Wildlife and fisheries markets are end-points in the supply chain of both legitimate and illegitimate or unregulated trade in species and natural products. Molecular ecology provides powerful tools for surveillance and estimation of this trade. Here, I review the application of these tools to market surveys and species in trade, including species identification and molecular taxonomy, population assignment and ‘mixed-stock’ analysis, genetic tracking and capture–recapture by individual identification. I consider the analogy of markets to natural populations and also the unique features that require novel analytical approaches and sampling design. In the most developed of these applications, the molecular ecology of market surveys and confiscated trade shipments has provided independent estimates of illegal, unregulated or unreported exploitation for sharks, elephants and whales. Although each study has taken advantage of information from trade records or official government reports concerning the ostensible levels of exploitation, it is telling that the truer measure of exploitation seems to arise from the market end-point of the supply chain.
The international component of global trade in wildlife and fisheries is estimated to be worth more than US$60 billion per year, or nearly US$160 billion if wild-sourced timber is included (Roe et al. 2002), of which US$5–15 billion is considered wildlife. The magnitude and diversity of reported international trade in wildlife species listed by the Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES) are astounding (Broad et al. 2003). From the years 1995–1999, CITES records show mean annual trade in more than 1.5 million live birds, 640 000 live reptiles, 1.6 million lizard skins, 1.1 million snake skins, 150 000 pelts from wild furbearers and 21 000 hunting trophies (Roe et al. 2002). The specifics of trade in the 5000 species of animals and 28 000 species of plants listed by CITES can be searched through the Web-based World Conservation Monitoring Centre CITES Trade Database (UNEP-WCMC CITES Trade Database 2008).
The magnitude of illegal international trade in wildlife, which by its nature largely escapes detection, is extremely difficult to estimate. In 1998, the United Nations Environment Programme (UNEP) valued this component at approximately US$5–8 billion per annum and some sources suggest the current value could be twice this large (US Department of State 2007). If so, illegal trade in wildlife is second in value only to the smuggling of drugs among the world's underground economies, and greater in value than the smuggling of weapons or people (Cook et al. 2002). The extent of illegal, unreported or unregulated (IUU) fishing is even more difficult to estimate, as it occurs across the full spectrum of small-scale to industrial fisheries, in zones of national jurisdiction as well as on the high seas (e.g. Lack & Sant 2001). The Food and Agriculture Organization (FAO) of the United Nations assumes that IUU fishing accounts for up to 30% of total catches in some important fisheries (Doulman 2001) but reliable estimates are rare. In an unusual admission, the government of Japan recently acknowledged systematic under-reporting of catches of southern bluefin tuna (Thunnus maccoyii) over the last 20 years. By comparing domestic trade records with reported catches, the Commission for the Conservation of Southern Bluefin Tuna estimates that the IUU component of Japan's catch during this time has been worth up to US$6 billion (Hayashi 2006). Perhaps the most egregious documented violation of an international wildlife or fisheries convention was exposed by Russian scientists who had worked aboard factory whaling vessels of the former Soviet Union following World War II (Yablokov 1994). For a period of nearly 30 years, the Soviet Union falsified its reports to the International Whaling Commission (IWC), under-reporting its true catch by more than 100 000 whales in the Southern Hemisphere alone (Clapham & Baker 2002). The full impact of this IUU exploitation is only now being understood, particularly in relationship to the variable recovery of some populations of humpback, Megaptera novaeangliae, and southern right whales, Eubalaena australis (Mikhalev 1997; Clapham et al. 2007; Jackson et al. 2008).
While attention has tended to focus on control of international trade (legal and illegal) as a threat to biodiversity, much of the world's commerce in wildlife and fisheries is domestic and largely unregulated. Markets in many parts of the world include local ‘bushmeat’ or ‘wild meat’ (Milner-Gulland et al. 2003), intermingled with domesticated species. Similarly, fisheries markets often include marine megafauna taken as incidental ‘bycatch’, for example, sea turtles, cetaceans, and pinnipeds (Lewison et al. 2004), intermingled with species targeted by commercial and traditional fisheries. The scale of this unregulated exploitation is massive; the bycatch of marine mammals, alone, is estimated to exceed 650 000 individuals a year worldwide (Read et al. 2006) and hunting for food is estimated to take 580 million terrestrial mammals a year in the Congo Basin and another 16 million in the Amazon Basin (Fa & Peres 2003). Although not all bushmeat or bycatch is destined for markets, the commercialization of traditional or subsistence hunting and fishing is increasingly fuelled by urbanization and global development (Mainka & Trivedi 2002). In coastal regions of the developing world, the supply and demand for bushmeat and bycatch are often linked (Clapham & Van Waerebeek 2007), reflecting the complex relationship of over-exploitation in both the terrestrial and marine ecosystems (Brashares et al. 2004; Rowcliffe et al. 2005).
There is a growing recognition that visual and molecular surveys of wildlife and fisheries markets can provide important information on the dynamics of both regulated and unregulated exploitation. Traditionally, efforts for documentation and control of this trade have been directed at the origin or source-point of the supply chain for products destined for markets. This under-represents the component of trade that arises from IUU exploitation, including much of the domestic trade in bushmeat and bycatch. Market surveys provide a measure of the outlet or end-point of this supply chain, including products originating from both legitimate and illegitimate or unregulated sources of exploitation. Surveys of large commercial markets can be an effective means of estimating true levels of exploitation or ‘takes’ where the sources of supply are diffuse, such as for many fisheries, or when reliable records of trade are lacking (e.g. Clarke et al. 2004). As the end-points for bushmeat and wildlife products used in traditional medicines, surveys of local or artisanal markets can be an effective means to ground-truth species diversity and monitor the status of rare species in remote regions (Fa et al. 2000; Rabinowitz 2002). When assisted by molecular methods, market surveys can provide accurate identification of species and estimation of the volume of trade even for products that have been butchered or processed before sale (Clarke et al. 2006a; Palumbi 2007).
Here, I review the application of molecular ecology to surveys of wildlife and fisheries markets and species in trade. Although the power of molecular tools has long been recognized in wildlife forensic genetics, market surveys pose a much wider range of questions analogous to those of interest for natural populations: What is the species origin of a product and what is the species composition or diversity of the market place? What is the geographical origin or population source of a product or specimen and by what route (i.e. supply chain) did it arrive in the market? What is the individual identity of a product and what is the abundance (i.e. volume) of those individuals in trade? To answer these questions, past studies have drawn on a range of methods shared with conservation genetics and molecular ecology (DeSalle & Amato 2004). However, a simple analogy of markets with natural populations can be misleading. Unlike individuals in natural populations, the distribution and ‘life history’ of market individuals are controlled by supply chain structures and the dynamics of both supply and demand, including the potential for long-term storage and long-distance transportation (see Box 1; Fig. 1). A truer measure of these markets will require an expanded application of standard molecular ecology and continued development of novel analytical methodologies to account for these differences.
Species identification and molecular taxonomy
Species identification remains one of the most important applications of molecular ecology to wildlife and fisheries markets. Before the expansion of interest in DNA taxonomy (Tautz et al. 2003), market surveys focused on the molecular identification of a relatively limited number of protected species, such as whales (Baker & Palumbi 1994), pinnipeds (Malik et al. 1997), turtles (Roman & Bowen 2000) and seahorses (Sanders et al. 2008), or specific products in regulated trade, such as caviar (DeSalle & Birstein 1996). Within the taxon-specific framework of each study, two somewhat different approaches have been used. In one approach, a short fragment of target DNA (usually 300–400 bp in length) is amplified and sequenced from a market product via the polymerase chain reaction (PCR). The sequence of the fragment, referred to as the ‘test’ or ‘query’ sequence, is aligned and compared to a database of homologous ‘type’ or ‘reference’ sequences of known species origin. A close match in sequence similarity, such as from a blast search of GenBank, or the nesting within a species-specific grouping in a phylogenetic reconstruction relative to the reference sequences of the taxon, is considered evidence of species identity (Baker et al. 1994). Where taxonomic sampling is incomplete, a test sequence that does not show a close match, or does not nest within a species-specific grouping of reference sequences, requires further investigation and could represent a new species (Baker et al. 1996). Although applied initially in the context of species in trade, the identification of a test sequence by distance-matching or tree-based reconstruction with homologous reference sequences is, in principle, the basis of most recent initiatives for molecular taxonomy (see below and The DNA barcode of life and ‘species registers’).
In the second approach to species identification, a data set of reference sequences is used to design PCR primers that anneal to species-specific or ‘diagnostic’ nucleotide substitutions (Bartlett & Davidson 1992; Amato & Gatesy 1994). The identity of a target species is assayed by a positive amplification (with negative controls), using a multiplex of species-specific primers (e.g. DeSalle et al. 1996; Chapman et al. 2003). Although a diagnostic PCR assay does not require direct sequencing to identify the target species, it provides little information on the identity of nontarget species (i.e. samples that do not amplify). Except for sharks, the preferred molecular marker for both sequencing and diagnostic assays has been mitochondrial DNA (mtDNA) because of its ease of amplification, relatively rapid rate of evolutionary divergence and the availability of published reference sequences. For sharks, species identification has relied on diagnostic differences in the nuclear ribosomal internal transcribed spacer 2 or ITS gene (Shivji et al. 2002; Magnussen et al. 2007), as well as sequences of mtDNA (Martin 1991; Hoelzel 2001).
The sequencing approach to molecular taxonomy of species in trade has been facilitated by the program dna surveillance (Ross et al. 2003). This web-based program aligns a user-submitted DNA sequence of unknown origin against a comprehensive database of homologous reference sequences. Unlike GenBank, the reference databases are curated or administered by species specialists (Baker et al. 2003). Evolutionary distances and a phylogenetic tree with bootstrap simulations are used to judge species identity of a ‘test’ sequence relative to the pre-aligned reference data set. dna surveillance is implemented currently for identification of the approximately 90 species of whales, dolphins and porpoises (order Cetacea) using sequences from the mtDNA control region and cytochrome b gene. The program has been used successfully for routine species identification of ‘whalemeat’ products purchased in surveys of markets in Japan and Korea (e.g. Endo et al. 2005; Baker et al. 2006), and for the description of new species of cetaceans (Dalebout et al. 2007). Taxon-specific databases are under development for other species in trade (e.g. Brazilian parrots, seahorses), as well as routine taxonomic identification for other purposes (e.g. ‘What rat is that?’, Robins et al. 2007).
Product authentication and ‘mock turtle’ syndrome
In the commercial sector, the application of molecular taxonomy for food ‘authentication’ is intended primarily to protect consumers against fraud (e.g. Lockley & Bardsley 2000; Gil 2007). However, mislabelling of wildlife and fisheries products can also mask sources of IUU exploitation and threats to protected species (e.g. Baker et al. 2002; Marko et al. 2004). Roman & Bowen (2000) coined the phrase ‘mock turtle’ syndrome to describe the false labelling of ‘turtle meat’ sold in some parts of the southeastern USA. In these markets, the sale of meat from marine turtles, which are now protected, has been replaced by meat from the alligator snapping turtle (Macroclemys temminckii), the largest freshwater turtle in North America (Roman, Bowen 2000). This species has, in turn, declined and is now protected in every state except Louisiana. To investigate concerns that the remaining legal trade in turtle products was providing a cover for sale of protected species, Roman & Bowen (2000) used mtDNA control region and cytochrome b sequences to identify the species origins of 36 ‘turtle-meat’ products purchased in Louisiana and Florida. Only one of the products was found to have originated from the alligator snapping turtle. The remainder originated from more common turtle species and from the American alligator (Alligator mississippiensis), a species on the increase in the region. Roman & Bowen (2000) described the replacement of an increasingly rare, but esteemed species, by more common species as the equivalent in wildlife trade to ‘fishing down the food web’ (Pauly et al. 1998).
Species discovery and ‘mock species’
Traditional markets are a rich source of discovery for new species. This is particularly true in the biodiversity ‘hotspots’ of Southeast Asia, an area experiencing an explosive growth in unregulated wildlife trade. In the last few years, four new species of ungulates have been discovered or rediscovered through collections of trophies from local hunters or displays in traditional markets of Vietnam or Laos (Groves et al. 1997; Giao et al. 1998; MacKinnon 2000; although see Robins et al. 2006), and a fifth in Burma/Myanmar (Amato et al. 1999). A new species of striped rabbit (Nesolagus timminsi) was first observed for sale in a market in a rural town in Laos (Surridge et al. 1999) and a new species of rodent, so distinct as to require its own family, was found for sale as wild meat in a traditional market in Laos (Jenkins et al. 2005). Artisanal and commercial fisheries markets are another likely source of new species, as with the discovery of the Indonesian coelacanth Latimeria sp. in a Sulawesi fish market (Erdmann et al. 1998).
Unlike species in natural communities, however, the molecular taxonomy of species in trade can be complicated by the forces of the market as well as by the forces of evolution. Trade in exotic species often places a priority on novelty, providing incentive to breed unusual specimens in captivity and fabricate collection data (Dalton 2003). Over the last two decades, 14 new species of freshwater turtles have been described from China, 10 of which were described from the morphology of specimens purchased through the Hong Kong wildlife trade (Stuart & Parham 2007). The descriptions of these new species added substantially to the approximately 260 species of turtles known previously worldwide (Ernst & Barbour 1989). However, attempts to discover the source populations of some newly described species have failed, and recent genetic analyses suggest that these ‘species’ are, in fact, intergeneric hybrids bred in captivity (Parham et al. 2001; Stuart, Parham 2007). An improved molecular taxonomy, including comprehensive reference databases for these taxa (see The DNA barcode of life and ‘species registers’), is urgently needed to control trade and direct conservation efforts for ‘true’ species under threat (Shiping et al. 2006; Fong et al. 2007).
Population assignment and mixed-stock analysis
In many cases, wildlife and fisheries are managed according to geographical populations or stocks, as well as by species (Dizon et al. 1992). Hunting, fishing or whaling may be allowed in an abundant stock but prohibited in another stock of the same species that has been depleted by past exploitation. However, it is rare for geographical populations or stocks to be phylogenetically distinct or to be delimited by diagnostic molecular characters. Instead, the ability to identify the geographical origin of a market product is likely to rely on differences in allele or haplotype frequencies among potential source populations (Manel et al. 2005).
Where a large collection of reference samples is available from across the distribution of a single species, multilocus assignment tests can be used to estimate the origin of an individual product or specimen in trade. In a survey of Atlantic cod (Gadus morhua), Nielsen et al. (2001) found strong genetic differentiation among three regulated populations or stocks using nine microsatellite loci. Individuals included in the reference samples and a set of ‘test’ samples not included in the reference were assigned back to their stock of origin with high reliability. Testing as few as two or three individuals could allow the unambiguous assignment of a catch, providing an important control for IUU fishing of these regulated stocks (Nielsen et al. 2001). Conversely, the potential to exclude a specific population of origin can be informative, even when assignment is limited by incomplete sampling of alternative source populations. In a case of suspected fraud in a salmon fishing competition, Primmer et al. (2000) used an assignment test based on surveys of microsatellite diversity in several putative source populations to show that the winning fish was highly unlikely to have originated from the competition location in southeast Finland. When confronted with this evidence, the offender confessed to purchasing the salmon at a local fish shop (Primmer et al. 2000).
For the control of trade and verification of catch records, the assignment of an individual product or specimen is often less important than estimating the origin for a collection of samples. In an effort to track illegal trade in African elephant ivory, Wasser et al. (2007) used a novel DNA assignment method to determine the geographical origin of a large shipment (> 6.5 tons) of contraband ivory seized by customs agents in Singapore. DNA was extracted from 37 of the tusks, considered to be representative of the shipment, and genotyped for up to 16 microsatellite loci (Comstock et al. 2003). The genotypes of the tusks were compared to an extensive reference collection of genotypes from tissue and faeces samples of both forest (Loxodonta africana cyclotis) and savannah (Loxodonta africana africana) elephants (Wasser et al. 2004). Rather than attempting to estimate the origin of each individual tusk, Wasser et al. (2007) developed an assignment method for jointly analysing multiple samples to distinguish between two alternative hypotheses: (i) the tusks came primarily from savannah elephants killed in or near Zambia and Malawi, the documented origin of the interdicted shipment; or (ii) the tusks originated from numerous locations across forest and savannah Africa, with stockpiles smuggled into Malawi before shipping. The joint assignment procedure supported the first hypothesis, pointing to a likely origin from a relatively narrow band of southern Africa, centred on Zambia. The allocation to this region of such a large volume of ivory called into question the government of Zambia's claim that only 135 elephants had been killed illegally over the last decade. As a result of this investigation, the Zambian government replaced its director of wildlife and began imposing harsher sentences for convicted ivory traffickers (Wasser et al. 2007).
Where source populations differ significantly in frequencies of mtDNA haplotypes, a mixed-stock analysis, similar to that developed for fisheries management (Pella & Milner 1987), can be used to estimate the contributions of each population to a market. This approach was used by Baker et al. (2000b) to estimate the IUU exploitation of North Pacific minke whale Balaenoptera acutorostrata scammoni, one of the most common species found in molecular surveys of Japanese markets (see Box 1). Products from this species originate from two sources (Baker et al. 2000b): (i) a relatively abundant stock in the offshore water of the western North Pacific (the ‘O'stock) that is hunted by Japan under special permit (i.e. ‘scientific whaling’); and (ii) a depleted stock in the Sea of Japan (the ‘J’ stock) that is subject to fisheries bycatch and some extent of illegal hunting by both Japan and Korea. The two stocks are known to differ markedly in frequencies of mtDNA haplotypes, although before the market surveys, characterization of the ‘J’ stock was based on a relatively small number of samples (Goto & Pastene 1997). Official reports of the Japanese scientific whaling programme provided information on the numbers of whales killed annually in this hunt (all of which were sold in Japanese markets, Normile 2000) and the frequencies of mtDNA haplotypes for the offshore or ‘O’ stock (Goto et al. 1997). The Fisheries Agency of Japan also reported a small number of minke whales killed as bycatch in coastal fisheries but no genetic information was collected from these whales.
Using haplotype frequencies from a sample of 81 North Pacific minke whale products purchased in Japanese market from 1993 to 1999, Baker et al. (2000b) found significant differences with the expected frequencies from the reported scientific catch, rejecting the assumption that all or most products originated from this source. Instead, the sample showed a greater than expected number of haplotypes shared with minke whale products purchased in Korean markets and, presumably, originating from the Sea of Japan or ‘J’ stock. Using haplotype frequencies of the Korean products to better characterize this alternative source, a mixed-stock analysis estimated that 31% (95% CI, 19–43%) of the Japanese market samples originated, in fact, from the protected ‘J’ stock. Given that the Japanese scientific catch was, at the time, about 100 whales/year and that bycatch from around the coast included both ‘J’ and ‘O’ stocks, Baker et al. (2000b) calculated that total IUU exploitation (take) in Japan must also be about 100 whales/year. This number is roughly four times greater than the average bycatch of 25 whales/year reported to the IWC by the government of Japan during the period of the surveys (Baker 2002). In 2001, Japan modified its regulations to fully legalize the commercial sale of whales killed as bycatch and to require collection of genetic samples from whales destined for the market (see Genetic tracking and ‘DNA registers’). Reporting of minke whale bycatch increased sharply after this change, to an average of more than 120/year for the years 2002–2006 (e.g. IWC 2006a).
Genetic tracking and ‘DNA registers’
If genetic samples are collected systematically as part of a regulated hunt or catch, individual identification can be used to track the origins of a product in trade and verify its legitimacy. One of the first efforts to track international trade of an individual involved whalemeat purchased from a Japanese market in 1993 and identified initially from mtDNA sequences as a blue whale (Balaenoptera musculus)— a species protected from hunting since 1966. However, an initial search of GenBank showed an exact match to the published sequence of a blue/fin hybrid (B. musculus/Balaenoptera physalus) killed during scientific whaling by Iceland in 1989 (Arnason et al. 1991). Using tissue archived from this hunt, a comparison of variable nuclear DNA introns confirmed that the Japanese product was derived from this hybrid individual, connecting the two ends of a very long supply chain (Cipriano & Palumbi 1999).
Even when genetic samples are not archived as part of a regulated hunt or catch, genetic tracking can be used to describe market dynamics, especially for large species that are butchered or flensed before transport. For whalemeat products purchased on markets in Japan and Korea, Dalebout et al. (2002) used mtDNA sequences to identify species and DNA profiling for up to six microsatellite loci to track replicate products derived from the same ‘market individual’. The distribution of replicate products provided information on the local supply chains connecting markets in different cities or prefectures, and the number of individuals provided a minimum ‘census’ of whales entering the market (Dalebout et al. 2002; see below).
A formal extension of genetic tracking involves establishing a ‘DNA register’ of all individuals destined for the market (Dizon et al. 2000; DeSalle, Amato 2004). The first DNA register for control of trade was initiated by the government of Norway in response to concerns about the continued sale of protected species and the poor control of whalemeat markets. Tissue from all whales killed in Norway's commercial hunt of North Atlantic minke whales (Balaenoptera acutorostrata acutorostrata, taken under objection to the current IWC moratorium) is archived and used for DNA profiling with 10 microsatellite loci, sequences of the mtDNA control region and sex (IWC 1998). The DNA profiles are stored on an electronic database, forming a searchable register of individuals intended for the market. The DNA profile of a market product can then be compared to the database; a market product that matches an existing profile would be legitimate, while a product that did not have a match in the register would be illegitimate or illegal (Dizon et al. 2000). Although the government of Japan has also committed to maintaining a DNA register for species taken in its scientific whaling programme (IWC 2005b), the situation is complicated by the sale of products from whales killed as fisheries bycatch and by the long-term storage of products predating the register (see Box 1).
Palsbøll et al. (2006) tested the effectiveness of the Norwegian DNA register with 20 whalemeat products purchased in Norwegian markets in 2002 and, as negative controls, two samples collected from minke whales beached in Denmark. In a collaborative validation exercise, genetic profiling of these samples was conducted in an independent laboratory and then submitted to the Norwegian DNA register for comparison to the 2676 individual profiles available from whaling during the years 1997–2001 (Palsbøll et al. 2006). The results demonstrated the ability to match the test profiles to the DNA register and to identify the ‘illegitimate’ (unregistered) samples from the two beached whales. However, the study also highlighted a number of technical problems with implementation of a DNA register, including interlaboratory standardization of allele binning and estimation of genotype errors, similar to those noted in microsatellite genotyping of living wildlife (e.g. Waits et al. 2001; Hoffman & Amos 2005). Palsbøll et al. (2006) concluded that these problems would need to be addressed before a DNA register could be a practical mechanism for the control of trade (see DNA registers and market trackability/traceability).
Modelling the market
To understand the full impact of species exploitation, molecular ecology will need to develop models for estimating the true volume of trade from market surveys. A recent study of ‘shark fin’ markets integrated molecular identification of species with commercial trade records to improve estimates of takes for several species in this global fishery (Clarke et al. 2004; Clarke et al. 2006a, b). Hong Kong is considered the centre of world trade in shark fins, controlling an estimated 50–85% of the total market (Clarke et al. 2004). Although trade in shark fins is largely unregulated, Hong Kong trading houses keep accounts of each auction, including volume and weight of fins according to loosely species-specific trade names and fin type (e.g. dorsal, pectoral, caudal). Clarke et al. (2004) obtained a sample of the Hong Kong trade records for the period October 1999 to March 2001, and used a Bayesian imputation procedure to estimate the total volume of trade, including the auctioned weights of fins by trade name. In a subsequent study, Clarke et al. (2006a) used molecular taxonomy to confirm the species-specific composition and proportions of the trade. The corrected species identification of trade names then allowed Clarke et al. (2006b) to convert fin weight to species-specific estimates of worldwide catches. The results provided the first ‘fishery-independent’ estimate for the true scale of shark catches, indicating that shark biomass in the fin trade is three to four times higher than the reported catches.
Where trade records are unavailable or intentionally withheld, more intensive molecular sampling will be required to describe market dynamics. As discussed in Box 1, the bycatch of whales, dolphins and porpoises supports a thriving market in the southeastern peninsula of Korea, where there has been a long history of whaling (Kang & Phipps 2000). Although fishermen are required to report bycatch of large whales to the Maritime Police and, subsequently, to the IWC (Kim 1999), there has been concern that this unregulated market acts as a cover for other IUU exploitation. To estimate the true number of individual minke whales sold in Korean markets, Baker et al. (2007) analysed 289 products purchased from a subset of these outlets during 12 surveys over a 5-year period, 1999–2003. Using DNA profiling, they found that the 289 products originated from 205 market individuals, some of which were represented by replicate products found in more than one survey or in more than one outlet. Previously, Dalebout et al. (2002) had suggested that the frequency of replicate products from market individuals could be used to improve estimates of the true number of whales in trade, similar to capture–recapture estimates of abundance for living individuals. However, this analogy is not strictly true because a market individual does not die suddenly, but instead, remains available for a time in decreasing quantities, rather like the exponential decay of a radioactive isotope. To account for this ‘product decay’, Baker et al. (2007) applied a novel capture–recapture model to first estimate that the average ‘half-life’ of the products was about 1.8 months. With the adjustment for this decay, the true market volume was estimated to be 827 whales, nearly two times greater than the 458 whales officially reported as bycatch for this 5-year period (Baker et al. 2007). Although the government of Korea had reported previously to the IWC on only a small number of infractions (e.g. IWC 2005a), local officials seem to have increased their vigilance of these markets. In January 2008, police seized 50 tonnes of whalemeat and questioned 70 people, including fishermen, distributors and operators of 46 whalemeat restaurants, as part of an investigation into illegal whaling in the Ulsan regions (ABC News 2008).
Future directions — towards a well-monitored market
Accessing genes and genomes in trade
A fundamental obstacle to a more general application of molecular ecology to wildlife and fisheries markets is the constraint on transport of any ‘native product’ derived from the 5000 species of animals and 28 000 species of plants listed by CITES. International transport of native products, including DNA, derived from species listed in CITES Appendix 1 requires exportation and importation permits and those listed in Appendix 2 require export permits. The processing of such permits is often a lengthy affair and either nation can deny a permit, effectively terminating any research considered politically sensitive or embarrassing. However, CITES regulations do not apply to ‘synthetic’ DNA created during in vitro amplification, assuming that the native DNA has been removed (Bowen & Avise 1994; Jones 1994). In market surveys of CITES-regulated species, such as whales and sea turtles, it has been necessary to assemble a portable PCR laboratory and conduct the initial PCR amplifications on location (Baker & Palumbi 1994). By labelling primers with a biotin molecule, PCR products can be bound to streptavidin-coated, 96-well plates, providing an effective procedure for washing away native DNA (Baker et al. 2006). Once the amplified DNA is isolated from the native product, it can be transported internationally for sequencing and final analysis.
To date, the portable PCR has been employed primarily for taxon-specific surveys using a relatively small number of pre-selected molecular markers (e.g. mtDNA and microsatellites; Dalebout et al. 2002). This limits the ability to conduct the full range of analyses necessary to characterize multispecies markets and to preserve the full genetic diversity of samples that might later prove to be of special interest (e.g. products representing new species or infractions of international agreements). This limitation can be overcome, in some cases, by whole genome amplification (WGA, Lasken & Egholm 2003). Several commercial kits are now available that produce microgram quantities of high molecular weight DNA (> 10 kb in length) from as little as 1 ng of genomic DNA template. Unlike PCR, which relies on sequence-specific primers and a thermal-stable Taq polymerase (or similar polymerases), WGA uses random primers (hexamers) and a high-processivity φ29 DNA polymerase to replicate genomic DNA by multiple-strand displacement amplification. Validation experiments show no discernable difference between WGA samples and the original DNA templates for downstream analyses such as amplification of microsatellites or single nucleotide polymorphisms (SNPs, Dean et al. 2002; Short et al. 2005). As with PCR primers, the random hexamers can be labelled with biotin and the synthetic copies of an entire genome can be separated from the native DNA. Although degraded genomic DNA is not a good template, many market products are sold relatively fresh or freshly frozen and could be suitable for WGA.
The DNA barcode of life and ‘species registers’
With proposals for a universal DNA barcode of life (Hebert et al. 2003) and its adoption for the Census of Marine Life (O'Dor 2004), there is the potential to vastly expand the sequence-based approach to routine identification of market species. The true diversity of species found in local or traditional markets is unknown but likely to reflect much of the natural biodiversity of a region. More than 50% of the 284 African forest mammals and nearly 30% of the 192 Amazon forest mammals are considered ‘game’ species (Fa & Peres 2003). Indigenous communities in North-East India hunt at least 134 species of wild animals for both subsistence and commercial trade (Hilaluddin et al. 2005). Many more could find their way to bushmeat markets as more valuable species decline due to overexploitation. The ability to visually identify such a large number of species poses a challenge to even trained observers. Larger species in particular are likely to be butchered (or flensed in the case of whales or dolphins) near the location of death and thus, unrecognizable when transported to markets. Fisheries markets represent an even greater challenge to visual monitoring. In the case of many cryptic or rare species, reliance on visual identification of even intact specimens is likely to misrepresent the true underlying taxonomy (Dalebout et al. 1998; Bickford et al. 2007). Given the known or suspected link between emerging diseases and wildlife markets (Bell et al. 2004; Karesh et al. 2005; Jones et al. 2008), molecular surveys could also undertake to identify species of pathogens, as well as the host species subject to trade.
Despite the appeal of a universal DNA barcode, the criteria for robust molecular identification of species and the implementation of these criteria in conservation biology remain controversial (e.g. DeSalle 2006; Rubinoff 2006; Rubinoff et al. 2006). For these reasons, Baker et al. (2003) have argued for a taxon-specific approach to developing databases of reference sequences for species threatened by exploitation or regulated by international agreements and, thus, subject to the standardization of practice required for monitoring and enforcement (Gerson et al. 2008). A comprehensive ‘species register’ provides a framework for verifying the accuracy of identification for known species and for interpreting patterns indicating inconsistent molecular taxonomy (e.g. Funk & Omland 2003), inconsistent organismal taxonomy or incomplete taxonomy (Birstein et al. 1998; Dalebout et al. 2007). The concordance between traditional and molecular taxonomy can then be validated by species specialists operating within the conventions of each taxon (e.g. for cetaceans, see Reeves et al. 2004). These conventions are often informed by different views on species concepts, which have important implications for species delimitation and identification (Sites & Marshall 2003; DeSalle et al. 2005). An additional component of validation should be to evaluate the sensitivity of alternative molecular markers (e.g. mitochondrial cytochrome c oxidase subunit 1 or cytochrome b, nuclear ITS) and analytical methodology (e.g. distance matching, strict or relaxed tree-based identification, diagnostic character-based identification) for resolving subspecies, species or genera. Where taxonomic groups are well characterized and species are genetically distinct, identification is likely to be robust to the choices of genetic markers or methodology (Ross et al. 2008). For less well-characterized groups or for closely related species, however, molecular taxonomy will require an integrative approach to validation of comprehensive sequence databases and the assessment of analytical methods for both routine identification and species discovery (Dalebout et al. 2004; Meyer & Paulay 2005; Hickerson et al. 2006; Ekrem et al. 2007).
DNA registers and market trackability/traceability
DNA registers of individuals taken under regulated hunting or fishing have great potential as a mechanism for control of trade, particularly for large, valuable species (e.g. whales, elephants, tuna).1
The questions of identification and verification are dealt with simultaneously; a product that matches a profile in the register is from an individual whale taken legitimately and a product that does not match a profile in the register is illegitimate. However, as confirmed in Palsbøll et al.‘s (2006) first test of such a register, the interlaboratory standardization required to establish identity by microsatellite profiling is not trivial (e.g. allele binning), limiting the potential for verification of market products by a second or third party. These requirements will increase with application of DNA profiling to market products that have degraded due to poor storage or preparation for consumption (as on Korean and Japanese markets, Dalebout et al. 2002). In the case of whaling, resolution of these technical difficulties is further confounded by political opposition to market surveys as a mechanism for control of illegal trade and the lack of agreement about transparency of DNA registers (IWC 2001a, b).
For a robust system of verification and identification, a DNA register should be integrated with a trade certification or catch documentation scheme, CDS (IWC 2005b). A CDS establishes a ‘paper trail’ for the identity of a product through a supply chain from the point of origin to the point of distribution. CDSs are now accepted ‘best practice’ in many international fisheries conventions (e.g. Agnew 2000). However, a CDS alone is subject to fraud by manipulation of records or mislabelling of products. A DNA register would establish the verifiable link between the ends of the supply chain documented by the CDS. With a CDS-assisted register, the technical demands of DNA profiling are greatly reduced by the limited question of verification; did the product originate from the species and specific individual indicated by the CDS? For a fully transparent system of trackability/traceability, all products could be labelled with an electronic barcode linked to the integrated CDS/DNA register and accessible through the Internet. A program for combining DNA profiling with electronic barcoding of individual Chinook salmon (Oncorhynchus tshawytscha) has been developed for real-time management decisions requiring information on the source stock of fish taken in coastal fisheries of Oregon (CROOS 2007). This could be extended to include product trackability/traceability and even to ‘eco-branding’ for promoting sale of abundant stocks of wild-caught salmon, while protecting depleted stocks.
Supply chain simulations and sampling design
Given the diversity of wildlife and fisheries markets, molecular surveys will need improved supply chain models and sampling design for accurate estimates of species diversity and volume of trade (i.e. total takes). As with visual surveys of bushmeat markets (e.g. Fa et al. 2004), a key question in the design of molecular surveys is likely to be the trade-off in the frequency of surveys, the duration of surveys and the number of samples collected on each survey. Unlike visual surveys, which rely on counts of carcases, molecular surveys will also need to account for the distribution of multiple products from butchered or flensed individuals (e.g. Clarke et al. 2006a; Baker et al. 2007). For the Korean whalemeat markets, Leaper & Cooke (2006) adapted an analytic framework developed for food supply chain networks to evaluate the bias and precision of the capture — recapture estimates of whales in trade. A simulation model allowed for variability in the sizes of whales, the seasonality of supply (as most whales are migratory), the distribution pathways and retail demand, as well as number of samples per survey and the frequency of surveys. The simulations confirmed that the capture —recapture analysis generally yielded reliable, but negatively biased, estimates of the number of whales in trade and the half-life of these products for the assumed structure of the Korean market. A dominant factor in the success of a simulated survey estimate was the relationship between the product half-life and the survey interval. For the estimated half-life of about 2 months, surveys conducted 9 weeks apart always performed better than those conducted twice as far apart, even with only half the sample size for each survey (Leaper & Cooke 2006). Simulation models incorporating available information from visual surveys or market records will be needed to design efficient sampling strategies for future molecular surveys of mixed-species markets (e.g. bushmeat) or markets with more complex supply chains (e.g. Japanese whalemeat).
The methods of molecular ecology are universally applicable to the monitoring and surveillance of fisheries and wildlife markets, including bycatch and bushmeat, despite the often different government agencies and international conventions regulating exploitation of these species. For species taken as part of a regulated hunt or fishery, molecular monitoring can complement source-point mechanisms of control, such as trade records, DNA registers and catch document systems. For the large component of trade that is unregulated, molecular surveys provide an empirical approach to surveillance and estimation of species diversity and market volume. The studies reviewed here show great promise in providing a truer measure of exploitation by applying both approaches to these end-points of the supply chain. To date, however, the power of molecular ecology has been applied to only a small fraction of markets, for a small fraction of the species in trade. To redress this deficiency will require a greater commitment by government and intergovernment agencies to implement molecular monitoring through formal requirements for control of regulated trade, and a greater effort by independent scientists and non-government organizations to initiate molecular surveillance of unregulated markets.
Although the focus of this review has been on wildlife and fisheries, it is worth noting that similar approaches are being applied to plants and timber (e.g. Deguilloux et al. 2002), offering a mechanism to control one of the largest components of illegal trade in natural products: illegal logging.
The market supply chain structure of ‘whalemeat’ products arising from whales taken as illegal, unregulated or unreported (IUU) exploitation in Japan and Korea.The potential differences in supply chain structure and dynamics of wildlife and fisheries markets can be illustrated by the distribution of products from the IUU exploitation of whales in Japan and Korea (Fig. 1; IWC 2006b). Although an international moratorium on commercial whaling has been in effect since 1986, Japan maintains a commercial market for whalemeat through sale of products from several species of large whales killed in its controversial programme of scientific research (‘scientific whaling’, Morell 2007). Whales killed as fisheries bycatch (or hunted illegally) enter a complex, nationwide supply chain shared with products from this scientific whaling programme and with products from small cetaceans killed in unregulated coastal hunting (Baker et al. 2000a). Whalemeat can be sold as fresh ‘sashimi’ or as processed products such as ‘bacon’. The long-term frozen storage of whalemeat (for more than 10 years, according to the whaling industry, Mills et al. 1997) is thought to be common. Products are often packaged but incomplete or inaccurate labelling of species origin is common.In Korea, information collected during market surveys since 1994, including the genetic tracking of products from individual whales (see Genetic tracking and ‘DNA registers’, Dalebout et al. 2002), suggests a relatively simple supply chain. Although South Korea has no programme of scientific or commercial whaling, whalemeat markets are supplied by whales killed as bycatch in coastal fisheries (Kang, Phipps 2000) and by some level of illegal hunting (see Modelling the market). The wholesale value of an adult minke whale is reportedly up to US$100 000 (Neff 2004), providing considerable incentive for this IUU exploitation. Whales are transported to one of three coastal cities, Busan, Ulsan or Pohang, where they are flensed and distributed by a relatively small number of wholesalers (~9–10) to as many as 100 outlets (shops and restaurants), some of which are nested within four large fisheries markets (e.g. Jagalchi market in Busan, IWC 2006c). Whalemeat is usually sold ‘freshly parboiled’ and long-term storage of frozen products is thought to be uncommon. Products are seldom packaged and almost never labelled by species.In both Japan and Korea, the largest component of market products from IUU exploitation originates from North Pacific minke whales (Balaenoptera acutorostrata scammoni), particularly from the coastal stock inhabiting the East Sea/Sea of Japan between Korea and Japan (referred to as the ‘J’ stock). The ‘J’ stock was depleted by commercial whaling before 1986 and is classified by the IWC as a protection stock (IWC 1997).Despite the differences in Japanese and Korean supply chain structure and dynamics, molecular surveys directed at commercial outlets in both countries have provided accurate estimates of total IUU exploitation of this depleted stock using standard mixed-stock and novel capture–recapture analyses (see Modelling the market). Under the combined pressure of this exploitation, population dynamic models used by the IWC predict a decline towards extinction of this genetically distinct, coastal population (Baker et al. 2000b).
I thank Don and Sue White of Earthtrust for initiating the first survey of whalemeat markets in Japan and Naoko Funahashi, Sidney Holt and Vassili Papastavrou of the International Fund for Animal Welfare (IFAW) and John Frizell of Greenpeace International for long-term support of surveys in both Japan and Korea. I am indebted to my collaborators, Steve Palumbi, Frank Cipriano, Gina Lento, Merel Dalebout, Yong-Un Ma, Justin Cooke, Debbie Steel, Shane Lavery and Vimoksalehi Lukoschek for their contributions to this work over the years. Our findings would not have received proper consideration by the International Whaling Commission without the interest of Mike Donoghue, Jim McLay, Sir Geoffrey Palmer and Bob Brownell. This manuscript benefited from constructive comments of four anonymous reviewers and the Associate Editor, Louis Bernatchez, as well as discussions with Brian Bowen, Phil Clapham, John Fa, Steve O’Brien and Steve Palumbi. Preparation of the manuscript was supported by the Endowment of the Marine Mammal Institute, Oregon State University.
Scott Baker's research includes both molecular and demographic approaches to the investigation of evolutionary pattern and process in whales and dolphins, particularly their abundance, population structure, genetic diversity, taxonomy and systematic relationships. He is a member of the Cetacean Specialist Group of the International Union for the Conservation of Nature (IUCN) and has acted as a delegate to the Scientific Committee of the International Whaling Commission (IWC) since 1994.