The molecular data from both Asian and African elephant populations, although still in early analysis, are already threatening to overturn the traditional systems of classification. At our present level of understanding of the genetics of Asian elephants, there is no support for the subspecies status of the Sri Lankan elephant population (Elephas m. maximus) and its differentiation from those in mainland Asia (E. m. indicus). These studies and unpublished data clearly indicate that several mitochondrial haplotypes found in Sri Lankan elephants are shared with those in the mainland, particularly southern India. On the other hand, there is some support for the differentiation of the Sumatran elephant (E. m. sumatranus) and the Malaysian elephant, although more work needs to be done.
The more revolutionary shift in elephant taxonomy seems to be happening with Loxodonta. The wide variation in body size (appendix 2) and morphology of African elephant populations has always brought demands for subspecies or species status for some of these forms. For decades, the legend of the pygmy elephant persisted in Central Africa. David Western investigated this legend in 1986 and found that, in regions where savanna elephants and forest elephants intermingled to a certain extent, the forest subspecies was naturally regarded as a "pygmy" relative to the much larger savanna animals. Based on morphological examination of skulls from savanna and forest elephant populations, Peter Grubb, along with Colin Groves and other associates, made a case in 2000 for giving species status to the savanna elephant (Loxodonta africana Blu-
menbach, 1797) and the forest elephant (Loxodonta cyclotis Matschie, 1900). They also recognized that some hybridization could be occurring where the two populations mingled.
In 1997, an article in the journal Science had hinted that the African forest elephant (L. a. cyclotis) may also be sufficiently differentiated in genetic terms from the savanna elephant (L. a. africana) to warrant distinct status as a species. The results of this work by Alfred Roca, Nicholas Georgiadis, and their associates, finally published in the same journal in 2001, also make a strong case for elevating the forest elephant to the status of a distinct species (Loxodonta cyclotis). The team sequenced four nuclear genes (1732 nucleotide base pairs) from 195 elephants in 21 populations (locations) across Africa to examine patterns of genetic divergence. Of this sample, 36 forest elephants from 3 locations (Dzanga-Sangha, Lope, and Odzala) and 121 savanna elephants from 15 locations formed two genetically distinct groups on the basis of each of the four genes as well as collectively. Within each of these two categories, it was difficult to differentiate individuals from their DNA sequences. For instance, the savanna elephant populations across the continents were indistinguishable, while the forest elephants of Dzanga-Sangha could not be differentiated from those at Lope. The forest elephants were also much more genetically diverse compared to the savanna elephants. At one site, Garamba, in the forest-savanna transition region of the Congo, the sampled elephants showed a combination of the otherwise distinct forest and savanna elephant genes, indicating a limited history of hybridization among their ancestors. The team concluded that the genetic distance between the forest elephant and the savanna elephant was 58% of the distance between Loxondonta and Elephas, and that the former had diverged about 2.6 My ago, which was sufficient time for differentiation at the species level.
One deficiency of the above study was the absence of samples from the scattered West African populations (see appendix 1), which include both the savanna-type and the forest-type elephants. When Lori Eggert, Caylor Rasner, and David Woodruff rectified this anomaly (by collecting samples from 10 locations there) and pooled the genetic results from the West African elephants from published results elsewhere in the continent, they came up with a far more complex picture of the phylogeography and taxonomy of Loxodonta. Their analyses included a portion of the mtDNA (from the standard cyt-b control region) and four microsatellite loci of the nuclear DNA.
Five clades emerged from the continent-wide mtDNA analysis, namely, (1) East, north Central, and South African savanna elephants, (2) West African forest and savanna elephants, (3) East, West, north Central, and South African savanna elephants, (4) Central African forest and savanna elephants, and (5) West African forest and savanna elephants. The microsatellite DNA analysis also showed rather similar clustering of populations from different regions of the continent. Of these divisions or clades, the Central African forest elephants were the most similar to the Asian elephant, perhaps not a surprising result. Eggert and associates suggested that the earliest members of modern Loxodonta
(africana) may have inhabited the Central African forests, about 5 My ago, and then repeatedly dispersed to other parts of the continent in response to episodes of climate change that resulted in contraction and expansion of forest or savanna. Three broad groups—the Central African forest elephants; the savanna elephants of central, eastern, and southern Africa; and the West African forest and savanna elephants—could be recognized at a minimum on the basis of genetics, geography, and ecology.
This study has obviously opened up a virtual Pandora's box of African elephant taxonomy. Thus, it seems premature at this time to differentiate Loxo-donta into two species. The IUCN/SSC African Elephant Specialist Group decided at their 2002 meeting to retain the traditional classification of two subspecies (L. a. africana and L. a. cyclotis) until further studies resolve the taxo-nomic uncertainties. Although rapid progress is being made, there is still some distance to go before the molecular genetic studies of elephant populations provide us with a satisfactory understanding of their phylogeny, differentiation, and diversity. The molecular studies are still based on unevenly distributed samples across the elephant's distributional range; the elephants of the Indian subcontinent are only now being analyzed. The nuclear DNA that constitutes the bulk of the genetic material in an organism has yet to be characterized in sufficient detail among elephant populations. The relationship between fossil and molecular dates for differentiation of populations is unclear; for instance, the population coalescent dates derived from mitochondrial haplotypes are not necessarily the same as the population divergence dates, as sometimes is assumed in genetic studies. More standard molecular genetic comparisons with larger samples of elephants, along with morphological data, would help to resolve these issues.
Was this article helpful?