F:\ALCES\Vol_39\p65\3901.PDF ALCES VOL. 39, 2003 SPEARS ET AL. - MOOSE BONE MARROW FAT 273 BONE MARROW FAT CONTENT FROM MOOSE IN NORTHEASTERN MINNESOTA, 1972-2000 Brian L. Spears1, William J. Peterson2, and Warren B. Ballard1 1Department of Range, Wildlife, and Fisheries Management, Box 42125, Texas Tech University, Lubbock, Texas 79409, USA; 2Minnesota Department of Natural Resources, Section of Wildlife, P.O. Box 115, Grand Marais, MN 55604, USA ABSTRACT: Percent fat in femur bone marrow has been used as an indicator of animal condition at time of death. However, femur bone marrow is not always available for collection. We used linear regression to examine relationships among marrow fat values for long bones (i.e., femur, tibia, mandible, humerus, radius, tarsal and carpal bones) of moose (Alces alces) from northeastern Minnesota during 1972-2000. Linear regressions for bone marrow fat in each set of bones (paired with femur) in calves and adults were significant and highly correlated (r2 = 0.83-0.99). Linear regressions for femur bone marrow fat for yearling moose were significant and highly correlated for tibia, humerus and radius bones (r2 = 0.86-0.93), and less so for tarsal bones (r2 = 0.63). Bone marrow fat deposition appeared first in proximal and distal bones and was mobilized last in distal bones. Calves had higher femur fat in fall and early winter than late winter and spring. Month, season, and year had no significant effect on femur marrow fat percent for yearlings or adults. Percent femur marrow fat was lower in accidentally killed calves than accidentally killed yearlings or adults. Adults killed by disease had lower percent femur fat than those killed by accident or wolves (Canis lupus). Amount of adult male femur fat was loosely correlated to a winter severity index for the previous winter. Our results suggest that fat deposition and mobilization were similar to that found in other studies and that bone marrow fat content may be a good indicator of relative moose health within a population. ALCES VOL. 39: 273-285 (2003) Key words: Alces alces, bone marrow, death, deposition, fat, femur, Minnesota, moose, wolves Bone marrow fat percentages have been widely used as an index of ungulate condition at time of death (Cheatum 1949, Baker and Leuth 1966, Neiland 1970, Franzmann and Arneson 1976, Peterson 1977, Peterson and Bailey 1984, Ballard et al. 1987, Mech et al. 1995). Marrow fat is mobilized after other body fat reserves are depleted or exhausted (Cheatum 1949, Smith and Jones 1961). Reduction in bone mar- row fat may therefore be a good indicator of decreased fitness of an individual. How- ever, the percentage of marrow fat deple- tion indicating a stressed animal has been debated (Mech and DelGiudice 1985, Ballard 1995). Moreover, what constitutes a healthy animal may be relative to other members of the population each year (Ballard 1995), and a high marrow fat content would not necessarily indicate an animal in good con- dition (Mech and DelGiudice 1985, Ballard et al. 1987). Despite discrepancies in con- clusions regarding stage of individual health, bone marrow fat depletion often remains the only indicator of individual condition at time of death (Ballard 1995). Fat levels from the femur have typically been used to draw conclusions regarding the condition of individual moose at time of death (Cheatum 1949, Franzmann and Arneson 1976, Peterson et al. 1982, Ballard et al. 1987, Hayes et al. 1991, Ballard 1995, MOOSE BONE MARROW FAT - SPEARS ET AL. ALCES VOL. 39, 2003 274 Mech et al. 1995). However, predators often consume proximal leg bones (Ballard et al. 1981, Peterson et al. 1982), and frost wedging may expose femur marrow to the air, rendering the sample useless (Peterson et al. 1982). Researchers must often settle for collection of other bones for analysis, and the correlation of marrow fat in these bones with that of the typically used femur is desirable. While investigating natural and human- caused mortality of moose (Alces alces) in Minnesota during 1972-2000, amount of marrow fat was estimated in femur, tibia, mandible, humerus, radius, tarsal, and car- pal bones of individual moose. We exam- ined correlations between femur marrow fat and marrow fat in the above bones for moose divided into 3 age classes to deter- mine their usefulness as indices of health as related to the femur marrow fat standard. We also present data on the annual cycle of bone marrow fat deposition and mobiliza- tion in moose, and examine differences in amount of femur bone marrow fat among seasons, years, cause of death, and age class. STUDY AREA AND METHODS The study was conducted from 1972- 2000 in Cook and Lake counties in north- eastern Minnesota. The majority of moose examined were within 65 km of Poplar Lake, located in Superior National Forest, Cook County, Minnesota. A description of the study area was provided by Nelson and Mech (1981). We sexed moose by examining sex or- gans or by the presence or absence of antlers or pedicles. Age of calves and yearlings was determined by tooth eruption and replacement. Dental cementum was used to determine adult ages. A femur, tibia, humerus, radius, tarsal, carpal, and mandible was collected from each moose carcass if possible, usually within several hours after death. Bone samples from each moose were stored frozen for up to a few months before being processed. Bones were broken with a hammer. Marrow samples were removed by hand from the central portion of the marrow tube. Mar- row fat content was determined by the oven drying technique described by Neiland (1970). We separated bone marrow fat data by age class of individual due to possible dif- ferences of marrow fat mobilization at dif- ferent stages of life (Ballard et al. 1981). Moose within 1 year of birth were classified as calves, 1-2 years of age as yearlings, and >2 years of age as adults. All moose were assumed born 1 June (Schwartz 1998). We classified causes of death as auto- mobile or train collision, natural accidental death, hunter-kill, killed intentionally as a danger to humans, diseased (brainworm [Parelaphostrongylus tenuis] confirmed or suspected, or heavy tick load with heavy hair loss), wolf (Canis lupus)-kill, or un- known. Marrow fat content means are given + SD. Fat was analyzed by month and season of death. Month of death was lumped into 1 of 4 seasons: 1 = December- February, 2 = March-May, 3 = June-August and 4 = September-November. We calculated regression lines and Pearson's Product-Moment correlations for each comparison between an individual's femur and other bones examined within each age class. We assumed the simple linear model: Y = B + MX + ε, where Y = marrow fat value of bone being correlated, X = marrow fat value of femur, B = the value of Y when X = 0, M = best-fit slope of comparison line, and ε = residual error unaccounted for by the model. F-tests were calculated to determine whether models of similar bone pair regressions differed sig- nificantly among age classes (Graybill 1976). F-statistics were calculated to examine differences in femur marrow fat content ALCES VOL. 39, 2003 SPEARS ET AL. - MOOSE BONE MARROW FAT 275 among age classes. F-statistics were also calculated to examine effects of month, season, and cause of death on bone marrow fat content in femurs within each age class. Where F-tests were significant, Student's t- test was used to determine differences in femur marrow fat means among age classes, causes of death, and seasons. Following Ballard et al. (1981) and Davis et al. (1987), we used paired t-tests to examine differ- ences in marrow fat content between fe- murs and all other bones examined, as well as humerus-tibia,humerus-carpal, and proxi- mal-distal bone pairs. Welch's approximate t values and associated degrees of freedom were used in cases where variances be- tween Student's t-test groups were not ho- mogeneous (Zar 1999). A winter severity index (WSI) was calculated for each winter from weather data collected at Poplar Lake in Cook County, Minnesota (Minnesota Department of Natural Resources, Section of Wildlife, Grand Marais, Minnesota). The WSI was defined as: sum of number of days <-17.8 degrees Celsius + number of days with >38.1 cm of snow. F-tests were calculated to examine relationships between femur marrow fat content and WSI the preceding winter. All computer regressions, t-tests, and F-test calculations were completed us- ing Statistica 2000 (StatSoft, Tulsa, OK). RESULTS One-hundred four adults (37M, 67F), 47 yearlings (26M, 21F) and 45 calves (26M, 19F) were examined. Calf marrows were obtained for all months except July, and marrows from only the femur and mandible were collected from October death sam- ples. Yearling samples were collected in all months except February, April, and De- cember. Adult marrows were collected in every month except March. Linear regressions for bone marrow fat in each set of bones (paired with femur) in calves were highly correlated and signifi- cant (Table 1). Linear regressions for marrow fat in yearling moose femurs were significant and highly correlated for tibia, humerus, and radius bones. Regressions in marrow fat between femur and mandible, Table 1. Regression equations for bone marrow fat percentage comparisons among bones in calf moose (<1 year old) from northeastern Minnesota, 1972-2000. Bone Pair X (SD) n pairs Regression Equation r 2 P Femur- 48.2 (24.7) 30 T = 4.8820 + 0.9976 (F) 0.94 <0.001 Tibia 53.0 (26.3) Femur- 50.0 (24.8) 35 M = 14.574 + 0.70435 (F) 0.91 <0.001 Mandible 50.0 (19.6) Femur- 48.2 (24.7) 30 H = 0.49474 + 1.0243 (F) 0.99 <0.001 Humerus 49.9 (25.6) Femur- 48.2 (24.7) 30 R = 0.2615 + 1.0579 (F) 0.96 <0.001 Radius 50.8 (27.2) Femur- 48.2 (24.7) 30 T = 15.012 + 0.80347 (F) 0.90 <0.001 Tarsal 53.8 (22.0) Femur- 48.2 (24.7) 30 C = 23.639 + 0.69442 (F) 0.85 <0.001 Carpal 57.1 (20.2) MOOSE BONE MARROW FAT - SPEARS ET AL. ALCES VOL. 39, 2003 276 tarsal, and carpal bones were significant but less correlated (Table 2). For adults, linear regressions for marrow fat in each set of bones were significant and highly corre- lated (Table 3; Figs. 1-6). Regression mod- els (i.e., either slopes, intercepts, or both) among adults, yearlings, and calves differed for all bone pair comparisons except for the yearling-calf comparison in the femur-hu- merus bone pair (Table 4). Mean marrow fat in calves was lower in the femur than tarsal, humerus, carpal, and tibia bones and lower in the humerus than carpal (Table 5). Mean marrow fat in yearlings was lower in the femur than tar- sal, humerus, and carpal bones, higher in the femur than mandible, and higher in the humerus than carpal (Table 6). Mean mar- row fat in adults was lower in the femur than tarsal and carpal bones, higher in the femur than mandible, and lower in the hu- merus than carpal (Table 7). Femur marrow fat did not differ among moose killed by vehicle, accident, hunters, or unknown causes in calves (F 2,29 = 0.195; P = 0.824), yearlings (F 3,35 = 1.081; P = 0.370), or adults (F 3,58 = 2.372; P = 0.080). These mortality categories were therefore combined as accidental deaths. No rela- tionship existed between age in years and femur fat in accidentally killed males (F 1,61 = 0.042; P = 0.838). Femur marrow fat in accidentally killed females increased with age (Y = 72.202 + 1.702(age); r2 = 0.112; F 1,67 = 8.429; P = 0.005). Femur marrow fat was lower in accidentally killed calves than accidentally killed yearling or adult moose (Fig. 7). Femur marrow fat and sample size in each age class by month is shown in Figure 8. Femur marrow fat in calf moose ( X = 52.500 + 24.035; range = 5-88) differed by month (F 10,31 = 4.117; P = 0.001) and sea- son (F 3,38 = 5.020; P = 0.005), but not year (F 20,21 = 1.342; P = 0.254). Femur fat content differed among months (F 10,31 = 4.117; P = 0.001) and seasons in calves killed accidentally. Calves had higher fe- mur fat in seasons 3-4 ( X = 64.250 + 19.461; n = 16) than 1-2 ( X = 42.269 + Table 2. Regression equations for bone marrow fat percentage comparisons among bones in yearling moose (1-2 years old) from northeastern Minnesota, 1972-2000. Bone Pair X (SD) n pairs Regression Equation r2 P Femur- 78.5 (11.1) 28 T = 16.739 + 0.815 (F) 0.81 <0.001 Tibia 80.7 (11.1) Femur- 78.7 (10.5) 31 M = 44.191 + 0.265 (F) 0.42 0.028 Mandible 65.0 (6.7) Femur- 78.9 (11.1) 27 H = 0.9519 + 0.899 (F) 0.93 <0.001 Humerus 80.4 (10.7) Femur- 78.9 (11.1) 27 R = 11.156 + 0.8759 (F) 0.86 <0.001 Radius 80.2 (11.3) Femur- 78.9 (11.1) 27 T = 49.722 + 0 .445 (F) 0.63 <0.001 Tarsal 85.0 (7.9) Femur- 78.8(11.1) 27 C = 68.058 + 0.24492 (F) 0.53 0.004 Carpal 87.4 (5.1) ALCES VOL. 39, 2003 SPEARS ET AL. - MOOSE BONE MARROW FAT 277 Table 3. Regression equations for bone marrow fat percentage comparisons among bones in adult moose (>2 years old) from northeastern Minnesota, 1972-2000. Bone Pair X (SD) n pairs Regression Equation r2 P Femur- 71.0 (26.3) 62 T = 5.2776 + 0.94322 (F) 0.96 <0.01 Tibia 72.3 (25.9) Femur- 71.3 (26.2) 63 M = 19.277 + 0.60666 (F) 0.86 <0.001 Mandible 78.4 (19.7) Femur- 70.2 (27.1) 70 H = 0.98791 + 0.99693 (F) 0.99 <0.01 Humerus 61.6 (18.8) Femur- 70.7 (26.4) 61 R = 4.3726 +0.94837 (F) 0.97 <0.01 Radius 71.4 (25.8) Femur- 71.3 (26.2) 63 T = 32.581 + 0.64246 (F) 0.85 <0.001 Tarsal 78.4 (19.7) Femur- 71.3 (26.2) 63 C = 36.891 + 0.59844 (F) 0.83 <0.001 Carpal 79.6 (19.0) Humerus = 0.92751 + 0.99875(Femur) Femur H um er us 0 20 40 60 80 100 120 0 20 40 60 80 100 120 Fig. 1. Relationship between percent marrow fat in the femur and humerus for adult moose from northeastern Minnesota, 1972-2000. Tarsal = 32.539 + 0.64380(Femur) Femur T ar sa l 0 20 40 60 80 100 0 20 40 60 80 100 120 Fig. 2. Relationship between percent marrow fat in the femur and tarsal for adult moose from northeastern Minnesota, 1972-2000. Mandible = 18.476 + 0.60800(Femur) Femur M an di bl e 0 20 40 60 80 100 120 0 20 40 60 80 100 120 Fig. 3. Relationship between percent marrow fat in the femur and mandible for adult moose from northeastern Minnesota, 1972-2000. Tibia = 6.4169 + 0.93866(Femur) Femur T ib ia -10 10 30 50 70 90 110 0 20 40 60 80 100 120 Fig. 4. Relationship between percent marrow fat in the femur and tibia for adult moose from northeastern Minnesota, 1972-2000. MOOSE BONE MARROW FAT - SPEARS ET AL. ALCES VOL. 39, 2003 278 24.029; n = 26) (t 40 = -2.664; P = 0.011). Calves killed by accident ( X = 60.563 + 18.371; n = 32) had more femur fat than those killed by disease ( X = 24.222 + 22.174; n = 9) (t 39 = 5.013; P <0.001). Only one calf killed by wolves was examined. Femur marrow fat for this individual was 49%. Femur marrow fat among all yearling moose killed ( X = 77.732 + 11.194 range = 46-92) did not differ by month (F 8,32 = 1.453; P = 0.213), season (F 3,37 = 0.634; P = 0.598), or year (F 14,26 = 0.898; P = 0.571). Femur fat did not differ among months (F 7,30 = 1.627; P = 0.166), seasons (F 3,34 = 0.598; P = 0.620), or years (F 14,23 = 0.825; P = 0.638) for yearlings accidentally killed. Only one yearling killed by wolves and one killed by disease were examined. Femur marrow fat for yearling moose killed acci- dentally averaged 78% + 11.4 (n = 38). Femur marrow fat content was 68% for the wolf-killed yearling and 73% for the year- ling dying of disease. For adults, number of moose examined Radius = 3.3440 + .96917(Femur) Femur R ad iu s -10 10 30 50 70 90 110 0 20 40 60 80 100 120 Fig. 5. Relationship between percent marrow fat in the femur and radius for adult moose from northeastern Minnesota, 1972-2000. Carpal = 29.937 + 0.68225(Femur) Femur C ar pa l -10 10 30 50 70 90 110 0 20 40 60 80 100 120 Fig. 6. Relationship between percent marrow fat in the femur and carpal for adult moose from northeastern Minnesota, 1972-2000. Table 4. F-test degrees of freedom and F-values for tests of equality of models in moose bone pairs from northeastern Minnesota, 1972-2000. All tests evaluated at alpha = 0.05. Adults vs. Yearlings Adults vs. Calves Yearlings vs. Calves Bone Pair df F df F df F Femur- 2, 85 5.50 2, 88 138.60 2, 53 0.041 Humerus Femur- 2, 97 592.45 2, 101 741.30 2, 62 11.81 Mandible Femur- 2, 208 2864.60 2, 114 999.70 2, 148 2381.40 Tibia Femur- 2, 86 94.3 2, 89 134.86 2, 53 321.37 Carpal Femur- 2, 86 83.55 2, 89 111.48 2, 53 159.23 Tarsal Femur- 2, 84 13.57 2, 87 141.73 2, 53 5.14 Radius 1Failed to reject H o : models tested are equal. ALCES VOL. 39, 2003 SPEARS ET AL. - MOOSE BONE MARROW FAT 279 Table 5. Means, standard deviations, differences between, and t-tests for mean differences of fat content in calf moose (<1 year old) bones from northeastern Minnesota, 1972-2000. Bone Pair n pairs X SD Diff SD Diff t P Femur- 30 48.233 24.713 Tarsal 53.767 21.970 -5.533 10.582 -2.864 0.008 Femur- 35 49.971 24.823 Mandible 49.771 19.629 0.200 11.552 0.102 0.919 Femur- 30 48.233 24.713 Humerus 49.900 25.578 -1.667 3.717 -2.456 0.020 Femur- 30 48.233 24.713 Radius 50.767 27.211 -2.533 7.678 -1.808 0.081 Femur- 30 48.233 24.713 Carpal 57.133 20.190 -8.900 13.044 -3.737 <0.001 Femur- 30 48.233 24.713 Tibia 53.000 26.265 -4.767 9.058 -2.882 0.007 Humerus- 30 49.900 25.578 Radius 50.767 27.211 -0.867 7.776 -0.610 0.546 Humerus- 30 49.900 25.578 Carpal 57.133 20.190 -7.233 13.723 -2.887 0.007 Table 6. Means, standard deviations, differences between, and t-tests for mean differences of fat content in yearling moose (1-2 years old) bones from northeastern Minnesota, 1972-2000. Bone Pair n pairs X SD Diff SD Diff t P Femur- 27 78.851 11.128 Tarsal 84.963 7.862 -6.111 8.657 -3.668 0.001 Femur- 31 78.677 10.537 Mandible 65.032 6.681 13.645 9.841 7.720 <0.001 Femur- 27 78.852 11.128 Humerus 80.444 10.714 -1.593 3.983 -2.078 0.048 Femur- 27 78.852 11.128 Radius 80.222 11.274 -1.370 5.832 -1.221 0.233 Femur- 27 78.852 11.128 Carpal 87.370 5.123 -8.519 9.456 -4.681 <0.001 Femur- 28 78.500 11.077 Tibia 80.679 11.049 -2.179 6.700 -1.721 0.097 Humerus- 27 80.444 10.714 Radius 80.222 11.274 0.222 5.199 0.222 0.826 Humerus- 27 80.444 10.714 Carpal 87.370 5.122 -6.926 9.583 -3.755 <0.001 MOOSE BONE MARROW FAT - SPEARS ET AL. ALCES VOL. 39, 2003 280 and mean femur marrow fat percentages by season, month, and cause of death are given in Table 8. Femur marrow fat percent in adult moose ( X = 69.477 + 26.202; range = 8-95) did not differ among months (F 10,75 = 1.47; P = 0.167), seasons (F 3,82 = 2.355; P = 0.078), or years (F 24,61 = 1.43; P = 0.133). Femur fat did not differ among months (F 9,52 = 1.325; P = 0.247), seasons (F 3,58 = 0.737; P = 0.534), or years (F 21,40 = 1.746; P = 0.064) for adults dying from accidental causes. Femur marrow fat was higher in adult moose dying accidentally ( X = 79.210 + 16.861; n = 62) than those dying from disease ( X = 45.380 + 28.409; n = 21) (t25 = 5.158; P <0.001). Femur marrow fat did not statistically differ between adult moose dying from accidental causes or those killed by wolves ( X = 37.00 + 42.673; n = 3) (t2 = 1.707; P = 0.228), or between those killed by wolves or disease (t 22 = 0.453; P = 0.655) (Fig. 9). Mean adult male femur fat within years was loosely related to yearly WSI (Fig. 10). Mean yearly adult female femur fat content was not related to yearly WSI (F 1,18 = 0.3692; P = 0.551). DISCUSSION The relationships among moose femur bone marrow fat and other examined moose Table 7. Means, standard deviations, differences between, and t-tests for mean differences of fat content in adult moose (>2 years old) bones from northeastern Minnesota, 1972-2000. Bone Pair n pairs X SD Diff SD Diff t P Femur- 63 71.254 26.175 Tarsal 78.413 19.727 -7.159 13.860 -4.100 <0.001 Femur- 70 70.157 27.103 Mandible 61.600 18.787 8.557 14.017 5.108 <0.001 Femur- 62 71.016 26.320 Humerus 71.855 26.522 -0.839 3.521 -1.875 0.066 Femur- 61 70.721 26.435 Radius 71.443 25.821 -0.721 6.330 -0.890 0.377 Femur- 63 71.254 26.175 Carpal 79.603 18.966 -8.349 14.917 -4.443 <0.001 Femur- 62 71.016 26.320 Tibia 72.290 25.895 -1.274 7.446 -1.347 0.183 Humerus- 62 71.790 26.472 Radius 71.758 25.729 0.032 6.724 0.038 0.970 Humerus- 63 72.143 26.406 Carpal 79.603 18.966 -7.460 15.169 -3.904 <0.001 Age class F em ur m ar ro w f at p er ce nt 52 58 64 70 76 82 88 Calf (n = 32) Yearling (n = 38) Adult (n = 62) ± 95% CI ± SE Mean F2,129 = 15.857; P <0.001 Fig. 7. Femur marrow fat percent and age class in moose kills classified as accidental from north- eastern Minnesota, 1972-2000. ALCES VOL. 39, 2003 SPEARS ET AL. - MOOSE BONE MARROW FAT 281 Table 8. Mean femur bone marrow fat percentages of adult moose (>2 years old) by season, month, and cause of death from northeastern Minnesota, 1972-2000 (standard deviations in parentheses; sample sizes appear below marrow fat percentages). Mortality All Accidental Disease Wolf kill Season Dec-Feb 56.55 (33.99) 79.78 (16.40) 35.44 (31.05) 47.00 (55.15) 20 9 9 2 Mar-May 76.83 (19.67) 83.60 (11.84) 43.00 (0.00) 6 5 1 Jun-Aug 73.76 (22.73) 81.56 (15.47) 43.83 (25.79) 33 27 7 Sep-Nov 71.04 (22.62) 74.37 (20.65) 65.60 (23.19) 17.00 (0.00) 27 19 5 1 Month Jan 46.71 (26.20) 76.20 (20.52) 15.00 (9.90) 8.00 (0.00) 8 5 2 1 Feb 55.20 (24.50) 88.00 (0.00) 46.20 (25.47) 6 1 5 Mar Apr 43.00 (0.00) 43.00 (0.00) 1 1 M a y 83.60 (11.84) 83.60 (11.84) 5 5 Jun 71.36 (24.45) 78.33 (18.26) 29.50 (3.54) 14 12 2 Jul 75.50 (21.60) 80.44 (15.81) 31.00 (0.00) 10 9 1 A u g 79.00 (23.01) 89.67 (2.25) 57.67 (32.89) 9 6 3 Sep 82.44 (11.61) 84.57 (6.92) 78.33 (18.90) 9 7 2 Oct 64.08 (26.55) 66.45 (23.97) 85.00 (0.00) 17.00 (0.00) 131 11 1 1 Nov 64.08 (26.55) 83.33 (6.43) 46.50 (14.85) 5 3 2 Dec 66.00 (35.06) 85.25 (11.07) 33.67 (42.74) 86.00 (0.00) 8 4 3 1 1Includes 2 unknown causes of death. MOOSE BONE MARROW FAT - SPEARS ET AL. ALCES VOL. 39, 2003 282 bones were significant and highly corre- lated for calves, yearlings, and adults. Our study corroborates findings by others that bone marrow fat in the mandible and leg bones can be just as useful as that of the femur in determining relative health of moose individuals (Snider 1980, Ballard et al. 1981). Unlike Davis et al.'s (1987) findings in cari- bou (Rangifer tarandus), we found signifi- cant differences among regression lines of adults, yearlings, and calves in all bone pair comparisons except for the yearling-calf comparison in the femur-humerus bone pair (Table 4). This is in agreement with Ballard et al. (1981), suggesting that calves, year- lings, and adults may deposit and mobilize bone marrow fat differently. Our study corroborated previous find- ings that bone marrow fat was deposited first and mobilized last in distal bones in ungulates (Cheatum 1949, Peterson et al. 1982, and Ballard 1995). Femur and hu- merus (proximal) fat was lower in all age classes than fat in tarsal and carpal (distal) bones, respectively, suggesting that fat is deposited in distal bone marrow first in calves, and depleted first from proximal bone marrow in older moose. However, femur fat did not differ from tibia fat in yearlings or adults, and humerus fat did not differ from radius fat in any age class. Peterson et al. (1982) indicated that while marrow fat withdrawal was sequential from proximal to distal bones in many ungulates, this pattern was not as marked in moose. Overall, calf femur marrow content in- creased after birth (1 June) and peaked in November, a trend also observed in moose in southcentral Alaska (Ballard and Whitman 1987). Femur fat significantly decreased following the first winter, and increased as moose became yearlings. Yearling femur marrow fat remained relatively constant throughout the year. Although not statisti- cally different, average femur marrow fat values seemed to increase in the spring and Fig. 9. Femur marrow fat percent and cause of death of adult moose from northeastern Min- nesota, 1972-2000. Cause of death F em ur m ar ro w f at p er ce nt -20 0 20 40 60 80 100 Accident (n = 62) Diseased (n = 21) Wolf kill (n = 3) ± 95% CI ± SE Mean Fig. 10. Yearly average adult male femur fat percent vs. the preceding winter's Winter Severity Index for moose from northeastern Minnesota, 1972-1999. Winter Severity Index F em ur f at p er ce nt 40 50 60 70 80 90 100 60 100 140 180 220 260 F1,13 = 3.365; P = 0.090 y = 78.321-0.092*x+eps r2 = 0.206 Fig. 8. Femur bone marrow fat percent by month from moose in northeastern Minnesota, 1972- 2000. Sample sizes are given in table below graph. ALCES VOL. 39, 2003 SPEARS ET AL. - MOOSE BONE MARROW FAT 283 summer months while declining throughout the winter months for adult moose, corre- sponding to the timing of increased and decreased availability of high-quality for- age (Schwartz and Renecker 1998). This trend was also observed in moose in Alaska (Franzmann and Arneson 1976). Previous authors suggested that bone marrow fat content can be used as an index of relative health for individuals within a population (Cheatum 1949, Ballard 1995). Assuming marrow fat values (as an index of animal condition) had no bearing on moose killed by accident, we can use marrow fat content in these random animals as an index for animals of normal health (or "healthy") within the population. Calf and adult moose we examined killed by disease had lower femur marrow fat than that of healthy indi- viduals, supporting the use of femur marrow fat as a relative health index. We were only able to examine 1 calf and 1 yearling killed by wolves, and were therefore unable to statistically compare femur marrow fat content in calves and yearlings killed by accident or disease and wolves. Using femur fat as an index of health for adults, we did not find statistical evidence that wolves hunting moose were singling out sick individuals from this popu- lation. Our results were similar to Franzmann and Arneson (1976) and Ballard et al. (1987), which found no difference in marrow fat values between wolf-killed and accidentally-killed adult female moose in Alaska. However, our low sample size of wolf-killed adult moose examined here (n = 3) may not be a good indication of moose- wolf interactions in this population. Several studies suggested universal fe- mur marrow fat content ranges that indicate animals that are healthy or are in poor condition. Bischoff (1954) classified most mule deer (Odocoileus hemionus) indi- viduals with 80-100% bone marrow fat as in fair or poor condition. Mech et al. (1995) suggested that caribou individuals with <70- 87% femur fat content had depleted muscle or fat reserves, and were in marginal con- dition. Franzmann and Arneson (1976) and Peterson and Bailey (1984) suggested that moose with <20% femur marrow fat were dying of starvation. Mean femur fat for normal representative moose calves we examined was 64% before their first winter and 42% after. Femur marrow fat content did not statistically change throughout the year for yearling or adult moose we exam- ined and used as normal representatives. Mean values for these animals were 77% and 79%, respectively. Given that these animals were representative of normal indi- viduals, our results are in contrast with suggestions by previous authors that these animals may be in an abnormal sub-healthy state. Ballard (1995) indicated that a de- clining trend in bone marrow fat content by late winter and early spring was common for many northern ungulate populations and relatively low values might be considered normal for many populations. Alterna- tively, the relative condition of this moose population as a whole may be diminished due to factors we did not examine, such as a possible lack of highly nutritious forage or other stressors. Marrow fat content has been observed at or near 100% in caribou, moose, and deer individuals (Bischoff 1954, Peterson et al. 1982, Davis et al. 1987, this study). How- ever, a baseline marrow fat content for healthy animals has not been established against which to judge individuals from different populations. A universal baseline for using marrow fat content in moose as an overall index of health may only be deter- mined through examination of total body fat depletion patterns in animals at different states of health (Ballard 1995). ACKNOWLEDGEMENTS Data for this paper were collected while author Peterson was employed as Area MOOSE BONE MARROW FAT - SPEARS ET AL. ALCES VOL. 39, 2003 284 Wildlife Manager at Grand Marais, MN by the Minnesota Department of Natural Re- sources, Section of Wildlife. He appreci- ates having been permitted the freedom of action to pursue this and other studies that were not covered by formal research pro- posals. We thank Rick Fields, Tim Webb, and Dave Ingebrigtsen who helped collect data, and Pam Coy who processed mar- rows in early years of the study. Special thanks are due Peterson's wife, Dale Peterson, who gave up many nights, week- ends, and holidays to help examine moose and collect samples in addition to putting up with having her oven used for drying mar- rows during the later years of the study. This is Texas Tech University, College of Agricultural Sciences and Natural Re- sources Publication T-9-977. REFERENCES BAKER, M. F., AND F. X. LEUTH. 1966. Mandibular cavity tissue as a possible indicator of condition of deer. Proceed- ing of the Annual Southeast Game and Fish Commissioners 20:69-74. BALLARD, W. B. 1995. Bone marrow fat as an indicator of ungulate condition-how good is it? Alces 31:105-109. , C. L. GARDNER, J. H. WESTLUND, AND S. M. MILLER. 1981. Use of mandible versus longbone to evaluate percent marrow fat in moose and cari- bou. Alces 17:147-164. , and J. S. WHITMAN. 1987. Marrow fat dynamics in moose calves. Journal of Wildlife Management 51:66-69. , , and C. L. GARDNER. 1987. Ecology of an exploited wolf population in south-central Alaska. Wildlife Mono- graphs 98. BISCHOFF, A. I. 1954. Limitations of the bone marrow technique in determining malnutrition in deer. Proceedings of the Western Association of State Game and Fish Commissioners 34:205-210. CHEATUM, E. L. 1949. Bone marrow as an index of malnutrition in deer. New York State Conservation 3:19-22. DAVIS, J. L., P. VALKENBURG, and D. J. REED. 1987. Correlations and depletion patterns of marrow fat in caribou bones. Journal of Wildlife Management 51:365- 371. FRANZMANN, A. W., and P. D. ARNESON. 1976. Marrow fat in Alaskan moose femurs in relation to mortality factors. Journal of Wildlife Management 40:336- 339. GRAYBILL, F. A. 1976. Theory and Applica- tion of the Linear Model. Duxbury Press, North Scituate, Massachusetts, USA. HAYES, R. D., A. M. BAYER, and D. G. LARSON. 1991. Population dynamics and prey relationships of an exploited and recovering wolf population in the southern Yukon. Yukon Fish and Wild- life Branch Final Report TR-91-1. Whitehorse, Yukon, Canada. MECH, L. D., and G. D. DELGIUDICE. 1985. Limitations of the marrow fat technique as an indicator of body condition. Wild- life Society Bulletin 13:204-206. , T. J. MEIER, J. W. BURCH, and L. G. ADAMS. 1995. Patterns of prey selec- tion by wolves in Denali National Park, Alaska. Pages 231-249 in L. S. Carbyn, S. Fritts, and D. Seip, editors. Ecology and Conservation of Wolves in a Chang- ing World. Canadian Circumpolar In- stitute, University of Alberta, Edmon- ton, Alberta, Canada. NEILAND, K. A. 1970. Weight of dried marrow as an indicator of fat in caribou femurs. Journal of Wildlife Manage- ment 34:904-907. NELSON, M. E., and L. D. MECH. 1981. Deer social organization and wolf predation in northeastern Minnesota. Wildlife Monographs 77. PETERSON, R. O. 1977. Wolf ecology and ALCES VOL. 39, 2003 SPEARS ET AL. - MOOSE BONE MARROW FAT 285 prey relationships on Isle Royale. U.S. National Park Service Science Mono- graph Serial 11. , D. L. ALLEN, and J. M. DIETZ. 1982. Depletion of bone marrow fat in moose and a correction for dehydra- tion. Journal of Wildlife Management 46:547-551. , and T. N. BAILEY. 1984. Wolves of the Kenai Peninsula, Alaska. Wild- life Monographs 88. SCHWARTZ, C. C. 1998. Reproduction, natality and growth. Pages141-171 in A. W. Franzmann and C. C. Schwartz, editors. Ecology and Management of t h e N o r t h A m e r i c a n M o o s e . Smithsonian Institution Press, Wash- ington, D. C., USA. , and L. A. RENECKER. 1998. Nutri- tion and energetics. Pages 441-478 in A. W. Franzmann and C. C. Schwartz, editors. Ecology and Management of t h e N o r t h A m e r i c a n M o o s e . Smithsonian Institution Press, Wash- ington, D. C., USA. SMITH, H. A., and T. C. JONES. 1961. Veterinary Pathology. Lea and Febiger, Philadelphia, Pennsylvania, USA. SNIDER, J. B. 1980. An evaluation of mandibular marrow fat as an indicator of condition in moose. Proceedings of the North American Moose Confer- ence and Workshop 16:37-50. ZAR, J. H. 1999. Biostatistical Analysis. Fourth edition. Simon and Schuster, Upper Saddle River, New Jersey, USA.