Marine strontium isotopes preserved in fossil shark teeth calibrate Neogene land mammal evolution

Many Cenozoic terrestrial fossil sites worldwide rely on land mammal evolution (biochronology) for temporal calibrations. Within marine depositional environments, strontium isotope compositions from invertebrate fossils are often used for enhanced temporal resolution. Here we demonstrate that strontium isotope ratios from fossil shark tooth enameloid can quantitatively calibrate ages at sites containing both marine and non-marine fossils. Analysis of REEs in the same samples provides data for understanding the taphonomic histories of the specimens and allows for identification of specimens that were likely reworked. Using Neogene shark teeth, we resolve a ~600,000-year age difference between two late Miocene Florida sites previously binned together within the latest Hemphillian (Hh4) North American Land Mammal Age (NALMA) based solely on biochronology. This refinement brackets the Nme2 sea level lowstand, providing further calibration for the Great American Biotic Interchange (GABI) between South and North America. These examples emphasize the significance of strontium isotope ratios preserved in fossil shark teeth for global Neogene calibrations, especially in sites with marine/non-marine intercalations.

Methods
Experimental design. The MFS and PFBV experienced sea level fluctuations, accounting for the marine and terrestrial fossils collected from both sites. To account for nearshore fluvial influx of strontium, teeth from neritic sharks were chosen for SIS and REE analysis, for dating and taphonomic analysis, respectively. While some of these species may inhabit estuarine environments (Compagno et al., 2005), analysis of REEs aid us in ruling out the influence of a fluvial source of strontium that could obscure the marine strontium isotopic signature. Additional samples taken from terrestrial and marine mammalian fossil bone were analyzed for REE composition as a frame of reference for the REE signatures found in the shark teeth. McArthur et al. (2020) point out that terrestrial strontium input is assumed to be negligible, and this assumption can be scrutinized with the addition of REEs as described below.

A total of 24 MFS shark teeth (5 species) and 25 PFBV shark teeth (5 species) were sampled from the University of Florida Vertebrate Paleontology (UF VP) collections (S1). Enameloid samples were collected via surface scratch technique using a dental grade carbide tip and Dremel tool, paying attention to avoid the dentine layer below. Previous work has demonstrated that the incorporation of more porous dentine can yield strontium values that deviate strongly from predicted values (Becker et al., 2008; Kocsis et al., 2016). Most samples were analyzed using approximately 20 mg of powdered enameloid. 

Samples were prepared in a controlled cleanroom environment within the UF Department of Geological Sciences using refluxed Teflon vials and split for strontium collection and REEs following established protocols. Per protocol, the powdered sample weights were calculated and 50% HNO3 added to dissolve and clean the samples. Evaporated sample weights were recorded and resuspended with 5% HNO3/Re-Rh (8ppb) and the final preparations split for Sr and REE analysis.

Strontium analysis. Samples were loaded into cation-exchange columns packed with strontium-specific resin (Eichrom Part #SR-B100-S) for extraction by loading 1 ml of 3.5 N HNO3 followed by 100 µl of each sample into their respective columns. For sample weights lower than 20 mg, 200 µl of sample were loaded into the columns. Columns were rinsed five times with 100 µl of 3.5 HNO3, followed by a final wash with 1 ml of 3.5 HNO3. Strontium was then collected in newly refluxed Teflon vials by running 1.5 ml of 4xH2O through the columns and then evaporated via hot plate.

Prepared samples were analyzed on a Nu Plasma multicollector inductively coupled plasma mass spectrometer (MC ICPMS) at the University of Florida, using 100 µl of each sample mixed with 1,000 ml of 2% HNO3. The standard, NBS 987, was loaded after every four to six samples. The average 87Sr/86Sr value and error calculated against standard NBS987 was 0.710240 ± 0.000020 (2 SD). The 87Sr/86Sr values were corrected for mass bias, presence of Rb, and any interference. Corrected values were converted into ages using the LOESS Version 6 Table calibrated to the GTS2020 (Table 1; Howarth and McArthur, 1997; Gradstein et al., 2020; McArthur et al., 2020). Mean 87Sr/86Sr ratios were calculated for each population (i.e., MFS, PFBV) and converted into numerical ages using the LOESS Version 6 Table calibrated to the GTS2020. Uncertainty is presented as 2s.e. (standard error) of the mean 87Sr/86Sr ratio for each site converted to numerical ages.

Rare Earth Element Analysis. For REE analysis, the prepped samples were run on a Thermo Finnigan ELEMENT 2 inductively coupled plasma mass spectrometer (University of Florida) and REE concentrations normalized to Post-Archean Australian Shale (PAAS) following McClellan’s protocol (McClellan, 1989). All but nine teeth (three from MFS and six from PFBV) were analyzed for REE composition. Additionally, samples were taken from the bones of mammals from each site (eight MFS terrestrial mammal samples, two from each of the four specimens, and one MFS cetacean; eight PFBV terrestrial mammal samples, two of which are from the same specimen, and two PFBV cetaceans) to compare against the REE signatures of the shark samples. Specific isotopes measured were 139La, 140Ce, 141Pr, 143Nd, 149Sm, 153Eu, 157Gd, 159Tb, 163Dy, 165Ho, 166Er, 169Tm, 172Yb, and 175Lu. Ce/Ce* and La/La* calculations followed the recent interpretations of Barret et al. (2023). All data analysis was conducted using R/R Studio programs (R Core Team, 2024), and the raw data and R code can be found in the Supplementary Materials section (Table S1 and Data S1–S4).