In total, we analyzed 79 ivory objects ( Supplementary Table 1 ), identified as elephant (44) or mammoth (35) by seizure agencies and sellers. Elephant ivory objects were acquired in seizures between 2007 to 2023 by the Agriculture, Fisheries and Conservation Department of the Hong Kong government (AFCD) in illegal imports to Hong Kong from African countries (Ethiopia, Gabon, Ghana, Ivory Coast, Nigeria, South Africa, and Zimbabwe), China, Thailand, and the United States of America. We acquired unworked mammoth ivory fragments from Siberian carvers, and additional mammoth ivory objects purchased between 2022 to 2024 in markets in China (Shanghai) and Hong Kong ( Supplementary Table 1 ), mostly by donations of WWF-Hong Kong. Unfortunately, the exact geographical origin and species identification is unknown. Samples were drilled between January and August of 2024, and stored in a dry-cabinet before isotope measurements.

We developed a protocol adapted from the literature to analyze ivory powder (Ziegler et al., 2016) with four main steps described below: 1) drilling, 2) cleaning, 3) weighting, and 4) measuring δ ¹³C, δ ²H, δ ¹⁵N, δ ¹⁸O, and δ ³⁴S. These steps were designed to optimize time and costs, with a minimum ivory powder amount required of 5 mg using the Continuous Flow-Isotope Isotope Ratio Mass Spectrometer (CF-IRMS) at the Stable Isotope Laboratory, University of Hong Kong (SIRMS-HKU).

Samples had a non-normal distribution and we tested for significant differences in the stable isotopic signatures with the Wilcoxon test and Principal Component Analyses (PCA) in RStudio 1.3.1093 (R Team, 2020).

Among the carved ivory objects acquired in illegal seizures and observed in Chinese and Hong Kong markets, we found a variety of samples, including bracelets, combs, chopsticks, earrings, neckless, and seals ( Figure 1 ; Supplementary Table 1 ). Two main types of mammoth ivory carved artefacts were being sold (staff of the Chinese market, personal communication): “bai jian (摆件)”, which are large items that can be displayed standing up, such as plates or a single carved tusk, and “accessories (饰 品)”, such as bracelets, pendants, earrings. We also observed that the markets in China and Hong Kong further classified mammoth ivory objects in four subjective categories related to the color and aspect of the ivory: (1) highest quality with shiny illustrious white ivory, with an oily sheen and not visible or delicate lines with narrow spaces; (2) milky white, oily sheen, lines rather delicate with narrow spaces; (3) yellowish white, waxy sheen, lines quite coarse with wide spaces; and (4) “coffee”, when the color is dull/dark yellow with earthy sheen, lines are coarse, with wide spaces.

Stable isotope data revealed statistically significant differences in the Wilcoxon tests for δ ²H and δ ¹⁸O of the elephant and mammoth samples (p-values = 3x10–¹⁴ and 3.5x10^-12, respectively). Although there was no overlap between the δ ²H values of the two ivory types, the δ ¹⁸O data of two elephant samples were within the value range of mammoth samples ( Supplementary Table 1 ; Figure 2 ). We also observed significant differences in the Wilcoxon tests between elephant and mammoth ivory samples for δ ¹³C and δ ³⁴S, while no significant difference was reported for δ ¹⁵N (p-values = 8x10^-2, 1.5x10^-8, and 0.64, respectively). Nevertheless, these three isotopes had overlapping values between elephant and mammoth samples. The δ ¹³C and δ ¹⁵N data of all mammoth samples were within the range values of elephant samples, while δ ³⁴S data of 67 mammoth samples were within the range of elephant samples ( Supplementary Table 1 ; Figure 3 ). The first two principal components of the PCA explained over 99% of the total variance ( Figure 4 ). The individual contributions to these eigenvalues were 96.9% for δ ²H, 2.7% for δ ¹⁸O, 0.23% for δ ³⁴S, 0.07% δ ¹³C, and 0.003% for δ ¹⁵N.

Our results showed that δ ²H is the most efficient element to distinguish elephant and mammoth ivory, as there were no overlapping values between the two sample types. ( Supplementary Table 1 ; Figure 2 ). Additionally δ ¹⁸O values were distinct between most elephant/mammoth samples. Thus, we recommend using a multi-elements SIA approach focusing on δ ²H and δ ¹⁸O to distinguish the two ivory types. Although our analysis omitted the extraction of collagen to optimize the protocol time, the isotopic values are potentially still comparable with elephant reference databases (e.g., www.ivoryid.org) using offsets discussed in the literature (Ziegler et al., 2016), and stable isotope ranges were similar as previously reported in the literature for elephants (van der Merwe et al., 1990; Codron et al., 2012; Ziegler et al., 2016; Ziegler, 2021) and mammoths (Bocherens et al., 1996; Ziegler, 2021). Due to their natural range, elephants consume water from tropical regions, and therefore have heavier δ ²H and δ ¹⁸O isotopic signatures compared to mammoths that ingested water from temperate environments of higher latitudes (Ziegler, 2021). On the other hand, δ ¹³C, δ ¹⁵N, and δ ³⁴S are associated with animals’ diet and trophic niche (Peterson and Fry, 1987; McCutchan et al., 2003; Boecklen et al., 2011; Jackson et al., 2011). The larger variation of these isotopes for elephant samples could be related to a more generalist feeding strategy (Codron et al., 2012) compared to mammoths, and/or a higher baseline variation of the food sources, as our sampling likely included elephant samples from across Africa and Asia. Thus, δ ¹³C, δ ¹⁵N, and δ ³⁴S are less useful when the focus is simply distinguishing between mammoth and elephant ivory, compared to δ ²H and δ ¹⁸O. Further work targeting samples of specific elephantid populations, and including SIA of their food sources such as grass, woody material, and fruits (Tchamba and Seme, 1993), will shed light on comparisons on the trophic niche of these megaherbivore species.

Although there are integrated approaches to understanding and mitigating human-elephant conflict (HEC) in Africa and Asia, HEC is a main issue for conservation strategies for all three elephant species (Dublin and Hoare, 2004; Gross et al., 2022; Saha and Soren, 2024). The most prominent human-elephant conflict usually arises over crop raiding incidents (Walpole and Linkie, 2007). There are many different approaches which are used to deter elephants, focusing on sensory deterrents, such as chili/beehive fences or loud noises (Chang’a et al., 2016; King et al., 2011; Enukwa, 2017). However, elephants are highly intelligent animals that often find ways around the obstacles, resulting in extensive HEC (Mumby and Plotnik, 2018). Therefore, farmers often support or are not opposed to poaching for ivory (Kiffner et al., 2021). To overcome this issue, recent popular programs targeting community attitudes and social change are currently in place to change the attitude paramount and alleviate the conflict (e.g.,Williams et al., 2024). Along with these initiatives, forensics tools, including SIA, could contribute to law enforcement and reducing poaching.

The trade ban in China and Hong Kong lead to an increasing market demand for harvesting mammoth material. Russia is the main export country of mammoth ivory, with more than 140 tons legally exported in 2023, most of which were destined to China, but also to Hong Kong and Netherlands (Potravny et al., 2024). The volume of illegal, and semi-legal exported ivory is thought to be three times more than that of legal ivory (Toropov, 2021). Siberian miners collect mammoths’ tusks and other carving material during summer when the permafrost, the ice layer below the topsoil, starts to melt. Due to current difficult economic conditions in the region, some ivory miners engage in illegal practices of soil erosion using powerful water pumps (Vasileva, 2022). Although this activity has considerable environmental impacts, carvers argue that the damage is not comparable to that of gas and oil mining activities (Toropov, 2021). Carvers also reported that mammoth ivory has a low market value because it often has a yellowish color and a negative perception associated with the product as it is from extinct animals (Toropov, 2021). Nevertheless, harvesting it has become one of the few sources of income available to these communities.

The trade of ivory flows across country borders, each of which have distinct social and economic environments. Conservation efforts for elephants are hampered by the current levels of governmental and localized corruption that makes it difficult to prevent the laundering of illegal ivory into legal markets (Bennett, 2015), including in the online trade (Venturini and Roberts, 2020). We suggest that the SIA approach described here is applied to distinguish between mammoth and elephant ivory, as the protocol is a relatively fast and economical tool (CF-IRMS measurements <USD25/sample). We highlight that although SIA will be facilitative in law enforcement, it is not suitable for presentation in court. In a potential framework to identify the ivory source, a batch of samples would be screened using SIA by an accredited laboratory. Fewer random objects, or any ivory samples with ambiguous results, could then be tested with more expensive and time consuming methods (e.g., genetics analyses and radiocarbon dating) to validate the results. In this way, governmental agencies would be able to analyze a higher volume of samples for a lower price and waiting time. We expect that a SIA database will be progressively constructed for this task, based on reports and publications. Besides individual government budgets, support to such a framework could be acquired from international multilateral foundations, United Nations programs, and non-governmental organizations, which have already financed ranger, customs and criminal justice functions to combat IUWT in several countries across Africa and Asia (UNODC, 2024).

While it is imperative to address the social issues related with the ivory trade, including reducing HEC (Jarungrattanapong and Olewiler, 2024; von Hagen et al., 2024; Williams et al., 2024) and offering financial support to Siberian carver communities (Potravny et al., 2024), alternatives for natural ivory should be a prioritized option in the markets, such as utilizing polished cattle bones (Sims et al., 2011) or developing materials for 3D printing with optical imitation of natural ivory (Rath et al., 2021). It is key to combine the investment in law enforcement with measures that address corruption and poverty (Hauenstein et al., 2019) and public engagement, such as education on the importance of elephants as key species to ecosystem health (van de Water et al., 2022; Williams et al., 2024), and on the Siberian permafrost to the ecosystem carbon balance (Bouchard et al., 2018).

Further studies will clarify how the isotopic signatures of elephant and mammoth ivory are influenced by biological factors, for instance, the animal’s age or the tusk portion analyzed (Codron et al., 2012) that could partly explain variations in some of the isotopes observed here. SIA applications in the field should also focus in estimating their isotopic niche (Jackson et al., 2011; Schwartz-Narbonne et al., 2019), a proxy of the trophic niche. Comparing the amount of isotopic niche overlap (Jackson et al., 2011) between different elephant populations, as well as mammoth populations, will contribute to revealing the resource partition and diet preferences of elephantids. Such analyses could also provide insights on the isotopic niche changes between wild and captive elephants. Compound specific isotopes analyses (CSIA), including amino acids and fatty acids, should also be used for a detailed comparison of their diets (McMahon and McCarthy, 2016). Although CSIA has a higher cost and is more time-consuming compared to SIA, such analyses could also be used to distinguish the two ivory types, when bulk isotope values are ambiguous. Lastly, it will be vital to facilitate networks between different countries’ institutes to combat illegal ivory trade.