The role of Quaternary glacial–interglacial intervals in shaping the diversity and distribution of Neotropical species has been the focus of considerable research. The Neotropics sustain the highest passerine diversity on Earth, but little is known about this region's historical biogeography based on fossils. To assess how passerine species were affected by Quaternary climate fluctuations, we identified 625 late Pleistocene fossils (individual fossilized bones) from the now arid and faunally depauperate Talara Tar Seeps in northwestern Peru. Of the 21 passerine species identified, only 2 likely live at the site now; the remaining 19 species require more mesic conditions. Species identified included members of the Thamnophilidae (antbirds), Melanopareiidae (crescentchests), Tyrannidae (flycatchers), Hirundinidae (swallows), Mimidae (mockingbirds), Thraupidae (seedeaters, “finches”), Emberizidae (sparrows), and Icteridae (blackbirds). Nearly half of the individual fossils and 8 of the 21 species were icterids, including 3 extinct species (1 previously described, 2 new). The late Pleistocene passerine community at Talara, which was nonanalog to any modern community, suggests that the site once supported savanna, grasslands, and forests during the last glacial interval, which are absent near Talara today. Quaternary climate change and the collapse of the community of large mammals had a major influence on the community composition and the geographic ranges of passerine species in northwestern Peru.
INTRODUCTION
The extraordinary biotic diversity found in the Neotropics is the result of numerous mechanisms that have acted over millions of years. The uplift of the Andes throughout the Cenozoic (past 65 million years) divided the ranges of taxa, changed soils and the direction of rivers, and led to the formation of numerous biomes (Graham 2009, Hoorn et al. 2010). These changes affected the entire South American continent, even lowland Amazonia, where plants and animals would have experienced novel atmospheric circulation and new riverine barriers (Rull 2011, Ribas et al. 2012). Over more recent time scales, the ~22 glacial–interglacial cycles during the Quaternary (2.6–0.0 mya) also have shaped modern diversity and distributions in the region (Colinvaux et al. 1997, 2000, Hoorn et al. 2010, Cohen and Gibbard 2011, Rull 2011).
Haffer (1969) hypothesized that Neotropical rainforest birds underwent allopatric speciation in habitat refugia (isolated areas of humid forest) during Pleistocene glacial periods, when dry vegetation was believed to have dominated and subdivided the lowland Amazonian rainforest. Allopatric populations of birds diverged and remained reproductively isolated, even after secondary contact when rainforest re-expanded during interglacial periods. However appealing, this hypothesis has received little support from phylogenetic studies across numerous taxa, which suggest that many (but not all) Amazonian plants and vertebrates diversified across both the Neogene (25.0–2.6 mya) and the Quaternary (Hoorn et al. 2010, Rull 2011).
Numerous lowland rainforest plant genera have been present in Amazonia since at least the late Miocene (Colinvaux and Oliveira 2001). Furthermore, pollen samples from sediment cores do not reveal lowland dry-habitat indicator plants in Amazonia during glacial periods, even though some cool-adapted Andean plants did occur in the western Amazon basin (Colinvaux et al. 1996a, 1996b, 1997, 2000). Nevertheless, pollen data show evidence of more expansive open habitats near the southern fringe of the Amazon Basin in glacial times (van der Hammen 1972, Absy et al. 1991). This presents the possibility that Quaternary climate fluctuations had a stronger influence on diversity and distributions in biogeographical regions outside Amazon Basin rainforests.
Relative to rainforests, Neotropical dry forests sustain lower diversity, yet greater levels of endemism (Janzen 1988, Stotz et al. 1996, Pennington et al. 2006). Despite hypotheses regarding the expansion and contraction of these dry forests and savannas during glacial–interglacial cycles (Prado and Gibbs 1993, Pennington et al. 2000), the role of Quaternary climate change on their biotic diversity and distributions is little studied, except for plants (but see Smith et al. 2012, Werneck et al. 2012). Determining the distributions of characteristic Neotropical dry forest and savanna organisms during glacial intervals would provide insight into the historical vs. modern connectivity of these habitats. Late Pleistocene fossils of Neotropical birds, for example, are tangible evidence of historical diversity, past species distributions, and environmental conditions (Steadman and Mead 2010, Steadman et al. 2015). Small vertebrate fossils, avian or otherwise, can be excellent indicators of past fine-scale habitat distributions, and furthermore often reveal nonanalog paleocommunities, i.e. sets of species that no longer exist sympatrically today, even when extinct species are excluded (Steadman et al. 1994, Stafford et al. 1999, Semken et al. 2010).
Songbirds (passerines; Passeriformes) constitute more than 50% of the world's living bird species and are hyperdiverse in the Neotropics (Stotz et al. 1996, Brumfield 2012), yet have a very limited fossil record. Here, we report 600+ passerine fossils from the Talara Tar Seeps in northwestern Peru. Today, the site is arid (desert) with the nearest dry forest ~50 km away. The passerine fossils from Talara present a unique opportunity to reconstruct not only the late Pleistocene songbird community but also the associated climatic conditions and habitats at the site, thereby elucidating the role of Quaternary glacial–interglacial cycles in shaping a distinctive Neotropical bird community.
METHODS
The Talara Tar Seeps (4°39′S, 81°08′W) are located at ~140 m elevation and ~20 km southeast of the coastal city of Talara (Figure 1). Today, the area around the site (<30 cm mean annual rainfall) sustains a depauperate flora and fauna. Dry forests, with considerable biotic diversity and a remarkable number of endemic taxa, are found northeast, east, and southeast of Talara at higher elevations ~50 km inland with higher precipitation (Parker et al. 1995, Stotz et al. 1996). Talara lies within the Tumbesian Region, which is an area of high endemism in northwestern Peru and southwestern Ecuador, characterized by gradients from coastal deserts to more inland dry and semideciduous forests.
FIGURE 1.
Location of the Talara Tar Seeps (4°39′S, 81°08′W) in northwestern Peru. The approximate distribution of Tumbesian tropical dry forest is shaded in dark gray. The Andes are shaded in light gray to white.
Based on conventional radiocarbon (14C) dating of associated pieces of wood, the fossils at Talara are late Pleistocene in age (last glacial interval; between 13,616 ± 600 and 14,418 ± 500 14C years before present [BP]; Churcher 1966). Using a standard calibration database (Heaton et al. 2009, Reimer et al. 2009), these two 14C determinations translate to 15,903 ± 808 and 17,054 ± 816 calendar years (cal) BP (~15,000 to 18,000 cal BP, or at the close of the last glacial maximum). Because stratigraphic mixing can occur in tar seeps as trapped animals sink while struggling in tar of varying viscosity (Shaw and Quinn 1986, Friscia et al. 2008), we attempted to augment the chronology of Talara by 14C-dating individual bird fossils from Talara. We were not successful because inadequate collagen remained in the bones. The age of the bird fossils from Talara is confidently late Pleistocene, centered on ~15,000–18,000 cal BP.
The Talara Tar Seeps fossils that we studied were excavated by A. G. Edmund and field crew in 1958 (Churcher 1959); they are housed at the Royal Ontario Museum (ROM), Toronto, Canada. While the nonpasserine fossils from Talara were studied by Campbell (1979), the passerine fossils were sorted by size categories and element but not examined further until our study. By “fossil,” we are referring to individual fossilized bones, e.g., humerus, coracoid, rostrum. We compared the fossils directly with modern skeleton specimens from the Florida Museum of Natural History, University of Florida (UF), Louisiana State University Museum of Natural Science (LSUMNS), and the National Museum of Natural History (USNM), and with fossils of extinct icterids housed in the UF Vertebrate Paleontology Division. If available, a series of skeletons of both males and females was used for identifications. Specimens listed in the Appendix represent average osteological characteristics for the specimens examined. The fossils also were evaluated using original descriptions of extinct species (Miller 1929, 1932, 1947). The osteological characters that we used for identification are given in the Appendix. All ROM catalog numbers are listed in Supplemental Material Table S12 (AUK-15-74_Supplemental Material Table S12.docx). Osteological nomenclature follows Baumel et al. (1993), supplemented by Howard (1929).
Information on habitats and geographic ranges of extant species is based on Schulenberg et al. (2007), Ridgely and Tudor (2009), and our own field observations. We visited the site twice in November 2011 and observed a very depauperate bird community (see Results). Because of the large number (~2,800) of passerine fossils recovered at Talara, we focused primarily on the most diagnostic skeletal elements of the skull, bill, shoulder girdle, and wing.
RESULTS
Based on scored osteological characters, we identified 625 fossils from the Talara Tar Seeps, representing up to 25 taxa (family level and below) and at least 21 species (Table 1; see Appendix Tables 2–11 for character matrices). We regard 3 of the species (all icterids; see below) as extinct. Only 2 of the 18 extant passerine species (Long-tailed Mockingbird [Mimus longicaudatus] and Cinereous Finch [Piezorina cinerea]) occur at the site today. Three other extant species recorded as fossils (Sulphur-throated Finch [Sicalis taczanowskii], Parrot-billed Seedeater [Sporophila peruviana], and Tumbes Sparrow [Rhynchospiza stolzmanni]) may also occur there sporadically, such as during El Niño events. The only other species that we recorded at Talara during our 2 visits was the Necklaced Spinetail (Synallaxis stictothorax), which we did not find as a fossil.
TABLE 1.
Passerine fossils identified from the Talara Tar Seeps, northwestern Peru. Modern habitat preferences are from Schulenberg et al. (2007), Ridgely and Tudor (2009), and our observations. “Tumbesian” refers to the region of arid and semi-arid habitats in northwestern Peru and southwestern Ecuador, west of the Andes. Taxa in brackets [] are not necessarily different from those identified with more taxon-specific resolution. If bracketed species are different from identified taxa, then the fossils represent 24 total species; if they are not different, then 21 species are represented. To be conservative, we assume that the bracketed species are not different species in the totals row under the “Occurs at site today” column. Extinct species are indicated by a dagger (†). N/A = not applicable (indeterminate).
The remaining 13 extant species certainly are extralocal. The Yellow-billed Cacique (Amblycercus holosericeus; Figure 2) shows the greatest range shift. In northwestern Peru, A. holosericeus lives in semideciduous forests inland and at higher elevations than Talara. Its habitat preferences in northwestern Peru differ from populations elsewhere, which tend to occur in the understory of secondary forest, bamboo thickets, and fields (Fraga 2011b). The Elegant Crescentchest (Melanopareia elegans), Collared Antshrike (Thamnophilus bernardi), Social Flycatcher (Myiozetetes similis), Baird's Flycatcher (Myiodynastes bairdii), and Yellow-tailed Oriole (Icterus mesomelas) also are indicators of dry, semideciduous, and gallery forests (Table 1, Appendix Figures 3 and 4). The Peruvian Meadowlark (Sturnella bellicosa), Scrub Blackbird (Dives warszewiczi), martin (Progne sp.), Tumbes Swallow (Tachycineta stolzmanni), Chestnut-collared Swallow (Petrochelidon rufocollaris), and Barn Swallow (Hirundo rustica) occupy grasslands, farms, savannas, and forest edges.
FIGURE 2.
The humerus (above) and coracoid (below) of fossil (left) and modern (right) specimens of blackbirds (Icteridae). (A) Dives warszewiczi fossil humerus ROM 70065, coracoid ROM 70085, modern LSUMZ 157416. (B) Sturnella bellicosa fossil humerus ROM 70238, coracoid ROM 70307, modern UF 47467. (C) Amblycercus holosericeus fossil humerus ROM 69934, coracoid ROM 69930, modern UF 33277 (old PB catalogue number 29362). PB = collection of Pierce Brodkorb; now in UF (Florida Museum of Natural History, University of Florida) collection. ROM = Royal Ontario Museum; LSUMZ = Louisiana State University Museum of Natural Science. Scale bars = 10 mm.
Eight species of icterid were recovered from Talara, 3 of which are extinct. Six of the 10 most numerous passerine species identified are icterids (Peruvian Meadowlark, n = 183; Shiny Cowbird [Molothrus bonariensis], n = 87; Scrub Blackbird, n = 41; the extinct Euphagus magnirostris, n = 36; the extinct troupial Icterus icterus s.l., n = 23; and the extinct Talara cowbird [Molothrus nov. sp.], n = 15). One of the extinct icterids (E. magnirostris) from Talara was first described from the Rancho La Brea Tar Seeps in Los Angeles, California, USA, based on a mandible larger and more robust than in either extant species of Euphagus (Miller 1929, Steadman et al. 2015). From Talara, we identified E. magnirostris mandibles and postcranial material sharing qualitative osteological characters with extant E. carolinus (Rusty Blackbird) and E. cyanocephalus (Brewer's Blackbird; Figure 5, Appendix Table 5). The extinct cowbird (Molothrus nov. sp.) is known thus far only from the fossils that we identified from Talara (Figures 6, 7). It is intermediate in size between the large Giant Cowbird (M. oryzivorus), a widespread species that occurs only in humid forests in the Tumbesian region, and the much smaller Shiny Cowbird, another widespread species known from the Tumbesian region, both living and fossil. The troupial fossils at Talara represent a new species or subspecies of Icterus icterus s.l. (Figure 8). The species and subspecies of troupial vary considerably in body size (Jaramillo and Burke 1999, Remsen et al. 2014) and osteological characters. The Talara troupial is larger than the Orange-backed Troupial (I. croconotus) found east of the Andes in Peru and Bolivia (I. c. croconotus and I. c. strictifrons) and is more similar in size to the Venezuelan Troupial (I. icterus) of Venezuela and Colombia. The 3 extant troupial species prefer dry habitats and gallery forests in the Amazonian lowlands in Peru and Bolivia, northern Venezuela, northeastern Colombia, and eastern Brazil (Ridgely and Tudor 2009).
FIGURE 5.
The humerus (A–C, upper), coracoid (A–C, lower), and mandible (D–F) of 3 species of Euphagus. (A, D) The extinct Euphagus magnirostris ROM catalog numbers 70107 (humerus), 70129 (coracoid), 70104 (mandible). (B, E) Euphagus carolinus UF 31110 (old PB catalogue number 26432). (C, F) Euphagus cyanocephalus UF 31114 (old PB catalogue number 38079). Scale bars = 10 mm.
FIGURE 6.
The humerus (above) and coracoid (below) of extinct and extant cowbird species. (A) Pandanaris convexa fossil (extinct; UF/PB 4221), Florida, USA. (B) Molothrus bonariensis fossil ROM 69963, Talara, Peru. (C) Molothrus bonariensis modern, LSUMZ 114332. (D) Molothrus nov. sp., fossil ROM 70030, Talara. (E) Molothrus oryzivorus modern, USNM (National Museum of Natural History) 344300. Scale bars = 10 mm.
FIGURE 7.
The rostrum of extinct and extant cowbirds in dorsal (A–D) and lateral (E–H) aspects. (A, E) Molothrus oryzivorus modern, USNM 344300. (B, F) Molothrus bonariensis modern, LSUMZ 114332. (C, G) Molothrus nov. sp., Talara fossil, ROM 70031. (D, H) Pandanaris convexa fossil, UF/PB 294. Scale bars = 10 mm.
FIGURE 8.
The humerus (above) and coracoid (below) of troupial (Icterus spp.) taxa. (A) Talara fossil Icterus icterus s.l. humerus ROM 70056, coracoid ROM 70042. (B) Modern Icterus icterus ridgwayi (male, Colombia) UF 30896 (old PB catalogue number 20737). (C) Modern Icterus croconotus strictifrons (male, Bolivia) LSUMZ 126183. (D) Modern Icterus croconotus croconotus (female, Peru) LSUMZ 49015. Scale bars = 10 mm.
The remaining passerine fossils at Talara represent various species of Thamnophilidae, Melanopareiidae, Tyrannidae, Hirundinidae, Mimidae, Thraupidae, and Emberizidae (Table 1). Numerous cranial elements represent either a new species of Sicalis (Thraupidae) or a slightly larger-billed variety of Sulphur-throated Finch (S. taczanowskii; Appendix Figure 9). Interspecific plasticity in shape and size of the bill is famously exemplified within related genera of thraupid finches and cardinalids such as Geospiza, Sporophila, and Passerina (Steadman 1982, Steadman and McKitrick 1982, Grant 1985). Pending successful extraction of DNA from tar seep fossils (Gold et al. 2014), it will be difficult to evaluate the species-level systematics of these Sicalis fossils, especially given that Sicalis species have extraordinarily similar postcranial osteological features (Appendix Table 6, Appendix Figure 9). Most of the postcranial fossils that have Sicalis characteristics are larger than modern S. taczanowskii, but overlap in size with S. flaveola (Saffron Finch). Therefore, we were able to identify these elements only to genus. A few postcranial fossils (n = 22) are probably S. taczanowskii. The Parrot-billed Seedeater (Sporophila pervuiana) also was found at Talara (a single fragmentary fossil rostrum; Appendix Figure 9). This species forages primarily on grass seeds (Jaramillo 2011) and is further evidence of the presence of grassy habitats at the site in the late Pleistocene.
Three species of martin (Progne; Hirundinidae) occur in northwestern Peru today. Each can be found near lakes, rivers, agricultural fields, and towns (Ridgely and Tudor 2009). We could not identify the Progne fossils to the species level because of a lack of comparative material from the region. Two species of swallow are represented as fossils at Talara, the Tumbes Swallow (Tachycineta stolzmanni) and the Chestnut-collared Swallow (Petrochelidon rufocollaris; Appendix Table 8, Appendix Figure 10). As with the martins, the modern habitat preferences of these 2 swallows are fairly ubiquitous and generalized. The Barn Swallow (Hirundo rustica) is a migratory species that breeds in North America and winters across Central and South America (Appendix Figure 10).
DISCUSSION
Late Pleistocene Climate and Habitats
The transition from the last glacial interval of the late Pleistocene to the modern interglacial interval (Holocene) dramatically changed the landscape of northwest Peru. Our findings from passerine bird fossils supplement the substantial information from sediments, archaeological remains, nonpasserine fossils, and megamammal fossils that argues for more diversity and more mesic climatic conditions in the region in the past (Lemon and Churcher 1961, Campbell 1976, 1979, 1982, Richardson 1978, Sandweiss 2003). Sedimentary evidence from 10 km east of the fossil site suggests the presence of permanent to semipermanent streams in this currently waterless region during the late Pleistocene (Lemon and Churcher 1961, Campbell 1979). A large number of the 6,200 nonpasserine bird fossils identified from the Talara Tar Seeps are from freshwater species, including grebes, herons, cormorants, ducks, geese, sandpipers, and plovers, not to mention other identified aquatic taxa such as amphibians, turtles, crocodilians, and mollusks (Campbell 1979). While other nonpasserine species found as fossils are characteristic of grassland or savanna (Black-chested Buzzard-Eagle [Geranoaetus melanoleucus], Short-eared Owl [Asio flammeus]), 2 species suggest dry forest at the site in the past (Pale-browed Tinamou [Crypturellus transfasciatus] and Crested Guan [Penelope purpurascens]). Also present at Talara in the late Pleistocene were large mammals characteristic of grassland, savanna, and dry forest, such as edentates (Scelidodon, Holmesina, Eremotherium), deer (Odocoileus, Mazama), proboscideans (Stegomastodon), dire wolves (Canis dirus), large jaguars (Panthera onca s.l.), and saber-toothed cats (Smilodon fatalis; Churcher 1959, Campbell 1979, Seymour 1983, Berta 1985). These findings led Lemon and Churcher (1961) and Campbell (1982) to propose that during the late Pleistocene the area near the site was predominantly grassland or savanna. Our results provide independent evidence to support this proposal, although the passerine fossils also argue for the past existence of dry forest and gallery forest. The late Pleistocene habitat matrix at or near the site probably did not include desert, which exists at Talara now.
The climatic mechanism(s) responsible for more mesic conditions and more diverse habitats at Talara during the late Pleistocene are unknown. On average, the Earth (including the tropics) was cooler and drier during glacial than interglacial intervals (Bush 1994, Peterson et al. 2000, Bromley et al. 2009). It is unlikely that meltwater from the Andes fueled the more mesic communities near Talara as there are few high peaks east of the site in the western Andean cordillera that would have sustained alpine glaciers during glacial intervals (Campbell 1979). Sea levels during the last glacial maximum (~25 to 18 ka) were ~120 m lower than they are today and reached modern levels by 6 ka (Clark et al. 2009). Given the steep continental shelf off the coast of Peru, the Pacific Ocean would have been only several tens of km farther from Talara during glacial times. Nevertheless, the lower sea levels and more distant coastline may have buffered Talara from the drying effects of the Humboldt Current during the late Pleistocene.
The Humboldt Current, found off the coast from southern Chile to northern Peru, fuels the nutrient-rich, cold upwelling that cools offshore air, which in turn causes the modern aridity of coastal Peru. Changes in trade winds and atmospheric pressure lead to a reduced temperature differential between the western and eastern Pacific Ocean that results in El Niño–Southern Oscillation (ENSO) events, which tend to occur on a 7–10 year cycle in modern times (Tudhope et al. 2001). During ENSO events, the upwelling ceases and surface waters warm, leading to a remarkable increase in precipitation that turns the desert of coastal northwestern Peru into an ephemeral shrubby grassland (Richter 2005, Muenchow et al. 2013). Perhaps longer-term, stronger, more frequent, or more persistent ENSO conditions resulted in the richer glacial-age floral and faunal communities in northwestern Peru in the past. It has been difficult, however, to reconstruct the periodicity and severity of ENSO events through time because different paleoclimate proxies yield different results. Sea surface temperature differentials have been stable since the Miocene (Zhang et al. 1998), which would have resulted in periodic ENSO cycles over the past 18 million years, although Wara et al. (2005) argue for permanent El Niño conditions during the Pliocene. Growth rings from corals in the western Pacific suggest that ENSO events have occurred since 130 ka, even during glacial periods, although the amplitude and duration of the events may have been more variable than in modern times (Zhang et al. 1998).
Icterids
At Talara, 296 of 625 (47%) of the passerine fossils are of icterids, with 3 of the 8 species being extinct. The fact that icterids are so common as late Pleistocene fossils may be due to their preference to form communal roosts near water (Oswald and Steadman 2011). These freshwater wetlands, whether more permanent lakes and marshes or more ephemeral ones lying on tar seeps after rains, would also serve as water sources for large mammals, thus augmenting their attractiveness to icterids. Four extinct species of Icteridae have been described from late Pleistocene sites in North America: Euphagus magnirostris (Miller 1929); Pyelorhamphus molothroides (Miller 1932); Pandanaris convexa (Miller 1947, Oswald and Steadman 2011); and Cremaster tytthus (Brodkorb 1959). The late Pleistocene extinctions of icterids included even very widespread species, such as Pandanaris convexa and Euphagus magnirostris (Oswald and Steadman 2011, Steadman et al. 2015). The losses were especially dramatic among large-billed, cowbird-like forms such as Pyelorhampus, Pandanaris, Molothrus nov. sp., and perhaps Cremaster. The extinction of multiple species with a unifying morphological trait (a large bill) suggests the loss of foraging niches that existed across the Americas during the Pleistocene.
In addition, icterids are unique among New World passerines in that many species have commensal relationships with large mammals such as horses, cows, and bison (Lowther 1993, Martin 2002). In Africa, extant megamammals such as elephants and zebras are “ecosystem engineers” that help to maintain the savannas and grasslands upon which many bird species depend (Laws 1970); it is likely that Pleistocene megamammals played comparable roles in maintaining habitats favorable to icterids in the Americas (Steadman and Martin 1984, Oswald and Steadman 2011). While both climate change and human impacts fueled the Pleistocene-to-Holocene faunal changes in the region, the loss of biotic interactions also may have played a significant role for icterids when their large mammal associates were lost. All extant cowbird species also are brood parasites of other songbird species. The host of the large, extinct Pleistocene species of Molothrus at Talara may have been the now-extirpated Yellow-billed Cacique or the extinct species of troupial.
The modern distribution of Euphagus is restricted to North America; both extant species (Rusty Blackbird [E. carolinus] and Brewer's Blackbird [E. cyanocephalus]) are migratory. Therefore, the E. magnirostris fossils discovered in northwestern Peru (and in the Inciarte Tar Seeps in northwestern Venezuela; Steadman et al. 2015) may represent migratory individuals. Alternatively, perhaps especially if E. magnirostris was a megamammal commensal, this species may have had an extensive resident distribution in both North and South America, alongside the various edentates, artiodactyls, equids, and proboscideans that were lost at the end of the Pleistocene (Anderson 1984). The only modern resident passerine birds with ranges that extend from the southern United States into South America are the Vermilion Flycatcher (Pyrocephalus rubinus; Farnsworth and Lebbin 2004) and Eastern Meadowlark (Sturnella magna; Fraga 2011a).
Nonanalog Bird Communities and Species' Responses to Change
Many Pleistocene biotic communities across the Americas are nonanalog compared with modern communities. In North America, for example, late Pleistocene rodent communities were composed of species that are allopatric today (Stafford et al. 1999, Blois et al. 2010, Semken et al. 2010). Pollen cores show that Andean plants once lived alongside tropical, lowland plant taxa in the Amazon Basin (Bush 1994, Colinvaux et al. 2000). Late Pleistocene nonanalog bird communities of the Americas reflect in part the loss of megafauna, which led to the loss of species or populations of large scavenging and predatory nonpasserine birds (vultures, condors, eagles, caracaras, etc.; Steadman and Martin 1984, Steadman et al. 2015). The nonanalog late Pleistocene passerine community that we discovered at Talara is what we might expect under a different climatic regime and with a very different community of large mammals. Considering that we could not radiocarbon-date the bird fossils themselves, we have no definitive evidence that all of the species identified in this study were contemporaneous. Nevertheless, in the late Quaternary sites with analog bird or mammal communities and a rigorous radiocarbon chronology, the time spans of the fossil deposits are measured in decades or at most a few centuries, rather than millennia (Stafford et al. 1999, Steadman et al. 2002, Semken et al. 2010). If the changing climatic conditions across the late Pleistocene to Holocene resulted in habitat changes and range shifts that led to high turnover of bird species composition at Talara, this suggests that biotic communities, whether birds or otherwise, may be continually in a state of flux, with the rate of change governed by the severity of the climatic (or anthropogenic) change. Because species have responded idiosyncratically to historical abiotic and biotic changes, the search for a single mechanism driving the changing diversity and distributions of tropical species may be futile.
Modern tropical dry forests in South America have a fragmented distribution in Andean valleys, on the Pacific slope of southwestern Ecuador and northwestern Peru (Tumbesian Region), and in a discontinuous ring around the Amazon Basin (Pennington et al. 2006, Särkinen et al. 2011). Very few species of bird are shared among these regions (Stotz et al. 1996). Our fossil evidence indicates that dry forest species had wider distributions in the past. The dearth of shared species among modern dry forest communities might reflect, in part, population extirpation since the last ice age. This hypothesis can be tested through more paleontological research.
ACKNOWLEDGMENTS
For access to fossil and modern specimens, we thank the Royal Ontario Museum (Kevin Seymour, Brian Iwama), the Louisiana State University Museum of Natural Science (Van Remsen, Steven Cardiff), and the Florida Museum of Natural History (Andrew Kratter, Tom Webber). We thank Jean-Noel Martinez for guiding us on a visit to Talara. Hayley Singleton kindly helped with figures. Mark Brenner, Rebecca Kimball, David Reed, Scott Robinson, Ryan Terrill, and Judit Ungvari-Martin kindly provided comments that improved the manuscript.
Funding statement: We thank the American Ornithologists' Union (grant 095819 to J.A.O.), the National Science Foundation (International Planning Grant 096747 to D.W.S. and J.A.O., and Graduate Research Fellowship 084142 to J.A.O.), and the Wilson Ornithological Society (to J.A.O.) for funding. No funders had any influence on the content of the submitted or published manuscript or required approval of the final manuscript prior to publication.
LITERATURE CITED
Appendices
APPENDIX. COMPARATIVE OSTEOLOGY
Osteological characters used to identify the Talara Tar Seep, Peru, passerine fossils. Royal Ontario Museum (ROM) catalog numbers for each identified fossil are found in Supplemental Material Table S12 (AUK-15-74_Supplemental Material Table S12.docx). All characters were scored relative to the character state of the reference skeletons. All specimens at table column headings are male unless otherwise noted, although female specimens (if available) also were included in osteological comparisons. Comparisons included more than 1 specimen (typically 5 to 8) per species. At the level of family or genus, a few of these characters are modified from those in Hamon (1964).
APPENDIX FIGURE 3.
Skulls and mandibles of Melanopareia elegans. The fossil (ROM 69700; A, D, G) has the cranium and mandible still articulated. The modern (UF 49342) cranium (B, E, H) and mandible (C, F, I) are shown below the fossil in different aspects (A–C, cranial; D–F, ventral; G–I, lateral). Scale bars = 10 mm.
APPENDIX FIGURE 4.
Postcranial elements of suboscines. For each pair, fossils are on the left, modern specimens on the right. (A) Fossil (ROM 69707) and modern (UF 49489) humerus of Myiodynastes bairdii. (B) Fossil (ROM 69701) and modern (UF 49356) humerus of Thamnophilus bernardi. (C) Fossil (ROM 69705) and modern coracoid of Myiozetetes similis (modern: M. s. columbianus, UF 46254). The fossil is likely M. s. grandis (see Appendix Table 10). Scale bar = 10 mm.
APPENDIX FIGURE 9.
Rostra and humeri of seedeaters. (A, B) Sicalis flaveola (UF 49190) in lateral and dorsal aspects. (C, D) Sicalis taczanowskii in lateral and dorsal aspects, with modern specimen (UF 49526) on left and fossil (ROM 69904) on right. (E, F) Sporophila peruviana in lateral and dorsal aspects, with modern specimen (LSUMZ 157412) on left and fossil (ROM 69926) on right. Humeri (G–I) of Sicalis species. (G) Sicalis flaveola modern, UF 49190. (H) Sicalis sp. fossil, Talara, ROM 69792. (I) Sicalis taczanowskii modern, UF 49526. Scale bars = 10 mm.
APPENDIX FIGURE 10.
Postcranial elements of martins and swallows. (A) Humeri (top) and coracoids (bottom) of Progne species, with Talara fossil (Progne sp. fossil; humerus: ROM 69712; coracoid: ROM 69708) on left, and modern Progne tapera tapera (UF 27837; old catalogue number 26568, Suriname) on right. (B) Humeri (top) and coracoids (bottom) of Petrochelidon species, with Talara fossil (Petrochelidon rufocollaris; humerus: ROM 69716; coracoid: ROM 69743) on left, modern Petrochelidon rufocollaris (LSUMZ 100541) at center, and modern Petrochelidon pyrrhonota pyrrhonota (UF 27803; old PB catalogue number 21812, New Brunswick) on right. (C) Humeri of Tachycineta species with Talara fossil (Tachycineta stolzmanni; ROM 69715) on left, modern Tachycineta stolzmanni (LSUMZ 90118) at center, and modern Tachycineta bicolor (UF 27878; old PB catalogue number 30242, Florida) on right. (D) Humeri of Hirundo rustica, with Talara fossil (ROM 69763) on left, and modern specimen (UF 27893; old PB catalogue number 29601, Indiana) on right. (E) Carpometacarpi of Petrochelidon species, with Talara fossil P. rufocollaris (ROM 69755) on left, modern P. rufocollaris (LSUMZ 100541) at center, and modern P. pyrrhonota pyrrhonota (UF 27803; old PB number 21812, New Brunswick) on right. PB = collection of Pierce Brodkorb; now in UF collection. Scale bars = 10 mm.
APPENDIX TABLE 2.
Osteological characters used to identify species of Icteridae of similar size and osteology. See Tables 3–5 for Icterus and Euphagus characters.
APPENDIX TABLE 3.
Icterus cranial characters referenced when identifying cranial elements preserved as fossils. All characters were scored relative to the character state of the elements considered. All specimens are male unless denoted by an “*” indicating that they are female. Note: No fossil cranial material from Talara was identified as belonging to Icterus. This table is included in the Appendix as a reference to the characters scored for comparably sized species of icterids in Table 2.
APPENDIX TABLE 4.
Osteological characters used to identify postcranial elements of Icterus. All characters were scored relative to the character state of the modern specimens. All specimens are male unless denoted by an “*” indicating that they are female.
APPENDIX TABLE 5.
Characters used to identify fossils in the genus Euphagus.
APPENDIX TABLE 6.
Osteological characters used to identify “seedeater” (Thraupidae) fossils. Referenced Talara fossil material: ROM catalog numbers 69791–69929.
APPENDIX TABLE 7.
Osteological characters used to identify martin (Hirundinidae) fossils.
APPENDIX TABLE 8.
Osteological characters used to identify swallow (Hirundinidae) fossils.
APPENDIX TABLE 8.
Extended.
APPENDIX TABLE 9.
Osteological characters used to identify Myiodynastes bairdii fossils using flycatcher species of similar size for comparison.
APPENDIX TABLE 10.
Coracoid measurements of Myiozetetes similis subspecies and related flycatchers of comparable size as the Myiozetetes cf. similis fossil from Talara.
APPENDIX TABLE 11.
Osteological characters used to identify Thamnophilus bernardi fossils.
