White-tailed deer are large, generalized herbivores with a widespread distribution across southern Canada, North America, and northern South America (Halls, 1984). While individual deer may live up to 20–25 years, they rarely live longer than 2–3 years in the wild (Halls, 1984; Smith, 1991). Primary food items include forbs, grasses, shrubs, fruit, acorns, and leaves (Vangilder et al., 1982; Fritzell, 2017; McGovern et al., 2020). Although they are dietary generalists, white-tailed deer will narrow their diet to preferred food species contingent on the season (Zagata and Haugen, 1973; Huegel et al., 1985; Fritzell, 2017). Deer from central Iowa also feed within agricultural fields, including cultivated crops during the growing season from June through October and agricultural residue during the winter. Major crops in Iowa include maize (Zea mays) and soybeans (Glycine max).

White-tailed deer tend to remain within a well-defined home range. The size of their home range and the extent of their mobility, however, depends on local habitat, water, and food availability and quality (Huegel et al., 1985; McGovern et al., 2020). While buck fawns and adult bucks tend to have a greater home range than female fawns and does, the general habitat range of deer in Iowa is typically around 80 hectares with daily movement of ~2 km, depending on the season (Zagata and Haugen, 1973). Scarcity of food and breeding activity increases home ranges for fall and winter, while they remain rather small in the spring and summer because of abundant food supplies and fawn rearing. Even though deer tend to remain in the area where they were born, spring extension is observed, with some individuals being documented to travel as far as 80 km away from their place of birth. Furthermore, deer are willing to move to a different core area if danger is felt, generally because of hunting, predator-prey dynamics, and infrastructural expansion (Zagata and Haugen, 1973; Huegel et al., 1985; Fritzell, 2017; McGovern et al., 2020).

Typically, only white-tailed deer males grow antlers. They occur in a variety of shapes and sizes from small and unbranched to large and branched (approximate range of length from 7 to 70 cm) (French et al., 1956; Grasman and Hellgren, 1993; Price et al., 2005; Dryden, 2016; Asleson et al., 2017). The size of antlers is determined by their phase of development, genetics, and by the age of the individual (Scribner et al., 1989). Generally, the size of antlers increases each year of an individual's life until about 5 years of age, at which point the antlers have reached their maximum size (Hewitt et al., 2014). Unlike horns that are made of bone covered by a layer of keratin and are permanently part of the animal, antlers are made of bone, temporarily covered in a velvet membrane, and are shed and regrown annually (Moen et al., 1999; Price et al., 2005; Kierdorf and Kierdorf, 2010). Trabecular bone (transporting nutrients and growth-regulating hormones) and cortical bone (compact bone that forms the hard outer structure) comprise the two types of bone within an antler (Banks, 1974; Gomez et al., 2013). Both bone types are composed of organic collagen and mineral hydroxyapatite.

The process of antler growth is defined as modified endochondral ossification and is described as the remodeling of cartilage by osteoclasts at the tip of each branch (Allen et al., 2002; Kierdorf and Kierdorf, 2010; Rössner et al., 2021). This regeneration process is later constituted of micro-lamellar bone and then primary osteons (Faucheux et al., 2001; Price et al., 2005). Antlers grow from pedicles as an extension of the frontal bone following an annual growth cycle starting and ending around late February to early March. During the growth phase, antlers are covered by a fine and soft membrane called velvet that supplies blood and nutrients to the antler. A steady antler growth from the tip (~1.3 cm per week) during April and May contrasts with the rapid growth that occurs in June and July (~5 cm per week). This inconsistent growth rate is marked by a circulating testosterone increase during the mineralization process from late summer to early fall, causing the shed of velvet (French et al., 1956; Fennessy and Suttie, 1985; Asleson et al., 2017). As the antler mineralizes, chondroclast cells resorb cartilage while osteoblast cells secrete new bone tissue. During the winter, osteoclast cells demineralize along the abscission line situated on the coronet, allowing the casting of antlers. Importantly, compact cortical bone is the most resistant to diagenesis, and thus mature antlers can preserve over long periods of time in the archaeological record (Moen et al., 1999).

Because white-tailed deer antlers are made of bone, they may be analyzed for stable isotope ratios of carbon, nitrogen, and oxygen in a similar manner as other skeletal elements. For carbon, the primary factor that produces variation in δ¹³C values within a given environment is the differences in photosynthetic pathways of the plants at the base of the ecosystem. Atmospheric CO₂ is fixed in plants through C3, C4, and Crassulacean acid metabolism (CAM) photosynthetic pathways. These photosynthetic pathways result in varying δ¹³C values within plant tissues, with C3 plants exhibiting relatively low values, C4 plants exhibiting relatively high (less negative) values, and CAM plants varying between these two depending on environmental circumstances (Bender, 1968; Smith and Epstein, 1971; Borland et al., 2011). Within C3 plants, additional variation in δ¹³C values can occur due to the influences of precipitation, humidity, temperature, amount of sunlight, and salinity of the soil (Tieszen, 1991; Kohn, 2010; Bonafini et al., 2013; Schoeninger et al., 2016). The δ¹³C values of herbivore bone reflect the δ¹³C values of consumed plants, where lower δ¹³C values are indicative of greater C3 dietary intake and higher δ¹³C values indicative of greater C4 and or CAM plant intake (DeNiro and Epstein, 1978). Stable carbon isotope values can be obtained from two phases of antler bone: the organic collagen phase (δ¹³C_col) and the mineral bioapatite phase (δ¹³C_apa). δ¹³C_apa values reflect the δ¹³C values of dietary inputs from all macronutrient sources (i.e., carbohydrate, protein, lipids), while δ¹³C_col values are biased toward the δ¹³C values of dietary protein sources (Ambrose and Norr, 1993; Froehle et al., 2010). Within white-tailed deer tissues, stable carbon isotope values in both collagen and apatite are thus the product of the amount of maize (predominant agricultural byproduct of Iowa) relative to non-maize plant consumption as well as the seasonal fluctuation in C3 plant δ¹³C values. While native C4 grasses do occur in Iowa, the vast areas of land dedicated to maize agriculture lead us to believe that any C4 signal detected within antler stable isotope values likely originated with consumption of agricultural maize.

δ¹⁵N values of mammalian bone are primarily influenced by the δ¹⁵N values from dietary protein sources (DeNiro and Epstein, 1981; Schoeninger and DeNiro, 1984). δ¹⁵N values of animal tissues reflect the position of an individual within the trophic system, with plants having the lowest values and carnivores having the highest values (Minagawa and Wada, 1984; Robinson, 2001; Bocherens and Drucker, 2003). For herbivores, stable nitrogen isotope values in skeletal tissue can be used to reflect aspects of the local environment (Somerville et al., 2018, 2020). δ¹⁵N values of organic tissues of herbivores reflect the δ¹⁵N values of the plants they consume, which are themselves influenced by soil δ¹⁵N values. Globally, δ¹⁵N values of soil and the plants that grow from it are influenced by temperature and precipitation (Murphy and Bowman, 2006, 2009; Hartman, 2011; Craine et al., 2015a,b), and such differences in δ¹⁵N values between environmental zones is incorporated into the tissues of organisms that consume them. Generally, organisms feeding in cool, wet environments exhibit the lowest δ¹⁵N values while organisms feeding in hot, dry environments exhibit the highest δ¹⁵N values (Ambrose, 1991; Murphy and Bowman, 2009; Hartman, 2011; Somerville et al., 2018). An additional source of variation in nitrogen values comes from human inputs to the soil. Plants growing out of soil supplemented by manuring and fertilizers show an increase in δ¹⁵N values, which can then be transferred to the tissues of animals that consume them (Fraser et al., 2011; Szpak et al., 2012).

δ¹⁸O values in mammalian bone tissue reflect the δ¹⁸O values of the water consumed by the organism (Luz et al., 1984; Ayliffe and Chivas, 1990; Huertas et al., 1995). Because the δ¹⁸O value of oxygen from the air is relatively constant worldwide (Kroopnick and Craig, 1972), the primary source of variation in δ¹⁸O values of mammalian skeletal tissue is the variation of δ¹⁸O values of consumed water, either imbibed or consumed with food (Huertas et al., 1995; Levin et al., 2006). Factors such as altitude, distance from the sea, temperature, precipitation, and relative humidity can influence the δ¹⁸O values of precipitation and hence meteoric water sources (Longinelli, 1984; Ayliffe and Chivas, 1990; Poage and Chamberlain, 2001). For meteoric water in a given location, the strongest influence on δ¹⁸O values is the local temperature, which implies that δ¹⁸O values in the skeletal tissue of obligate drinking organisms will correlate with local temperature variation (Dansgaard, 1964; Iacumin et al., 1996; Hallin et al., 2012). For non-obligate drinking organisms that obtain water primarily from the plants they consume, skeletal δ¹⁸O values correlate more strongly with local relative humidity and precipitation levels than temperature as leaf water is subject to the additional oxygen isotope fractionation process of evapotranspiration that results in ¹⁸O enrichment in leaf water during dryer/warmer periods (Burk and Stuiver, 1981; Ayliffe and Chivas, 1990; Levin et al., 2006). The magnitude of the change in oxygen isotope ratios is influenced by relative humidity and precipitation, with higher humidity and greater precipitation leading to lower δ¹⁸O values in leaf water and hence tissues of animals that consume them (Huertas et al., 1995; Levin et al., 2006; Somerville et al., 2018). White-tailed deer's primary source of water is leaf water, but they will drink surface water on occasion, particularly when temperatures are high (Michael, 1968). δ¹⁸O values from deer bone then are influenced both by changes in both local moisture and temperature (Luz et al., 1990).

Previous studies have demonstrated the potential of the isotopic ratios of C, N, and O in bone and enamel of cervids to reflect aspects of diet and the environmental context in which they lived. For δ¹³C values, early analyses of bone collagen of modern Odocoileus virginianus specimens from North America found a positive relationship between mean temperature and δ¹³C values of bone collagen, presumably due to the greater consumption of C4 plants during warm summer months (Cormie and Schwarcz, 1996). Research on red deer (Cervus elaphus) populations in European environments that do not contain C4 plants, however, observed a negative relationship between temperature and δ¹³C values (Stevens et al., 2006). Despite the ambiguous general relationship between temperature and stable carbon isotope values, studies on red deer δ¹³C values across the late Pleistocene to early Holocene find significantly higher values during the cool/dry Pleistocene than in the warm/wet Holocene, demonstrating the ability of these values to reflect general changes in properties of the landscape over time (Drucker et al., 2003; Hedges et al., 2004).

For δ¹⁵N values, a negative correlation between precipitation levels and δ¹⁵N values in bone collagen of modern Odocoileus virginianus of North America was observed, but only when individual deer had consumed > 10% C4 plants (Cormie and Schwarcz, 1996). Analyses of δ¹⁵N values from red deer bone collagen from Europe, however, did not find a significant relationship with precipitation levels; instead they found a strong positive correlation between δ¹⁵N values and temperature from each site (Stevens et al., 2006). Such varying results highlight the complexity of interpreting mammalian nitrogen isotope values and the challenges inherent to cross-regional comparisons.

In addition to bone, previous studies have documented the potential of cervid enamel to serve as an archive of paleoenvironmental information. Both studies of white-tailed deer (Rivera-Araya and Pilaar Birch, 2018) and red deer (Stevens et al., 2011) have found correlations between patterns of change in δ¹⁸O values from intra-tooth sub-samples and those of seasonal change in δ¹⁸O values of local precipitation, which are themselves strongly influenced by changes in temperature. In deer, enamel mineralizes during the first 1–3 years of development and captures the isotopic signature of the food and water consumed during that time. This process, however, necessitates caution when interpreting enamel δ¹⁸O values as developmental changes such as nursing and weaning can be captured in the enamel signature and thus obscure environmental signals, particularly in the first molar (Malasek et al., 2023).

Isotopic values of cervid antlers, like bone and enamel, reflect aspects of their diet and environment. Research on white-tailed deer with known diets shows a consistent offset between diet and antler δ¹³C and δ¹⁵N values of +4.5 ± 0.16‰ and +4.29 ± 0.29‰, respectively (Darr and Hewitt, 2008). Research that conducted inter-antler analysis of collagen stable carbon and nitrogen isotope values found that bone isotope values and antler isotope values are not in equilibrium and that considerable variation could exist within antlers (Stevens and O'Connell, 2016). The present study attempts to explain to what extent intra-antler variation of white-tail deer antlers can be explained by seasonal variation in diet, temperature, and precipitation variables.

The sampled population is comprised of four antlers from wild deer from central Iowa (Figure 1). Three of the specimens were collected by the Iowa Department of Natural Resources (DNR) and one was collected opportunistically from a deer discovered by the authors. Each of the sampled antlers was collected during the fall months and was still attached to the skull at the time of the death. The locations of specimens are displayed in Figure 1. The first antler, PEL-0090, was collected from a hunted white-tailed deer in Jasper County, Iowa. The length (measured along the posterior side) and weight are 50.7 cm and 191 g, respectively. The deer was a juvenile of ~1.5 years of age at its time of death (December 5th, 2016). The antler was donated to this research project by the Iowa DNR.

Antler PEL-0114 was collected from a hunted white-tailed deer in Wright County, Iowa. The length (measured along the posterior side) and weight are 50.5 cm and 185 g, respectively. The deer was a juvenile of ~1 years old at its time of death (December 3rd, 2016). The antler was donated to this research project by the Iowa DNR.

Antler PEL-0164 was collected from a hunted white-tailed deer in Boone County, Iowa. The length (measured along the posterior side) and weight are 68.6 cm and 366.7 g, respectively. The deer was an adult of ~2.5 years of age at its time of death (late October 2019). The antler was donated to this research project by the Iowa DNR.

Finally, antler PEL-0165 was collected from a discovered dead white-tailed deer from Story County, Iowa by ADS during November 2020. The length (measured along the posterior side) and weight are 21 cm and 63.2 g, respectively. The deer was a juvenile at its time of death.

Ten samples of collagen and apatite were obtained for each antler, for a total of 80 samples. Collagen and apatite samples each yielded two isotope variables (δ¹³C_col and δ¹⁵N_col for collagen and δ¹³C_apa and δ¹⁸O_apa for apatite), for a total of 160 data points that were investigated in this study. To extract powdered antler tissue, the outer layer of the antler that was in direct contact with the external environment was first removed by ablating the surface with a hand-held Dremel tool equipped with a diamond engraving bit before sample extraction. All samples for analysis were then collected from these cleaned areas by extracting powdered samples using a Dremel tool with a clean diamond engraving bit along the posterior side of the antlers from the proximal end to the distal end.

The samples were prepared for the procedure of carbon and nitrogen isotope analysis of collagen by modifying established methods (Longin, 1971; Sealy et al., 2014). Samples of ~50 mg of cortical antler bone were powdered and weighed out into 15 mL polypropylene tubes. Approximately 10 mL of 0.25 M HCl was added to each tube for demineralization, which were then mixed on the touch mixer for 45–60 s. The caps were then loosened, and the tubes sat at room temperature for 48 h. Samples were then rinsed with ultrapure H₂O to neutralize. Because all samples were modern, the alkaline soak in sodium hydroxide (NaOH) solution and the filtration steps were skipped. Tubes with collagen gelatin were placed with closed caps in an oven at 85°C for 24 h to solubilize. Finally, the collagen samples were frozen and lyophilized in a freeze dryer for 24 h.

The samples were prepared for the procedure of carbon and oxygen isotope analysis of carbonate in bone mineral apatite (Ca₁₀([PO₄, CO₃])₆(OH)₂) by following a modified version of established procedures that are commonly used on archaeological specimens (Koch et al., 1997; Crowley and Wheatley, 2014). Although the antler specimens are not likely to have suffered diagenetic alteration, we used the same treatment protocol as archaeological specimens to make data from potential future studies of ancient specimens comparable to those of this modern sample. Sample sizes of ~30 mg were weighed. Then, 1,200 μL of 2% bleach (NaOCl) was added to the samples, mixed, and left to sit at room temperature for 24 h. The bleach was then decanted and samples were rinsed three times with ultrapure water. One thousand two hundred microliter of 1 M acetic acid (CH₃COOH) buffered to a pH of 5 was added to the samples, which were then mixed and left to sit at room temperature for 24 h. The water rinse process was once again applied and then the samples, placed in a tray uncapped, were put in a laboratory oven to dry at 60°C for 24 h.

Mass spectrometry of collagen and apatite samples occurred in the Stable Isotope Laboratory in the Department of Geological and Atmospheric Sciences at Iowa State University. Collagen samples were analyzed for carbon and nitrogen isotopes using a Costech Elemental Analyzer coupled to a ThermoFinnigan Delta Plus XL mass spectrometer. Reference standards [Caffeine(USGS-62), Caffeine (IAEA-600), Cellulose (IAEA-CH-3), and Acetanilide (laboratory standard)] were used for isotopic corrections, and to assign the data to the appropriate isotopic scale. Corrections were done using a three-point regression method and isotope results are reported in parts per thousand (per mil, ‰). Percent concentration (%) was calculated using the peak area of the sample. The combined uncertainty (analytical uncertainty and average correction factor) for δ¹³C is ± 0.1‰ (VPDB) and δ¹⁵N is ± 0.2‰ (AIR). For the carbonate apatite samples, carbon and oxygen isotope ratios were analyzed with a Gas Bench with a CombiPAL autosampler coupled to the ThermoFinnigan Delta Plus XL mass spectrometer. Reference standards (NBS-18, IAEA 603) were used for isotopic corrections, and to assign the data to the appropriate isotopic scale. Corrections were done using a two-point regression method and isotope results are reported in parts per thousand (per mil, ‰). The analytical uncertainty for δ¹³C is ± 0.1‰ (VPDB) and δ¹⁸O is ± 0.1‰ (VPDB).

As part of our intra-antler analysis of the relationship between seasonal climatic variables and stable isotope data, we tested correlations between two scales of analysis. First, we used the available monthly climate data from local weather stations made available by the National Oceanic and Atmospheric Administration (NOAA). We obtained monthly mean precipitation, and monthly mean maximum, average, and minimum temperature using weather data from NOAA. This dataset from weather stations was provided by the NOAA climate data online under the local climatological data search tool. Access to the historical weather data allowed us to associate the unique seasonal conditions of the habitat of each individual in the study (NOAA, 2021). The Des Moines international airport station historical data of 2016 is associated with PEL-0090; the Fort Dodge Duncombe school area station historical data of 2016 is associated with PEL-0114; the Fort Dodge RGNL Station historical data of 2019 is associated with PEL-0164 and PEL-0165 (Steremberg et al., 1995). Antlers PEL-0164 and PEL-0165 were found in relatively proximity to each other and therefore share the same local weather station and thus local climate data. Because the maximum and minimum temperature variables were highly correlated with the mean monthly temperatures, we narrowed the analysis to only mean monthly temperature for the temperature variables. Second, we assessed the 50-year mean monthly data from each antler location because monthly rainfall totals can be quite variable at the local level (Figure 2). We obtained the multi-decadal mean monthly precipitation totals from the study regions using the geographic information system software DIVA-GIS (Hijmans et al., 2001, 2005) and the publicly available WorldClim dataset (Hijmans et al., 2005). The WorldClim data represent the interpolated mean monthly values across the period of AD 1950–2000 at each antler location.

Analysis of variance tests (ANOVA) were used to compare the mean stable isotope values between antlers, and linear regression models were used to analyze the relationship between the dependent isotope values and the independent variables of temperature and precipitation. Each antler had 10 evenly spaced samples taken from along the main beam. Resulting stable isotope values from each of these sub samples were associated with monthly mean environmental variables obtained from the closest weather stations (Figure 1). Based on the knowledge of antler growth patterns, which begin in February, and the knowledge of the date of death of each deer specimen, we matched the 1–10 sample sequence from each antler with the calendar dates of February-November to assess how seasonal changes in the local climate are reflected in the isotopic values of the antlers. This association of a month of development with a stable isotope value is conducted with the acknowledgment that this is not a precise pairing because antler velvet tends to be shed a month or two prior to November. Nevertheless, we believe that assessing the data in this way captures the patterns of seasonal change through time, with cooler periods at the beginning and ending of antler development. By analyzing the data in this manner we may assess how the extreme seasonal changes in environmental conditions of central Iowa might influence the intra-antler stable isotope data. Indeed, all analyzed antlers should have the earliest (most proximal) values representing climatic conditions of around February, samples from the middle of the antlers should correlate to the summer months, and the most distal samples should represent the early fall months (see Figure 2). Statistical analyses and vizualizations were conducted in the R computing environment (version R-4.0.2) (R Core Team, 2021). In addition to statistical tests of mean differences and intra-antler isotope/climate correlations, we also qualitatively assess the isotopic patterning of the isotopic data within the antlers. If seasonal changes strongly influence deer isotope values, then we would expect to see a similar U-shaped pattern within each antler reflecting the beginning and end of the summer season.