Reconstructing the reproductive history of female polar bears using cementum patterns of premolar teeth

Transcription

1 DOI /s z ORIGINAL PAPER Reconstructing the reproductive history of female polar bears using cementum patterns of premolar teeth Sarah Medill Æ Andrew E. Derocher Æ Ian Stirling Æ Nick Lunn Received: 8 February 2009 / Revised: 30 June 2009 / Accepted: 7 July 2009 Ó Springer-Verlag 2009 Abstract Premolar teeth collected from 220 adult female polar bears (Ursus maritimus) from western Hudson Bay, Canada, were examined to determine whether past reproductive events are recorded in cementum. The widths of annular cementum growth layer groups (GLGs) were measured and compared as proportional width index (PWI) values to correct for age and body size bias. Known reproductive states (pregnant, with cubs, or with yearlings) were used to confirm and calibrate cementum annuli. Significant differences in PWI were observed between GLGs formed the year females were pregnant versus when accompanied by cubs or yearlings. The probability of a female having produced a cub in adulthood was determined by fitting a logistic regression model between the DPWI of females when pregnant and with their cubs. Logistic regression of DPWI (b 0 =-0.229, b 1 = , G 2 = 46.55, df = 1, P \ 0.001) correctly classified the presence or absence of cubs in 72% of GLGs. Cementum width did not vary between different litter sizes. Observations of females with early litter loss suggests that longer periods of lactation contributes to decreased cementum S. Medill A. E. Derocher I. Stirling Department of Biological Sciences, University of Alberta, Edmonton, AB T6G 2E9, Canada I. Stirling N. Lunn Wildlife Research Division, Science and Technology, Environment Canada, Street, Edmonton, AB T6H 3S5, Canada Present Address: S. Medill (&) Department of Environment, Government of Nunavut, Box 209, Igloolik, NT X0A 0L0, Canada smedill@gov.nu.ca width and therefore cementum may record a minimum age of litter survival. Predictions of litter production rate (0.43 litters/female/year) derived from cementum were similar to field observations; however, age at first parturition was underestimated by 1 year. We conclude that patterns of cementum deposition may be useful to determine individual reproductive histories and establish course estimates of reproductive parameters when regular field observations are not feasible. We also conclude that reproductive parameters derived from cementum are not adequate on their own for monitoring populations which are in decline or under stress and field observation should not be replaced under these conditions. Keywords Cementum Growth layer group Incremental line Recording structure Reproduction Ursus maritimus Introduction Changes in reproductive parameters of female polar bears have been documented during long-term research on the population ecology of Western Hudson Bay polar bears. In the 1970s and early 1980s, females often had a 2-year interbirth interval and weaned offspring successfully at 1.5 years of age (Ramsay and Stirling 1982, 1988). Between 1985 and 1992 fewer independent yearlings were observed, and offspring were weaned at 2.5 years of age resulting in the interbirth interval increasing to 3 years (Derocher and Stirling 1995). After 1993, body condition continued to decline which in turn influenced natality rates, survival of cubs, and the proportion of independent yearlings in the annual capture sample (Stirling et al. 1999; Regehr et al. 2007).

2 Dental tissues are highly mineralized; the patterns remain largely unaltered throughout life and post-mortem (Schroeder 1986; Armitage 1991; Lieberman and Meadow 1992; Wedel 2007). Observing layers in dental tissues has long been used to determine the age of wildlife (Eidmann 1932; Scheffer 1950; Laws 1952). Distinct patterns of deposition in cementum and dentine have also been attributed to a number of physiological, environmental, and life history events: growth, fasting, tooth movement, physical condition, climate, food quality, and reproduction (Carrick and Ingham 1962; Craighead et al. 1970; Grue and Jensen 1979; Klevezal 1980; Polson et al. 1984; Coy and Garshelis 1992; Carrel 1994; Lieberman 1994; Klevezal 1996; Cipriano 2002; von Biela et al. 2008). Reconstructing life history events from the deposition of dentine and cementum provides researchers access to both current and historical population parameters. In addition, it could reduce the number of visits to the field and increase the information retrieved from hunter harvested samples. Reproduction and lactation are energetically costly processes (Linzell 1972; Bauman and Currie 1980; Oftedal 1984; Zera and Harshman 2001). Changes in the widths of annual cementum growth layer groups (GLGs; Fig. 1) may be indirectly caused by the physiological demands of reproduction and lactation, or influenced through hormones (Sergeant 1967; Kolb 1978; Rogers 1978). Disruption to regular cementum deposition during pregnancy and lactation has been detected in humans, walrus (Odobenus rosmarus), and ringed seal (Phoca hispida; Mansfield 1958; Stewart et al. 1996; Kagerer and Grupe 2001). A reduction in the width of cementum annuli signifies the production of offspring in sea otter (Enhydra lutris), and black bears (Ursus americanus; Coy and Garshelis 1992; Carrel 1994; Hristienko et al. 2004; von Biela et al. 2008). In the cementum of brown bears (U. arctos) reproductive events could not be identified; however, the authors methods targeted cementum locations which displayed the least variation in width (Matson et al. 1999). Kirkegaard (2003) suggested that decreased width of cementum GLGs in female polar bears (U. maritimus) represented past reproductive events but was unable to verify with field observations of reproductive histories. Multiple captures of female bears from the Western Hudson Bay population of bears allowed the assignment of known reproductive, physiological, and environmental conditions to our interpretations of empirical measurements of cementum GLGs. Based on previous observations of polar bears, and other ursids, females accompanied by cubs of the year should produce narrow GLGs (Coy and Garshelis 1992; Carrel 1994; Kirkegaard 2003). Because cementum formation in a number of species appears correlated to resource availability (Klevezal 1980; Lockyer 1993; Lieberman 1994; Klevezal 1996) annual GLGs Fig. 1 Longitudinal section of polar bear (Ursus maritimus) lower first premolar, decalcified, stained (Toluidine Blue), and viewed using transmitted light (a). Cementum growth layer groups (GLG), representing annual deposition (b). C cementum, D dentine. PDL periodontal ligament, DCJ dentinocemental junction should be wider when females are in good physical condition and narrow when in poor physical condition. Cementum annuli with known reproductive states were used to construct a predictive model to classify the presence or absence of cubs. These predictions were then used to reconstruct complete reproductive histories and estimate reproductive parameters. In this paper, we examine the accuracy of using variation in the structure of GLGs to reconstruct reproductive histories and estimate reproductive parameters such as litter production rate, litter loss, and age of primiparity by comparing cementum-derived values to observed rates and trends. Materials and methods The Western Hudson Bay population of polar bears occupies the sea ice of western Hudson Bay, Canada, and coastal areas of Manitoba, northwestern Ontario, and southern reaches of continental Nunavut. Samples and data collected during the long-term monitoring of this population of polar bears (1965-current) include information on the physiological and reproductive status of polar bears. First premolars were collected from bears during capture, and prepared for age determination (Stirling et al. 1977a, b, 1989; Calvert and Ramsay 1998). GLGs were identified and measured by S. Medill. To prevent misidentification of GLGs, only teeth in which cementum annuli were clear and readily matched either known ages (bear marked at \1-year-old) or the age estimated by experienced technicians were included in any analysis or prediction. Investigation into sources of measurement bias indicated that combining measurements taken from upper and lower first premolars introduced error (Medill et al. 2009). Evaluation was restricted to lower first premolars as they were

3 preferentially sampled in the field and therefore more numerous. Estimate of width for a GLG was the mean of 20 sample measurements obtained from digital images of the central distal aspect of the root (Medill et al. 2009). Standardizing GLG width estimates into Proportional Width Indices (PWI) removes the age-effect of decreased cementum width with increase in age, and allows comparison between individuals and different ages (Medill et al. 2009). Proportional width (PW) for a GLG is calculated as follows: PW i ¼ x i P i 0 x i where x = width (lm) and i = specific GLG or age. The gender-specific mean for each GLG i of all bears in the sample is determined. An individual s PW i divided by the gender-specific mean creates the index value, indicating where an individual s GLG i exists compared to other individuals at that age. To normalize the data for statistical testing individual PWI i is calculated: PWI i ¼ Log10 PW i PW i þ 1 These transformations are similar to the relative width indices created by Carrel (1994) but modified to meet assumptions for statistical testing and large sample sizes (Medill et al. 2009). Reproductive state of the females was confirmed by observation of dependent cubs, yearlings, or 2-year olds or recaptured independent offspring. GLGs were categorized into four reproductive states: Cub if they were known to have given birth during that year, Pregnant for the year immediately prior to the Cub year, Yearling if survival of the cub(s) was confirmed to exceed 12 months, and Unknown where observations were unavailable. In instances where females produced cubs every second year, the yearling year GLG was also the pregnancy year for the subsequent reproductive effort. Because subsequent reproduction may be unknown for some females, females observed with yearlings could potentially be pregnant. The GLG coinciding with the year of capture, and tooth removal, could not be used in analyses because the concurrent annulus would not have been completely formed. Comparison of GLG widths between observed reproductive states Growth layer group widths, represented by PWI values, were compared between observed Pregnant, Cub, and Yearling GLGs of 104 female polar bears. Eight knownaged females were evaluated independently from those whose age was estimated by counts of cementum annuli to monitor potential bias from incorrect age assignment. Cub survival was confirmed either from autumn sightings of the female with offspring or from recapture of offspring as independent yearlings or adults. Mortality of cubs was confirmed by cub absence during autumn recapture of the female, or from the presence of new cubs the following year. GLG for which the survival of the cubs could not be confirmed were excluded from this analysis. Confirmation of yearling survival was based on observation of the yearling with its mother or recapture as an independent bear. Predicting parturition A binary variable Cub (GLG formed when cub(s)\1-yearold are present) or Non-Cub (GLG formed during preceding year of pregnancy) was assigned to GLGs of females with observed reproductive states. GLGs formed when females were accompanied by yearlings were disregarded during model building. For model creation and statistical testing one single reproductive effort per female was considered; a Non-Cub:Cub pair was randomly selected from females with multiple litters. From other studies, the amount of change in the GLG widths of an individual was most indicative of the presence of offspring (Coy and Garshelis 1992; Carrel 1994; von Biela et al. 2008). Therefore, the difference [DPWI = PWI (i) - PWI (i-1) ] was calculated for each GLG. Logistic regression of DPWI for the binary variables Cub and Non-Cub was used to determine the probability of cub presence for any value of DPWI. Probabilities for the presence of cubs, as calculated from the logistic regression model, were applied to the remaining GLG differences (DPWI) where reproductive status was Unknown and to any additional observations of Pregnant, Cub, and Yearling GLGs from females included in the model. The proportion of correctly classified reproductive states based on the threshold probability value (the minimum probability at which DPWI is classified as Cub) was determined at increments of The probability at which sensitivity (correct classification of cub presence) and specificity (correct classification of cub absence) are equal was selected as the probability at which to assign the Cub or Non-Cub classification (Fielding and Bell 1997). Predictions were limited to females C4-years-old; the minimum-age females have been observed with cubs in western Hudson Bay (Ramsay and Stirling 1988; Derocher and Stirling 1995). To assess whether patterns of cementum deposition were similar between males and females, and the possibility of using cementum to determine gender, probabilities were calculated for the DPWI from 119 male bears C4-years-old. The expected proportion of Cub classification should be

4 0; if Cub is predicted it is an indication that similar fluctuations in GLG width occur within the cementum of male polar bears. Evaluation of estimated reproductive parameters The classification of Cub and Non-Cub, based on the probabilities determined from logistic regression, was used to predict age at first parturition, litter production rate, and interbirth intervals from GLGs of all sampled females of reproductive age (C4 years). Predictions of consecutive Cub GLGs were considered an indication of total litter mortality; litter production in consecutive years has been observed in the field when total litter loss occurs before the end of the breeding season and females become pregnant. The first GLG of any prediction of consecutive Cub was considered as a first parturition, but predictions of consecutive Cubs were not considered as interbirth interval. Observations were grouped into 4 periods, , , , and , to allow comparison with published data on litter production rates for the same population (Derocher and Stirling 1995). Factors influencing GLG width of females with cubs Additional information about the individuals from which these teeth were collected permitted investigation into other factors that may influence cementum GLG widths. Litter size, and the nutritional demands of lactation for multiple offspring, may be correlated to differences in cementum width. The number of cubs successfully reared was compared to the PWI for the annuli produced during that reproductive event (t test). Body condition may also influence variation in GLG width of females with cubs; this was examined using linear regression between PWI and a body condition index. The body condition index was created using an estimated mass for the individual obtained from axillary girth measurements (Kolenosky et al. 1989) subtracted from the expected mass as determined by a von Bertalanffy growth curve for polar bears from the Western Hudson Bay population (Derocher 1991). predictions of accompanying cubs (Fig. 2). The eight females of known-age showed significantly reduced GLG widths (PWI) when accompanied by cubs than did those that were pregnant (paired t test, t = 3.73, df = 7, P = 0.007). This trend was also present in the full sample of females with confirmed cub survival. PWIs of cementum GLGs when Pregnant were on average PWI larger than the subsequent Cub GLG (paired t test, t = 8.16, df = 103, P \ 0.001). GLGs from Cub years were PWIs narrower than the following Yearling year when offspring survived C21 months (paired t test, t =-3.88, df = 31, P \ 0.001). GLG width during pregnancy were not significantly different from widths of the corresponding Yearling cementum (paired t test, t = 1.31, df = 31, P = 0.20). There were five recorded instances where entire litters died within 8 months (Table 1). A significant relationship existed between DPWI and the presence or absence of cubs (Logistic regression, b 0 = , b 1 = , G 2 = 46.55, df = 1, P \ 0.001; Fig. 3). Correct classification of Cub or Non-Cub was maximized, without compromising sensitivity or specificity, at a threshold probability equal to The presence or absence of cubs was predicted using DPWI with 72% accuracy (Fig. 4). GLGs deposited in three of the five females that lost their litter prior to 8 months of age were classified as Cub (Table 2). GLGs of litters in which cubs may have survived, but which could not be confirmed, had a lower percentage Cub classifications (65%, n = 17). From 220 females of reproductive age, 594 litters were predicted within 1,387 GLGs (42.8%). When applied to male bears C4 years of age (mean age ± SE = 8.6 ± 0.14 years, Results PWIs were calculated for 1,387 cementum GLGs from 220 females C4 years of age (mean age ± SE = 9.4 ± 0.06 years, range = 4 21 years). Observations of 135 GLGs assigned as Cub were from 104 females (mean age ± SE = 9.8 ± 0.11 years, range = 5 19 years). Small PWIs are associated with reduced GLG widths and lead to Fig. 2 Distal cementum of 11-year-old, known-aged female polar bear (Ursus maritimus). Observed with cubs-of-the-year when growth layer group (GLG) 6, 8, and 10 were formed, cub presence predicted for GLG 6 and 8. PWI proportional width index, observed cubs black arrow, predicted and observed cubs gray arrow

5 Table 1 Year, age, and body condition of female polar bears (Ursus maritimus) when loss of litter was confirmed Bear Year Age Body condition index (spring) Female rebred Probability of cub presence X * X Y X Y X * X Y Probability of females accompanied by cub(s) \1-year-old determined from logistic regression of known cub and non-cub observations against DPWI * No cub present at autumn capture, not known if female rebred range = 4 20 years), 262 out of 645 GLGs (40.6%) had DPWI indicating similar reductions in GLG width are formed in male bears. The range for age of primiparity predicted using cementum annuli was 4 9-years-old with mode = 4 years (mean = 4.9 years, n = 195). For females known to have successfully raise cubs at age 5 and 6 years, only 38% of those litters were predicted as their first parturition. A first parturition event was not predicted for 25 of the females; these were mostly young individuals that may not have given birth before tooth removal but also some older females which should have parented offspring (age 4 n = 6, age 5 n = 9, age 6 n = 5, age 7 n = 3, age 8 n = 1, and age 13 n = 1). There was no difference in the predicted age of first parturition across the first three periods encompassing ; however, there was a significant increase in the proportion of first parturition predicted at age six in (Tukey; q = 3.373, P \ 0.05; Table 3). Predictions of cubs (i.e. litters) in consecutive years occurred 127 times (21% of predictions). A significantly higher proportion of consecutive litter predictions were observed in young females, aged 4 9 years (n = 103), than Probability of predicting COY PWI Fig. 3 Distribution of DPWI for cementum growth layer groups classified as cub (1) and non-cub (0), and the probability of predicting Cub presence based on logistic regression on 104 female polar bears (Ursus maritimus) from western Hudson Bay, Canada females 10 years of age and older (n = 24, v 2 c 0 ¼ 8:17, df = 1, n = 127, P = 0.004; Table 4). About 52% of the consecutive litter predictions followed the predicted age of primiparity. Proportion of consecutive litter predictions were unequal across the four periods with (0.31) greater than (0.17; q = 3.653, P \ 0.05). There was no difference in the proportion of consecutive cub predicted between , , and , nor from , , and The mean litter production rate (lp x ) across all age groups and periods was 0.43 litters/female/year. No difference in litter production rates between sampling periods was predicted using cementum (Table 5). Comparisons of lp x predicted from cementum and those calculated from field observations were not equal when looking at specific age groups but were similar for the population of reproductiveaged females as a whole. Predicted interbirth interval (1/lp x ) from cementum for young adults (age 4 9 years) was 2.3 years, and for adults in their prime (age years) 2.4 years. Proportion correctly classified Probability threshold Fig. 4 Sensitivity (solid line) and specificity (dotted line) for probability threshold, increasing at 0.01 intervals, to classify presence or absence of cubs based on the difference in cementum width (DPWI) of female polar bears (Ursus maritimus) from western Hudson Bay, Canada

6 Table 2 Correct prediction of cub (female accompanied by cub(s)\1-year-old) or non-cub (females not accompanied by cub(s)\1-year-old) of cementum growth layer groups of female polar bears (Ursus maritimus) Reproductive Condition n Proportion correctly classified False positive or false negative error rate Cub: survival confirmed Non-cub: pregnant and yearling Cub: confirmed lost Cub: survival not confirmed Yearling: survival not confirmed Classification based on probability threshold 0.50 for classifying reproductive condition from DPWI Table 3 Predictions of age at first parturition by cementum annuli from female polar bears (Ursus maritimus) from western Hudson Bay, Canada ( ) * = (Tukey; q = 3.373, P \ 0.05) Age (years) v 2 df = 3 n = 58 n = 35 n = 70 n = 32 v 2 P * ? Table 4 Predictions based on cementum annuli width (DPWI) of consecutive years of parturition by female polar bears (Ursus maritimus) in western Hudson Bay, Canada Age No. of GLGs Proportion of GLGs with predictions of consecutive litters Litter size did not influence the width of cementum annuli. No difference in PWI was observed between females accompanied by 1 or 2 cubs (t =-0.252, df = 134, P = 0.801), and there were insufficient observations to compare the PWI in females that had a litter size of 3 (n = 5). Body condition was positively correlated with the PWI when females were accompanied by cubs (PWI = (body condition)? 0.274; F 1,53 = 7.54, P = 0.008, r 2 = 0.12). Discussion To be a useful tool for evaluating past reproductive history, GLG widths need to reliably reflect a female s reproductive history without supporting data from field observations. Our results indicate that a decrease in the width of a cementum GLG in female polar bears can be a strong indication of the production of offspring. The record of cub rearing could be predicted from GLG measurements with 72% accuracy and estimates of litter production rates were overall very similar to field observations. However, predictions of fine scale parameters, including age at primiparity and litter production rates for specific periods of time and age groups, were different from what had been determined from field observations. The discrepancies may be the result of calculating reproductive rates from autumn field observations while the record of cub rearing within cementum may be formed earlier in the year and therefore not account for late litter loss. Dark staining, narrow, incremental lines are formed when activity, feeding, and growth are reduced, while the wide light-staining cementum is produced during rapid gains in mass and structural growth (Craighead et al. 1970; Grue and Jensen 1979; Klevezal 1980). For polar bears in western Hudson Bay, the incremental line is formed

7 Table 5 Litter production rate (lp x ; litters/female/year) of female polar bears (Ursus maritimus) by age-class and sampling period for Western Hudson Bay population Age (years) Source G test lp x n lp x n lp x n lp x n G P 4 9 GLG Capture GLG Capture GLG Capture Overall GLG Capture Source: GLG = predictions based on cementum growth layer group widths (this study). Capture = field observations (reported in Derocher and Stirling 1995) between late summer and autumn, coinciding with periods of fasting, weight loss, and reduced activity associated with the ice-free season (Stirling and Archibald 1977; Stirling et al. 1977b; Watts and Hansen 1987; Ramsay and Stirling 1988; Ramsay et al. 1991; Arnould and Ramsay 1994; Atkinson et al. 1996). The change in cementum width from a wide GLG during the year females become pregnant to the thin cementum GLG produced when accompanied with cubs is potentially linked to a number of environmental and physiological variables. Pregnant females, instead of returning to the sea-ice when it forms in autumn, enter maternity dens where implantation, gestation, birth, and initial lactation will occur without access to food and water (Ramsay and Dunbrack 1986; Watts and Hansen 1987; Nelson 1987). Pregnant females must put on significant weight before they come on to land in order to withstand this eight months fast and the additional stresses of cub rearing (Derocher et al. 1992; Atkinson and Ramsay 1995; Derocher and Stirling 1995). Our research indicated a positive correlation between body condition and the production of broader cementum annuli. Additionally, female polar bears increase bone mass during their year of pregnancy to buffer against loss to disuse, fetal development, and lactation (Lennox and Goodship 2008). Similarities in insulin-like growth factor-1 and insulin-like growth factors concentrated in the cementum (Ikezawa et al. 1997) suggest a possible connection between an increase in bone formation and cementum deposition during pregnancy (Naylor et al. 2000; Lennox and Goodship 2008). Calcium demands of lactation are likely the greatest contributing factor to creating reduced cementum annuli when cubs of the year are present. Females caring for cubs (\1-year-old) transfer nearly 4 times more milk energy per day than those accompanied by yearlings or 2-year-olds (Arnould and Ramsay 1994). The period of time pregnant females remain inactive during late winter and early spring, in addition to the demands of gestation and lactation, may delay the apposition of light cementum and facilitate the formation of reduced annuli (Rausch 1969; Rogers 1978; Klevezal 1980). Litter size was not significantly correlated with cementum annuli width; therefore, cementum will reflect only the presence or absence of whole litters. The change in cementum width for two of three females that lost litters early enough for the female to rebreed were classified as Non-Cub, and the third was not a strong probability of cubs present (Table 1). Research on black bears indicated that a reduction in the GLG width in female cementum occurs only when cubs are nursed for a prolonged period (Coy and Garshelis 1992; Carrel 1994; Hristienko et al. 2004). Cub mortality at B4 months of age resulted in Non-Cub cementum patterns (Coy and Garshelis 1992). Early litter loss, before cub rearing is recorded in cementum, could explain the lower percentage of Cub predicted for females with unconfirmed litter survival. Our data also provide evidence that a minimum period of cub rearing may be required before female polar bears develop a detectable difference in GLG width. The litter production rate derived from cementum was similar to the estimate using field capture data; however, the observed trend of decreased litter production rates from 1985 to 1992 was not detected (Table 5; Derocher and Stirling 1995). It is possible that the litter production rates determined using cementum are only similar to autumn field observations when cub survival is high but overestimates recruitment when cub survival is low. Derocher and Stirling (1995) calculated litter production based on the observations of litters present during autumn captures and would have missed incidents of whole litter loss occurring prior to capture. Whole litter loss would also increase the interbirth interval based on autumn capture observations. If cementum is reduced by 4 5 months of cub rearing, litter

8 mortality occurring after that time would result in cementum widths that indicate cub presence. Shorter interbirth intervals predicted by cementum GLGs do not contradict the observations of increased litter loss in the last two decades, and agree with observations of stable pregnancy rates measured from serum progesterone levels (Derocher et al. 1992; Derocher and Stirling 1995; Stirling et al. 1999; Regehr et al. 2007). Evaluating cementum GLG widths was not successful in identifying age at first parturition. The predicted age of primiparity from cementum identified that most bears gave birth first at four years of age. Observations of 4-year-old females rearing cubs in western Hudson Bay exist; however, most females were observed to have their first litter at 5 years (Ramsay and Stirling 1982, 1988; Derocher et al. 1992; Derocher and Stirling 1995, 1998). Different methods of measuring and comparing cementum widths have detected age at primiparity in sea otters, but it was not evaluated whether complete reproductive histories could be reconstructed (von Biela et al. 2008). Young polar bears (0 4 years) produce wider cementum layers with greater variation in widths compared to when they reach maturity (Medill et al. 2009). PWIs are designed to account for agerelated changes, but the transition between high and low variation that exists between bears at 3 and 4 years of age may influence DPWI during this critical time. Predictions of consecutive Cub (i.e. litters) would occur when cementum annuli remain consistent or decrease in actual width and could be interpreted as early litter mortality and rebreeding by the female. However, this is complicated by observations that prolonged lactation is likely required to decrease GLG widths, and that females with nursing cubs would not be receptive to rebreeding. It is possible that other factors, along with litter loss, could explain the high proportion of consecutive cub predictions. Dental trauma, skeletal trauma or renal disease can directly influence cementum deposition and have the potential to contribute to the predictions of cub presence (Polson et al. 1984; Kagerer and Grupe 2001). Predictions of consecutive litters were most frequent in young adult females where body condition is typically lower; this could reduce cub survival, but also limit the ability of females to deposit wide layers of cementum (Table 4; Derocher 1991; Derocher et al. 1992; Atkinson and Ramsay 1995; Derocher and Stirling 1996; Polischuk et al. 2001). Several studies warn of the risk of mistaking false or accessory incremental lines, which can be present in both males and females, for true incremental lines (Rogers 1978; Grue and Jensen 1979; Coy and Garshelis 1992; Klevezal 1996; Kirkegaard 2003; Christensen-Dalsgaard 2006). In research involving the measurement of GLGs, mistaking false annuli as true incremental lines would result in an incorrect assignment of age and/or year to a GLG, and an incorrect estimate of width. Many false incremental lines are fainter than primary incremental lines, inconsistent around the surface of the root, merge, and/or fade (Coy and Garshelis 1992; Klevezal 1996; Calvert and Ramsay 1998). For these reasons, false incremental lines are easier to detect in the longitudinal sections of teeth. Male ursids produce false incremental lines in the cementum more commonly than females; these false annuli range from being faint to well defined, but should not be considered as two separate GLGs (Coy and Garshelis 1992; Klevezal 1996; Kirkegaard 2003). In this study, the person measuring (S. Medill) was alert for false incremental lines, and these should not have influenced GLG width. We detected a reduction in GLG width that creates similar patterns in males as in females. False incremental lines have been hypothesized to reflect mating effort in male polar bears, similar to a rutting line observed in ungulates (Low and Cowan 1963; Mitchell 1967; Reimers and Nordby 1968; Coy and Garshelis 1992; Christensen- Dalsgaard 2006). During the mating season, male polar bears may reduce hunting and food intake while courting females and develop blood urea creatinine ratios similar to a fasting individual (Ramsay et al. 1991). The mating season for polar bears can extend from mid-february into July, with peak activity occurring March-May (Rosing- Asvid et al. 2002; Spady et al. 2007). If feeding by male bears is compromised by mate seeking, courting, defending females from other males, or if skeletal trauma or injury occurs during these tasks, the formation of false incremental lines recording these periods of stress could explain fluctuations in the overall width of cementum annuli (Lockyer 1993; Kagerer and Grupe 2001). Data regarding mating activity of male polar bears have not been collected to allow comparisons of mating effort, success, injury, or fasting duration, against cementum GLG width. Additionally, false incremental lines have been observed in juvenile bears (Lønø 1970; Hensel and Sorensen 1980) which cannot be explained by reproductive stress. Nutritional stress during intense periods of growth caused by periods of low food availability or underdeveloped hunting skills could potentially explain the development of these lines (Mitchell 1967; Stirling and Archibald 1977; Lockyer 1993). Management programs often incorporate the collection of teeth to monitor population demographics either from live capture or hunter harvested samples. Teeth are submitted from hundreds of polar bears harvested by native hunters in Canada (Lee and Taylor 1994; Schliebe et al. 1999). Many specimens are from regions outside of those regularly sampled by researchers. Standardizing the collecting, processing, and sampling of teeth is required to obtain the most accurate information from cementum (Medill et al. 2009). Improvements in technology and

9 software, for both micro-imaging and measurement, have made it easier to collect morphometric data from sectioned teeth. A reduced width of cementum annuli can be associated with the rearing of cubs by female polar bear. Similar to black bears, and dolphins (Genus Stenella), prolonged lactation is likely necessary to record cub rearing in the dental tissues of polar bear teeth (Klevezal and Myrick 1984; Coy and Garshelis 1992; Carrel 1994). Predicting the presence and absence of cubs from changes in cementum width is possible, but not error free. Cementum formation is regulated by a number of endogenous and exogenous factors which may counteract and mask the evidence of cub rearing in cementum. Several of the finer parameters, such as age of primiparity and whole litter loss were not adequately reflected in cementum patterns; additionally, trends which have indicated a population decline for Western Hudson Bay polar bears were not detected (Derocher and Stirling 1995). For this reason, the use of cementum annuli to estimate reproductive parameters should be routinely verified with field observations. Management decisions based on cementum-derived reproduction parameters should proceed with a high degree of caution and the knowledge that cementum does not reflect recruitment of adult individuals. Acknowledgments We thank the many personnel from the Canadian Wildlife Service and Manitoba Conservation who collected data and teeth from the Western Hudson Bay polar bear population since the initiation of the monitoring program. Special acknowledgment to D. Andriashek, W. Calvert, M. Kay, and C. Spencer for their assistance in the laboratory, preparation of samples, and management of the tooth archive and data, and the initial age determination of bears. Additionally, SM is grateful to Canadian Wildlife Service for the access to the archived teeth collection, associated data, laboratory space, and equipment. Additional financial support for this project came from the University of Alberta, Canadian Circumpolar Institute, (SM), and from the Natural Sciences and Engineering Research Council (AED). Funding and support of the long-term research in western Hudson Bay has been provided by the Canadian Wildlife Federation, Environment Canada, Manitoba Conservation, Manitoba Sustainable Development Innovations Fund, ArcticNet, Care for the Wild International, the Nunavut Wildlife Research Trust Fund, Parks Canada Agency, Polar Bears International, and World Wildlife Fund Canada, and World Wildlife Fund International Arctic Programme. References Armitage GC (1991) Cementum. In: Bhaskar SN (ed) Orban s oral histology and embryology, 11th edn. Mosby Year Book, St. Louis, pp Arnould JPY, Ramsay MA (1994) Milk production and milk consumption in polar bears during the ice-free period in western Hudson Bay. Can J Zool 72: Atkinson SN, Ramsay MA (1995) The effects of prolonged fasting of the body composition and reproductive success of female polar bears (Ursus maritimus). Funct Ecol 9: Atkinson SN, Nelson RA, Ramsay MA (1996) Changes in the body composition of fasting polar bears (Ursus maritimus): the effect of relative fatness on protein conservation. Physiol Zool 69: Bauman DE, Currie WB (1980) Partitioning of nutrients during pregnancy and lactation: a review of mechanisms. J Dairy Sci 63: Calvert W, Ramsay MA (1998) Evaluation of age determination of polar bears by counts of cementum growth layer groups. Ursus 10: Carrel WK (1994) Reproductive history of female black bears from dental cementum. Int Conf Bear Res Manage 9: Carrick R, Ingham SE (1962) Studies on the southern elephant seal, Mirounga leonina (L.). CSIRO Wildl Res 7: Christensen-Dalsgaard SN (2006) Temporal patterns in age structure of polar bears (Ursus maritimus) in Svalbard, with special emphasis on validation of age determination, M.Sc. Thesis. University of Tromsø, Norway Cipriano A (2002) Cold stress in captive great apes recorded in incremental lines of dental cementum. Folia Primatol 73:21 31 Coy PL, Garshelis DL (1992) Reconstructing reproductive histories of black bears from the incremental layering in dental cementum. Can J Zool 70: Craighead JJ, Craighead FC, McCutchen HE (1970) Age determination of grizzly bears from fourth premolar tooth sections. J Wildl Manage 34: Derocher AE (1991) Population dynamics and ecology of polar bears in western Hudson Bay, Ph.D. Thesis. University of Alberta, Canada Derocher AE, Stirling I (1995) Temporal variation in reproduction and body mass of polar bears in western Hudson Bay. Can J Zool 73: Derocher AE, Stirling I (1996) Aspects of survival in juvenile polar bears. Can J Zool 74: Derocher AE, Stirling I (1998) Maternal investment and factors affecting offspring size in polar bears (Ursus maritimus). J Zool (Lond) 245: Derocher AE, Stirling I, Andriashek D (1992) Pregnancy rates and serum progesterone levels of polar bears in western Hudson Bay. Can J Zool 70: Eidmann H (1932) Alterserscheinungen am Gebiß des Rothirsches (Cervus elaphus L.). Mitt Forstwirt Forstwissen 3: Fielding AH, Bell JF (1997) A review of methods for the assessment of prediction errors in conservation presence/absence models. Environ Conserv 24:38 49 Grue HE, Jensen B (1979) Review of the formation of incremental lines in tooth cementum of terrestrial mammals. Dan Rev Game Biol 11:1 48 Hensel RJ, Sorensen FE Jr (1980) Age determination of live polar bears. Int Conf Bear Res Manage 4: Hristienko H, Pastuck D, Rebizant KJ, Knudsen B, Connor ML (2004) Using reproductive data to model American black bear cub orphaning in Manitoba due to spring harvest of females. Ursus 15:23 34 Ikezawa K, Hart CE, Williams DC, Narayanan AS (1997) Characterization of cementum derived growth factor as an insulin-like growth factor-1 like molecule. Connect Tissue Res 36: Kagerer P, Grupe G (2001) Age-at-death diagnosis and determination of life-history parameters by incremental lines in human dental cementum as an identification aid. Forensic Sci Int 118:75 82 Kirkegaard M (2003) History of selected immunological organs in east Greenland polar bears in relation to levels of organohalogens and age determination and use of dental structures to determine female reproductive history in east Greenland polar bears (Ursus maritimus), M.Sc. Thesis. Royal Veterinary and Agricultural University, Copenhagen, Denmark

10 Klevezal GA (1980) Layers in the hard tissues of mammals as a record of growth rhythms of individuals. Rep Int Whal Comm Spec Issue 3:89 94 Klevezal GA (1996) Recording structures of mammals: determination of age and reconstruction of life history. A.A. Balkema, Rotterdam Klevezal GA, Myrick AC Jr (1984) Marks in tooth dentine of female dolphins (Genus Stenella) as indicators of parturition. J Mammal 65: Kolb H (1978) The formation of lines in the cementum of premolar teeth in foxes. J Zool 185: Kolenosky GB, Lunn NJ, Greenwood CJ, Abraham KF (1989) Estimating the weight of polar bears from body measurements. J Wildl Manage 53: Laws RM (1952) A new method of age determination for mammals. Nature 169: Lee J, Taylor M (1994) Aspects of the polar bear harvest in the Northwest Territories, Canada. Int Conf Bear Res Manage 9: Lennox AR, Goodship AE (2008) Polar bears (Ursus maritimus), the most evolutionary advanced hibernators, avoid significant bone loss during hibernation. Comp Biochem Physiol A 149: Lieberman DE (1994) The biological basis for seasonal increments in dental cementum and their application to archaeological research. J Archaeol Sci 21: Lieberman DE, Meadow RH (1992) The biology of cementum increments (with an archaeological application). Mammal Rev 22:57 77 Linzell JL (1972) Milk yield, energy loss in milk, and mammary gland weight in different species. Dairy Sci Abstr 34: Lockyer C (1993) A report on patterns of deposition of dentine and cement in teeth of pilot whales, genus Globicephala. Rep Int Whal Comm Spec Issue 14: Lønø O (1970) The polar bear (Ursus maritimus Phipps) in the Svalbard area. Norsk Polarinst Skrifter 149:1 103 Low WA, Cowan IM (1963) Age determination of deer by annular structure of dental cementum. J Wildl Manage 27: Mansfield AW (1958) The biology of the Atlantic walrus Odobenus rosmarus rosmarus (Linnaeus) in the eastern Canadian Arctic. Fish Res Board Can Rep Ser 653:1 145 Matson GM, Casquilho-Gray HE, Paynich JD, Barnes VG Jr, Reynolds HV, Swenson JE (1999) Cementum annuli are unreliable reproductive indicators in female brown bears. Ursus 11: Medill S, Derocher AE, Stirling I, Lunn N, Moses RA (2009) Estimating cementum width in polar bears: identifying sources of variation and error. J Mammal 90(5) (in press) Mitchell B (1967) Growth layers in dental cement for determining the age of red deer (Cervus elaphus L.). J Animal Ecol 36: Naylor KE, Iqbal P, Fledelius C, Fraser RB, Eastell R (2000) The effect of pregnancy on bone density and bone turnover. J Bone Mineral Res 15: Nelson RA (1987) Black bears and polar bears still metabolic marvels. Mayo Clin Proceed 62: Oftedal OT (1984) Milk composition, milk yield, and energy output at peak lactation: a comparative review. Symp Zool Soc Lond 51:33 85 Polischuk SC, Hobson KA, Ramsay MA (2001) Use of stable-carbon and -nitrogen isotopes to assess weaning and fasting in female polar bears and their cubs. Can J Zool 79: Polson A, Caton J, Polson AP, Nyman S, Novak J, Reed B (1984) Periodontal response after tooth movement into intrabony defects. J Periodontol 55: Ramsay MA, Dunbrack RL (1986) Physiological constraints on life history phenomena: the example of small bear cubs at birth. Am Nat 127: Ramsay MA, Stirling I (1982) Reproductive biology and ecology of female polar bears in western Hudson Bay. Nat Can (Que) 109: Ramsay MA, Stirling I (1988) Reproductive biology and ecology of female polar bears (Ursus maritimus). J Zool (Lond) 214: Ramsay MA, Nelson RA, Stirling I (1991) Seasonal changes in the ratio of serum urea to creatinine in feeding and fasting of polar bears. Can J Zool 69: Rausch RL (1969) Morphogenesis and age-related structure of permanent canine teeth in brown bear, Ursus arctos L., in arctic Alaska. Z Morph Tiere 66: Regehr EV, Lunn NJ, Amstrup SC, Stirling I (2007) Effects of earlier sea ice breakup on survival and population size of polar bears in western Hudson Bay. J Wildl Manage 71: Reimers E, Nordby Ø (1968) Relationship between age and tooth cementum layers in Norwegian reindeer. J Wildl Manage 32: Rogers LL (1978) Interpretation of cementum annuli in first premolars of bears. Eastern Workshop Proc Black Bear Manage Res 4: Rosing-Asvid A, Born EW, Kingsley MCS (2002) Age at sexual maturity of males and timing of the mating season of polar bears (Ursus maritimus) in Greenland. Polar Biol 25: Scheffer VB (1950) Growth layers on the teeth of pinnipedia as an indication of age. Science 112: Schliebe SL, Evans TJ, Fischbach AS, Cronin MA (1999) Using genetics to verify sex of harvested polar bears: management implications. Wildl Soc Bull 27: Schroeder HE (1986) The periodontium. Springer-Verlag Berlin Heidelberg, Germany Sergeant DE (1967) Age determination of land mammals from annuli. Z Saugetierkd 32: Spady TJ, Lindburg DG, Durrant BS (2007) Evolution of reproductive seasonality in bears. Mammal Rev 37:21 53 Stewart REA, Stewart BE, Stirling I, Street E (1996) Counts of growth layer groups in cementum and dentine in ringed seals (Phoca hispida). Mar Mammal Sci 12: Stirling I, Archibald WR (1977) Aspects of predation of seals by polar bears. J Fish Res Board Can 34: Stirling I, Archibald WR, DeMaster D (1977a) Distribution and abundance of seals in the eastern Beaufort Sea. J Fish Res Board Can 34: Stirling I, Jonkel C, Smith P, Robertson R, Cross D (1977b) The ecology of the polar bear (Ursus maritimus) along the western coast of Hudson Bay. Can Wildl Serv Occ Paper 33:1 64 Stirling I, Spencer C, Andriashek D (1989) Immobilization of polar bears (Ursus maritimus) with Telazol in the Canadian arctic. J Wildl Dis 25: Stirling I, Lunn NJ, Iacozza J (1999) Long-term trends in the population ecology of polar bears in western Hudson Bay in relation to climatic change. Arctic 52: von Biela VR, Testa JW, Gill VA, Burns JM (2008) Evaluating cementum to determine past reproduction in northern sea otters. J Wildl Manage 72: Watts PD, Hansen SE (1987) Cyclic starvation as a reproductive strategy in polar bear. Symp Zool Soc Lond 57: Wedel VL (2007) Determination of season at death using dental cementum increment analysis. J Forensic Sci 52: Zera AJ, Harshman LG (2001) The physiology of life history tradeoffs in animals. Annu Rev Ecol Syst 32:95 126