Proposal (639) to South American Classification Committee
Split extralimital R. l. crepitans group from Rallus longirostris
Synopsis: The following is the full proposal submitted to NACC on revising species limits in the Rallus longirostris/R. crepitans group, sections of which are not directly relevant to SACC. The key issue is that the taxon we call Clapper Rail, R. longirostris, is not the sister species to the North American taxa currently including in that species. This proposal has been approved and adopted by NACC. The only taxa present in the SACC area are members of the Rallus longirostris group sensu stricto. A YES vote means that the only change to the SACC list would be the change in English name of more narrowly defined Rallus longirostris to “Mangrove Rail”, and a NO vote means no change.
Description of the Problem:
In the most recent checklist (AOU 1998:131), there is a discussion of hybridization between Rallus longirostris and R. elegans in the eastern and southern United States. There is a suggestion to merge the entire complex into a superspecies, but phylogenetic and detailed hybridization studies of the group have not been published until now (Maley 2012; Maley and Brumfield 2013). Taxonomy has always been difficult in this group given plumage variation, morphologically distinct allopatric populations, and uncertainty in the degree of hybridization between populations currently in contact. For example, R. elegans of the eastern US are bright rufous ventrally and breed in freshwater marshes, whereas R. longirostris of the eastern US are duller ventrally and breed in saltmarshes. Very similar allopatric birds of the southwestern US and northwestern Mexico are bright rufous ventrally and breed primarily in saltmarshes, making their classification into either species difficult (Olson 1997).
A phylogenetic study using mitochondrial and nuclear markers found discordance between genetic relationships and current classification (Maley and Brumfield 2013). Rallus elegans, as currently recognized, is paraphyletic with respect to R. longirostris. Genetic lineages correspond roughly to geography instead of current species limits. The R. l. obsoletus subspecies group found in California, Arizona, and northwestern Mexico was discovered to be sister to R. e. tenuirostris of the highlands of Mexico instead of previously suggested sister relationships to either R. l. crepitans or R. e. elegans of eastern North America (Hellmayr and Conover 1942; Ripley 1977; Olson 1997). Additionally, the lineages of the R. l. crepitans group and R. e. elegans, which are known to hybridize in eastern North America (Olson 1997), are in the same clade (Maley and Brumfield 2013). This pattern of hybridization apparently also occurs on Cuba (Olson 1997) between members of these same two lineages (Maley and Brumfield 2013). This clade also includes birds from throughout the Caribbean (Fig. 1). Detailed investigations of hybridization using morphological, ecological, and genetic (mitochondrial and nuclear) characters in Louisiana reveal that strong selection against hybrids is likely preventing the fusion of these lineages (Maley 2012). Members of the nominate R. l. longirostris group of South America were found to be genetically distinct and sister to Caribbean and eastern North American birds (Fig. 1B). In the study the authors were unable to obtain samples of R. l. longirostris, instead sampling two members of the group R. l. cypereti and R. l. phelpsi. The following recommendations would remove R. longirostris from the checklist, because members of this subspecies group have not been documented in North America.
Figure 1. Maximum Clade Credibility gene tree of ND2 inferred in Beast (Drummond and Rambaut 2007). The labels above nodes are the posterior probability followed by the bootstrap support value (if greater than 65) for that node. The labels below nodes are the posterior probability for that node in the estimate of the species tree; this label is not included if the value was below 0.95. Each of the three major clades is outlined and labeled by geography, with clade A comprising eastern North American and Caribbean birds, clade B comprising South American birds, and clade C comprising birds of western North America, including Mexico.
Two members of the complex are in extensive secondary contact in eastern North America and Cuba, but have not fused despite hybridization (Olson 1997). The morphological and genetic characterization of the hybrid zone in Louisiana (Maley 2012) found that it is very narrow (~ 4.2 km wide), with selection against hybrids acting to prevent fusion. These data suggest there is strong, albeit incomplete, reproductive isolation between these species in Louisiana. There is no evidence of population genetic structure within R. longirostris in the eastern US, and very little within R. elegans, so we extrapolate these results for the entire distribution in the eastern and southeastern US. Extending these results to the remaining taxa and considering the differential level of morphological, ecological and genetic divergence between previously identified subspecies groups, we conclude that at least five species should be recognized in this complex. This treatment would be consistent with recent genetic analyses of other members of the family showing similar levels of divergence (Tavares et al. 2010; Goodman et al. 2011). The most divergent clade within the complex, according to mtDNA data, represents a pair of subspecies groups from both currently recognized species (R. l. obsoletus group and R. e. tenuirostris). This pair shares the same pattern observed in the birds of eastern North America, where individuals of one group are relatively smaller than those of the other and are found primarily in saltmarshes (R. l. obsoletus group), whereas the other is relatively larger, brighter, and found in freshwater habitats (R. e. tenuirostris, Olson 1997).
We propose species rank for five members of the complex described below. These taxonomic recommendations are based primarily on two factors: 1) there is strong but incomplete reproductive isolation between parapatric populations based on hybrid zone analyses, 2) that each of the species represents a morphologically and genetically distinct group within the complex that is at least as distinct from other members of the complex as members that are currently in contact but showing evidence of reproductive isolation.
We propose recognizing as a species the nominate form R. longirostris Boddaert, 1783, plus the subspecies phelpsi Wetmore, 1941, margaritae Zimmer and Phelps, 1944, pelodramus Oberholser, 1937, cypereti Taczanowski, 1877, and crassirostris Lawrence, 1871. These birds are relatively very small, dull-breasted, robust-billed, and restricted to mangroves (Eddleman and Conway 1998), which is why we propose to give them the English name Mangrove Rail. This lineage is morphologically, genetically, and vocally distinct from all other members of the complex, and far more distinct from the rest of the complex than the members that are currently in contact are from one another.
The second species we propose is R. tenuirostris Ridgway, 1874, which includes the population of birds inhabiting the highland freshwater marshes of Mexico. Individuals are large, very bright rufous ventrally, and have diffuse flank banding (Meanley 1992). They are found almost entirely within the former Aztec Empire and are not the only member of the complex found in Mexico; thus we propose the English name Aztec Rail. They are distinct morphologically, genetically, and ecologically from their closest relative, in that they breed exclusively in freshwater marshes as opposed to saltmarshes, which is the same reproductive isolating mechanism as found in other lineages within the complex.
The third species we propose is R. obsoletus Ridgway, 1874, which includes the populations that occur along the Pacific Coast of North America. This species would include the subspecies levipes Bangs, 1899, beldingi Ridgway, 1882, yumanensis Dickey, 1923, rhizophorae Dickey, 1930, and nayaritensis McLellan, 1927. This group is characterized by their relatively small body size (although larger than South American birds), by a bright rufous breast, and by their occurrence primarily in saltmarshes (Eddleman and Conway 1998). Because Robert Ridgway contributed a significant amount of work on the complex, including describing R. l. obsoletus and R. l. beldingi, we propose the English name Ridgway’s Rail in his honor. We propose species rank using a comparative approach: because this lineage is as distinct morphologically, genetically, and ecologically from its closest relative (R. e. tenuirostris) as are other members of the complex in contact known to be reproductively isolated.
The fourth species in the complex we propose is R. elegans Audubon, 1834, comprised of two subspecies, R. e. elegans and R. e. ramsdeni Riley, 1913, while excluding R. e. tenuirostris (as described above). We propose retention of King Rail as the English common name. This species is distinct from its closest relatives ecologically, morphologically, and genetically. Despite hybridization, they are reproductively isolated from their closest relative in contact, members of the R. l. crepitans group, apparently due to ecological differences (Maley 2012).
The fifth species proposed is R. crepitans Gmelin, 1789, comprised of the eastern North America group of R. l. crepitans, including the subspecies waynei Brewster, 1899, scotti Sennett, 1888, insularum Brooks, 1920, and saturatus Ridgway, 1880, as well as the birds of the Caribbean and Yucatan, including R. l. caribaeus Ridgway, 1880, pallidus Nelson, 1905, grossi Paynter, 1950, belizensis Oberholser, 1937, leucophaeus Todd, 1913, and coryi Maynard, 1887. These birds are intermediate in size, and the breast spans a range of colors from very dull, silvery-gray, to dull rufous. They breed in saltmarshes and salt-meadows of the Atlantic and Gulf Coasts of North America, as well as mangroves in the Yucatan, extreme southern Florida, Louisiana, and Texas, and throughout the Caribbean (Eddleman and Conway 1998). We propose to retain Clapper Rail as the English common name to avoid confusion. They are distinct morphologically, genetically, and ecologically. Despite hybridization, they are reproductively isolated from the other members of the complex they are in contact with, R. e. elegans and R. e. ramsdeni (Maley 2012).
Goodman, S. M., M. J. Raherilalao, and N. L. Block. 2011. Patterns of morphological and genetic variation in the Mentocrex kioloides complex (Aves: Gruiformes: Rallidae) from Madagascar, with the description of a new species. Zootaxa 2776:49-60.
Maley, J. M., and R. T. Brumfield. 2013. Mitochondrial and next-generation sequence data used to infer phylogenetic relationships and species limits in the Clapper/King rail complex. Condor 115:316-329.
Olson, S. L. 1997. Towards a less imperfect understanding of the systematics and biogeography of the Clapper and King rail complex (Rallus longirostris and R. elegans). in The Era of Allan R. Phillips: A Festschrift. (R. W. Dickerman, Ed.). Horizon Communications, Albuquerque, NM.
Tavares, E. S., G. H. J. de Kroon, and A. J. Baker. 2010. Phylogenetic and coalescent analysis of three loci suggest that the Water Rail is divisible into two species, Rallus aquaticus and R. indicus. BMC Evolutionary Biology 10:226.
James M. Maley and Robb T. Brumfield, August 2013
Comments from Remsen: “YES I reviewed the paper at an early stage and consider the authors’ taxonomic arrangement to be the one that matches best the existing data. The elegans and crepitans groups have extensive, multiple contact zones, yet hybridization is limited by apparent selection against hybrids; thus, they have to be treated as separate species. Ripley’s treatment (in his Rallidae monograph) of them as conspecific is incorrect. Given that the other two groups, including our R. longirostris group, are successively more distantly related to the two for which we have a test of sympatry, the logical taxonomic treatment is to consider them each also as separate species.
“The Maley-Brumfield name Mangrove Rail is a good one for reasons stated in their paper. Although the usual policy is to christen each daughter species from a split with new names to avoid confusion, phylogenetically this is not a case of splitting a species into two daughters: in the case of R. crepitans and R. longirostris, they are not close to being sister taxa, so Maley & Brumfield, followed by NACC, retained the long-established “Clapper Rail” for the R. crepitans group. What NACC calls extralimital R. crepitans is actually irrelevant to SACC, but I mention this here in case the issue is raised.”
Comments from Zimmer: “YES” to change the English name of the more narrowly defined Rallus longirostris to “Mangrove Rail”. Given the data showing strong selection against hybrids from the various contact zones in North America, the splits already adopted by the North American committee appear justified. Because our South American longirostris-group is morphologically, genetically, vocally and ecologically distinct from all other members in the complex, and these distinctions are greater than those between sympatric/parapatric members of the complex, treatment of the South American populations as a distinct species seems pretty straightforward. The proposed English name of “Mangrove Rail” is perfect for these birds.”
Comments from Stiles: “YES. The only part of the proposal that affects SACC is the English name Mangrove Rail, which is eminently reasonable.”
Comments from Stotz: “YES. This split is clear. The North American committee spent way too much time thinking about English names on this one and came up with Mangrove Rail for the more limited longirostris. I think Mangrove Rail is a good name and so favor the split and the English name.”