VICARIANT SPECIATION OF CURASSOWS (AVES, CRACIDAE): A HYPOTHESIS BASED ON MITOCHONDRIAL DNA PHYLOGENY

Auk, The, Jul 2004 by Pereira, S�rgio Luiz, Baker, Allan J

ABSTRACT.-

The curassows comprise 14 species of sedentary Neotropical birds classified in four genera (Crax, Nothocrax, Mitu, and Pauxi) in the family Cracidae. Congeneric species have a striking pattern of allopatric distributions that might be attributable to vicariance, dispersal, or a combination of the two. To test those biogeographic hypotheses, a strongly supported phylogeny was needed, so that existing problems of taxonomic rank could be solved and a better understanding of the group's evolutionary history attained. We therefore estimated the phylogenetic relationships of all 14 species, on the basis of 6,929 sites of six different mitochondrial DNA regions, and reassessed the status of the four genera. Sequences from the ND4 gene favored a tree that was highly incongruent with the tree recovered using the other five gene regions. However, when the ND4 sequences were concatenated with the sequences of the other genes, the optimal phylogeny was unchanged from that derived for the other genes. That combined tree was divided into two well-supported clades: one containing the seven species of Crax and the other containing the monospecific genus Nothocrax, as sister to a clade of the Mitu and Pauxi species. Mitu and Pauxi are not reciprocally monophyletic, which appears to be attributable to a distant hybridization event and a transfer of Mitu mtDNA into P. unicornis. We estimated divergence times; the diversification of curassow seems to have occurred from the Middle Miocene to the end of the Pliocene (9.5 to 1.6 Ma). Vicariance -following marine transgressions, the rise of the Andes, and subsequent changes in river basins in South America-seems to be the major mode of isolation that favored allopatric speciation in the group. Received 20 January 2003, accepted 23 February 2004.

RESUMEN.-El grupo de los pavones ? paujiles incluye 14 especies de aves neotropicales sedentarias clasificadas en cuatro gener�s (Crax, Nothocrax, Mitu y Pauxi) en Ia familia Cracidae. Las especies cogenericas presentan un llamativo patron de distribuciones alop�tricas, que podria ser atribuido a vicarianza, dispersion ? a una combinacion de los dos procesos. Para poner a prueba estas hipotesis biogeogr�ficas, resultaba necesario contar con una filogcnia fuertemente respaldada de modo que los problemas de rango taxonomico existentes pudieran resolverse y se alcanzara un mejor entendimiento de la historia evolutiva de] grupo. For Io tanto, estimamos las relaciones fllogen�ticas entre las 14 especies con base en 6,929 posicioncs nucleotidicas de seis regiones mitocondriales diferentes, y revaluamos el estatus de los cuatro gener�s. Las secuencias del gen ND4 sugirieron un �rbol que fue sustancialmente incongruente con el �rbol recobrado usando las otras cinco regiones. Sin embargo, cuando las secuencias de ND4 se concatenaron con las de los dem�s g�nes, la filogcnia optima no difirio de la derivada a partir de �stos. Este �rbol combinado estuvo dividido en dos clados con buen respaldo, uno incluyendo las siete especies de Crax y el otro incluyendo al g�nero monotipico Nothocrax como hermano de un clado compuesto por las especies de Mitu y Pauxi. Mitu y Pauxi no son grupos reciprocamente monofil�ticos, Io que parece ser atribuible a un evento distante de hibridacion y a la transferencia de ADNmt de Mitu a P. unicornis. Tambi�n estimamos tiempos de divergencia; la diversificacion de los pavones parece haber sucedido desde el Mioceno medio hasta el final del Plioceno (9.5 a 1.6 millones de anos atr�s). Los procesos de vicarianza causados por transgresiones marinas, el levantamiento de los Andes y los cambios subsecuentes en las cuencas de los nos de Sur America parecen representar el principal modo de aislamicnto que favorecio la especiacion alop�trica en el grupo.

BIOLOGISTS HAVE LONG debated the relative importance of dispersal and vicariance in explaining disjunct geographic distributions and speciation of organisms (Zink et al. 2000, Crisci 2001). Proponents of dispersal (e.g. Udvardy 1969) hypothesize that new species originated after individuals of an ancestral population crossed geographic barriers and colonized an area isolated from the original population. Proponents of vicariance (e.g. Nelson and Platnick 1981) argue that speciation commonly occurred when a continuous ancestral population was fragmented by a geographic event, such as uplift of a mountain chain or change in a river's course; provided that gene flow was restricted between ancestral and derived populations, the latter accumulated genetic incompatabilities and eventually became new species. Vicariance and dispersal are not mutually exclusive propositions, however, and both are likely to have occurred in the historical biogeography of major clades of organisms (e.g. Ronquist 1997).

Availability of DNA sequences has provided a new opportunity to test the roles of dispersal and vicariance in biological diversification; sequences not only can be used to construct hypotheses of phylogenetic relationships but also allow estimation of approximate times of divergence of taxa. If allopatric taxa are older than the barrier that separates them, vicariance is the most parsimonious explanation for their divergence; if the barrier is older, dispersal is most parsimonious. Although phylogenies of multiple lineages can be useful in checking whether patterns of distribution correspond with major geographic barriers (Cracraft and Prum 1988, Prum 1988, Zink et al. 2000, Hovenkamp 2001), that comparative approach is presently limited by availability of well-supported phylogenies and estimates of times of divergence.