6 Spring 2005 Molecular Genetic Analysis Reveals Six Living Subspecies of Tiger, Panthera tigris Stephen J. O’Brien, Shu-Jin Luo, Jae-Heuoup Kim and Warren E. Johnson 1 T igers historically inhabited much of Asia and likely num- bered near 100,000 as recently as a century ago (Fig. 1). Today’s tiger census is much lower, and is estimated by various sources to be around 7000 individuals in the wild (Nowell & Jackson,1996; Dinerstein et al 1997; Kitchener & Dugmore 2000). Tigers have been traditionally classified into eight subspecies (Fig. 1), three of which (P. t. sondaica-Ja- van tiger; P. t. balica-Bali tiger and P. t. virgata-Caspian tiger) were lost to extinction in the mid to late 20th century. The challenge to preserve the existing tiger populations has become a major goal of conservation efforts throughout their range. As with many endangered species, tigers have been classified into subspe- cies - natural geographically separate populations - for purposes of recogni- tion and conservation. The subspecies concept is controversial, but many sub- species including those for tigers are considered as specific units of conser- vation and are protected by treaties and organizations that are concerned with the management and stewardship of species. For this reason alone, the esta- blishment of a formal subspecies defini- tion, an explicit basis for subspecies re- cognition, and an understanding of the implications of subspecies assignment become critically important. In 1991, in collaboration with the distinguished evolutionary biologist Ernst Mayr, one of us (SJOB) proposed some working guidelines for subspecies considerati- ons (O’Brien & Mayr 1991). In that essay we defined subspecies as “geographically defined aggregates of local populations which differ taxo- nomically from other species subdivi- sions.” To help government regulators recognize subspecies we suggested gui- delines. Members of a subspecies share a unique geographical range or habitat, a group of recognizable genetically con- trolled characteristics, morphological or molecular, and a unique natural history as compared to other subspecies. Since subspecies are not distinct species, they are reproductively compatible and will periodically interbreed with adjacent subspecies. All subspecies have the po- tential to acquire suitable adaptations to their specific ecological habitat and the longer they are separated the more cumulative adaptation we might expect. All subspecies also have the potential to one day evolve into new species as sug- gested by Charles Darwin in “On The Origin of Species” in 1886. These two potentials, which are unfortunately not certain for any individual subspecies, nonetheless provide compelling rationa- le for their conservation management. Recognition and pronouncement of a subspecies require the description of objective heritable characters that eve- ry individual of the subspecies carries, which are in effect diagnostic for the subspecies; that is, they are found only in that subspecies and not in other po- pulations within the same species. Avise and Ball (1991) and we suggested that valid criteria for subspecies include concordant distribution of multiple in- dependent genetic traits. These can be morphological or molecular or both. Traditional morphology-based assess- ment (body size, skull characters, pe- lage coloration, and striping patterns) have been applied to tiger subspecies but have been equivocal in adequately describing tiger subspecies (Kitchener, 1999; Herrington,1987; Mazak,1981). Previous molecular studies (Wentzel et al. 1999; Hendrickson et al 2000; Cracraft et al. 1998) have also been dis- appointing in affirming the commonly accepted tiger subspecies designations (Fig. 1). In December 2004, the culmination of a twenty-year long study to charac- terize living tiger populations and sub- species differentiation using molecular genetic approaches was published in the new free-online-access journal Pu- blic Library of Science PLoS-Biology (Luo et al. 2004: http://biology.plos- journals.org/archive/1545-7885/2/12/ pdf/10.1371_journal.pbio.0020442- S.pdf). In that study, we and twenty additional authors (including members of the IUCN-Cat Specialist Group and pioneers in tiger ecology, behavior, and conservation: Joelle van der Walt, Janice Martenson, Naoya Yuhki, Dale G. Miquelle, Olga Uphyrkina, John M. Goodrich, Howard B. Quigley, Ronald Tilson, Gerald Brady, Paolo Martelli, Vellayan Subramaniam, Charles Mc- Dougal, Sun Hean, Shi-Qiang Huang, Wenshi Pan, Ullas K. Karanth, Melvin Sunquist, and James L. D. Smith) ex- amined “voucher specimens” (biolo- gical blood and skin materials) from 134 tigers born in the wild at a known location or descended directly from pa- rents of known geographic origins. The paper described the phylogeography patterns using three distinct families of variable genetic markers: 1.) 4000 nu- cleotide letters of mitochondrial DNA sequence; 2.) a highly variable nucle- ar DNA sequence FLA-DRB (an im- mune response gene within the tiger’s major histocompatibility complex) and a group of short repetitive nuclear ele- ments called microsatellites. This was a rather large data set analyzed with the most advanced population genetic and phylogenetic computational algorithms available to molecular geneticists. The results were interpreted together and converged on a rather illuminating and in most cases statistically robust (me- aning high confidence) picture of the tiger’s natural history and subspecies recognition. Here is what the data sho- wed about living tigers. First, there was strong genetic evi- dence for the separation and recognition of four of the five traditional subspecies: (1) Amur tiger P. t. altaica; (2) Indochi- nese tiger P. t. corbetti (3) Sumatran ti- ger P. t. sumatrae and (4) Bengal tiger P. t. tigris. (Fig. 1). Second, Indochinese tiger P. t. corbetti