VARIATION IN METATARSAL MORPHOLOGY AMONG SUBGROUPS OF NORTH AMERICAN MOOSE (Alces alces) William J. Silvia1, Rolf O. Peterson2, John A. Vucetich2, William F. Silvia1, and Alexander W. Silvia1 1Department of Animal and Food Sciences, University of Kentucky, Lexington, Kentucky, 40546-0215; 2School of Forest Resources and Environmental Science, Michigan Technological University, Houghton, Michigan, 49931 ABSTRACT: The objectives of this study were to characterize variation in dimensional data from the metatarsus of 4 different subpopulations of North American moose (Alces alces) that are known to dif- fer in stature, and to determine if specific metatarsal width measurements (proximal, middle, distal) can be used to accurately predict metatarsal length in these subpopulations. We found that subpopula- tions differ in the dimensions of their metatarsal bones. Alaskan moose (A. a. gigas) are significantly larger in the length and width of the metatarsus than non-Alaskan moose. Moose from Isle Royale have significantly shorter metatarsal bones than the other groups which is associated with a propor- tional reduction in the middle metatarsal width; the ratio of middle width:length was similar across groups in contrast to the proximal: and distal width:length ratios. These dimensions were not reduced proportionally in Isle Royale specimens as these ratios were greater in the Isle Royale moose than in other groups. Predictive equations for estimating metatarsal length from each of the 3 width measure- ments were developed. The length could be predicted accurately from each of the width measurements if separate predictive equations were developed for specimens collected from Isle Royale versus the other subgroups. These data indicate that considerable variation exists in the dimensions of a single bone, the metatarsus, in subgroups of the same species. Valid predictive equations developed using data sets from one subgroup may not provide accurate predictions when applied to other subgroups of the same species. ALCES VOL. 50: 159–170 (2014) Key Words: Alces, metatarsus, moose, morphology, variation Estimating the body size of individuals is an important part of any population assess- ment. Direct measures (e.g., shoulder height, heart girth, body weight) of large species are often difficult to obtain in the field, and esti- mates of body size are often made from extrapolations of other body parts. Foot length is correlated with live or carcass weight in many ungulate species (Bandy et al. 1956, McEwan and Wood 1966, Rose- berry and Klimstra 1975, Martin et al. 2013) including moose (Alces alces) (Franzmann et al. 1978, Lynch et al. 1995, Jensen et al. 2013). For ungulates, both living and recently deceased, this is most often mea- sured along the plantar surface from the cal- caneal protuberance to the tip of the longest toe. For animal remains that are collected after significant decomposition, it may be more convenient and consistent to measure the length of the metatarsus itself, commonly referred to as the cannon bone. The length of the metatarsus is correlated with body size across mammalian species (McMahon 1975, Alexander et al. 1979). For example, the length of the metatarsus is correlated with body weight and growth rate in cattle (Coble et al. 1971b), and length and width William J. Silvia, Department of Animal and Food Sciences, 409 WP Garrigus Bldg, University of Kentucky, Lexington, Kentucky 40546-0215 159 of the metatarsus were smaller in female than male cattle (Coble et al. 1971a), a clear indi- cation of sexual dimorphism. The length of the metatarsus is an excellent indicator of fetal age in sheep (Santucci et al. 1993), the length of the metatarsus in growing lambs is directly related to maternal nutrition dur- ing gestation (Pálsson and Vergés 1952), and the heritability of metatarsal dimensions is relatively high (Coble et al. 1971a). The length of the metatarsus has been used as an indirect measure of body size in moose (Alces alces; Peterson 1977). In the field, it is quite common to find metatarsal bones from moose that have been broken or damaged in such a way that an accurate length cannot be determined. However, por- tions of the metatarsus are often intact per- mitting accurate measurement of the width at some point along the length of the bone. Recognizing that metatarsal dimensions are of great utility in field research with moose and that there is considerable size variation among subpopulations of moose, our first objective was to characterize the length and 3 specific width measurements of metatarsal bones collected from 4 groups of moose: 1) Isle Royale National Park (subspecies undetermined, either A. a. americana or A. a. andersoni), 2) extant Alaskan moose (subspecies A. a. gigas), 3) fossilized Alas- kan moose (subspecies undetermined), and 4) mainland, excluding Alaska (includes subspecies A. a. americana, A. a. andersoni, A. a. shiras). Our second objective was to determine if specific metatarsal widths (proximal, middle, distal) can be used to accurately predict metatarsal length of North American moose, and to determine if the relationships between length and specific widths vary among the 4 subgroups. METHODS Quantitative measurements of metatarsal morphology of adult North American moose were made on 4 subgroups. The first sub- group consisted of 420 moose from Isle Roy- ale National Park (48°06’N, 88°30’W; Peter- son 1977) located in Lake Superior approximately 30 km from the Ontario, Canada coastline. The precise origin of moose on Isle Royale is unknown, but the founding animals could be either A. a. amer- icana orA. a. andersoni subspecies; however, Isle Royale moose are morphologically dif- ferent from both subspecies (Peterson et al. 2011). These metatarsal specimens are cur- rently housed at Michigan Technological University’s (MTU) Ford Center in Alberta, Michigan. The second group of specimens was from 170 modern Alaskan moose and included specimens housed at 1) the Museum of the North, University of Alaska, Fairbanks, Alaska (collected from Denali National Park; 63°20’ N, 150°30’ W; n = 65), 2) the MTU Ford Center (collected in the Kenai National Wildlife Refuge [KNWR] at 60° 20’ N, 150°30’ W; n = 95), 3) the American Museum of Natural History (AMNH), New York, New York (collected throughout Alaska; n = 6), and 4) the Field Museum of Natural History (FMNH), Chicago, Illinois (collected throughout Alaska; n = 3). The third group of 49 metatarsal bones was fossil material from the late Pleistocene age that was collected from several sites 10–35 km north of Fairbanks, Alaska (65° N, 147°40’ W) (Frick 1930, Wilkerson 1932) and was part of the Frick collection at the AMNH (n = 49); these are presumed from the subspecies A. a. gigas. The fourth set of 34 specimens, referred hereafter as mainland moose, was collected from a vari- ety of sites in Canada and the United States (excluding Alaska) and included subspecies A. a. americana, A. a. andersoni, and A. a. shiras. These specimens are housed at the 1) MTU Ford Center (collected by the Michigan and Minnesota Departments of Natural Resources (n = 20), 2) the AMNH (n = 7), 3) the FMNH (n = 1), 4) Brown 160 METATARSAL DIMENSIONS IN ALCES – SILVA ET AL. ALCES VOL. 50, 2014 University (n = 1), 5) Harvard University (n = 2), and 6) the University of Kentucky (n = 3). All specimens were from moose either killed by hunters or vehicular collisions. Specimens collected in Isle Royale National Park, Denali National Park, or the KNWR were obtained from animals that died of natural causes. On Isle Royale, the majority resulted from predation by wolves (Peterson 1977); as a result, animals that were more susceptible to predation (due to age, injury, disease) may be overrepresented. The sex of specimens was determined from examination of soft tissue (when present) and morphological characteristics of the associated skull (when present). A general age was determined by the size of the remains and the complement of deciduous and permanent teeth (Peterson et al. 1983). When necessary, tissue from the metatarsal bones was removed manually with a knife and/or by prolonged immersion in hot water (>80 °C). Quantitative measurements of the can- non bone were made using 2 sizes of manual vernier calipers. The length was measured using a 24-inch, Cen-Tech aluminum caliper (Harbor Freight Tools Inc., Camarillo, Cali- fornia, USA) that was modified by adding a vertical fence to each side, extending the height to approximately 2.2 cm (Fig. 1). The width of each metatarsus was measured at the proximal end, midpoint, and distal end with a standard 5-inch manual caliper (Helios, Germany; Fig. 2a). The width at the proximal end was measured at the widest point, typically within 1 cm of the end (Fig. 2b). The width at the distal end was also measured at the widest point, but the precise location varied; in some, it was very close to the end at the lateral and medial edges of the corresponding articular condyles, and in others it was proximal to the condyles, in the approximate location of the epiphyseal plate (Fig. 2c). These measurements were used to calculate width:length ratios for each width (i.e., proximal, middle, and dis- tal). The condition of the epiphyseal plate was classified as either unfused or fused. The unfused classification included speci- mens in which the 2 portions of the metatar- sus were separable, and specimens in which the 2 portions were not separable but a dis- tinct suture was clearly visible (Fig. 3). Spe- cimens were classified as coming from adults only if the distal epiphyseal growth plate was no longer visible. The effects of subgroup and/or sex on quantitative measurements (length and width of the metatarsus, width:length ratio) were evaluated with analysis of variance using the GLM procedure of SAS (1985). The rela- tionships between the length of the cannon bone and the 3 width measurements were evaluated with linear regression using the REG procedure of SAS (1985). The accu- racy of the regression equations in predicting metatarsal length from width measurements was evaluated with paired T-test using the MEANS procedure of SAS (1985). Fig. 1. The modified vernier caliper used to measure the length of the cannon bone. Note that vertical fences were added to each of the ‘jaws’ of the caliper to extend the height. ALCES VOL. 50, 2014 SILVA ET AL. – METATARSAL DIMENSIONS IN ALCES 161 RESULTS Metatarsal Length and Width The length of the metatarsus was differ- ent for each of the subgroups (P < 0.01; Fig. 4a). The fossil metatarsal bones from Alaskan moose were the longest, followed by those of modern Alaskan moose, main- land moose, and lastly Isle Royale moose. The width of the metatarsus at the proximal end was greater in the 2 Alaskan subgroups than in the other subgroups (P < 0.01; Fig. 4b); the Alaskan subgroups did not differ (P = 0.10), nor did the non-Alaskan sub- groups (P = 0.64). The ratio of proximal metatarsal width:metatarsal length was dif- ferent among groups (P < 0.01; Fig. 4c). Among all subgroups the ratio was highest Fig. 2. Dorsal view of the cannon bone with points of measurement indicated. Panel a shows the measurement of length (dashed line) and proximal, middle, and distal widths (vertical arrows); 2 possible points for measure of the distal width are indicated. Panel b shows a detailed view of the proximal end indicating the point of measurement more precisely. Panel c shows a more detailed view of the distal end and the 2 possible points for measurement. Fig. 3. The dorsal view of the distal end of the cannon bone from an adult (panel a) and a juvenile animal (panel b). Although not separable, the epiphyseal plate is clearly visible on the juvenile specimen (arrow). 162 METATARSAL DIMENSIONS IN ALCES – SILVA ET AL. ALCES VOL. 50, 2014 in specimens from Isle Royale (P < 0.05); the ratio among the other 3 subgroups did not differ (P > 0.70). To further examine the relationship between proximal width and length, the effect of width and subgroup on metatarsal length was determined (n = 285). Proximal width had a significant effect (P < 0.01), but subgroup did not (P = 0.06). There was an interaction between proximal width and subgroup on metatarsal length (P = 0.01). Since the ratio width:length appeared to be different for Isle Royale moose compared to the other subgroups, a second analysis was conducted without Isle Royale moose. Again, the effect of proximal width was evi- dent (P < 0.01), but not subgroup (P = 0.27) or the interaction term of proximal width and subgroup (P = 0.20), implying that the Alas- kan and mainland subgroups are similar and a different relationship exists for Isle Royale moose. Middle (n = 226) and distal width mea- surements (n = 224) were not available from the fossil specimens; therefore, com- parisons could only be made among the modern subgroups. The width of the metatar- sal at the midpoint was greater in Alaskan moose than those from non-Alaskan sub- groups (P < 0.01; Fig. 5a); this width was similar in Isle Royale and non-Alaskan sub- groups (P > 0.30). The ratio middle width: length was not different among subgroups (P = 0.44; Fig. 5b). As with the proximal metatarsal width, the distal metatarsal width was greater in the Alaskan subgroup than non-Alaskan subgroups (P < 0.01; Fig. 5c). The ratio distal metatarsal width:metatarsal length also differed among subgroups (P < 0.01; Fig. 5d). The distal width:length ratio was greater for Isle Royale moose than the other groups (P < 0.01); the other groups did not differ (P = 0.12). The effect of sex on metatarsal dimen- sions was analyzed with all specimens in which sex could be determined, which excluded the fossil subgroup. The length of the metatarsus was greater in males than females (P < 0.01; Fig. 6a). As in the first analysis, the length of the metatarsus differed among subgroups (P < 0.01), and metatarsal length was longer in males than females in all subgroups. A significant interaction Fig. 4. The effect of subgroup (Isle Royale [ISRO], mainland [non-ISRO from the lower 48 contiguous United States and Canada], Alaska, and fossil Alaska) on metatarsal length (a), proximal metatarsal width (b), and the ratio of proximal metatarsal width to metatarsal length (c). Bars with different letter superscripts are different (P < 0.05). ALCES VOL. 50, 2014 SILVA ET AL. – METATARSAL DIMENSIONS IN ALCES 163 between sex and subgroup was also found (P < 0.01). This interaction was strongest in Alaskan moose that had the longest metatar- sal length and largest difference between males and females. The effect of sex on the relationship between each of the 3 width measure- ments and metatarsal length was examined Fig. 5. The effect of subgroup (Isle Royale [ISRO], mainland [non-ISRO from the lower 48 contiguous United States and Canada] and Alaska) on middle metatarsal width (a), the ratio of proximal metatarsal width to metatarsal length (b), distal metatarsal width (c), and the ratio of distal metatarsal width to metatarsal length (d). Bars with different letter superscripts are different (P < 0.05). Fig. 6. The effect of sex and subgroup on metatarsal dimensions: length of the meta- tarsus (panel a) in which effects of sex, subgroup, and their interaction were ob- served (P < 0.01). Proximal, middle, and distal width of the metatarsus is shown in panels b, c, d. Effects of sex and subgroup on all 3 width measurements were observed (P < 0.01) but not of their interaction (P ≥ 0.09). 164 METATARSAL DIMENSIONS IN ALCES – SILVA ET AL. ALCES VOL. 50, 2014 separately in the Isle Royale and non-Isle Royale subgroups. In both cases, the effect of proximal width was evident (P < 0.01; Table 1). Sex had no effect on length that was not already accounted for by proximal width (P > 0.21). The interaction term of sex with proximal width on metatarsal length was also not significant (P > 0.20) in either the Isle Royale or non-Isle Royale subgroups. Similarly, there were no effects of sex or the interaction of sex and width on the relation- ship between middle or distal width on the length of the metatarsal (Table 1). Predictive Equations for Metatarsal Length Based on Widths A quantitative description of the rela- tionship between proximal width and metatarsal length was investigated with lin- ear regression. Separate regression analyses were conducted for the Isle Royale and non-Isle Royale subgroups using the follow- ing simple model: metatarsal length ¼ m� proximal metatarsal width þ b þ e ð1Þ where: m = slope, b = y-intercept, and e = error term. Comparison of the estimates of slope and y-intercept for the two groups (Isle Roy- ale versus non-Isle Royale) indicated sub- stantial difference (Table 2, Fig. 7). These relationships explained a high percentage of the variation in metatarsal length for Isle Royale (r2 = 0.47) and non-Isle Royale speci- mens (r2 = 0.66) (Table 2). The accuracy of the regression lines in predicting metatarsal length from proximal width was evaluated by comparing measured lengths to estimated lengths from specimens not used to derive the regression equations; specimens from both groups were included in this test. The length of metatarsal bones from both groups was more accurately predicted using the separate regression equations derived from the respective data sets (Table 3). The same analytical procedures were used to examine the relationships between middle width and metatarsal length, and dis- tal width and metatarsal length; middle and distal widths were not available from fossil Alaskan moose. There was no effect of sub- group or the width by subgroup interaction term, indicating consistency across all sub- groups. Subsequently, regression analysis was used and prediction equations developed with the combined subgroup data (Fig. 8, 9, Table 4). Again, width measurements accounted for a large percentage of the varia- tion in metatarsal length (r2 = 0.55 and 0.53 for middle and distal widths, respectively). Table 1. The effect of sex on the relationship between the width of the metatarsal at 3 points of measurement (proximal, middle, and distal) and the length of the metatarsal in specimens from Isle Royale and non-Isle Royale locations including Alaska (modern and fossil), Canada, and the 48 contiguous United States (excluding Isle Royale). Isle Royale non-Isle Royale proximal width n 102 120 width P < 0.01 P < 0.01 sex P > 0.21 P > 0.78 sex × width interaction P > 0.20 P > 0.83 middle width n 65 121 width P < 0.01 P < 0.01 sex P > 0.89 P > 0.79 sex × width interaction P > 0.83 P > 0.84 distal width n 65 120 width P < 0.01 P < 0.01 sex P > 0.82 P > 0.89 sex × width interaction P > 0.86 P > 0.83 ALCES VOL. 50, 2014 SILVA ET AL. – METATARSAL DIMENSIONS IN ALCES 165 As with proximal width measurements, a reasonably accurate estimate of metatarsal length was obtained from either middle or distal width (Table 5). As expected, the length of the Isle Royale specimens tended to be overestimated. Although not justified based on the initial analysis, a more accurate estimate of metatarsal length was developed using sepa- rate equations derived from the Isle Royale and non-Isle Royale data (Table 6, Fig. 8, 9). Accuracy was substantially improved for the Isle Royale and Alaskan subgroups (Table 7), but not mainland moose that was a small heterogeneous group representing 3 subspecies. Table 3. Comparison of measured and predicted metatarsal lengths (mm) in Isle Royale and non-Isle Royale moose using separate predictive equations developed from proximal metatarsal lengths (mm). Non-Isle Royale moose include modern and fossil specimens from Alaska, and modern specimens from the 48 contiguous United States (excluding Isle Royale) and Canada. Isle Royale Alaska mainland Sample size 6 6 6 Ave. proximal metatarsal width 53.6 57.7 51.6 Ave. metatarsal length 384.5 414.3 389.8 Predicted metatarsal length from proximal metatarsal width (Isle Royale) 388.2 398.9 382.7 Difference between, range, and probability that true and predicted lengths differ (Isle Royale) −3.7 −13.6−9.8 15.5 5.4−33.4 7.1 −0.4−18.8 P = 0.40 P = 0.02 P = 0.04 Predicted metatarsal length from proximal metatarsal width (non-Isle Royale) 403.7 420.2 395.5 Difference between, range, and probability that true and predicted lengths differ (non-Isle Royale) −19.2 −28.0−5.4 −5.9 −14.7−9.0 −5.6 −16.7−2.1 P = 0.01 P = 0.15 P = 0.08 Table 2. The regression parameters describing the different relationship between proximal metatar- sal width and metatarsal length in specimens from Isle Royale compared to other populations in Canada and the United States including Alaska. Isle Royale non-Isle Royale Sample size 110 159 Significance level P < 0.01 P < 0.01 Adjusted r2 0.47 0.66 Slope (SE) 2.67 (0.27) 4.09 (0.23) y-intercept (SE) 245 (14) 185 (13) Fig. 7. Scatter plot depicting the relationship between the proximal metatarsal width and metatarsal length for the 4 subgroups (Isle Royale [ISRO], mainland [non-ISRO from the lower 48 contiguous United States and Canada], Alaska, and fossil Alaska). Regres- sion lines for the Isle Royale (dotted) and non-Isle Royale (solid line) groups are shown. 166 METATARSAL DIMENSIONS IN ALCES – SILVA ET AL. ALCES VOL. 50, 2014 DISCUSSION The measurements of metatarsal length and width indicated that the Alaskan sub- groups are larger in relative size. The Isle Royale subgroup is different from the other subgroups with shorter metatarsal length and correspondingly larger proximal: and distal width:length ratios. The length of the metatarsus was shorter in Isle Royale moose than the other subgroups and may reflect the trend for large herbivores to experience a reduction in size when isolated on small islands (Peterson et al. 2011), which con- forms to the ‘island rule’ (Van Valen 1973, Lomolino 2005). Given this hypothesis and the short history of Isle Royale moose, these data demonstrate the remarkable speed at which this phenomenon can occur. The metatarsal length:width ratios also provide insight into the biological mechan- ism by which reduction in metatarsal size occurred on Isle Royale. Long bones, includ- ing metatarsals, initially form in 3 parts, the proximal epiphysis (proximal articular sur- face), diaphysis (shaft), and distal epiphysis. Growth ceases when the cartilaginous epi- physeal plates separating these portions ossify. It appears that the reduced size in Isle Royale specimens is limited to the dia- physis with both the length and width of the diaphysis affected proportionally. The widths at the proximal and distal epiphyses do not appear to be reduced, particularly when compared to the mainland group. Thus, the shortening effect appears to be mediated solely through the diaphysis and this isolated effect may facilitate the identifi- cation of specific genes mediating such evo- lutionary action. The length of the metatarsus could be predicted accurately from each of the width measurements, particularly if separate predic- tive equations were developed for specimens from Isle Royale versus other subgroups. The greatest deviation between predicted and actual metatarsal length was only 4.3% using the specific equations; refinements to these predictive equations are presumably possible. For example, the distal width mea- surement was taken either at the distal Fig. 8. Scatter plot depicting the relationship between middle metatarsal width and meta- tarsal length for 3 subgroups (Isle Royale [ISRO], mainland [non-ISRO from the lower 48 contiguous United States and Canada], and Alaska). Regression lines derived from ISRO specimens (dotted line), non-ISRO specimens (solid line), and for all specimens combined (dashed line) are shown. Fig. 9. Scatter plot depicting the relationship between distal metatarsal width and meta- tarsal length for 3subgroups (Isle Royale [ISRO], mainland [non-ISRO from the lower 48 contiguous United States and Canada], and Alaska). Regression lines derived from ISRO specimens (dotted line), non-ISRO specimens (solid line), and for all specimens combined (dashed line) are shown. ALCES VOL. 50, 2014 SILVA ET AL. – METATARSAL DIMENSIONS IN ALCES 167 epiphysis or at the distal articular condyle, whichever was wider; however, a more accu- rate equation might be developed with a sin- gle, consistent measurement. Predictive equations based on middle and distal widths for the non-Isle Royale subgroups improved the accuracy of prediction for the Alaskan, but not mainland group, possibly reflecting the potential heterogeneity within the main- land group. It may indicate that separate equations need to be developed for subpopu- lations within. Finally, possible differences in the method of sample collection among data sets should be considered. The majority of mainland specimens were collected by hun- ters or the result of vehicular accidents, whereas specimens from Isle Royale, Kenai National Wildlife Refuge, and Denali National Park were collected from moose presumably dying of natural causes. Animals that were particularly susceptible to preda- tion may be overrepresented in these groups. A more robust sample size reflecting consis- tent sampling and population variation would presumably improve the relationships presented in this paper. ACKNOWLEDGEMENTS The authors would like to thank the field workers who collected metatarsal specimens on Isle Royale and contributed to the collec- tion at Michigan Technological University. We are indebted to the officials at the United States Department of the Interior, National Table 4. The regression parameters describing the relationship between middle metatarsal width (mm) and metatarsal length (mm), and distal metatarsal width (mm) and metatarsal length. Common equations were developed using speci- mens from all subgroups. middle metatarsal distal metatarsal Sample size 210 208 Significance level P < 0.01 P < 0.01 Adjusted r2 0.55 0.53 Slope (SE) 4.81 (0.30) 3.38 (0.22) y-intercept (SE) 242 (10) 172 (15) Table 5. Comparison of measured metatarsal length (mm) and predicted metatarsal length (mm) in Isle Royale and non-Isle Royale moose based on middle and distal metatarsal widths (mm). Common predictive equations were derived using specimens from all subgroups. Isle Royale Alaska mainland Sample size 6 6 6 Ave. metatarsal length 384.5 414.3 389.8 Ave. middle width of metatarsal 32.5 35.2 31.8 Predicted metatarsal length from middle metatarsal width 397.8 411.2 394.7 Difference between, range, and probability that true and predicted length differ (mm) −13.3 −30.0−6.2 3.2 −5.2−15.9 −4.8 −9.3−0.1 P = 0.08 P = 0.34 P = 0.01 Ave. distal width of metatarsal 66.9 70.9 64.3 Predicted metatarsal length from distal metatarsal width 398.3 411.7 389.7 Difference between true length and predicted length (mm) and range (below) −13.8 −23.3−0.0 2.7 −13.6−11.9 0.2 −11.7−12.0 Probability that the true length and the predicted length are different P = 0.04 P = 0.53 P = 0.96 168 METATARSAL DIMENSIONS IN ALCES – SILVA ET AL. ALCES VOL. 50, 2014 Parks Service, and Isle Royale National Park for granting access to the park and permitting collection of specimens. We would also like to thank Mr. P. Poore for the modification of the vernier calipers used to measure the length of the cannon bone. This research was supported in part by the Kentucky Agri- cultural Experiment Station and is published with the approval of the director (publication number 14-07-014). REFERENCES ALEXANDER, R. M., A. S. JAYES, G. M. O. MALOIY, and E. M. WATHUTA. 1979. Allo- metry of the limb bones of mammals from shrews (Sorex) to elephant (Loxo- donta). Journal of Zoology 189: 305–314. BANDY, P. J., I. M. COWAN,W. D. KITTS, andA. J. WOOD. 1956. A method for the assess- ment of the nutritional status of wild Table 7. Comparison of measured and predicted metatarsal lengths (mm) from middle and distal metatarsal widths (mm) in Isle Royale moose with those from non-Isle Royale moose. The Alaska and mainland equations were derived from measurements from non-Isle Royale moose that included modern and fossil specimens from Alaska, and modern specimens from the 48 contiguous United States (excluding Isle Royale) and Canada. Isle Royale Alaska mainland Sample size 6 6 6 Ave. metatarsal length 384.5 414.3 389.8 Ave. middle width of metatarsal 32.5 35.2 31.8 Predicted metatarsal length from middle metatarsal width 388.3 414.5 400.5 Difference, range, and probability that true and predicted lengths differ −3.8 −16.2−13.7 −0.2 −7.3−14.5 −10.7 −15.5−5.5 P = 0.50 P = 0.96 P < 0.01 Ave. distal width of metatarsal 66.9 70.9 64.3 Predicted metatarsal length from distal metatarsal width 388.0 415.5 396.5 Difference, range, and probability that true and predicted lengths differ −3.5 −13.5−11.0 −1.2 −16.6−10.4 −6.7 −16.8−6.4 P = 0.47 P = 0.78 P = 0.10 Table 6. The regression parameters describing the relationship between proximal metatarsal width (mm) and metatarsal length (mm) in specimens from Isle Royale compared to non-Isle Royale moose. Non-Isle Royale moose included modern and fossil specimens from Alaska, and modern specimens from the 48 contiguous United States (excluding Isle Royale) and Canada. Middle metatarsal width Distal metatarsal width Isle Royale Non-Isle Royale Isle Royale Non-Isle Royale Sample size 71 139 71 137 significance P < 0.01 P < 0.01 P < 0.01 P < 0.01 adjusted r2 0.41 0.54 0.50 0.54 slope (SE) 3.30 (0.47) 4.07 (0.32) 2.38 (.29) 2.91 (0.23) y-intercept (SE) 280 (15) 270 (11) 228 (19) 209 (16) ALCES VOL. 50, 2014 SILVA ET AL. – METATARSAL DIMENSIONS IN ALCES 169 ungulates. Canadian Journal of Zoology 34: 48–52. COBLE, D. S., L. L. WILSON, J. P. HITCHCOCK, H. VARELA-ALVAREZ, and M. J. SIMPSON. 1971a. Sire, sex and laterality effects on bovine metacarpal and metatarsal charac- ters. Growth 35: 65–77. ——— , ———, M. J. SIMPSON, H. VARELA- ALVAREZ, J. P. HITCHCOCK, J. H. ZIEGLER, J. D. SINK, and J. L. WATKINS. 1971b. Relation of bovine metacarpal and meta- tarsal characters to growth and carcass characters. Growth 35: 79–89. FRANZMANN, A. W., R. E. LERESCHE, R. A. RAUSCH, and L. L. OLDEMEYER. 1978. Alaskan moose measurements and weights and measurement–weight rela- tionships. Canadian Journal of Zoology 56: 298–306. FRICK, C. 1930. Alaska’s frozen fauna. Nat- ural History (the journal of the American Museum of Natural History) 30: 71–80. JENSEN, W. F., J. R. SMITH, J. J. MASKEY JR., J. V. MCKENZIE, and R. E. JOHNSON. 2013. Mass, morphology and growth rates of moose in North Dakota. Alces 49: 1–15. LOMOLINO,M. V. 2005. Body size evolution in insular vertebrates: generality of the island rule. Journal of Biogeography 32: 1683–1699. LYNCH, G. M., B. LAJUENESSE, J. WILLMAN, and E. S. TELFER. 1995. Moose weights and measurements from Elk Island National Park, Canada. Alces 31: 199–207. MARTIN, J. G. A., M. FESTA-BIANCHET, S. D. COTE, and D. T. BLUMSTEIN. 2013. Detect- ing between-individual differences in hind-foot length in populations of wild mammals. Canadian Journal of Zoology 91: 118–123. MCEWAN, E. H., and A. J. WOOD. 1966. Growth and development of the barren ground caribou. I. Heart girth, hind foot length and body weight relationships. Canadian Journal of Zoology 44: 401–411. MCMAHON, T. A. 1975. Allometry and bio- mechanics: limb bones in adult ungu- lates. American Naturalist 109: 547–563. PALSSON, H., and J. B. VERGES. 1952. Effect of the plane of nutrition on growth and the development of carcass quality in lambs. Part I. The effects of high and low planes of nutrition at different ages. Journal of Agricultural Science 42: 1–92. PETERSON, R. O. 1977: Wolf ecology and prey relationships on Isle Royale. U.S. National Park Service Scientific Mono- graph Series 11. U.S. Government Print- ing Office, Washington D. C., USA. ——— , C. C. SCHWARTZ, andW.B. BALLARD. 1983. Eruption patterns of selected teeth in three North American moose popula- tions. Journal of Wildlife Management 47: 884–888. ———, J. A. VUCETICH, D. BEYER, M. SCHRAGE, and J. RAIKKONEN. 2011. Phe- notypic variation in moose: The island rule and the moose of Isle Royale. Alces 47: 125–133. ROSEBERRY, J. L., and W. D. KLIMSTRA. 1975. Some morphological characteristics of the Crab Orchard deer herd. Journal of Wildlife Management 39: 48–58. SAS. 1985. User’s Guide: Statistics. SAS Institute Inc., Cary, North Carolina, USA. SANTUCCI, V. L., J. A. KULLER, A. F. BATTELLI, S. A. LAIFER, and D. I. EDELSTONE. 1993. Fetal metatarsal length: an accurate pre- dictor of gestational age and weight in the ovine fetus. Gynecologic and Obste- tric Investigation 35: 76–79. VAN VALEN, L. 1973. A new evolutionary law. Evolutionary Theory 1: 1–33. WILKERSON, A. S. 1932. Some frozen deposits in the goldfields of interior Alaska: a study of the Pleistocene deposits of Alaska. American Museum Novitates 525: 1–22. 170 METATARSAL DIMENSIONS IN ALCES – SILVA ET AL. ALCES VOL. 50, 2014 VARIATION IN METATARSAL MORPHOLOGY AMONG SUBGROUPS OF NORTH AMERICAN MOOSE (Alces alces) METHODS RESULTS Metatarsal Length and Width Predictive Equations for Metatarsal Length Based on Widths DISCUSSION ACKNOWLEDGEMENTS REFERENCES