Overcoming the non-specificity of histological accentuated growth lines in hard tissues is an ongoing challenge. Identifying season at death and reproductive events has profound implications for evolutionary, ecological and conservation studies. Dental cementum is a mineralized tissue with yearly periodicity that continues deposition from tooth formation until death, maintaining a record spanning almost the entire life of an individual. Recent work has successfully employed elemental analysis of calcified incremental tissues to detect changes in extrinsic conditions such as diet and climate and to identify two important life history milestones: weaning and sexual maturity. Here, we employ field-emission scanning electron microscopy and energy-dispersive X-ray analysis to measure the relative concentrations of calcium, phosphorous, oxygen, magnesium and sodium in the cementum of 34 teeth from seven male and female rhesus macaques with known medical and life history information. We find that changes in relative magnesium concentrations correspond with reproductive events in females and breastfeeding in infants. Additionally, we observe seasonal calcium patterns in 77.3% of the samples.

All data were collected using a Bruker Quantax 200 XFlash 6160 EDS detector (Bruker, Billerica, MA, USA), coupled with a Zeiss Gemini-300 FE-SEM (Carl Zeiss Microscopy, White Plains, NY, USA). We acquired data in high vacuum mode, with an accelerating voltage of 15kV and an electron beam aperture of 60μm, in high current mode and with a working distance of 8.5mm, which is the analytical working distance of this setup.

We chose our regions of interest (ROI) on the cervical third of the root. For each specimen we acquired line scan spectra across the entire thickness of the cementum, starting from the cementum-dentine junction (CDJ). Prior to collecting data, we confirmed that the surface was free of topographic relief. The spot size of the line scan was set to match the radius of the interaction volume of the beam, which was consistently ~0.75μm. The scan dwell time was 64μs, and the spectrometer throughput 275Kcps. To have at least 10,000 counts per peak, scan time was manually determined for each specimen since they varied in cementum thickness. We used the zero kV peak to align the spectra and the P/B-ZAF standardless analysis quantification model in interactive mode to validate each peak. We only accepted as peaks those that were at least 3 standard deviations greater than the background. Carbon (C), oxygen (O), sodium (Na), P, Ca and Mg were detected with sufficient counts per peak in all samples. For each point across the cementum transect, using the software Bruker Esprit 2.3, we recorded the total net (NETTO) counts for each element. 

Data analysis was performed using R v.3.6.3. Within each datafile we added the variables derived from the recorded life histories (Supplementary Methods): Age, Event / Event Category (Table S2), Month and Season. The age at each point was derived by scaling the total thickness of the cementum to the range of years during which cementum was deposited and then adding the age at cementum initiation specific to the chosen ROI (reported for each tooth in Table S3).

Since our samples were carbon-coated the elemental percentages are altered, allowing only the collection of relative proportions of the elements. Therefore, we pre-processed our data in two different ways (both datasets are available here), aimed at addressing different questions:

1. For predictions 1 and 4: Within each data file, we normalized the data for each element to a 0-1 range. Instead of providing information regarding the relationships between elements at each point in time, it provides information regarding the behavior of a single element over time.
2. For predictions 2 and 3: Within each data file, we eliminated the values for C and then calculated the percent concentration for each element at each acquisition point. The resulting relative percent concentrations reveal how changes in the concentration of one element are correlated with those of the others. This allows for the analysis of the simultaneous behavior of several elements as their numerical relationship is maintained, even if the values are increased by the absence of C.

None