Deciphering Mammoth Life Histories to Test Causes of Mammoth Extinction and Advance Elephant Conservation

Abstract

Mammoths went extinct in the face of human hunting and habitat changes, the same factors that now have driven elephants to the brink of extinction. Mammoths had tusks that grew throughout their lives and contain mineral and organic components reflecting body chemistry. Density variations during tusk growth result in a series of zones similar to tree rings that serve as timestamps for reconstructing body chemistry and growth rate through life. Prior measurements of growth increments in computed tomography (CT) scans showed cycles of faster and slower growth in females that likely reflect reproductive episodes. Environmental deterioration can increase time between reproductive episodes, or calving interval (CI), and human hunting can decrease CI; but the extent to which each contributed to mammoth extinction across their range is debated. Using growth increments to reconstruct reproductive life history offers a means of testing these competing hypotheses. More efficient and accurate methods of analyzing mammoth dentition to extract CIs in the lab and field are needed, along with population models to set clear expectations for hypothesis testing in mammoths and conservation of elephants. Here, I first use a Michigan mammoth molar to demonstrate computed tomography (CT) methods that improve visibility of increments to interpret life history. Then, I examine a poorly preserved male mammoth cranium from Minnesota with adhered microcrystalline pyrite and gravel matrix. Laser ablation induction-coupled plasma mass spectrometry (LA-ICP-MS) on tusk slabs shows that Sr and Ba replace Mg post-burial, and that impregnation of Fe and Mn in exposed parts of the tusk causes overprinting of original signals. Serially-sampled carbonate and collagen isotopes and trace elements nonetheless provide a record of seasonality and suggest a healthy animal near the time of its death. CIs in well-preserved female Russian permafrost mammoths are reported to be similar to those in elephants (~5y). Models show non-hunted populations affected by changes in habitat go extinct with longer CIs (~10y). Multiple population models averaged over time and space show extinction is possible at shorter average calving intervals (~7y) with combinations of hunting and habitat deterioration. Interpretations of growth increments in nine tusks from Chukotka, Siberia range between ~4y and ~8y CIs, but more tusks are needed to assess whether reproductive life history changed or remained stable during the process of extinction. Stable reproductive life history traits are an assumption of extinction risk models for elephants; the validity of this assumption may be tested with more mammoth tusks analyzed in this way. Measurement error and confounding causes of differential growth diminish confidence in these CT-based interpretations of CIs, though, and multiple chemical proxies, including the female sex hormone progesterone, may provide greater certainty. In 20 years at different times of life in a permafrost tusk from the Taimyr Peninsula, collagen isotopes, Zn, Mn, and sometimes Ba fluctuate seasonally. Pregnancy may be marked by alignment of lower Zn and higher Mn with periods of higher and less-variable levels of progesterone measured with enzyme-linked immunosorbent assays (ELISAs). However, these proxies do not always align with each other or with expected changes in growth rate. More tusks and different proxies are needed to provide more confidence in interpretations of mammoth life histories. Ultimately, incorporating what we learn from mammoths into models of elephant populations will allow more accurate assessments of extinction risk to better protect these endangered animals.