Two living species of coelacanths?
The discovery of a living coelacanth off the coast of South Africa in 1938 (1) surprised the scientific community, be- cause the fish is a member of a lineage that was thought to have gone extinct almost 80 million years ago (MYA) (2). A subse- quent 14-year search for a second specimen eventually revealed the ‘‘true’’ home of Latimeria chalumnae in the Comores Islands in the western Indian Ocean (3). Since that discovery in 1952, more than 200 coelacanths have been captured in or near the Comores, and by 1994 the population size was thought to have dwindled to a few hundred animals (4). The discovery of coelacanths off the coast of Manado Tua Island, Sulawesi, Indonesia, was reported in 1998 (5). These coelacanths are the first individuals recorded from a location outside of the western Indian Ocean, and extensive interviews with Indonesian fisher- men have revealed a history of catches in the north Sulawesi area (6).[unique_solution] These facts and the general pattern of current flow from north Sulawesi toward the Comores (7) imply that the Indone- sian specimens cannot be regarded as waifs from the Comorean population (8). Rather, the Indonesian and African Latimeria may represent members of a widespread population, members of separate but conspecific populations (perhaps with one popu- lation being recently established via long-distance dispersal), or two separate species. Ideally, a wide range of systematic and ecological information would be used to distinguish among the three possibilities listed above. Unfortunately, the preferred habitat of the coelacanth at depths exceeding 150 m (9) precludes the regular collection of relevant populational or ecological data. Detailed knowledge of the ecology of these fish in the Comores is limited largely to the findings from six submersible expeditions (4), and the natural history of the fish in Indonesia is virtually unknown. In lieu of in-depth knowledge of population structure of two populations, systematists typically use measures of divergence in morpholog- ical and molecular characters to gauge whether populations merit specific status. However, there are several caveats to consider. In considering differences in morphology, such char- acters may be influenced heavily by environmental difference between two localities. Although molecular characters are rarely affected by environment, molecular divergences reflect the phylogenies of alleles, and this does not always correspond to the history of the populations in which the genes are found. Fur- thermore, the number of molecular differences that are expected between two sibling species varies widely depending on the gene studied and the taxa under consideration. Finally, although molecular divergence data may be useful for dating the timing of separation between two populations, rates of molecular evolu- tion vary across lineages, so estimates of the time of divergence of two alleles must consider a range of substitution rates. Based on a preliminary comparison of external morphological measurements from the Sulawesi specimen with those of L. chalumnae from literature reports, Erdmann et al. (6) concluded that the Indonesian coelacanth is morphologically extremely similar to L. chalumnae. Those authors suggested further that the specific status of the Indonesian fish would be resolved only after careful consideration of the results of both an ongoing genetic comparative analysis and a detailed morphological examination planned by the Indonesian Institute of Sciences. Pouyaud et al. (10) described the Indonesian coelacanth as a new species, Latimeria menadoensis, based on both molecular and morphological grounds. Briefly, those authors sequenced 1,829 bp from parts of two mitochondrial genes of the Indonesian coelacanth and reported those sequences to be 4.85% divergent from cytochrome b and 2.85% divergent from 12S rRNA sequences published from a Comorean specimen (11). They estimated the time since the last common ancestor to be 1.22 million years based on cytochrome b and 1.42 million years based on 12S rRNA sequences. Additionally, Pouyaud et al. (10) reported nine morphological traits that purportedly differentiate L. menadoensis from L. chalumnae (2). We have collected 4,823 bp of mitochondrial DNA sequence from the same specimen sampled by Pouyaud et al. (10) and have examined the purported morphological differences between the Indonesian and Comorean fishes. Methods Gill tissue from the second known coelacanth captured off Manado Tua [MZB10003; Coelacanth Conservation Council (CCC) no. 175] was preserved in ethanol immediately after death. DNA was obtained by digestion of the sample with proteinase K in sodium chlorideyTrisyEDTA and 1% SDS. The lysate was purified by two extractions with phenol and chloro- form followed by two extractions with chloroform. Extracted DNA was precipitated by using NaCl and ethanol. DNA was