Recent dating of extinct Atlantic gray whale fossils, (Eschrichtius robustus), Georgia Bight and Florida, western Atlantic Ocean

Fossil localities of the Georgia and Florida gray whales

The Georgia locality is J-Reef a low exposure of shell beds about 16 km north of the Gray’s Reef (Fig. 1).

At J-Reef the shell beds are in a conformable relationship with finer grained sediment that dates to no earlier than the late Pleistocene (Garrison et al., 2012).

The dentary and three putative dentary fragments shown in Fig. 2 were recovered from the fine sediment or adjacent to the same outcrop on the sea floor. During subsequent dives, in fall 2006, the large fossil dentary was discovered partially embedded in the coquina (Fig. 2B). This coquina sand, typically deposited on a beach or in a nearshore marine environment, was probably the original sediment, like that of the Hobe Sound skull, in which it was preserved before being exposed.

The following details on the discovery and exact location of the gray whale skull from Hobe Sound (Fig. 3, Specimen UF 69000) were provided by Burkett S. Neely, Jr. (in litt., 6 June 1983) of the U. S. Fish and Wildlife Service (USFWS). The skull was discovered by a Mr. and Mrs. Kornit on 19 January 1983 at the edge of the surf on the northern end of Jupiter Island about 10 km south of St. Lucie Inlet, on the Hobe Sound National Wildlife Refuge (NWR), Martin County, southeastern Atlantic Coast of Florida (Fig. 1). Exact coordinates for the Hobe Sound skull are 27°07′40″ North latitude and 80°08′45″ West longitude. The skull was collected later that same day by USFWS personnel. The discovery of the Hobe Sound skull apparently coincided with extensive beach erosion that occurred during a strong winter storm. Through the generosity of Burkett S. Neely, Jr. and the USFWS, the Hobe Sound gray whale skull was placed on permanent loan in the vertebrate paleontology collection of the Florida Museum of Natural History (FLMNH), University of Florida (UF), Gainesville (catalogue number UF 69000). Mead & Mitchell (1984, 48) stated that “The specimen consists of most of the cranium of what looks like an adult…” However, upon closer examination, this skull consists only of a braincase, including both periotics (Fig. 3), and is from a very young individual or calf probably less than a year old. Robert K. Bonde (in litt, 19 July 1983), likewise, estimated that this specimen was from a very young calf, probably less than 3 months old. The Hobe Sound skull represents the southernmost record of E. robustus in the Atlantic Ocean, based on distributional maps in Mead & Mitchell (1984). Attached to the internal portion of the periotic (Fig. 3) of this juvenile skull is a small sample of sediment, consisting of medium to coarse quartz sand and fragments of mollusk shells, forming a semi-indurated “coquina.” This coquina sand, possibly deposited on a beach or in a nearshore marine environment, was probably the original sediment in which the skull was preserved before being dislodged and washed up on the beach.

The second gray whale skull from Florida (FLMNH Specimen UF 99000) was collected during the 1970s by Jesse S. Robertson of Jacksonville University on the beach at Jacksonville Beach, Duval County, northeastern Florida (approximate coordinates, 30°17′N, 81°23′W) (Fig. 1). Although the collector was unable to remember the specific details surrounding the discovery of this fossil, he did recollect it was found after a strong storm. The Jacksonville Beach gray whale skull consists of a braincase of a juvenile, compared to that of the younger calf from Hobe Sound.

Carbon isotope analysis and AMS dating

For the stable isotope analysis of samples we used both a Finnigan MAT 252 mass spectrometer, which is a dual-inlet mass spectrometer as well as a National Electrostatics Corporation Model 1.5SDH-1 Pelletron 500 kV compact AMS) unit for precise analyses of carbon isotopes ¹²C, ¹³C, and ¹⁴C. The Finnigan spectrometer is a double collector gas source mass spectrometer and allows for the measurement of two isotopes of oxygen and three isotopes of carbon with the double collector. It measures the ratio on the sample and the standard by alternating dual inlets. By contrast the Model 1.5SDH-1 Pelletron 500 kV only analyzed for the three isotopes of carbon using Faraday cup collectors at the end of the beam line array.

All samples were quite poorly preserved and did not retain sufficient collagen for AMS analysis, so we applied the bioapatite fraction dating technique (Cherkinsky, 2009). Specimens were chemically pretreated to remove all secondary carbonate and organics, but leave bioapatite in the bone structure. Bioapatite belongs to the group of hexagonal calcium phosphates, of which hydroxyapatite, (Ca₁₀(PO₄)₆(OH)₂, is closest in structure to biological apatite. It differs from geological apatite by a high degree of isomorphic substitutions and absorption of carbonate and small crystal size, properties that each result in a poorer crystallization of bone bioapatite (LeGeros & LeGeros, 1984).

The carbonate substituted within bioapatite maintains its carbon isotope signature in both stable and radioactive isotopes. Carbonate occurs in several locations in the crystals, as absorbed ions on the surface and within the crystals. Substitutions are mainly in the phosphate position and most likely in the hydroxyl position. The absorbed carbonates are more labile but substituted ones are much more stable (they are actually structure carbonates) and contribute to preserving the original isotope composition. Thus, they could be used for radiocarbon dating.

The specimens were treated with a weak acetic acid (CH₃COOH) bath under vacuum. Once washed and dried the samples were treated with phosphoric acid (H₃PO₄) in reaction tubes and heated to ensure removal of all carbonates. The bone samples generated CO₂ in the vacuum system and this gas was then analyzed in the mass spectrometer against a Vienna Pee Dee Belemnite standard (VPDB) with the per mil [‰] values of δ¹³C reported against VPDB and δ¹⁸O reported against VPDB and Vienna Standard Mean Ocean Water (VSMOW).

We have used the internal standards which are:

• δ13C, ‰ δ18O, ‰

• Fisher −0.64   −14.90

• A 1296 +2.56   −O.60

These standards were calibrated with National bureau of Standards (now NIST) NBS 19

• δ13C, ‰ = +1.95 and δ18O, ‰ = +28.60.

For AMS isotopic analyses carbon dioxide was cryogenically purified from the other reaction products and catalytically converted to graphite using the method of Vogel et al. (1984). Graphite ¹⁴C/¹³C ratios were measured by the Model 1.5SDH-1 Pelletron 0.5 MeV AMS. The sample ratios were compared to the ratio measured from the Oxalic Acid I (NBS SRM 4990). The sample ¹³C/¹²C ratios were measured separately using a stable isotope ratio mass spectrometer and expressed as δ¹³C with respect to VPDB, with an error of less than 0.1‰.

Collagen peptide mass fingerprinting

We attempted to taxonomically identify five specimens using collagen PMF: the three dentary fragments from JY Reef, Georgia Bight (GMNH 00-28-09, 00-28-10, 00-28-13 and the two Florida skull specimens (UF 69000, UF 99000). Collagen PMF, also known as ZooMS, is a rapid and cost-effective technique, whereby taxonomic groups are discriminated based on difference in the collagen protein sequence. Robust species identification can be accomplished by comparing collagen peptide fingerprints with the fingerprints from known samples using mass spectrometry (Collins et al., 2010). The success of this method has already been demonstrated for ancient North Atlantic cetacean species, including Atlantic gray whale (Kirby et al., 2013; Buckley et al., 2014).

Sample preparation, mass spectrometry and data analysis followed that described in Buckley et al. (2014) and Evans et al. (2016) at the BioArCh centre, University of York. Briefly, between 10 and 30 mg of bone powder was fully demineralized through immersion in 0.6M hydrochloric acid, followed by gelatinization in 100 μl of 50 mMol ammonium bicarbonate at 65 °C for 1 h. The resulting collagen was incubated with 0.4 μg of trypsin overnight at 37 °C, acidified to 0.1% trifluoroacetic acid and purified using a 100 μl C18 resin ZipTip® pipette tip (EMD Millipore). The collagen extract was spotted in triplicate on a 384 spot MALDI target plate, with calibration standards and run on a Bruker ultraflex III MALDI TOF/TOF mass spectrometer. mMass software (Strohalm et al., 2008) was used to average spectra replicates from each specimen, and compare to the list of m/z markers for marine mammals presented in Kirby et al. (2013), Buckley et al. (2014) and Hufthammer et al., 2018.

Paleontology

If the craniae and dentary elements are that of a gray whale, then systematic paleontology can be:

• SYSTEMATIC PALEONTOLOGY

• Class MAMMALIA Linnaeus, 1758

• Order CETACEA Brisson, 1762

• Suborder MYSTICETI Gray, 1864

• Family ESCHRICHTIIDAE Ellerman & Morrison-Scott, 1951

• Genus ESCHRICHTIUS Gray, 1864

• ESCHRICHTIUS ROBUSTUS.

• Eschrichtius cf. E. robustus (Lilljeborg, 1867).

We compared the two Florida gray whale skulls to descriptions and photographs of modern skulls of E. robustus (True, 1983, 1904; Barnes & McLeod, 1984) and a late Pliocene skull from Japan referred to Eschrichtius sp. (Ichishima et al., 2006). Because both Florida skulls consist only of the braincase, our comparisons are limited to characters present in posterior portion of the skull (Fig. 3; Table 1), as well as several characters of the periotics (see next paragraph). Characters the Florida skulls share with E. robustus include: triangular-shaped occipital shield with prominent paired occipital tuberosities (see Fig. 3A, UF 99000); large occipital condyles; concave exoccipitals lateral to occipital condyles; large and posteriorly oriented paroccipital processes; short, massive postglenoid processes of the squamosals that are parabolic in shape; and presence of a squamosal cleft.

Figures 3G and 3H are external (= medial = tympanic) and internal (= lateral = cerebral) views, respectively, of the left periotic from the juvenile gray whale skull from Hobe Sound, Florida (UF 69000). The right periotic of this specimen is still preserved intact in the skull (Fig. 3F). The posterior process of the right periotic (= posterior petrotympanic process of Ichishima et al., 2006) is still attached to the skull in the second Florida fossil, from Jacksonville Beach (UF 99000; Fig. 3C). The periotics from the Hobe Sound skull are very similar to periotics from modern skulls of E. robustus and to the periotic of a Pliocene skull of Eschrichtius from Japan (Ichishima et al., 2006). Specifically, the periotic of UF 69000 is similar to that of E. robustus in the short, triangular-shaped anterior process, comparatively large cochlear portion (as large as the body and anterior process of the periotic but excluding the posterior process), relatively fibrous posterior process, and the confluence of the fenestra cochleae and the aperture for the cochlear aqueduct. The anterior process of the perotic is proportionally small relative to the cochlear portion (pars cochlearis) in the fossil and Eschrichtius, compared to balaenopterids in which the anterior process is much larger.

On the basis of the similarities in the cranial morphology, periotics and ZooMS results, we confidently refer the two Florida skulls to E. robustus. Both Florida gray whale skulls represent immature individuals based on the unfused basioccipital/basisphenoid joint or suture. Walsh & Berta (2011) examined the closure of the bones in the occipital region in 20 skulls of E. robustus, including 18 calves under 1 year of age. They found that the basioccipital/basisphenoid joint becomes completely ossified between 7 months and 1 year of age. The cranial measurements (Table 1), together with the unfused basioccipital/basisphenoid joint, indicate that the Hobe Sound skull (UF 69000) is most likely a newborn calf, whereas, the somewhat larger Jacksonville Beach skull (UF 99000) was probably between 7 months and a year old. Both Holocene skulls of gray whales from Florida are juveniles less than a year of age, lending credence to our suggestion that the east coast of Florida may have been a breeding ground for the western Atlantic population of E. robustus (see below under Discussion).

Comparisons with previously described and measured fossils skulls of Eschrichtius, as well as modern skulls of E. robustus, provide further clarification on the age at death of the two Florida Holocene gray whales. A juvenile fossil gray whale cranium from Cape Lookout, North Carolina measured 80 cm in width across the zygomatic processes (Mead & Mitchell, 1984). This compares to a width across the zygomatic processes in the two juvenile skulls from Florida of 76 cm in UF 99000 and 60 cm in UF 69000 (width of squamosals in Table 1). An adult fossil gray whale cranium from Corolla, North Carolina measured between 130 and 140 cm across the zygomatics (Mead & Mitchell, 1984). The width of the occipital condyles in the adult from Corolla is 31 cm, compared to 24 cm in UF 99000 and 22 cm in UF 69000 (Table 1). A juvenile Pliocene gray whale skull (Eschrichtius sp.) from Japan measured 54 cm across the paraoccipital processes (Ichishima et al., 2006), compared to widths across the paraoccipitals of 59 cm (UF 99000) and 45 cm (UF 69000) in the two Florida juvenile skulls (Table 1). Ichishima et al. (2006) noted that the paraoccipital measurement of the Japanese Pliocene fossil was comparable with the same measurement in a modern gray whale skull from an individual 7.25 m in body length and an age of approximately 30 weeks (Sumich, Goff & Perryman, 2001). Comparative data from modern gray whales, including both cranial measurements (Sumich, Goff & Perryman, 2001) and fusion of cranial sutures (Walsh & Berta, 2011), support our assertion that the Hobe Sound skull was from a calf, younger than 30 weeks based on comparative measurements of the paraoccipitals, whereas, the larger Jacksonville Beach skull was from an animal older than 30 weeks but less than a year old.

Our assessment together with earlier studies, are in line with that of two recently reported Holocene-to-Pleistocene aged gray whale finds in the western Pacific, off Taiwan (Tsai et al., 2014) and Japan (Kimura, Hasegawa & Kohno, 2018). The former study identified the paired tuberosities on the occipital shield as diagnostic of Eschrichtiidae sp. (supra). Both Florida gray whale skulls demonstrate these paired tuberosities, in particular UF 99000 (Fig. 3).

Furthermore, the collagen PMF results confirm the anatomical identification of Gray Whale for the two Holocene-aged Florida skulls (Fig. 4). The two skulls exhibited diagnostic gray whale specific markers presented in Kirby et al. (2013) and Hufthammer et al. (2018) (specifically, the presence of diagnostic collagen peptide F at 2899). The application of this technique to our other three Pleistocene-aged specimens, however, failed due a lack of preserved collagen in these ancient bones—a result mirrored in the lack of organic fraction available for radiocarbon dating (see below, and also Harvey et al., 2016).

Radiocarbon (¹⁴C) dates

We dated the samples using the organic fraction and mineral fraction of bioapatite. The overall morphology of the bones was very well preserved, however, the collagen fraction was almost completely destroyed and the concentration of organic carbon was about 0.1% or lower. In all four cases the bioapatite fraction radiocarbon age estimate was older (33–37.5 ka) than that from the organic, collagen-like, fractions (8.3–23 ka). The per mil (‰) values of δ¹³C reported against VPDB and δ¹⁸O reported against VPDB and VSMOW and are shown in Table 2, and the AMS dates are displayed in Table 3.

Without doing a full statistical analysis, the results shown in Table 2 for the δ¹³C of sample A (−5.11 ± 0.04‰) and sample B (−5.04 ± 0.04‰) seem to be more strongly related and are probably from the same whale, while sample C (−4.80 ± 0.04‰) differs slightly from the other samples. Sample C could also be from another individual, but because the difference from the other samples is not more than 1‰ it is likely that all three of the mysticete whale bone fragments are from the same individual.