Portal de Periódicos da CAPES

Artigo Acesso aberto Produção Nacional Revisado por pares

BIBTEX RIS

Solving the South Atlantic puzzle requires more data, not more speculation

2003; Wiley; Volume: 30; Issue: 9 Linguagem: Inglês

10.1046/j.1365-2699.2003.00952.x

1365-2699

J.‐C. Joyeux, Sergio R. Floeter, Carlos E. L. Ferreira, João Luiz Gasparini,

Fish Biology and Ecology Studies

Almost 2 years after the publication of the study on biogeographic patterns of South Atlantic reef fishes (Joyeux et al., 2001), Moura (2003) chose to answer it not as a new article, where he could really explore ‘interesting ideas’ and test hypotheses, but as a correspondence. There, he argued we wrongly compiled data, superficially read and ambiguously quoted the literature, imagined phylogenetic relationships, and described flawed biogeographic patterns. Here, we propose to demonstrate that Moura (2003), in order to condemn such ‘unethical behaviour’, greatly misrepresented our data and hypotheses and used much ambiguous or wrongly quoted literature. In his critical comment, Moura (2003) seems to be worried about the introduction of erroneous data in the current literature. However, divergent interpretations of the literature may generate the so-called ‘unsubstantiated patterns’. A clear example can be found in his own comment: Moura (2003) cites Williams & Smart (1983) in his second paragraph to illustrate new species discoveries in Brazil during the 1980s. However, Williams & Smart (1983) did not discover a new species; they provided new information about Starksia brasiliensis, a species that had been previously described by Gilbert (1900). Moura (2003) also states that the Brazilian Sparisoma atomarium (Poey 1861) is an undescribed species but, contrarily to his claimed principles against ‘unsubstantiated distributional and phylogenetic patterns’, partly bases his statement on his 2001 paper (Moura et al., 2001), which does not explicitly say so. In his words, S. atomarium is ‘considered widespread in the west Atlantic’ and ‘appear[s] to be absent from north-eastern Brazil and [is] restricted to the continental area between Espírito Santo and São Paulo’. Phylogenetic relationships of Sparisoma, proposed by G. Bernardi and colleagues (unpublished), were simply not available at the time of our writing. Further, when Moura (2003) indicates the presence of ‘Chaetodon striatus (Linnaeus 1758) and B. pulchellus in Atol das Rocas’ [sic], he corroborates his statement about our ‘incomplete distributional database’ by citing his own unpublished and inaccessible information. Indeed, Chaetodon striatus was independently detected in Atol das Rocas, but seems to be very rare (C. Buitrón, pers. comm.). The maps in our study (Fig. 1 in Joyeux et al., 2001) were never intended to exactly represent the distribution of all indicated species, but to best synthesize distributional information available at the time across a series of unrelated taxa. Such multi-specific patterns are always described in detriment to a high precision that is generally not useful for the detection and comprehension of general patterns. It is possible, for example, that the deep water species Halichoeres bathyphilus (Beebe & Tee-Van 1932) does not present an antitropical distribution since one individual was recorded in Bahia 12°S (A. Carvalho-Filho, pers. comm.). Thus, we agree with Moura (2003) that the term used in the legend of the figure may have been too vague to adequately cover or describe all presented distributions. A better expression for both Chromis flavicauda (Günther 1880) and H. bathyphilus (Fig. 1c in Joyeux et al., 2001) would be ‘anti-Caribbean’ (but see Cervigón, 1993). However, ongoing meristic and genetic studies by L. Rocha and colleagues with C. flavicauda and J. L. Gasparini and colleagues with the Brazilian H. bathyphilus may reveal that the northern and southern components of these disjunct distributions are indeed different species. Regarding the presence of Epinephelus niveatus (Valenciennes 1828) along the north-eastern coast, new records were published by Ferreira & Cava (2001) for Pernambuco (08°S) and various specimens were caught or sighted in Paraíba (L.A. Rocha, pers. comm.) and in Amapá (A. Carvalho-Filho, pers. comm.) (07°S and 00°N, respectively). At the time of our writing (Joyeux et al., 2001), the unpublished information (P.S.A. Costa, pers. comm.) indicated that the northern limit of distribution of this species was southern Bahia, 16°S. This led us to state that it was restricted to the southern coast, a limit that may be politically and geographically incorrect but that seemed justified at the time. Holacanthus tricolor (Bloch 1795) was considered resident at Trindade due to its abundance (Gasparini & Floeter, 2001). Sightings of 10–20 individuals per dive were common. Dermatolepis inermis (Valenciennes 1833) is very rare on the coast and this prompts us to consider the species vagrant (i.e. ‘exceedingly rare’; Joyeux et al., 2001). The terms ‘ecological replacement’ or ‘competitive exclusion’ were probably too strong, but it is interesting to note that Pomacanthus paru (Bloch 1787) is common at St Paul's Rocks (Feitoza et al., in press), Atol das Rocas (L. Candisani, pers. comm.) and Fernando de Noronha (S.R. Floeter, pers. obs.), whereas H. tricolor is rare or absent at these locations (Floeter et al., 2001). However, H. tricolor is common at Trindade Island and on the Vitória–Trindade Chain, locations where P. paru and Holacanthus ciliaris (Linnaeus 1758) were never recorded to date (Gasparini & Floeter, 2001). The reason why some of these spongivores are common in some islands and absent in others needs further investigation. Since the publication of our paper (‘Jouyeux et al., 2001’ [sic] in Moura, 2003), the idea of prevailing south to north colonization across the Amazon delta was further reinforced by new records of species previously considered Brazilian endemics at the southern tip of the Caribbean (Rocha, in press). Although Chromis scotti Emery 1968 has been collected from a few sites in north-eastern Brazil (information already presented in our paper; e.g. Fig. 1g in Joyeux et al., 2001), it is still not known southwards from the hump of Brazil. Moura (2003) also made a mistake stating that Halichoeres radiatus (Linnaeus 1758) does not occur south of the Equator, as the species is present in Atol das Rocas and Fernando de Noronha (3°50′S) (Rocha & Rosa, 2001b) and strays (i.e. vagrants) were found in the NE Brazilian coast (L.A. Rocha, pers. comm.). East-to-west migration was hypothesized (e.g. the expressions ‘most probably crossed the other way’ and ‘putative colonisers’ in Joyeux et al., 2001, p. 837; see also p. 832) for Acanthurus monroviae (Steindachner 1876), Epinephelus marginatus (Linnaeus 1758) and Aulostomus strigosus Valenciennes 1842. We assume their eastern origin based on the large geographic area that these species occupy in the eastern Atlantic vs. their narrow distributional range in the western Atlantic (for an extreme example, only a few vagrant A. monroviae were recorded so far; O.J. Luiz-Júnior et al., submitted to Journal of Fish Biology). This same criterion was used by Briggs (1974, 1995) when asserting a western origin of other amphi-Atlantic fish species, and Rocha (in press) when discussing the Brazilian origin of a few species that are, in the northern part of their range, restricted to the southernmost areas of the Caribbean. Actually, most amphi-Atlantic fishes are believed to originate from the western Atlantic (Briggs, 1974). Of ninety-two reef-associated amphi-Atlantic teleost species, only four (3.68%) seem to have migrated from east to west (S.R. Floeter, pers. obs.): the three discussed above and Parablennius pilicornis (Cuvier 1829; Blenniidae) (O.J. Luiz-Júnior et al., submitted). We anxiously await the ‘genetic or long-term ecological studies’ that may support or falsify our hypothesis. Concerning the absence of south-western Atlantic endemic genera, both Joyeux et al. (2001) and Moura (2003) are wrong. There are (at least) two endemic monotypic genera in the south-western Atlantic: Ribeiroclinus Pinto 1965 (R. eigenmanni [Jordan 1888]; from Rio de Janeiro to Argentina) and Storrsia Dawson 1982 (S. olsoni Dawson, 1982; apparently endemic to Fernando de Noronha). These two unique and atypical species (from a pool of more than 400 Brazilian reef fish) do not invalidate the fact that all other reef-associated species are from genera also found in the Caribbean. They can be nicely incorporated in Fig. 1i of Joyeux et al. (2001), although neither is large or conspicuous. Finally, concerning the use of subspecies, we followed Nelson (1999) who considered the subspecies concept ‘useful in studies of biodiversity or conservation where we wish to draw attention in formal taxonomy to cases where there has been some recognisable divergence […] within a species’. Many of the phylogenetic relationships we hypothesized (e.g. ‘PUTATIVE sister species’ in Table 1 inJoyeux et al., 2001) proved to be right. New molecular studies (G. Bernardi and collaborators, unpublished) confirmed Thalassoma noronharum (Boulenger 1890) and Thalassoma bifasciatum (Bloch 1791) as sister species and raised the chances for a specific status for the Brazilian population of S. atomarium (species currently being studied by Gasparini and collaborators). Similarly, a phylogenetic study of Pomacanthidae using allozymes (Chung & Woo, 1998) established that Pomacanthus paru and P. arcuatus (Linnaeus 1758) are closely related, as are H. tricolor and H. ciliaris (a relationship we were not aware of). In all cases, we based our suppositions on general likeliness, extreme similarity in colour pattern and/or total overlap in meristic data for all counts [e.g. Stegastes fuscus (Cuvier & Valenciennes 1830) and S. imbricatus Jenyns 1840 – for photographs, see Allen, 1991, p. 175, and Rocha, 2002, p. 464; Gasparini et al., 1999; G. Allen, pers. comm.]. Rocha (in press) used similar criteria to analyse speciation in the western Atlantic by considering similar species as closely related pairs: ‘…those pairs are very similar in general morphology, colour and ecological requirements (being herein tentatively considered sister species)…’. The case of Aulostomus is complex and unresolved, mainly because the population genetics study of the ring species (Bowen et al., 2001) was not intended as a phylogenetic study and lacks an outgroup. Moura (2003) stated that we presented no evidence of differentiation of Chromis cf. multilineata (Guichenot 1853) at St Paul's Rocks. It, therefore, seems appropriate to comment that an aberrant coloured C. multilineata was recently photographed, just like some fishes described by Lubbock & Edwards (1981) 22 years ago (B. Feitoza, pers. comm.). Also, the blue colour morph of H. ciliaris from St Paul's Rocks was recently photographed by Yataka Niino at a Japanese aquarium, leaving no doubt that the aquarium collectors are now operating even at this remote islet (R. Kuiter, pers. comm.). In conclusion, it has been 2 years since we presented our opinions as hypotheses to be further tested (Joyeux et al., 2001). Instead of attempting to demolish our work, Mr Moura should have dedicated his time and energy to use his own work and analyses to falsify (or confirm) our hypotheses. Moreover, his unpublished information on distributional records would greatly improve future analyses if it were made available to the ichthyological community. However, the only comprehensive Brazilian reef fish lists published are those by Lubbock & Edwards (1981), Rocha et al. (1998), Ferreira & Cava (2001), Gasparini & Floeter (2001) and Rocha & Rosa (2001a). Neither minor mistakes in distributional data nor new data published since our paper invalidate the hypotheses presented by us. We thank Luiz A. Rocha, Osmar J. Luiz-Júnior, Vicente V. Faria and Peter Wirtz for discussions and helpful comments on earlier drafts of this comment. Unpublished information was kindly provided by Gerry Allen (Conservation International), Giacomo Bernardi (University of California, Santa Cruz), Cristina Buitrón (Bimini Biological Field Station, Bahamas), Luciano Candisani (Candisani Fotografia, São Paulo), Alfredo Carvalho-Filho (Fish Bizz, Ltd, São Paulo), Bertran Feitoza (Universidade Federal da Paraíba, João Pessoa), Rudie Kuiter (Zoonetics erBooks, Australia), Osmar J. Luiz-Júnior (ProDivers Scuba Dive Centre, São Paulo), Paulo A.S. Costa (Universidade do Rio de Janeiro) and Luiz A. Rocha (Smithsonian Tropical Research Institute, Panama). Dr Jean-Christophe Joyeux is Professor of Ecology at the Universidade Federal do Espírito Santo and is currently leading various projects on larval fish ecology. Sergio R. Floeter has been granted a post-doctoral fellowship by the National Center for Ecological Analysis and Synthesis, Santa Barbara, California, USA, to further analyse the distributional database for the tropical Atlantic reef fishes. Dr Carlos Eduardo L. Ferreira works as marine researcher at IEAPM Institute (Brazilian Navy), and has been dealing with various aspects of the ecology of Brazilian reefs and fishes for more than a decade. João Luiz Gasparini is affiliated to the Universidade Federal do Espírito Santo, and has been working for the last 15 years on the taxonomy, systematics, and natural history of Brazilian reef fishes.