Northeastern Naturalist Vol. 25, No. 2 C.M.T. Vu, D.W. Piniewski, O.A.P. McLean, and D.J. McCabe 2018 319 2018 NORTHEASTERN NATURALIST 25(2):319–332 Use of Point-and-Shoot Photography to Compare Regional Differences in Canis latrans (Coyote) Skull Size Catherine M.T. Vu1, Devan W. Piniewski1, Osrica A.P. McLean1, and Declan J. McCabe1,* Abstract - Increasing interest in digital resources for zoological study have resulted in the creation of several online collections of specimens with varying degrees of complexity and sophistication. We illustrate how an inexpensively created archive of digital photographs can be used to test hypotheses of general interest to evolutionary biologists. We used conventional and digital measurements from Canis latrans (Coyote) crania, to show that northeastern Coyote skulls are larger than skulls in our collection from all other areas. Furthermore, we reject Bergmann’s rule by demonstrating that Coyote skulls from Texas are larger than skulls from Alaska and Washington. Measurements taken from calibrated digital photographs support conclusions drawn using conventional hands-on measurements from the same specimens. We employed simple point-and-shoot photography to make all images used. We have placed our images on a wiki platform without copyright restrictions; thus, they are available for use in any manner, and the digital archive can be expanded by others. We employed Bland–Altman plots to demonstrate an approach to image-quality control that can be employed to identify and replace images that could lead to erroneous measurements. We describe how digital archives shared in this manner could aggregate data from isolated specimens and small collections to make these otherwise obscure specimens available to the broader scientific community and the general public. Introduction Museum specimens are being digitized for educational and research applications (Ang et al. 2013, Blagoderov et al. 2012, Digimorph 2017, Vollmar et al. 2010). Numerous authors have called for expanded availability of digitally archived museum specimens (Ang et al. 2013, Balke et al. 2013, Blagoderov et al. 2012, Heerlien et al. 2015). Broader access to digital collections will democratize science by making images of rare and delicate specimens available to scientists and teachers (Maschner and Schou 2013) to further enhance the value of museum collections in a time of vanishing biodiversity (Beaman and Cellinese 2012). Carefully photographed images with scale bars can yield quantitative data for comparative study (Takahashi et al. 2006). Calibrated specimen-images require tests to confirm accuracy, ultimately leading to the refinement of photographic techniques (Johnson et al. 2013). Simplicity of image capture will facilitate crowdsourced digitization of museum collections (Blagoderov et al. 2012) and permit the addition of privately held specimens to digital archives. 1Biology Box 283, Saint Michael’s College, One Winooski Park, Colchester, VT 05439. *Corresponding author - dmccabe@smcvt.edu. Manuscript Editor: John Litvaitis Northeastern Naturalist 320 C.M.T. Vu, D.W. Piniewski, O.A.P. McLean, and D.J. McCabe 2018 Vol. 25, No. 2 Saint Michael's College (Colchester, VT) students assembled a digital museum collection of C. latrans Say (Coyote) skulls, as a resource for researchers, educators, and students (http://wikieducator.org/Digital_Coyote; McCabe and Vu 2014). Here, we demonstrate that such a collection can yield data with 95% accuracy in gross- and fine-scale measurements. The photographic archive is suitable for research applications; we demonstrate this attribute by showing that Northeastern Coyote skulls are larger than Western Coyote skulls (Thurber and Peterson 1991) and by testing Bergmann’s (1847) rule. Bergmann (1847) predicted increasing size of homoeothermic animals with latitude (Blackburn et al. 1999, Rensch 1938). While Bergmann’s rule and its thermoregulatory mechanism (as translated by James 1970) has been questioned (Ashton et al. 2000, Geist 1987), patterns consistent with the rule are common (Ashton et al. 2000, James 1970, Meiri and Dayan 2003). Whether Bergmann’s rule applies to Coyotes remains controversial (Meachen and Samuels 2012, Meiri et al. 2004). Western Coyotes colonized southern Canada and the northeastern US in the early 20th century, reaching New England and New York by the 1930s (Parker 1995). Eastern Coyotes are larger than their western counterparts (Parker 1995, Thurber and Peterson 1991, Way 2013), and proposed mechanisms for this difference include responses to prey and hybridization with Canis lupus L. (Wolf; Kays et al. 2010, Thurber and Peterson 1991). Coyote size differences are reflected in skull measurements (Elbroch 2006, Kays et al. 2010), and museum specimens have been used to test Bergmann’s rule (Meiri et al. 2004); thus, skulls represent an ideal system to test measurement techniques. We collected geographically diverse Coyote skulls to test Bergmann’s rule and compared northeastern and western Coyote skulls, using photographic and caliper measurements. Methods Canis latrans skull collection We obtained Coyote skulls between 2011 and 2014 from trappers, taxidermists, farmers, and pest-control companies. The collection currently includes 97 Coyote skulls from a range of locations including Alaska, Washington, Texas, and the northeastern region (Connecticut, Maine, Massachusetts, Vermont, and New York) of the US. It is important to note that the age and gender of these specimens is unknown. It is well established that both age and gender influence skull size in canids (Elbroch 2006) and that male Coyotes are larger than females (Way 2007). We have no reason to believe that gender or age is in any way geographically biased in our sample but must acknowledge that variance introduced by these factors reduces statistical power to detect patterns. To address age-based size bias, we have complemented gross cranial measurements with measurements of the canine teeth. Tooth size is established prior to tooth eruption, correlates with body size, and does not change with animal maturation (Koch 1986). We also ranked Coyotes by maturity based on closure of basicranial sutures. Specifically, we Northeastern Naturalist Vol. 25, No. 2 C.M.T. Vu, D.W. Piniewski, O.A.P. McLean, and D.J. McCabe 2018 321 scored the basispheno-presphenoid, basispheno-basioccipital, and pterygopalate sutures as open, closing, or closed (Geiger and Haussman 2016) to establish 4 age categories: all sutures open (youngest), 2 of 3 sutures open, 2 of 3 sutures closed, and all sutures closed (oldest). Photography and measurements We used a Canon EOS-10D DSLR with a Quantaray zoom lens or a pointand- shoot Canon PowerShot A590IS to obtain images. We did not systematically compare results between cameras. We mounted the cameras on a copy stand 80 cm from the skulls to reduce perspective distortion (Takahashi et al. 2006) and we used the zoom features to crop the images and include the skull and a ruler for scale. We used modeling clay and cardboard to cradle each skull with the sagittal or palatine plane horizontal and to provide a platform at mid skull depth for the scale (Takahashi et al. 2006). We draped the cradle with black fabric before adding the skull and scale. We used larger f-stop numbers corresponding to smaller apertures to increase depth of field. We photographed superior, inferior, left, and right anatomical views of each skull; inferior views were taken without the mandibles. We measured condylobasal length (CB), from the most anterior point of the rostrum (excluding teeth) to the anterior edge of the occipital condyle, and greatest length (GL) from the most anterior point of the rostrum (excluding teeth) to the most posterior point of the sagittal crest on the interparietal bone of each skull (Elbroch 2006). Skulls missing relevant structures were excluded from these analyses. We measured canine-tooth height (CTH) and mesiodistal tooth width (MTW) of an upper canine tooth from each skull. We made these measurements from the left canine tooth whenever possible. In rare cases, when the left canine was damaged or missing, we used the right canine. We excluded from this portion of analyses any skulls missing both upper canines. We obtained all measurements using traditional methods and then repeated them using high-resolution (3072 x 2048 pixels) digital photographs. For GL, we used aluminum sculptor’s calipers referenced to standard metric rulers. We used stainless steel digital calipers to measure teeth. Measurements from digital photographs were taken using ImageJ (Schneider et al. 2012). We carried out all analyses and graphing using measurements from left-side photographic images, with the exception of a small number of cases, when only right teeth were available. We downloaded all photographs from the digital Coyote archive and copied them into an ImageJ window; the software was recalibrated for each image using the photographed scale that accompanies each skull view. Analysis We confirmed normality of all measured datasets within categories (Northeast, West, Texas, Alaska, Washington) using Shapiro–Wilks tests before proceeding with standard parametric statistical tests. We quantified overall variability of measurements using coefficients of variability. Analyses were performed on CB, GL, CTH, and MTW measured by caliper and by photograph. We compared skulls from Northeastern Naturalist 322 C.M.T. Vu, D.W. Piniewski, O.A.P. McLean, and D.J. McCabe 2018 Vol. 25, No. 2 the Northeast (Connecticut, Maine, Massachusetts, Vermont, and New York) to skulls from the West (Washington State and Alaska) using two-tailed Student’s ttests assuming equal variances, and skulls from Alaska, Washington, and Texas via one-way analysis of variance. To test Bergmann’s Rule, we compared skulls from Texas with skulls from Alaska using two-tailed Student’s t-tests assuming equal variances. We expressed differences between compared groups from all t-tests as standardized effect size (SES; Cohen’s d [Cohen 1988]) and measured the 95% confidence interval of SES using Wuensch’s (2011) modification of Smithson’s (2011) SPSS scripts. Cohen’s d is calculated as follows: d = (X̅1 - X̅2) / s where X̅1 and X̅2 are the group means and s is the standard deviation. Cohen (1988) did not specify which standard deviation to use. The group variances were quite homogenous in all cases; thus, we followed Cumming’s (2013) recommendation and used the pooled-sample standard deviation. We ran linear regressions to determine how well photographic measurements replicated caliper measurements. To check for size-related and/or systematic differences between the techniques, we generated Bland–Altman (Tukey mean difference) plots. Each data point on these plots represented the difference in measurements of a structure between photographic and caliper techniques and deviations from zero indicated disagreement between techniques. We used Student’s t-tests to compare measurements made using the different techniques. We made 2 x 4 (Northeast/West versus age rank) and 3 x 4 (TX/WA/AK versus age rank) contingency tables and used chi-square analysis to evaluate the null hypothesis that skulls of different ages were evenly distributed among the locations of interest. Digital archive We uploaded all images to Wikimedia Commons under an Attribution-Share Alike Creative Commons agreement (CC-BY-SA-3.0). We linked sets of images associated with particular skulls together on pages in Wikieducator (http://wikieducator. org/Digital_Coyote). Thumbnails images link to low-resolution (800 x 533 pixels) images that, in turn, link to high-resolution images (3072 × 2048 pixels). Specimen pages include collection location and approximate date of collection. We provided notes on calibration and accuracy. When photographs yielded inaccurate measurements, we rephotographed the specimens in question and provided updated images. It is important to note that occasionally, photographs with shadows, glare, or with slightly out-of-focus rulers yielded the most accurate measurements and were retained over more aesthetically pleasing images. Ongoing quality control continues as specimens are added and less-accurate images are replaced. Each skull in the archive is tagged using wiki categories to gather the specimens into useful taxonomic and geographical groups. Northeastern Naturalist Vol. 25, No. 2 C.M.T. Vu, D.W. Piniewski, O.A.P. McLean, and D.J. McCabe 2018 323 Results All measured variables (GL, CB, CTH, and MTW) were normally distributed within the compared groups (Alaska, Texas, Washington; Northeast, West; Shapiro– Wilks test; P ≥ 0.22 in all cases). Coefficients of variability ranged between 0.06 and 0.08. Northeastern skulls and teeth were significantly larger than western specimens for all measured traits except tooth height measured by caliper (t values for photographic/caliper measurements: GL: 3.73/4.22, CB: 4.44/6.11, CTH: 1.95/3.41, and MTW: 2.98/3.82; P < 0.01 in all cases except tooth height by caliper measurement: P = 0.057; Fig. 1). One-way ANOVA did not detect differences in the greatest length of skulls from Alaska, Washington, and Texas (F = 1.36; P = 0.26; Fig. 2). When measured from photographs, mesiodistal tooth width (MTW) was larger in skulls from Texas than those from Alaska (t = 2.49; P = 0.019; Fig. 3) as was condylobasal length (CB; t = 2.08, P = 0.046; caliper: t = 1.5, P = 0.029). Other variables (GL, CB, and CTH) did not differ between Alaskan and Texan skulls using standard inferential statistics, regardless of measurement technique (t values for caliper measurements: CB: 1.51; GL: 0.55, CTH: 0.62, and MTW: 1.85, and for photographic: GL: 1.54 CTH: 1.31; P > 0.05 in all cases; Fig. 3). Figure 1. Condylobasal length of northeastern and western Canis latrans skulls measured from photographs using ImageJ and directly from specimens using calipers. Northeastern specimens are from 5 northeastern states; western specimens are from Alaska and Washington. Horizontal bars represent medians. Northeastern Naturalist 324 C.M.T. Vu, D.W. Piniewski, O.A.P. McLean, and D.J. McCabe 2018 Vol. 25, No. 2 Figure 2. Condylobasal length of Canis latrans skulls from Texas, Washington, and Alaska measured from photographs using ImageJ and directly from specimens using calipers. Horizontal bars represent medians. All measurements from northeastern skulls, other than CTH measured using calipers, were greater than those from western specimens, with large standardized effect-sizes (Cohen’s d) exceeding 0.8 standard deviations (Fig. 3). CB measurements of skulls from Texas were greater than those from Washington with a large effect-size of 1.1 standard deviations. Skulls from the Texas were also larger than those from Alaska (CB; Cohen’s d = 0.75; medium effect-size). With the exception of MTW measured from photographs, the differences between Texan and Alaskan Coyote skulls were not significant using standard inferential statistics and confirmed by the fact that the 95% CI of Cohen’s d overlapped the zero-difference line (Fig. 3). Measurements made using calipers and digital photographs were tightly correlated (linear regression; P < 0.001; R2 = 0.84 for GL and MTW; Figs. 4a and 4b). The least-consistent measurement between techniques was tooth height (P less than 0.001; R2 = 0.77; Fig. 4c). Bland–Altman plots showed that differences between techniques did not change with the size of the specimen measured (Fig. 5). Greatest lengths of skull measurements from photographs were an average of 3 mm less than GL measurements using calipers; this difference was statistically significant (t = 2.07; P = 0.04; Fig. 5a). We detected no differences between capliper or photographic measurements of CTH nor MTW (t = 0.082 and 0.401; P = 0.93 and 0.67, respectively; Fig. 5b, c). Results of our chi-square analysis confirmed that skulls Northeastern Naturalist Vol. 25, No. 2 C.M.T. Vu, D.W. Piniewski, O.A.P. McLean, and D.J. McCabe 2018 325 in age categories are distributed in an unbiased manner among the geographical locales represented in our collection (northeast vs. west comparison: χ2 = 0.87, P = 0.83; AK, TX, and WA comparison: χ2 = 10.93, P = 0.09). Discussion Our finding that Coyote skulls from the Northeast are larger than western skulls is consistent with recent literature (Gompper 2002; Way 2007, 2013). Bekoff and Gese (2003) listed Coyote weights noted in 13 studies from 11 localities; the largest Coyotes were found in Minnesota and Oklahoma, with Coyotes from Quebec and Kansas 3rd and 4th largest, respectively. However, published eastern Coyote weights exceeding those in Bekoff and Gese (2003) have been reported for New Hampshire (Silver and Silver 1969), Maine (Richens and Hugie 1974), and the Northeast (Moore and Millar 1986). Parker (1995) qualified his assessment of northeastern Coyote size, and Hilton (1978) cautioned against overestimation. Way’s (2007) study included Coyote weights from 33 sites spanning North America, and provided evidence that northeastern Coyotes were larger than Coyotes in their western range. Gompper (2002) found that southern Coyote sizes were comparable to those in the Northeast; this finding is consistent with Alabama Coyote skulls we acquired measuring 203.9 mm and 189 mm, thus straddling the average of northeastern skulls (Fig. 1). By showing that Coyote skulls form the Northeast are larger than Figure 3. Standardized effect-sizes (Cohen’s d) of pairwise differences between northeastern and western skulls, and skulls from Texas compared to skulls from Alaska and from Washington. Horizontal error bars are 95% confidence intervals of d. Effects sizes in the pale grey are considered small; dark grey equates to moderate effects, and large effects are greater than 0.8 standard deviations and are beyond the dark grey. Northeastern Naturalist 326 C.M.T. Vu, D.W. Piniewski, O.A.P. McLean, and D.J. McCabe 2018 Vol. 25, No. 2 those from the West (P < 0.05; Fig. 1), our data support the conclusion of Way (2007, 2013) and Kays et al. (2010) that northeastern Coyotes or “coywolves” are larger than western Coyotes. That Bergmann’s rule has been declared invalid (Geist 1987) has not diminished interest in the topic (Feldman and Meiri 2014, McCoy 2012, Penniket and Cree 2015, Teplitsky and Millien 2014). In Coyotes, the rule has been supported (Meiri et al. 2004, Murray and Larivière 2002) and rejected (Meachen and Samuels 2012, Rosenzweig 1968, Thurber and Peterson 1991). Inferential statistics did not detect skull-size differences between Texas, Washington, and Alaska with our sample sizes (n = 18, 17, and 12, respectively). However, standardized effect-size analyses (Cohen’s d) of size differences between Texas and Alaska showed small to large effects opposite to the predictions of Bergmann’s rule, and running contrary to Murray and Larivière’s (2002) conclusion that Coyote size increased with latitude. The trends we detected are consistent with earlier cranial data (Young and Jackson Figure 4. Comparison of photographic and caliper measurements of (A) greatest length of skull, (B) mesiodistal canine-tooth width, and (C) canine-tooth height in Canis latrans; n = 87. Dashed lines represent caliper measurement plus or minus 5%. Northeastern Naturalist Vol. 25, No. 2 C.M.T. Vu, D.W. Piniewski, O.A.P. McLean, and D.J. McCabe 2018 327 1951). Importantly, Coyotes colonized Alaska in the early 20th century (Young and Jackson 1951); thus, there has been little time for selection to act. Additionally, Alaskan Wolves may represent a negative selective force on Coyote size. Figure 5. Bland–Altman (Tukey mean-difference) plots examining differences between caliper and photographic measurements. The sizes of differences between techniques are represented on the Y axis. Mean differences are represented by the dashed lines and the dotted lines represent 95% confidence limits of the overall means. (A) greatest length of skull, (B) mesiodistal canine-tooth width, and (C) canine-tooth height in Canis latrans; n = 87. Northeastern Naturalist 328 C.M.T. Vu, D.W. Piniewski, O.A.P. McLean, and D.J. McCabe 2018 Vol. 25, No. 2 Koch (1986) suggested measuring teeth as a correlate of body size because tooth size is established before eruption and is independent of developmental stage. Animal ages in our collection are unknown; thus, tooth measurements are a particularly valuable index of the relative sizes of animals from different populations. Tooth measurements confirmed the large Northeast–West cranial-size difference, and the smaller differences between Texas and Alaska, and χ2 analysis of our suture-closure data support a lack of age-based bias among the regions in our study. Photographic collections are more useful if measurements can be taken from the specimens. We tested measurements across a range of scales from greatest length of skull (mean = 190 mm) to mesiodistal canine tooth width (mean = 9 mm). We calibrated photographs and adjusted values based on skull length, so it was unsurprising that we could usually achieve 5% measurement accuracy for that metric. Our accuracy in measuring tooth width and skull length was similar with the only obvious systematic error being a slight (1%, or 3-mm) underestimate of skull length. We attribute lack of precision in tooth-height measurement (Fig. 4C) to subjectivity and judgements made in determining where exactly to measure when skulls had damage and bone loss around tooth sockets. Importantly, none of the measurement differences between the techniques substantively altered the study conclusions. Validity of future digital collections will require similar quality-control testing (Johnson et al. 2013). Expansive research collections are beyond the reach of many investigators and educators, including those at small Vermont undergraduate colleges. Our physical collection currently includes 144 specimens from a small number of species. Meiri et al. (2004) utilized skulls from 29 collections in 14 countries. The Idaho Museum of Natural History was not part of Meiri et al. (2004) study, but that collection is being digitized and could be part of future studies (Maschner and Schou 2013). Digital collections lower barriers to research and educational opportunities and can revolutionize zoological research (Betts et al. 2011). Virtual museum collections consist of simple digital photographs (McCabe and Vu 2014), complex CAT-scanned cross-sections (Digimorph 2017), or 3-D scanned models with measurement-enabled online software (Betts et al. 2011). By using photographic images and free software (Schneider et al. 2012), our archive illustrates how just a camera and internet access can create a community resource. We utilized Wikimedia Commons to host images and Wikieducator as our user interface to encourage contributions from others. Both sites use Creative Commons licensing to manage copyright during upload, and thus guarantee unrestricted access. We incorporated images contributed by Will Higgs (Higgs 2012) as our proof of concept. The contributor provided caliper measurements and we digitally adjusted the scale to match the caliper measurements of the skull to eliminate the need to rephotograph specimens. Post-photography digital-scale recalibration can simplify addition of off-site collections. The value of museum collections is beyond doubt (Suarez and Tsutsui 2004). The importance of museums for biodiversity assessment (Ponder et al. 2001), conservation biology (Drew 2011), and as a source of historic DNA samples (Besnard et al. 2015) is also clear. Small collections are valuable in aggregate, but are hard to locate, at higher risk of loss, and rarely included in published studies (Casas-Marce Northeastern Naturalist Vol. 25, No. 2 C.M.T. Vu, D.W. Piniewski, O.A.P. McLean, and D.J. McCabe 2018 329 et al. 2012). Digital collections can combine information from small, scattered collections into accessible metacollections (Balke et al. 2013). Our study illustrates the application of digital collections to questions of scientific importance. This collection was assembled by undergraduate students, demonstrating that with brief training, and inexpensive equipment, accurately calibrated images can be created. We have shown that photographic measurements concur with traditional methods. Our conclusions that northeastern Coyotes are larger than their western counterparts and that Bergmann’s rule does not apply to Coyotes support previously published studies. Digital archives can aggregate scattered specimens and contribute to broader research and educational efforts. Acknowledgements This work was supported by indirect funding provided by the Office of the Vice President of Academic Affairs at Saint Michael’s College to the Biology Department of Saint Michael’s College. We wish to thank Dr. Karen Talentino for her ongoing support of undergraduate research at Saint Michael’s College. In addition to photographs taken by the authors, several images were captured by Michael Gordon and Sarah Burridge, and an additional set was contributed by Will Higgs. We are grateful to Dr. Kristian Omland for his suggestions regarding the Bland–Altman plots, and to Heather McCabe for help with the figures. D. McCabe’s research is supported by Vermont EPSCoR’s Grant NSF EPS Award #1556770 from the National Science Foundation. We thank 2 anonymous reviewers, whose contributions improved this manuscript. Literature Cited Ang, Y., J. Puniamoorthy, A.C. Pont, M. Bartak, W.U. Blanckenhorn, W.G. Eberhard, N. Puniamoorthy, V.C. Silva, L. Munari, and R. Meier. 2013. A plea for digital reference collections and other science-based digitization initiatives in taxonomy: Sepsidnet as exemplar. Systematic Entomology 38:637–644. Ashton, K.G., M.C. Tracy, and A. de Queiroz. 2000. Is Bergmann’s rule valid for mammals? The American Naturalist 156:390–415. Balke, M., S. Schmidt, A. Hausmann, E. Toussaint, J. Bergsten, M. Buffington, C.L. Häuser, A. Kroupa, G. Hagedorn, A. Riedel, A. Polaszek, R. Ubaidillah, L. Krogmann, A. Zwick, M. Fikáček, J. Hájek, M.C. Michat, C. Dietrich, J. La Salle, B. Mantle, P. KL Ng, and D. Hobern. 2013. Biodiversity into your hands: A call for a virtual global natural history “metacollection”. Frontiers in Zoology 10:1–9. Beaman, R.S., and N. Cellinese. 2012. Mass digitization of scientific collections: New opportunities to transform the use of biological specimens and underwrite biodiversity science. ZooKeys 209:7–17. Bekoff, M., and E.M. Gese. 2003. Coyote (Canis latrans). Pp. 467–481, In G.A. Feldhamer and B.C. Thompson (Eds.). Wild Mammals of North America: Biology, Management, and Conservation. Johns Hopkins University Press, Baltimore, MD. 481 pp. Bergmann, C. 1847. Über die Verhältnisse der Wärmeökonomie der Thiere zu ihrer Grösse. Göttinger Studien 3:595–708. Besnard, G., J.A. Bertrand, B. Delahaie, Y.X. Bourgeois, E. Lhuillier, and C. Thébaud. 2015. Valuing museum specimens: High-throughput DNA sequencing on historical collections of New Guinea Crowned Pigeons (Goura). Biological Journal of the Linnean Northeastern Naturalist 330 C.M.T. Vu, D.W. Piniewski, O.A.P. McLean, and D.J. McCabe 2018 Vol. 25, No. 2 Society 117:71–82. Betts, M.W., H.D. Maschner, C.D. Schou, R. Schlader, J. Holmes, N. Clement, and M. Smuin. 2011. Virtual zooarchaeology: Building a web-based reference collection of northern vertebrates for archaeofaunal research and education. Journal of Archaeological Science 38:755–762. Blackburn, T.M., K.J. Gaston, and N. Loder. 1999. Geographic gradients in body size: A clarification of Bergmann’s rule. Diversity and Distributions 5:165–174. Blagoderov, V., I.J. Kitching, L. Livermore, T.J. Simonsen, and V.S. Smith. 2012. No specimen left behind: Industrial-scale digitization of natural history collections. ZooKeys 209:133–146. Casas-Marce, M., E. Revilla, M. Fernandes, A. Rodríguez, M. Delibes, and J.A. Godoy. 2012. The value of hidden scientific resources: Preserved animal specimens from private collections and small museums. BioScience 62:1077–1082. Cohen, J. 1988. Statistical power analysis for the behavioral sciences. Routledge Academic, Hillsdale, NJ. 490 pp. Cumming, G. 2013. Understanding the New Statistics: Effect Sizes, Confidence Intervals, and Meta-Analysis. Routledge, New York, NY. 519 pp. Digimorph. 2017. Digital morphology: A National Science Foundation digital library at The University of Texas at Austin. The High-resolution X-ray-computed Tomography Facility at The University of Texas at Austin, Austin, TX. Available online at http:// digimorph.org/. Accessed 15 December 2017. Drew, J. 2011. The role of natural history institutions and bioinformatics in conservation biology. Conservation Biology 25:1250–1252. Elbroch, M. 2006. Animal Skulls: A Guide to North American Species. Stackpole Books, Mechanicsburg, PA. 740 pp. Feldman, A., and S. Meiri. 2014. Australian snakes do not follow Bergmann’s rule. Evolutionary Biology 41:327–335. Geiger, M., and S. Haussman. 2016. Cranial suture closure in Domestic Dog breeds and its relationships to skull morphology. The Anatomical Record 299:412–420. Geist, V. 1987. Bergmann’s rule is invalid. Canadian Journal of Zoology 65:1035–1038. Gompper, M.E. 2002. The ecology of northeast Coyotes. Wildlife Conservation Society 17:1–47. Heerlien, M., J. Van Leusen, S. Schnörr, S. De Jong-Kole, N. Raes, and K. Van Hulsen. 2015. The natural history production-line: An industrial approach to the digitization of scientific collections. Journal on Computing and Cultural Heritage (JOCCH) 8:289–294. Higgs, W. 2012. Will’s skull page. Available online at http://www.skullsite.co.uk/. Accessed 6 October 2017. Hilton, H. 1978. Systematics and ecology of the eastern Coyote. Pp. 209–228, In M. Bekoff (Ed.) Coyotes: Biology, Behavior, and Management. Academic Press, New York, NY. 384 pp. James, F.C. 1970. Geographic size-variation in birds and its relationship to climate. Ecology 51:365–390. Johnson, L., B.L. Mantle, J.L. Gardner, and P.R. Backwell. 2013. Morphometric measurements of dragonfly wings: The accuracy of pinned, scanned, and detached measurement methods. ZooKeys 276:77–84. Kays, R., A. Curtis, and J.J. Kirchman. 2010. Rapid adaptive evolution of northeastern Coyotes via hybridization with wolves. Biology Letters 6:89–93. Koch, P.L. 1986. Clinal geographic variation in mammals: Implications for the study of Northeastern Naturalist Vol. 25, No. 2 C.M.T. Vu, D.W. Piniewski, O.A.P. McLean, and D.J. McCabe 2018 331 chronoclines. Paleobiology 12:269–281. Maschner, H.D., C.D. Schou, and J. Holmes. 2013. Virtualization and the democratization of science: 3-D technologies revolutionize museum research and access. Pp. 265, In Digital Heritage International Congress. DigitalHeritage DOI:10.1109/ DigitalHeritage.2013.6744763. McCabe, D.J., and C.M. Vu. 2014. Digital Coyote: Examining geographical variation using a virtual museum collection. Tested Studies for Laboratory Teaching. Proceedings of the Association for Biology Laboratory Education 35:369–375. McCoy, D.E. 2012. Connecticut birds and climate Change: Bergmann’s rule in the 4th dimension. Northeastern Naturalist 19:323–334. McNab, B.K. 2010. Geographic and temporal correlations of mammalian size reconsidered: A resource rule. Oecologia 164:13–23. Meachen, J.A., and J.X. Samuels. 2012. Evolution in Coyotes (Canis latrans) in response to the megafaunal extinctions. Proceedings of the National Academy of Sciences 109:4191–4196. Meiri, S., and T. Dayan. 2003. On the validity of Bergmann’s rule. Journal of Biogeography 30:331–351. Meiri, S., T. Dayan, and D. Simberloff. 2004. Carnivores, biases, and Bergmann’s rule. Biological Journal of the Linnean Society 81:579–588. Moore, G.C., and J.S. Millar. 1984. A comparative study of colonizing and longer-established eastern Coyote populations. The Journal of Wildlife Management 48:691–699. Murray, D., and S. Larivière. 2002. The relationship between foot size of wild canids and regional snow conditions: Evidence for selection against a high footload? Journal of Zoology 256:289–299. Parker, G. 1995. Eastern Coyote: The Story of Its Success. Nimbus Publishing, Halifax, NS, Canada. 264 pp. Penniket, S., and A. Cree. 2015. Adherence to Bergmann’s rule by lizards may depend on thermoregulatory mode: Support from a nocturnal gecko. Oecologia 178:427–440. Ponder, W., G. Carter, P. Flemons, and R. Chapman. 2001. Evaluation of museum-collection data for use in biodiversity assessment. Conservation Biology 15:648–657. Rensch, B. 1938. Some problems of geographical variation and species formation. Proceedings of the Linnean Society of London, Wiley Online Library 150:275–285. Richens, V.B., and R.D. Hugie. 1974. Distribution, taxonomic status, and characteristics of Coyotes in Maine. Journal of Wildlife Management 38:447–454. Rosenzweig, M.L. 1968. The strategy of body size in mammalian carnivores. American Midland Naturalist 80:299–315. Schmitz, O., and D. Lavigne. 1987. Factors affecting body size in sympatric Ontario Canis. Journal of Mammalogy 68:92–99. Schneider, C.A., W.S. Rasband, and K.W. Eliceiri. 2012. NIH Image to ImageJ: 25 years of image analysis. Nature Methods 9:671–675. Silver, H., and W.T. Silver. 1969. Growth and behavior of the Coyote-like canid of northern New England with observations on canid hybrids. Wildlife Monographs 17:3–41. Smithson, M.J. 2011. Workshop on: Noncentral confidence intervals and power analysis. Page SPSS scripts for statistical calculations. Available online at http://psychology3. anu.edu.au/people/smithson/details/CIstuff/CI.html. Accessed 24 March 2011. Suarez, A.V., and N.D. Tsutsui. 2004. The value of museum collections for research and society. BioScience 54:66–74. Takahashi, H., M. Yamashita, and N. Shigehara. 2006. Cranial photographs of mammals on the web: The Mammalian Crania Photographic Archive (MCPA2) and a comparison of Northeastern Naturalist 332 C.M.T. Vu, D.W. Piniewski, O.A.P. McLean, and D.J. McCabe 2018 Vol. 25, No. 2 bone-image databases. Anthropological Science 114:217–222. Teplitsky, C., and V. Millien. 2014. Climate warming and Bergmann’s rule through time: Is there any evidence? Evolutionary Applications 7:156–168. Thurber, J.M., and R.O. Peterson. 1991. Changes in body size associated with range expansion in the Coyote (Canis latrans). Journal of Mammalogy 72:750–755. Vollmar, A., J.A. Macklin, and L. Ford. 2010. Natural-history specimen digitization: Challenges and concerns. Biodiversity Informatics 7:93–112. Way, J.G. 2007. A comparison of body mass of Canis latrans (Coyote) between eastern and western North America. Northeastern Naturalist 14:111–124. Way, J.G. 2013. Taxonomic implications of morphological and genetic differences in northeastern Coyotes (Coywolves) (Canis latrans × C. lycaon), western Coyotes (C. latrans), and eastern wolves (C. lycaon or C. lupus lycaon). Canadian Field-Naturalist 127:1–16. Wikimedia Commons. 2017. Available online at https://commons.wikimedia.org. Accessed 15 December 2017. Wikieducator. 2017. Available online at https://wikieducator.org. Accessed 15 December 2017. Wuensch, K.L. 2011. Using SPSS to obtain a confidence interval for Cohen’s d. Available online at http://core.ecu.edu/psyc/wuenschk/SPSS/CI-d-SPSS.pdf. Accessed 12 August 2016. Young, S.P., and H.H. Jackson. 1951. Clever Coyote. Stackpole Books and Wildlife Management Institute, Harrisburg, PA. 411 pp.