Introduction

During the nineteenth century, East Africa became a major source of elephant ivory for a range of rapidly expanding industries, including cutlery, comb, piano and billiard-ball manufacturing, reflecting broader changes in leisure activities and patterns of consumption. As Thorbahn remarks, 'Ultimately this slaughter and wastage was accomplished to supply Hindu brides with bangles and New Englanders with combs' (1979: 32). The scale was enormous. Between 1840 and 1875, British demand alone rose from approximately 200 000kg to over 800 000kg per annum, with even conservative estimates suggesting that this could have equated to between 4000 and 17 000 elephants per year being killed for the trade (Reference BeacheyBeachey 1967; Reference SpinageSpinage 1973; Reference SheriffSheriff 1987; Reference HåkanssonHåkansson 2004). The net effect was that the size of the African elephant populations fell from an estimated 24 million to 4 million in less than a century. Once the large herds had gone the grasslands transformed to forest because elephants no longer trampled young trees and browsed 22kg of biomass per day (UNEP/GEMS 1989). The loss of open grassland would have had further effects for hunter-gatherers, pastoralists and agriculturalists (Reference ThorbahnThorbahn 1979; Reference HåkanssonHåkansson 2004).

Unfortunately, estimating the level of elephant hunting in East Africa during the nineteenth century relies on incomplete records of ivory exports (Reference SpinageSpinage 1973; Reference ParkerParker 1979). Since there is no finite resolution on where all this ivory was obtained it is difficult to identify the areas that were most affected and assess when these ecological impacts began and how long they lasted. The research presented here — part of a broader investigation of the historical ecology of East African landscapes funded by the European Union (Reference LaneLane 2010) — employs isotope analysis of historic and modern pieces of ivory, collected in the USA, the UK and East Africa, to determine the geographical locations in East Africa of ivory procurement in the nineteenth and twentieth centuries.

Although this research is historical in nature, it has significance for current elephant studies in East Africa and beyond. Recently, news surrounding elephants and the ivory trade have included a controversial legal sale of ivory from Namibia, Botswana, South Africa and Zimbabwe sending more than 100 000kg of CITES-sanctioned ivory to China and Japan (Reference BlackBlack 2008), culling of elephant populations in South Africa (Reference WhyteWhyte 1993; Garaï et al. 2004; Reference LangeLange 2008) and parks across Kenya tagging their elephants to track illegal poachers (Reference QuammenQuammen 2008). One method to control the illegal trade has been isotope analysis, matching the chemical signatures of illegally obtained tusks to specific regions of Africa, thus helping governments establish their provenance. The method has been proven to work on a regional scale, but there is a need for more baseline data from different regions. As van der Merwe et al. (1990a: 15) note, 'There is an urgent need to build up a comprehensive index of ivory (or bone) isotope signatures for all regions where elephants still exist in Africa and Asia...' The research presented here not only provides historic isotope data from elephants, but it is also supported by modern samples from locations across East Africa, adding to the existing database of isotope values of ivory from elephants across the continent.

Method

In this research, isotope analysis is used as a method for reconstructing the diets of historic elephants hunted for their ivory as well as a tool for sourcing where these historic elephants roamed. This is possible thanks to the way animal bone and dentine such as ivory grow, as the chemical elements incorporated into the bone are directly affected by the type of environment in which the animal lives, its diet, and the geology of the bedrock in its habitat range.

In this study I analysed δ¹³C, δ¹⁵N, δ¹⁸O and ⁸⁷Sr/⁸⁶Sr from hair, bone and teeth. In the case of δ¹³C and δ¹⁵N, the isotopes primarily derive from food, whilst ⁸⁷Sr/⁸⁶Sr derive from a combination of food and water. Differences in vegetation, hydrology and soil chemistry (itself reflecting bedrock geology) will therefore all influence one or more of the isotope values measured. Used together these isotopes have potential for regional characterisation and help to 'fingerprint' elephant tissue (van der Merwe et al. 1990a & b; Vogel et al. 1990a & b; Koch et al. 1995; Reference Cerling, Omondi and MachariaCerling 2007).

Collections

In order to build up a database of material that would help to isotopically characterise the region it was important to collect baseline data, ideally from elephants known to have lived in specific areas of East Africa. Baseline data include any East African elephant biological material (bone, molar, tusk, hair) securely sourced to a specific location and of known date of death or collection. This was gathered by sampling specimens from a variety of museum collections, including the Smithsonian National Museum of Natural History, the University of California-Berkeley Museum of Vertebrate Zoology, the Los Angeles Museum of Natural History, the Natural History Museum in Vienna, the Field Museum, the Quex (Powell-Cotton) Museum, the National Museums of Kenya, and the Tanzanian Wildlife Division/Tanzania National Parks. Unprovenanced, worked ivory used to test this baseline data came from two collections: the Hawley Collection, a nineteenth and early twentieth-century cutlery collection in Sheffield, England and Ivoryton, Connecticut (USA), an industrial piano key making centre in the nineteenth and twentieth centuries.

Results

The material in the Powell-Cotton collection deserves special mention, because tissue is extremely well preserved, all the specimens have a specific record identifying the date and place where the elephant was shot and tail hairs from the collection allow a detailed and specific time-scale isotope analysis for the life of individual elephants to be undertaken. By sampling the tail hairs sequentially it is possible (using modern elephant growth rates) to work out when the tail was growing; these samples therefore reflect short-term changes, for example in the individuals' eating preferences or roaming patterns. It was thus possible to analyse how much dietary change elephants experienced in the nineteenth century in specific places by using δ¹³C and δ¹⁵N to determine how much C₃ or C₄ plant material was in the diet. As expected, tail hairs from elephants in forested regions, such as the Ituri Forest, Congo, showed a much more stable C₃ diet than that of savannah elephants relying on mixed vegetation (Figure 1). Tail hairs were also analysed for ⁸⁷Sr/⁸⁶Sr (Figure 2). Two key observations could be made: first, that elephants from different areas have widely differing ⁸⁷Sr/⁸⁶Sr compositions, which means that this is useful for provenancing; and secondly, that there is little change in the ⁸⁷Sr/⁸⁶Sr composition over the period of 6–18 months. These results show that our historic elephants did not move over sufficiently large areas to cross the major geological boundaries which characterise the East African region. This correlates well with what we know about modern populations, as most elephant movement is controlled by access to resources: they tend to move within specific corridors, as for example the elephants from Amboseli in Kenya which seem to move within 50km of a specific basin (Western & Lindsay 1984). And a more recent study by Wittemyer et al. (2007) shows that during the dry season elephant groups in the Samburu and Buffalo Springs National Reserves in northern Kenya seldom moved more than 1km away from a permanent water source, and that movement primarily had to do with access to resources as well as protection. Though today elephants are hemmed in by national park boundaries and more restricted due to human land use than they were in the past, these preliminary results on historic elephant tail hairs do not show these four elephants moving between different geological substrates. So, based on these initial results as well as further analysis on historic elephant material it is possible to use a combination of isotopes to define East African elephant populations by vegetation and geological zones, which means historically it is possible to characterise elephants along nineteenth century caravan trade routes. There is a definite separation between elephants living further in the interior towards the Great Lakes region and those right along the coast, as the older, basement geology combined with the mixed and forested vegetation of the interior contrast with the younger geology and C₄-dominated vegetation of the coast. This separation could therefore enable a better understanding of where elephants were extracted in the nineteenth century, specifically in relation to hypotheses put forward by historians regarding the sustainability of elephant numbers in the coastal hinterland to supply the number of tusks being exported during this time (Reference ThorbahnThorbahn 1979; Reference HåkanssonHåkansson 2004).

Figure 1.

δ¹⁵N and δ¹³C: diet and environment. Two tail hairs analysed in sequence along the length of the hair every 10mm, using Wittemyer et al. (2009) average growth rate for male tail hair of 0.56±0.08 mm per day. Sub-sampling in this way allows for specific time depth for both elements, noting the Congo forest dweller, which has a wholly C₃ diet, and stable δ¹⁵N signal, compared to the arid Somalia dweller, which has a much more varied diet of primarily grass, mixed with browse in a possible wetter environment as evidenced by a more depleted δ¹⁵N signal.

Figure 2.

⁸⁷Sr/⁸⁶Sr: elephant movement. Four tail hairs from elephants shot during Major Powell Cotton's nineteenth-century East African hunting trips plotted with an inset of the varying geologies of this region, which produced distinguishable ⁸⁷Sr/⁸⁶Sr values. The range of values (0.71813-0.70542) reflects the varying geologies where these animals roamed. The Precambrian basement geology of the Congo produced expected high ⁸⁷Sr/⁸⁶Sr values (~0.71813) while the young, active volcanic of Mt. Elgon, Uganda region, produced low ⁸⁷Sr/⁸⁶Sr values (~0.70542), which are consistent with values published for this region (0.70314–0.70604) (Simonetti & Bell 1995) (image: M. Collins).

Outlook

Isotope analysis is useful for provenancing historic and archaeological ivory, and is an ecological tool for understanding diet and habitat changes in historic and modern elephant populations living in the same regions. However, more work to test data and retrieve baseline data is required to use this methodology routinely for provenancing ivory on a large scale. This research will create a database of isotope values from historic and modern East African elephants, of value to archaeologists and ecologists as well as wildlife monitoring groups attempting to control the current illegal trade of African ivory. Future work on these samples could include other molecular techniques such as DNA analysis (successfully used by Wasser et al. [2007] for modern ivory) in combination with isotope results.