Geographical distributions of Uca (sensu lato)

Our survey reveals a lack of information on species in the north-eastern States of Piauí, Alagoas and Sergipe. The geography of species' distributions would be enhanced by collections and reports from these States. Nevertheless, we can draw several conclusions as to the biogeography of Uca (sensu lato) in Brazil. We recognized four fundamental patterns of geographical distribution: (1) Amapá to Santa Catarina (AP–SC); (2) Maranhão to Santa Catarina (MA–SC); (3) primarily southern coast between Pernambuco and Rio Grande do Sul (PE–RS); and (4) primarily north-eastern coast between Amapá and Pernambuco (AP–PE). The ten species of fiddler crabs collected are distributed among these four patterns (Table 1).

Four species clearly exhibit the AP–SC distribution pattern: U. (U.) maracoani; U. (M.) mordax; U. (B.) thayeri; and U. (M.) rapax. They occupy four of the five coastal biomes of Brazil (Biomes I–IV). Uca (M.) mordax was difficult to locate on the Ponta do Calcanhar between Rio Grande do Norte (RN) and SE where previous records indicate it to be sporadic (Table 1). The finding of U. (M.) mordax along the Rio Mampituba, Torres, RS (Biome V) gives it the most southerly distribution of all species except U. (L.) uruguayensis. The other three species encounter their southern limits of distribution at more northerly latitudes, around Florianópolis in SC.

Two species, U. (M.) burgersi and U. (L.) leptodactyla, although rare in AP and Pará (PA) are common from MA to SC (Biomes II to V: Tables 1 & 2). Previous records for U. (M.) burgersi by Holthuis (Reference Holthuis1959, Reference Holthuis1967), von Hagen (Reference Hagen1970a, Reference Hagenb, Reference Hagen1987) and Crane (Reference Crane1975) include Trinidad and the Caribbean shore of Venezuela but not the peripheral areas of the Golfo de Paria, Venezuela, and areas south to PA, Brazil. Likewise, U. (L.) leptodactyla occurs on the Caribbean shores of Venezuela and Trinidad and farther south in Ceará (CE), Brazil (Crane, Reference Crane1975). To our knowledge, there are no records for this species between the Orinoco River and the mouth of the Amazon River. Such a disjunct pattern can be linked to silting sediments from the Amazon River along the shores of AP, PA and the coasts of the Guianas. Consequently, the sandy habitats preferred by U. (M.) burgersi and U. (L.) leptodactyla may be shifting, unavailable or overlain by fine sediments in this region (Swennen et al., Reference Swennen, Duiven and Spanns1982). Nevertheless, six species of Uca exhibit distributional ranges from either AP or MA to SC.

The third pattern (Table 1: PE–RS) is seen in U. (M.) victoriana and U. (L.) uruguayensis. Uca (M.) victoriana is distributed mainly from CE to SP with a distribution centre in ES. Primarily, U. (M.) victoriana inhabits Biome III (Table 3). Endemic to the tropical Brazilian coast, it is unclear why this species is so restricted in range and found no farther north or south. Uca (L.) uruguayensis has been reported only from areas south of Cabo Frio, RJ to Argentina. It is found primarily in Biomes IV and V. It is equally unclear why this species does not occur at more northerly latitudes close to the equator like U. (M.) victoriana. The southern range limits for both species are most likely determined by the effect of cold water on larval development (Sanford et al., Reference Sanford, Holzman, Haney, Rand and Bertness2006). Interestingly, the distributions of the two species overlap for only a few hundred kilometres in RJ between Cabo Frio and Baía de Sepitiba.

The ranges of the remaining two species, U. (M.) vocator and U. (L.) cumulanta, are primarily along the north-eastern coast of Brazil (Table 1: AP–PE). These species are unusual in also having a few individuals or populations along the southern coast. According to historical records, U. (M.) vocator was common from Trinidad to SP, and is considered one of the main species from the Guianas to PA (Holthuis, Reference Holthuis1959; von Hagen, Reference Hagen1970a, Reference Hagenb; Swennen et al., Reference Swennen, Duiven and Spanns1982; Diele et al., Reference Diele, Kock, Saint-Paul and Schneider2010). Specimens identified by Luederwaldt at the MZUSP (numbers 298 (1895), 301 (1913), 320 (1920), 361 (1913) and 892 (1915)) indicate that U. (M.) vocator was abundant as far south as Santos, SP in the late 19th and early 20th Centuries. As recently as 1999 and 2000, more than 2600 were collected from three localities in SP (Colpo & Negreiros-Fransozo, Reference Colpo and Negreiros-Fransozo2004) which we re-visited in 2009 and 2010 but encountered very few U. (M.) vocator. Bedé et al. (Reference Bedé, Oshiro, Mendes and Silva2008) found only 30 individuals at Itacuruçá over the course of one year. We collected a total of 152 specimens of U. (M.) vocator. While 90% were taken between AP and BA, only 10% were captured in ES, RJ and SP, emphasizing the current scarcity of the species along the southern Atlantic coast. Although substantial along the north-east coast (Biomes I–II), U. (M.) vocator appears to be disappearing from southern estuaries and perhaps should be considered an endangered or threatened species in that region.

In contrast, U. (L.) cumulanta appears to be opportunistic and expanding. It has been studied thoroughly along the north-east coast of South America (Holthuis, Reference Holthuis1959; Ahmed, Reference Ahmed1976; Swennen et al., Reference Swennen, Duiven and Spanns1982; Diele et al., Reference Diele, Kock, Abrunhosa, de Farias-Lima, Brito-Simith, Saint-Paul and Schneider2010) and occurs from Venezuela to Mamanguape (USNM 25700: Crane, Reference Crane1975) and Itamaracá, PE (Coelho & Coelho-Santos, Reference Coelho and Coelho-Santos1989) in northern Brazil. In the south, specimens were found on Ilha de Paquetá, RJ (USNM 71171 and 138875), as early as 1925. In 2008, Pralon & Negreiros-Fransozo reported populations at Patitiba beach, Paraty, RJ and Bedé et al. described another population at Itacuruçá. Consequently, U. (L.) cumulanta is firmly established in the Baía de Sepetiba. Since there are few records of the species between PE and RJ (Table 1: Almeida et al., Reference de Almeida, Souza, Boehs and Bezerra2010), we consider this species to exhibit a disjunct distribution between northern Brazil and RJ. As for the Kingston Bay, Jamaica population (Rathbun, Reference Rathbun1918; Barnwell, Reference Barnwell1986), the species may have been introduced into this remote location by oceanic vessels where it successfully established founder populations over the last few hundred years. We hope to test this hypothesis using molecular biology methods (Schubart, Reference Schubart2003; Cunningham, Reference Cunningham2008).

Ecological preferences of each Uca species

The surface coastal waters of Brazil derive from the warm Equatorial Current (Peterson & Stramma, Reference Peterson and Stramma1991) and most of the species discussed here are tropical or subtropical. Only U. (L.) uruguayensis is considered a temperate zone species. Consequently, except for species in Biome V, daily and/or seasonal variation in air/water temperatures along the Brazilian coast is not expected to determine distribution. Rather, four other factors appear to be more significant: (1) position in the intertidal zone; (2) substrate composition; (3) salinity; and (4) behaviour (particularly larval migration and adult feeding). Data on zonation, and substrate and salinity preferences for each species are presented in Table 3 and Figure 3.

The tidal regimen along the east coast of South America is primarily regular and semi-diurnal (Doty, Reference Doty1957), and fiddler crabs here express tidal rhythms in their physiology and behaviour (Barnwell, Reference Barnwell1963). In the South Atlantic, tidal amplitude is routinely 1 to 3 m but along the north-eastern coast, amplitude is usually 3 to 6 m. In AP, amplitude is normally 5 to 6 m. Such large amplitudes create very strong tidal currents with each successive ebb and flood, which also produces a very short slack water period when the current turns. The megalopae of developing crabs ride the flood tide and settle into the adult habitat during transitional slack water (Behum et al., Reference Behum, Brodie and Stanton2005; Simith et al., Reference Simith, Diele and Abrunhosa2010). In the region from AP to PA, the large tide, strong currents and silty substrates may act in synergism with other factors to limit successful colonization by certain species of fiddler crabs.

Substrate is important for fiddler crabs in two ways: burrows and feeding. Species like U. (U.) maracoani, U. (M.) vocator, U. (L.) cumulanta and U. (B.) thayeri are able to construct burrows in soft mud near mangroves. Likewise, U. (M.) rapax, U. (M.) victoriana and U. (M.) mordax can excavate tunnels in firmer sandy- or silty clay habitats. However, Uca (M.) burgersi, U. (L.) leptodactyla and U. (L.) uruguayensis burrow into mud substrates with a greater sand component. Besides burrowing, fiddler crabs utilize substrate as a food source, which their mouthparts are structured to process (Miller, Reference Miller1961). There is a clear relationship between buccal anatomy and habitat granulometry. Since the c-index has been calculated previously for eight of the Brazilian Uca (von Hagen, Reference Hagen1970b; Thurman, Reference Thurman1984, Reference Thurman1987), those for U. (L.) uruguayensis and U. (M.) victoriana are now added in Table 3. For primarily sand feeding species (i.e. U. (L). leptodactyla and U. (L.) uruguayensis) the c-index is >90.0. For the sand–mud feeders (i.e. U. (M.) rapax, U. (M.) burgersi, U. (M.) mordax, U. (M.) victoriana and U. (L.) cumulanta), the c-index is between 60 and 90. For species preferring silty mud (i.e. U. (U.) maracoani, U. (B.) thayeri, U. (L.) cumulanta and U. (M.) vocator), the c-index lies between 80 and 50.

The morphology of the setae on the 2nd maxillipeds likewise reflects the substrate feeding preference (Miller, Reference Miller1961; Crane, Reference Crane1975). If a species feeds primarily on sandy soils, there is an abundance of spoon-tipped setae or hairs covering the edge of the 2nd maxillipeds. In species that prefer silt or fine-grained mud, there is a greater abundance of woolly hairs and the number of spoon-tipped setae is reduced. Here we add the morphology of the 2nd maxillipeds of U. (M.) victoriana to the list of those crabs possessing setae for processing substrate particles.

Although species transplantation and imposed substrates are yet to be explored, the relationship between mouthparts and substrate appears to be flexible. Uca uruguayensis from southern Brazil shows considerable feeding plasticity depending on the presence of congeneric species (Vernberg & Vernberg, Reference Vernberg and Vernberg1967). North of Florianópolis (SC), U. uruguayensis cohabits with U. rapax and U. leptodactyla on sandy substrates in the intertidal zone where it is limited mainly to shaded intertidal sand bathed by meso- to eusaline waters (Figure 3). However, south of Florianópolis where it is the sole species, U. uruguayensis occupies diverse habitats ranging from silt to mud to pure sand substrates, and from oligo- to hypersaline waters, constituting an opportunistic species that occupies niches vacated by the more tropical and sub-tropical species. Apparently, when interspecific competitive pressures relax, U. uruguayensis expands into less suitable habitats and survives. When confronted by congeners at more northerly latitudes, U. uruguayensis selects a narrower range of habitat characteristics that it occupies with greater feeding efficiency (Sayão-Aguiar et al., Reference Sayão-Aguiar, Pinheiro and Colpo2012) and osmoregulatory ability. Consequently, mouthpart structures appear to assure survival by allowing specialization under competitive pressures.

Different species inhabiting littoral regions often select habitats with very different osmotic properties (Teal, Reference Teal1958). Partitioning of the intertidal zone has been documented in fiddler crabs from Texas, Mexico and the northern Caribbean (Thurman, Reference Thurman1985, Reference Thurman1987; Thurman et al., Reference Thurman, Hanna and Bennett2010). In part, carapace physiology and morphology reduce desiccation in the more terrestrial fiddler crab species (Thurman, Reference Thurman1998; Yoder et al., Reference Yoder, Tank, Rellinger, Moore and Gribbins2007, Reference Yoder, Bosley, Chambers, Hart, Gribbins and Collier2010; Hopkins & Thurman, Reference Hopkins and Thurman2010). Those living low in the littoral zone usually survive meso- and eusaline conditions. In contrast, species inhabiting estuaries exhibit preferences for more dilute, oligosaline conditions (Teal, Reference Teal1958). In Brazil, species were found to be distributed along similar intertidal salinity transects (Table 3). In the meso- to eusaline zones, U. (U.) maracoani is found lowest in the zone, being covered every few hours by the tide. From the mid to upper littoral, U. (L.) uruguayensis, U. (L.) cumulanta and U. (B.) thayeri are common. The most common supralittoral species are U. (L.) leptodactyla and U. (M.) rapax. However, in oligosaline or riverine habitats, U. (M.) vocator, is found lower in the intertidal than U. (M.) burgersi, U. (M.) victoriana, or U. (M.) mordax. Figure 3 presents Brazilian fiddler crab habitats discriminated using two important physiographical components: substrate grain size and habitat osmolality (=salinity). Based on these characteristics, fiddler crab species in Brazil can be sorted across a simple two-dimensional niche array. Consequently, patterns of species diversity along the Brazilian coast are undoubtedly influenced to a great degree by heterogeneity in substrate and water/salt sources.

Areas of high diversity

Tropical mangroves and salt marshes are high diversity communities in Brazil (Por, Reference Por1994). Owing to their rapid degradation on a global scale, these ecosystems must be conserved before diversity is irretrievably lost (Valiela et al., Reference Valiela, Bowen and York2001; Lourie & Vincent, Reference Lourie and Vincent2004; Amaral & Jablonski, Reference Amaral and Jablonski2005). The mangrove environment itself is of wide interest to Brazilian society owing to its role as a very visible nursery area for larvae of many commercial fish, and as a specific habitat for commercially exploited, large mangrove crabs such as Ucides cordatus and Cardisoma guanhumi. Other notable inhabitants include mammals like the crab-eating raccoon Procyon cancrivorus, and once rare birds like the spectacular scarlet ibis, Eudocimus ruber. The São Paulo State government has made significant investments in conservation programmes to preserve its tropical diversity (Joly et al., Reference Joly, Rodriquez, Metzger, Haddad, Verdade, Oliveria and Bolzani2010). In São Paulo alone, a network of 64 state parks and preserves has been set aside to establish restoration and conservation.

Our findings reveal two regions along the South Atlantic coast of Brazil that warrant preservation based on their high species diversity. Presuming that an elevated number of species occurring in a fairly small area indicates environmental heterogeneity, these areas may be important for maintaining both intra- and interspecific variation. Further, for those species depending on oceanic movements for larval and genetic dispersal, such sanctuaries constitute sources of up-stream recruitment for the global ecosystem. Based on these criteria, conservation efforts should be focused: (1) along the north-eastern coast between Fortaleza, CE and Recife, PE; and (2) along the southern coast from Barra de Guaratíba, RJ, to Ubatuba, SP. In the former, we recovered nine species from habitats along the course of the Rio Cocó. From the latter, we recorded ten Uca species from the banks of a canal and tidal ditch in Barra de Guaratuba at the east end of the Baía de Sepetiba. Bedé et al. (Reference Bedé, Oshiro, Mendes and Silva2008) reported the same ten species in mangroves a short distance away at Itacuruçá.

There are four subgenera and 20 species of Uca (sensu lato) on the western perimeter of the Atlantic Ocean. Along the Brazilian coast from AP to SC, all Atlantic subgenera (100%) are represented in the Uca fauna. Brazilian gamma diversity for Uca (sensu lato) is 50% of the total diversity for the western Atlantic (Table 4, N = 20). At Barra de Guaratíba, RJ, alpha diversity is equal to gamma diversity (Whitaker, Reference Whitaker1972). Corroborating findings by Bedé et al. (Reference Bedé, Oshiro and Melo2007, Reference Bedé, Oshiro, Mendes and Silva2008), we submit that the finding of ten species of Uca (sensu lato) along a 2-m wide, 8 to 10-m transect surpasses the phenomenal level of alpha biodiversity reported by Barnes (Reference Barnes2010) who found up to six species of Uca (sensu lato) within 2-m² quadrants along a 25-m intertidal transect in Sulawesi (Indonesia). This may be the largest known number of Uca species occurring in any small area. Wallacea is an important source of biological diversity in the South Pacific and the region is strictly conserved. For these same reasons, we suggest that the coast of the Baía de Sepetiba and Baía de Paraty between Rio de Janeiro and Trindade, RJ, seriously be considered for conservation and protection.

Table 4.

Distribution of twenty Uca spp. on the tectonic plates of the western Atlantic Ocean.

+, known; −, uncommon; *, both colour morphs included.

Phylogeographical considerations

All the factors controlling the geography of littoral species and their cohorts may never be known. However, their distributions are determined, in part, by: (1) the availability of suitable habitat and climate; (2) the geomorphology and history of the potential colony; (3) the hydrodynamics of ocean currents and tides; (4) patterns of larval dispersion; and (5) the biotic composition of the area available for colonization. Patently, once a littoral species locates and settles in a suitable environment, it must be anatomically and physiologically competent for survival and reproduction. The overall dispersal potential of a species or cohort is a combination of all these capabilities together with other more subtle physical, chemical and biological factors. Previous sections addressed some of the habitat, climate and geomorphological characteristics circumscribing the apparent ecological niches of each of the ten species of Uca in Brazil. However, since the ranges of several of these species extend into the Caribbean Basin, Gulf of Mexico, and the Atlantic coast of the southern United States (Table 5; Figure 4) it is necessary to consider factors that regulate distributions on a larger geographical scale.

Fig. 4.

Tropical western Atlantic (redrawn from Rocha, Reference Rocha2003): general scheme of sea currents and their flows along the eastern coast of the Americas and Caribbean. CCS, Caribbean Current System; FC, Florida Current; GC, Guiana Current; GS, Gulf Stream; NBC, North Brazil Current; SBC, South Brazil Current; SEC, South Equatorial Current; STC, Subtropical Current; YC, Yucatan Current. Grey shading indicates approximate coastal area influenced by freshwater outflow from the Amazon and Orinoco Rivers.

Table 5.

Estimated oceanic distances over which larval Uca spp. (subgenus/species) are transported by the North Brazil and the Guiana Currents. Oceanic distance (km/generation) is the product of pelagic duration (zoea + megalopa) in days and the mean current velocity (0.038 m s⁻¹, Bourlés et al., Reference Bourlès, Gouriou and Chuchla1999). All other values are given in days.

Table 5 lists the developmental durations of zoeae and megalopae and passive dispersal potentials for five species of Brazilian Uca (sensu lato). Since there is no information available on larval development in U. (U.) maracoani, data for U. (U.) tangeri are used as an equivalent substitute. Based on the average speed of the North Brazil and Guiana Currents (Bourlès et al., Reference Bourlès, Gouriou and Chuchla1999), potential dispersal varies from 75 to 175 km/generation depending on species. However, this mechanism can be modified significantly by adult and larval behaviour (Salmon & Zucker, Reference Salmon, Zucker, Chelazzi and Vannini1988; Christy, Reference Christy2011).

It is clear that Uca zoeae behaviourally select tidal currents to control their dispersal (Lopez-Duarte & Tankersley, Reference Lopez-Duarte and Tankersley2007, Reference Lopez-Duarte and Tankersley2009). In estuarine species from North and Central America (N = 6 species), zoeae migrate with a tidal rhythm to catch the ebb tide while coastal species (N = 9 species) are arrhythmic. Such behaviour would assure the export of Uca living upstream versus local retention for the more marine species near the estuary mouth (Lopez-Duarte et al., Reference Lopez-Duarte, Christy and Tankersely2011). One case, the development and ecophysiology of larval U. (M.) vocator from north-eastern Brazil, is insightful in this regard. Along the shores of the Rio Caetá near Bragança, PA, populations of this species reproduce during the rainy season (250 to 300 cm yr⁻¹) with peak numbers of ovigerous females in March and June (Koch et al., Reference Koch, Wolff and Diele2005). A combination of rainfall and freshwater outflow from the Amazon River reduces salinity to near 0 psu in the adult habitat, creating a strong hyposmotic challenge and potential larval mortality. Neither zoeae nor megalopae survive ≤15 psu. Consequently, it is unlikely that larvae will develop in the estuary near the parent crabs (Simith et al., Reference Simith, Diele and Abrunhosa2010, Reference Simith, de Souza, Maciel, Abrunhosa and Diele2011).

For the western Atlantic, the two oceanic currents determining the distribution of intertidal and shallow-water marine species at higher latitudes are the North and South Brazil Currents (Briggs, Reference Briggs1974: Figure 4). From the Ponta do Calcanhar (RN), the Southern Brazil Current carries water and planktonic larvae to the southern Atlantic. Since the southern coasts of Brazil, Uruguay and Argentina are mainly rectilinear, the relationship between current and species' distributions, coastal geomorphology and climate is fairly straightforward at least for Uca (sensu lato). Given that the locations of each species from PE to SP or SC are highly connected by larval distribution driven by the South Brazil Current, populations and species should be genetically uniform since they are controlled primarily by habitat quality and climate. Genetic studies of another sympatric brachyuran, Ucides cordatus, support this notion (Oliveira-Neto et al., Reference Oliveira-Neto, Boerger, Pue, Ostrensky and Hungaria2007).

The impact of the North Brazil Current is less clear (Figure 4). Since the North Brazil Current and the Guiana Current flow northward, the north-eastern coast of Brazil may have a significant long-term influence on the Caribbean, Central America, Gulf of Mexico and temperate North Atlantic (Briggs, Reference Briggs1974). The geography and hydrology of the intertidal zone along the north-eastern coast of Brazil presents a formidable habitat for colonization by shallow-water marine organisms. Here, the tides range from 2.3 to 6.3 m; consequently, with their turn every 6.2 hours, a tremendous volume of water ebbs and flows rapidly across the littoral zone. From Macapá, AP north to Trinidad, the surface water over the shelf is diluted by the world's largest freshwater discharge from the Amazon and Orinoco Rivers (Curtin, Reference Curtin1986) producing extensive soft bottoms and turbid water for about 2000 km. This freshwater surface plume extends downward to a depth of 30 m over the shelf and up to 500 km out to sea (Lentz, Reference Lentz1995). Most sediment from the Amazon River is deposited along the shores of Amapá and the Guianas forming extensive silt and clay banks (Figure 4). Fresh water and sediment discharged from the Orinoco River are carried north along the Atlantic shore and Golfo de Paria by the Guiana Current. Eventually, 70% of the Guiana Current flows around Barbados, Trinidad and Tobago into the Caribbean (Borstad, Reference Borstad1982).

Potentially, the nine species of Uca (sensu lato) found along the north-eastern coast of Brazil could be distributed into the Caribbean, Mexico and southern United States (Table 4). Based on laboratory studies (Table 5), the zoeae of Brazilian species could remain in the plankton for 25 to 30 days while the swimming megalopae are aquatic for an additional 12 to 20 days. Assuming that suitable settling habitat is available, larvae could move a maximum of 80 to 170 km northward along the coast with each generation. Consequently, moving approximately 2000 km across the mouths of the Amazon and Orinoco Rivers into the Caribbean would take approximately 14 generations (≈15 to 18 fiddler crab years). This pilgrimage to the Caribbean requires the larvae to survive in dilute, nearshore waters and settle in silty habitats (Swennen et al., Reference Swennen, Duiven and Spanns1982). Species such as U. (L.) leptodactyla and U. (M.) burgersi are likely unsuited for this task given their ecological requirements. However, as supported by our collections from Amapá and museum and historical literature records, a high degree of population connectivity probably exists for U. (M.) mordax, U. (M.) rapax, U. (M.) vocator and U. (B.) thayeri between north-eastern Brazil, Caribbean islands and the shores of the Panama isthmus. Until better collecting records are available across the southern Caribbean between Panama and the Windward Islands, U. (L.) cumulanta and U. (U.) maracoani appear to be naturally restricted to latitudes south of Trinidad and the Golfo de Paria. Curiously, both U. (L.) leptodactyla and U. (M.) burgersi are common throughout the northern Caribbean (Thurman et al., Reference Thurman, Hanna and Bennett2010) and are known from southern Florida and the Yucatan peninsula (Barnwell & Thurman, Reference Barnwell and Thurman1984). Clearly, populations of these species south of the equator are disjunct from their Caribbean relatives.

Tectonic plate movements undoubtedly influenced the evolution of Uca (sensu lato) subgenera and species in the western Atlantic basin (Salmon & Zucker, Reference Salmon, Zucker, Chelazzi and Vannini1988; Beinlich & von Hagen, Reference Beinlich and Hagen2006). In the Neotropics, north-western Africa and Europe, mangrove communities, which are associated with fiddler crabs, appeared during the late Eocene (≈40 Mya) (Plaziat, Reference Plaziat, Bosence and Allison1995; Hogarth, Reference Hogarth2004). By the early Miocene (≈22 Mya), fiddler crabs were genetically distinct from other cosmopolitan Ocypodidae (Levinton et al., Reference Levinton, Strumbauer and Christy1996; Sturmbauer et al., Reference Strumbauer, Levinton and Christy1996) and members of the subgenus Uca (sensu stricto) were distributed from Brazil to Honduras (Brito, Reference Brito1972; Dominguez-Alonso, Reference Dominguez-Alonso2008). The subgenus persists in western North America through the Pliocene (5 Mya) as indicated by the fossil U. (U.) oldroydi (Rathbun, Reference Rathbun1926). Other Pacific and American clades (i.e. subgenera) appeared later during the middle to late Miocene (≈17 Mya: Rathbun, Reference Rathbun1926; Sturmbauer et al., 1996). Closure of the trans-Panama seaway in the late Pliocene (3.5 to 2.8 Mya: Ibaraki, Reference Ibaraki1997) resulted in additional radiation within the four American subgenera (Rathbun, Reference Rathbun1926; Sturmbauer et al., 1996).

From three extinct and ten extant species, we can speculate on the vicariance and early dispersal history of the subgenus Uca (sensu stricto). Based on distribution centres, the apparent range of each ancestral narrow-fronted species is coincident with plate tectonics (von Hagen, Reference Hagen1968; Crane, Reference Crane1975; Prahl, Reference Prahl1982; Prahl & Toro, Reference Prahl and Toro1985; Dominguez-Alonso, Reference Dominguez-Alonso2008). Among the extant forms, the most widely distributed in terms of latitude are U. princeps, U. insignis, U. maracoani and U. tangeri (Crane, Reference Crane1975). Uca major is limited to the Caribbean basin and the Bahamas but probably occupies the greatest unit area. Isolated on the African plate, we suspect that U. tangeri separated from the Neotropical Uca about 50 Mya (Sanmartín & Ronquist, Reference Sanmartín and Ronquist2004). On the Miocene shores of the Neotropics, cohorts (or predecessors) of the subgenus Uca (sensu stricto) appeared to be distributed on each tectonic plate—on the South American plate, U. maracoani, U. antigua Brito, Reference Brito1972 (fossil), U. marinae Dominguez-Alonso, Reference Dominguez-Alonso2008 (fossil), U. insignis (Milne-Edwards, 1852) and U. ornata (Smith, Reference Smith1870); on the Caribbean plate, U. heteropleura (Smith, Reference Smith1870), U. major, U. intermedia Prahl & Toro, 1985 and U. stylifera (Milne-Edwards, 1852), with U. princeps (Smith, 1970), U. monilifera (Rathbun, 1914) and U. oldroydi (fossil) on the western shores of the North American plate. As the seaway connection closed approximately 2.8 to 3.1 Mya (Lessios, Reference Lessios2008), the ancestor of U. major was isolated in the Caribbean from those of U. intermedia, U. heteropleura and U. stylifera in Honduras, Costa Rica, Panama and Colombia along the Pacific coast. Across the northern edge of South America, the isthmus of Panama separated the Atlantic ancestors of U. maracoani and U. antigua from those of U. insignis, U. ornata and U. marinae in the Pacific. Ensuing glaciation over the last 1.0 Mya likely produced variation among U. princeps and U. monolifera as well as the extinction of U. oldroydi on the North American plate. Radiation of U. intermedia, U. heteropleura and U. stylifera on the western slopes of Central America, and the divergence of U. ornata and U. insignis on the Pacific coast of the South American plate may also be linked to palaeoclimate changes.

Since we have no fossil record or extant species from the subgenus Uca (sensu stricto) on the Atlantic shore of the North American plate, this area may well have been a refuge only for species in the subgenera Leptuca and Minuca. Beinlich & von Hagen (Reference Beinlich and Hagen2006) suggested a second Uca invasion of the Caribbean by a Paraleptuca-like predecessor from the Mediterranean–Arabian Basin during the late Miocene or early Pliocene (22 to 5 Mya) giving rise to the American clades prior to the closure of the Panama seaway. The fossil Uca (M.) hamlini Rathbun, Reference Rathbun1926 from California demonstrates that rapax-like Minuca had evolved by the Pliocene (Crane, Reference Crane1975). From distributional information on all western Atlantic subgenera (Table 4), it is clear that three Minuca (burgersi, rapax and vocator) and the Boboruca (thayeri) are distributed by oceanic currents across the South American, the Caribbean and the North American plate. The geminate species, U. (B.) umbratila, occurs from El Salvador to Colombia (Bott, Reference Bott1973; Crane, Reference Crane1975; Prahl, Reference Prahl1982) but has not been reported from the Pacific rim of the North American plate, supporting the notion that the predecessor of the subgenus Boboruca is of South Atlantic origin. However, there are four Minuca and four Leptuca endemic to the North American plate. Expansion of the subgenera Minuca and Leptuca into the Caribbean and South America may have been prevented by the prevailing directions of ancient Atlantic currents (Poulsen et al., Reference Poulsen, Seidov, Barron and Peterson1998).

Owing to their morphological similarity, a wide-ranging, meso- to eusaline species similar in morphology to U. (M.) rapax may have given rise to the oligosaline sister species U. (M.) victoriana and U. (M.) marguerita Thurman, 1981 by dispersal and recent (<3 Mya) sympatric speciation (Knowlton, Reference Knowlton1993). Closely related eusaline Leptuca (U. panacea, U. pugilator and U. subcylindrica) may be recent products of speciation wrought by a similar mechanism through the effects of glaciation and seismic activity in North America; these Leptuca lack true morphological counterparts on the shores of the Caribbean, Central America and South Atlantic. However, ecologically they are similar to U. (L.) leptodactyla and U. (L.) uruguayensis in the Caribbean and South America. Likewise, the closely related oligosaline Minuca, U. minax, U. longisignalis and U. pugnax, may be the products of recent North American glaciation events (Barnwell & Thurman, Reference Barnwell and Thurman1984) isolating them from U. (M.) mordax and U. (M.) vocator. The endemic, North American, oligosaline species U. (L.) spinicarpa is morphologically and physiologically similar to U. (L.) cumulanta from South America. Other authors (Crane, Reference Crane1975; Salmon et al., Reference Salmon, Ferris, Johnston, Hyatt and Whitt1979; Rosenberg, Reference Rosenberg2001) have suggested that U. (L.) spinicarpa is closely related to the eusaline species Uca (L.) speciosa (Ives, 1891). It is revealing that Rathbun (Reference Rathbun1918) confused specimens of U. (L.) cumulanta from Jamaica (USNM 22313) and Brazil (USNM 25700) with U. (L.) spinicarpa but not with U. (L.) speciosa. A cumulanta–spinicarpa distinction was not clear until Crane (1943) described U. (L.) cumulanta. Based on their morphology and distributions, there are two possible implications. Either an ancestor of U. (L.) cumulanta and U. (L.) spinicarpa was distributed from north-eastern Brazil to the northern Gulf of Mexico, or a relative on the Pacific shore of Central America was the link between the two species of Atlantic Leptuca.

Although these scenarios are speculative, they are rooted in geological history and the fossil record as well as in the distributions of extant species. The evolutionary origins of Neotropical fiddler crabs are most likely connected to changing climates and continental shores over the last 40 to 60 million years. Based on genetic divergence, at least two events or invasions seem likely: one occurring more than 22 Mya, the other more recent, about 17 Mya (Levinton et al., Reference Levinton, Strumbauer and Christy1996; Sturmbauer et al., 1996). These events established the ancestors of various subgenera. The evolution of the various sibling species and greater diversification within subgenera of Neotropical Uca (sensu lato) probably took place during the last 3 million years.