Sample procurement and data analysis
To establish a diachronic dataset of Galápagos tortoise dietary stable isotope ecology, we selected samples from five sources (see Supplemental Text): the American Museum of Natural History, New York, New York, (2) the California Academy of Sciences, San Francisco, California, (3) the Natural History Museum, London, England, (4) the National Museum of Natural History, Smithsonian Institution, Washington, D.C., and (5) the Thompson’s Cove (CA-SFR-186H) archaeological site in San Francisco, California. We provide details regarding sample provenience information and date-of-death as supplemental information. From these collections, we obtained single or multiple isotope samples from a total of 57 individual tortoises representing the following subspecies (n = 10) and islands: five C. n. abingdonii (Pinta Island), one C. n. becki (Volcán Wolf, Isabela Island), five C. n. chathamensis (San Cristóbal Island), four C. n. darwini (Santiago Island), thirteen C. n. duncanensis (Pinzón Island), four C. n. guentheri (Sierra Nega, Isabela Island), six C. n. hoodensis (Española Island), one C. n. microphyes (Volcán Darwin, Isabela Island), four C. n. niger (Floreana Island), nine C. n. porteri (Western Santa Cruz Island), one C. n. vicina (Cerro Azul, Isabela Island), one unknown Isabela Island tortoise, two C. n. vicina tortoises which were transported, lived and collected on Rabida Island, and one unknown tortoise (Chelonoidis niger ssp.; unknown Island—the San Francisco Gold Rush sample). The two earliest collected tortoises in our sample date to1833 and the latest tortoise is from 1967, representing a period of 134 years.
To understand tissue-specific isotopic variation and fractionation for the purposes of reconstructing long-term dietary ecology, we sampled tortoise bone collagen (n = 57), bone apatite (n = 23), scute keratin (n = 8) and skin (n = 2) for carbon (δ13Ccollagen and δ13Capatite), nitrogen (δ15N), hydrogen (δD) and oxygen (δ18Oapatite) stable isotopes. All samples were drilled or cut using a Dremel rotary tool with either a blade or diamond spherical bit attachment and were transported to the University of New Mexico, Center for Stable Isotopes (UNM-CSI), Albuquerque, NM, for preparation and analysis. All statistical and metric data analysis and visualization occurred in R (4.0.4) and RStudio (2022.02.4). We provide reproducible source code supplemental to the text35.
Bone collagen δ13C, δ15N and δD
Analysis of bone collagen, skin and scute keratin for carbon, nitrogen and hydrogen stable isotopes followed standardized protocols (e.g., see36). For bone collagen, we cut and demineralized a small portion of bulk bone in 0.5 N hydrochloric acid (HCl) at 5 °C for 24 h prior to rinsing all samples to neutrality using deionized water. For lipid extraction, we immersed the samples in a solution of 2:1 chloroform:methanol (C2H5Cl3) for 24 h (repeated three times) while also sonicating samples for 15 min to ensure complete chemical saturation. Preparation of skin and scute keratin samples was only included this during the later lipid extraction step (i.e., no demineralization required). After 72 h we rinsed all samples to neutrality and lyophilized the tortoise samples for another 24 h. We then measured approximately 0.5–0.6 mg of bone collagen/skin/scute tissue into tin capsules for carbon (δ13Ccollagen) and nitrogen (δ15N) stable isotope analysis. We also measured approximately 0.2–0.3 mg of bone collagen/skin/scute tissue into silver capsules for hydrogen (δD) isotope analysis. We report isotope values in delta (δ) notation, calculated as: ((Rsample/Rstandard) − 1) × 1000, where Rsample and Rstandard are the ratios (e.g., 13C/12C, 15N/14N) of the unknown and standard material, respectively. Delta values are reported as parts per thousand (‰).
Carbon and nitrogen samples were measured on a Costech 4010 elemental analyzer (Valencia, California, USA) coupled to a Scientific Delta V Plus isotope ratio mass spectrometer by a Conflo IV, and hydrogen samples were measured on a Finnigan high-temperature conversion elemental analyzer (TC/EA) coupled to a Thermo Scientific Delta V Plus mass spectrometer by a Conflo IV at UNM-CSI (see37 for details on the high temperature conversion method for hydrogen analysis). All nitrogen and carbon isotope data are reported relative to atmospheric N2 and V-PDB, respectively. The data were corrected using lab standards with values of δ15 N = 6.4‰ and δ13C = − 26.5‰ (casein protein), and of δ15N = 13.3‰ and δ13C = − 16.7‰ (tuna muscle) that have been calibrated relative to the universally accepted standards: IAEA-N1, USGS 24, IAEA 600, USGS 63, and USGS 40.
To ensure equilibrium between the exchangeable hydrogen in tissue samples and local atmosphere38, we weighed hydrogen standards and samples into silver capsules and allowed both to sit in the laboratory for at least 2 weeks before analysis. Hydrogen data were corrected using three UNM-CSI laboratory keratin standards (δDnon-ex = − 174‰, − 93‰, and − 54‰) of which the δDnon-ex values were previously determined through a series of atmospheric exchange experiments. These standards were also calibrated to USGS standards CBS and KHS values of − 178.8‰ and − 47.5‰, respectively (see39,40 for details and updated values). To quantitate any error imparted to our collagen data through correction with keratin standards, a UNM-CSI cow (Bos taurus) bone collagen standard was analyzed in every run over a 6-month period (July 2017–January 2018) and gave an inter-run standard deviation of 3.9‰, suggesting the difference in percent exchangeable hydrogen between collagen and keratin tissues did not significantly impact our results. All hydrogen isotope data are reported relative to Vienna-Standard Mean Ocean Water (V-SMOW). The H3 factor was between 8 and 8.5 for all runs.
Collagen precision (standard deviation; SD) for within-run analyses is < 0.14‰ for δ13Ccollagen and δ15N, and ≤ 0.54‰ δD. Given that weight percent C:N ratios provide a measure of collagen contamination41, and our samples ranged between 2.7 and 3.5, this suggests intact and preserved collagen.
Bone apatite δ13C and δ18O
To understand bone apatite (structural carbonate) carbon and oxygen stable isotopes we homogenized a small sample of tortoise bone powder for analysis following standardized protocols (e.g., see36). This homogenized powder was cleansed with 3% hydrogen peroxide (H2O2) for 24 h to remove organics and then was rinsed to neutrality using deionized water and centrifugation. A second treatment of 0.1 M buffered acetic acid (CH3COOH) for 30 min (followed by rinsing to neutrality) occurred to remove labile carbonates. After drying, we measured approximately 8.0–10.0 mg of homogenized bone apatite powder into exetainer vials, flushed those vials with He to remove atmospheric CO2, and then reacted the sample with phosphoric acid at 50 °C for at least 6 h. This reaction produced CO2 for carbon (δ13Capatite) and oxygen (δ18Oapatite) stable isotope analysis. All samples were analyzed at UNM-CSI on a Thermo Scientific GasBench (Bremen, Germany) coupled to a Delta Plus isotope ratio mass spectrometer with a Conflo II. An in-house Carrara marble standard (δ13C = 2.0 and δ18O = − 1.8) was analyzed in every run and used to correct the data. All data are reported relative to Vienna Pee Dee Belemnite (V-PDB). Within-run standard precision (SD) is < 0.25‰ for δ13C and δ18O.
Radiocarbon
We submitted radiocarbon samples to Beta Analytic Testing Laboratory in Miami, Florida. Beta Analytic prepared sub-sampled bulk bone collagen specimens from AMNH-46422 (Beta-599427), AMNH-46424 (Beta-599428), and USNM-84294 (Beta-599425), using internationally accepted protocols prior to measurement on an Accelerator Mass Spectrometer. Specific details relating to sample preparation and results are available in the Supplemental Text. Radiocarbon determinations are standardized to the Libby 5568-year half-life.
Stable isotope mixing model
Estimation of the proportion of C3 and C4 plants in Galápagos tortoise diets occurred through a Bayesian Monte Carlo fitted model using consumer (tortoises), sources (plants) and, and trophic enrichment factors (see27,42). Our dataset of non-corrected (see below) Galápagos tortoise δ13C and δ15N stable isotopes functioned as the consumer data for the model. We then extracted plant δ13C and δ15N from Gibbs et al.17. Collection of these modern plants occurred on three islands—Pinta, Santa Fe and Española (see Supplemental Table S2)—but we use these data as a proxy for vegetation stable isotope ratios across the archipelago and thus compare them against all tortoise subspecies. Plants within this dataset fall into two distinct groups (C3 and C4), and we take the average δ13C and δ15N ratios from these groups for our source input values. We obtained tortoise trophic enrichment factors from turtle controlled feeding studies43,44. Future such studies specific to Galápagos tortoises are required to fully evaluate these factors (see45,46).
Suess corrections
Anthropogenic changes in atmospheric carbon (13CO2 to 12CO2) due to the burning of fossil fuels impact δ13C ratios; addressing these requires a Suess correction47. Given the short temporal range of most of our samples, in this study we apply the correction only to those samples pre-dating 1833, that is, the Floreana Island tortoise δ13C ratios. For all other tortoise and plant samples, we did not Suess correct the data, and it is presented in unaltered form (see Supplemental Table S1 for details). We note that for wild, non-captive tortoises, the greatest calendrical age in our samples is between 1833 (Floreana and Santa Cruz) and 1929–1934 (Santa Cruz), an approximately 100-year period, and the Suess correction between 1833 and 1934 is 0.3‰. As future research focuses on analysis of more recent Galápagos tortoises, it will be necessary to correct our ratios to account for changes between, for example, the nineteenth and twenty-first centuries.
Source: Ecology - nature.com
