes RS, Gannon WL. The animal care and use Committee of the American Society of Mammalogists. Guidelines of the American Society of Mammalogists for the use of wild mammals in research. J Mammal. 2016;97:663–88.

    Article  PubMed  PubMed Central  Google Scholar 

  74. Chadwick J, Fazio B, Karlin M. Effectiveness of GPS-based telemetry to determine temporal changes in habitat use and home-range sizes of red wolves. Southeast Nat. 2010;9:303–16.

    Article  Google Scholar 

  75. Dellinger JA, Proctor C, Steury TD, Kelly MJ, Vaughan MR. Habitat selection of a large carnivore, the red wolf, in a human-altered landscape. Biol Conserv. 2013;2013(157):324–30.

    Article  Google Scholar 

  76. Gotelli NJ, Ellison AM. A primer of ecological statistics. Sunderland: Sinauer Associates Inc.; 2004.

    Google Scholar 

  77. Honaker J, King G, Blackwell M. Amelia II: a program for missing data. J Stat Softw. 2011;45:1–47.

    Article  Google Scholar 

  78. Dray S, Josse J. Principal component analysis with missing values: a comparative survey of methods. Plant Ecol. 2015;216:657–67.

    Article  Google Scholar 

  79. McGarigal K, Cushman S, Stanford S. Multivariate statistics for wildlife and ecology research. New York: Springer; 2000.

    Book  Google Scholar 

  80. Karlin M, Vaclavik T, Chadwick J, Meentemeyer R. Habitat use by adult red wolves, Canis rufus, in an agricultural landscape, North Carolina, USA. Mammal Study. 2016;41:87–95.

    Article  Google Scholar 

  81. Seaman DE, Millspaugh JJ, Kernohan BJ, Brundige GC, Raedeke KJ, Gitzen RA. Effects of sample size on kernel home range estimates. J Wildl Manag. 1999;77:739–47.

    Article  Google Scholar 

  82. Kranstauber B, Kays R, LaPoint SD, Wikelski M, Safi K. A dynamic Brownian bridge movement model to estimate utilization distributions for heterogeneous animal movement. J Anim Ecol. 2012;81:738–46.

    Article  PubMed  Google Scholar 

  83. Kranstauber B, Smolla M. Move: Visualizing and analyzing animal track data. R package version 1.1.387. 2013. https://cran.r-project.org/web/packages/move/index.html. Accessed 01 Mar 2020.

  84. R Core Team. R: A language and environment for statistical computing. Vienna: R Foundation for statistical Computing. 2020. https://www.r-project.org/. Accessed 01 Mar 2020.

  85. Manly BF, McDonald LL, Thomas DL, McDonald TL, Erickson WP. Resource selection by animals: statistical analysis and design for field studies. 2nd ed. Boston: Kluwer Academics; 2002.

    Google Scholar 

  86. Northrup JM, Vander Wal E, Bonar M, Fieberg J, Laforge MP, Leclerc M, et al. Conceptual and methodological advances in habitat-selection modeling: guidelines for ecology and evolution. Ecol Appl. 2022;32:e02470.

    Article  PubMed  Google Scholar 

  87. Homer C, Dewitz J, Yang L, Jin S, Danielson P, Xian G, et al. Completion of the 2011 National Land Cover Database for the conterminous United States–representing a decade of land cover change information. Photogramm Eng Remote Sensing. 2015;81:345–54.

    Google Scholar 

  88. U. S. Geological Survey. The National Map. 2020. https://www.usgs.gov/programs/national-geospatial-program/national-map. Accessed 01 Oct 2020.

  89. Johnson DH. The comparison of usage and availability measurements for evaluating resource preferences. Ecology. 1980;61:65–71.

    Article  Google Scholar 

  90. Bates DM, Mächler M, Bolker B, Walker S. Fitting linear mixed-effects models using lme4. J Stat Softw. 2015;67:1–48.

    Article  Google Scholar 

  91. Burnham KP, Anderson DR. Model selection and multimodel inference: a practical information-theoretic approach. New York: Springer; 2002.

    Google Scholar 

  92. Boyce MS, Vernier PR, Nielsen SE, Schmiegelow FKA. Evaluating resource selection functions. Ecol Model. 2002;157:281–300.

    Article  Google Scholar 

  93. Therneau TM, Grambsch PM. Modeling survival data: extending the Cox model. New York: Springer; 2000.

    Book  Google Scholar 

  94. Fieberg J, DelGiudice GD. What time is it? Choice of time origin and scale in extended proportional hazards models. Ecology. 2009;90:1687–97.

    Article  PubMed  Google Scholar 

Download references

Acknowledgments

We appreciate the support of past and present members of the USFWS Red Wolf Recovery Program. We also thank collaborators in the Alabama Department of Conservation and Natural Resources (ALDCNR), Georgia Department of Natural Resources – Wildlife Resources Division (GADNR), and South Carolina Department of Natural Resources (SCDNR) for their support. We thank Weyerhaeuser Company and numerous landowners for providing access to their properties. We thank D. Eaton and R. Johnson for their assistance in trapping coyotes in Alabama, Georgia, and South Carolina. We thank A. Shutt for providing archived photos of melanistic red wolves. The findings and conclusions in this article are those of the authors and do not necessarily represent the views of the USFWS, ALDCNR, GADNR, SCDNR, USEPA, and Weyerhaeuser Company. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the US Government.

Funding

This research was funded by the Alabama Department of Conservation and Natural Resources, Georgia Department of Natural Resources—Wildlife Resources Division, South Carolina Department of Natural Resources, the School of Renewable Natural Resources at Louisiana State University, Weyerhauser Company, and Warnell School of Forestry and Natural Resources at the University of Georgia.

Author information

Authors and Affiliations

  1. Wolf Conservation Center, 7 Buck Run, South Salem, NY, 10590, USA

    Joseph W. Hinton

  2. Washington Department of Fish and Wildlife, 1111 Washington Street SE, Olympia, WA, 98501, USA

    Kyla M. West

  3. United States Environmental Protection Agency, Great Lakes Toxicology and Ecology Division, 26 Martin Luther King Drive West, Cincinnati, OH, 45268, USA

    Daniel J. Sullivan

  4. Department of Environmental and Forest Biology, State University of New York College of Environmental Science and Forestry, 1 Forestry Drive, Syracuse, NY, 13210, USA

    Jacqueline L. Frair

  5. Warnell School of Forestry and Natural Resources, University of Georgia, 180 E Green Street, Athens, GA, 30621, USA

    Michael J. Chamberlain

Authors
  1. Joseph W. Hinton
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Kyla M. West
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Daniel J. Sullivan
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Jacqueline L. Frair
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Michael J. Chamberlain
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

JWH conceived the project, designed the study, organized and did field work, data analysis, and drafted the manuscript; KMW contributed intellectually, assisted with data analysis and drafting the manuscript; DJS contributed intellectually, assisted with data analysis and drafting the manuscript; JLF assisted with drafting the manuscript; MJC contributed to the project design, secured funding, provided equipment, contributed intellectually, and assisted with drafting the manuscript. All authors reviewed, edited, and approved the manuscript.

Corresponding author

Correspondence to Joseph W. Hinton.

Ethics declarations

Ethics approval and consent to participate

This study was approved by the Institutional Animal Care and Use Committees at the University of Georgia (A2014 08–025-R2) and Louisiana State University (AE2009–19), and our capture and handling of animals followed guidelines approved by the American Society of Mammalogists. Because the red wolf is listed as critically endangered under the United States Endangered Species Act and by the International Union for Conservation of Nature, we operated under a cooperative agreement with the USFWS Recovery Program to assist in their trapping and handling of red wolves. Permits to trap and handle coyotes were acquired from Alabama Department of Conservation and Natural Resources, Georgia Department of Natural Resources, North Carolina Wildlife Resources Commission, and South Carolina Department of Natural Resources. Red wolf-coyote hybrids existed under the same legal protections as coyotes and were trapped and handled under permits granted by the North Carolina Wildlife Resources Commission. The study was carried out in compliance with the ARRIVE guidelines.

Consent for publication

This manuscript does not contain any individual person’s data, and further consent for publication is not required.

Competing interests

The authors declare that they have no competing interests.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Hinton, J.W., West, K.M., Sullivan, D.J. et al. The natural history and ecology of melanism in red wolf and coyote populations of the southeastern United States – evidence for Gloger’s rule. BMC Zool 7, 33 (2022). https://doi.org/10.1186/s40850-022-00138-5

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s40850-022-00138-5

Keywords

  • Canis latrans
  • Canis rufus
  • Coyote
  • Gloger’s rule
  • Habitat selection
  • Melanism
  • Red wolf

BMC Zoology

ISSN: 2056-3132

Contact us