Prehistoric Exploitation and Biogeography of Birds in Coastal and Andean Ecuador

Material Information

Prehistoric Exploitation and Biogeography of Birds in Coastal and Andean Ecuador
Copyright Date:


Subjects / Keywords:
Birds ( jstor )
Bones ( jstor )
Female animals ( jstor )
Forests ( jstor )
Genera ( jstor )
Humerus ( jstor )
Osteology ( jstor )
Scapula ( jstor )
Species ( jstor )
Ulna ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Markus Patricio Tellkamp. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
Resource Identifier:
495638873 ( OCLC )


This item is only available as the following downloads:

Full Text




Copyright 2005 by Markus Patricio Tellkamp


This dissertation is dedicated to our daught ers Doménica and Daniela. May they inherit from us a beautifully diverse and healthy planet earth.


iv ACKNOWLEDGMENTS This work would not have been possible without the help of many people. I thank my advisory committee, especially my advisor David Steadman, for their patience and enthusiastic support when my dissertation t ook an unexpected turn and a side project became the main focus of my research. I also thank David Steadman for unrestricted access to the specimen collection of birds at th e Florida Museum of Natural History and a great deal of inspiration. Brian McNab, No rris Williams, Susan deFrance, and Michael Binford all provided great comments that im proved the quality of this work. The recent acquisition of specimens from Peru by Andrew Kratter has been in strumental in the success of this study. Lauran Bianco assisted in taking morphometric measurements. My visit to archaeological field sites in coas tal Ecuador was greatly facilitated by Ana Agreda. I thank my wife Ivonne Ulloa for he lping with so many aspects of this study, from cleaning bones, entering data, to providi ng companionship into the wee hours of the night. For help with my study on the phylogeogra phy of Andean birds, although not the subject of this dissertation, I need to thank the Department of Zoology, Ginger Clark, Ivonne Ulloa, Lauran Bianco, Alexandra R udnik, Phuong Pham, Robert Gotshalk, Fabián Cupuerán, Miguel Burgos, Fabián Granda, Fundación Arco Iris, Martha Mondragón, Juan Fernando Freile, Fundación La Otonga , and the Ecuadorian ornithological foundation (CECIA).


v Financial support came from a Grinter Fellows hip, the generosity of Günter and Irene Tellkamp, and my work with CECIA. My parents, Günter and Irene Tellkamp, also provided moral support when ever it was needed. Above all, I thank my wife for her total s upport, help, love, and sacrifice in sharing my dream of obtaining a PhD degree. Our daughters are a constant inspiration and motivation to preserve the wonderful diversit y of life and all the ecosystems that sustain it (and us).


vi TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES.............................................................................................................ix LIST OF FIGURES..........................................................................................................xii ABSTRACT....................................................................................................................... xv CHAPTER 1 INTRODUCTION: THE HOLOCENE ENVIRONMENTS OF ECUADOR.............1 Introduction................................................................................................................... 1 Palynology....................................................................................................................4 Eastern Lowlands..................................................................................................4 High Andes............................................................................................................6 Western Lowlands.................................................................................................8 Sedimentology............................................................................................................10 Paleontology...............................................................................................................14 Archaeology................................................................................................................18 Biogeographical Patterns............................................................................................23 Historical Accounts....................................................................................................26 Synthesis.....................................................................................................................2 8 Environmental Change as the Result of the Interaction between Culture and Nature.....................................................................................................................28 2 COMPARATIVE OSTEOLOGY OF ECUADORIAN BIRDS FROM ARCHAEOLOGICAL DEPOSITS............................................................................32 Introduction.................................................................................................................32 Methods......................................................................................................................34 Results........................................................................................................................ .35 Identifications.............................................................................................................35 Discussion.................................................................................................................199


vii 3 EARLY HOLOCENE MANGROVE AND WETLAND BIRDS ON THE ARID SANTA ELENA PENI NSULA, ECUADOR..........................................................201 Introduction...............................................................................................................201 Methods....................................................................................................................204 Results.......................................................................................................................2 07 Systematic List..........................................................................................................210 Discussion.................................................................................................................220 4 REMARKABLY HIGH NUMBERS OF SONGBIRDS AND GROUND DOVES AT THE EL AZÚCAR ARCHAEOLOGI CAL SITE, COASTAL ECUADOR: AN UNUSUAL HUNTING ADAPTATION TO AN ARID ENVIRONMENT?..226 Introduction...............................................................................................................226 Methods....................................................................................................................230 Study Site...........................................................................................................230 Excavation.........................................................................................................230 Analysis of Bird Bones......................................................................................232 Results.......................................................................................................................2 37 Systematic List..........................................................................................................239 Discussion.................................................................................................................263 Guangala Fishers, Farmers, H unters, and Bird Trappers?.................................265 The Paleoenvironment of El Azúcar.................................................................266 Conclusion................................................................................................................269 5 A PREHISTORIC BIRD ASSEMBLAGE FROM TH E ANDEAN TREE LINE AT LA CHIMBA, NORTHERN ECUADOR..........................................................271 Introduction...............................................................................................................271 Methods....................................................................................................................273 Study Site...........................................................................................................273 Bone Identification............................................................................................275 Results.......................................................................................................................2 77 Systematic List..........................................................................................................281 Discussion.................................................................................................................298 6 PREHISTORIC EXPLOITATION AND BIOGEOGRAPHY OF BIRDS IN ANDEAN AND COASTA L ECUADOR................................................................305 Introduction...............................................................................................................305 Prehistoric Exploitation of Birds in Ecuador............................................................309 Prey availability.................................................................................................310 Prey density.......................................................................................................311 Hunting Technique............................................................................................311 Taboos...............................................................................................................312 Hunting Regulation...........................................................................................312 Human Habitat Alteration.................................................................................313


viii Comparison with other Neotropical Sites.................................................................314 Historical Biogeography of Birds in We stern Ecuador and the Northern Andes.....325 Coastal Sites......................................................................................................325 Northern Andes.................................................................................................328 Migratory Species..............................................................................................330 Lessons for Bird Conservation.................................................................................330 Conclusion................................................................................................................333 APPENDIX LIST OF MUSEUM SPECIMENS USED.............................................335 LIST OF REFERENCES.................................................................................................349 BIOGRAPHICAL SKETCH...........................................................................................364


ix LIST OF TABLES Table page 1-1. Fauna from Chobshi Cave* with modern habitat associations†.................................22 1-2. Summary of climatic changes during the Holocene as suggested for the three major regions in Ecuador....................................................................................................29 2-1. Character matrix for the tibiotarsus in the Family Tinamidae....................................51 2-2. Unrotated factor loadings for four meas urements in a principal components analysis on the distal end of the cor acoid in several species of Anas (see Figure 2-3)..........70 2-3. Characters and character states for the ulna of medium-sized Accipitridae...............84 2-4. Character matrix for the ulna of medium-sized Accipitridae.....................................85 2-5. Characters and character states for the femur in the family Cracidae........................97 2-6. Character matrix for seven traits found in the femur of the Cracidae........................98 2-7. Characters and character states for the dist al end of the tibiotar sus in the Cracidae..99 2-8. Character matrix for the distal end of the tibiotarsus in four species (three genera) of Cracidae..................................................................................................................101 2-9. Characters and character states for the humerus in the family Scolopacidae...........111 2-10. Character matrix for the humerus in 8 species (8 genera) of Scolopacidae...........112 2-11. Characters and character states for th e coracoid of simila rly-sized species of Larus , summarized from Campbell (1979).......................................................................114 2-12. Character matrix for the coracoid in six species of Larus . See Table 2-11 for definition of character states..................................................................................115 2-13. Characters and character states for th e humerus of similarly-sized species of Larus , summarized from Campbell (1979).......................................................................115 2-14. Character state matrix for the humerus in six species of Larus ..............................116


x 2-15. Characters and character states for th e carpometacarpus of similarly-sized species of Larus , summarized from Campbell (1979)........................................................118 2-16. Character matrix for the carpometacarpus in six species of Larus .........................119 2-17. Diagnostic characters and character st ates for the coracoid in the Columbidae....124 2-18. Character matrix for the coracoid in nine species of Columbidae.........................125 2-19. Character state matrix for os sa cranii of the Psittacidae.........................................136 2-20. Characters for the tarsomet atarsus of large species of Ara (Psittacidae)...............148 2-21. Character matrix for six traits found in the tarsometatarsus of the Psittacidae......149 2-22. Diagnostic characters and character states for the proxim al end of the scapula in the Cuculidae................................................................................................................164 2-23. Character state matrix for diagnostic ch aracters of the cucu lid scapula in five species....................................................................................................................165 2-24. Unrotated factor loadings for four measurement in a principal components analysis on the distal end of the coracoid in seve ral species of Corvidae and Icteridae (see Figure 2-6)..............................................................................................................177 2-25. Diagnostic characters and character states for the prox imal end of the humerus in the Passeriformes....................................................................................................180 2-26. Character state matrix for the passe rine humerus regarding three diagnostic characters in 18 species of Passeriformes..............................................................181 2-27. Character states of the suboscine ulna....................................................................184 2-28. Character states for the ulna in nine suboscine genera...........................................185 2-29. Diagnostic characters and character states for the carpometacarpus (Order Passeriformes)........................................................................................................188 2-30. Character state matrix for the ca rpometacarpus in the Passeriformes....................191 3-1. Minimum number of individuals (MNI) a nd number of individua l specimens (NISP) for all species at the OGSE-80 archaeol ogical site, Santa Elena, Ecuador............208 3-2. Minimum number of individuals (MNI) a nd number of individua l specimens (NISP) for two broad habitat classes for three de pth categories, at Site 80, Santa Elena Peninsula, Ecuador.................................................................................................221


xi 4-1. Minimum number of individuals (MNI) a nd number of individua l specimens (NISP) for all species identified from the El Azúcar archaeological site, Santa Elena Peninsula, Ecuador, divided by “lot.”....................................................................234 4-2. Minimum number of individuals (MNI) a nd number of individua l specimens (NISP) for all species identified from the Valdivia Valley archaeological sites, just north of Santa Elena Peninsula, Ecuador.............................................................................242 4-3. Minimum number of individuals (MNI) and number of individual specimens (NISP) for all species at the Real Alto archaeological site, Santa Elena Peninsula, Ecuador...................................................................................................................248 5-1. Minimum number of individuals (MNI) a nd number of individua l specimens (NISP) for each species at the La Chimba archaeological site, by cultural periods...........278 5-2. Densities of MNI, NISP, and sp ecies per 100 liters of sediment.............................281 6-1. Summary statistics for the six archaeological sites studied in Chapters 3 to 5, providing age range, cultural period, elevation, potential current vegetation (after Sierra et al. 1999b), a nd key references.................................................................306 6-2. Species identified at archaeological sites OGSE-80 ( O80 ), El Azucar ( EA ), Valdivia Valley ( VV ), and La Chimba ( LC ), Ecuador.........................................................318 6-3. Number of species (percentage) in five broad habitat categories for all Ecuadorian sites combined, combined coastal sites, and La Chimba.......................................327


xii LIST OF FIGURES Figure page 1-1. Map of Ecuador showing the lo cations mentioned in the text.....................................3 1-2. Chronology for major cultural periods ment ioned in the text (s implified after Stahl 2004).........................................................................................................................2 0 2-1. The distal end of the tarsometatarsus of Ixobrychus exilis (UF 42030) in dorsal (a) and plantar (b) aspect...............................................................................................63 2-2. Dorso-medial aspect of coracoid, showing the width of impressio ligamenti acrocoracohumerale of Anas georgica (a), archaeological specimen (cf. A. georgica ) from OGSE-80 (b), and A. bahamensis (c)..............................................68 2-3. A principal components analysis on four variables of the dist al end of coracoid, suggesting that most archaeogica l specimens tend to be stout.................................69 2-4. Cranial aspect of the three specimens of cracids, representing Aburria pipile (a), a large specimen from La Chimba probably belonging to Aburria aburri (b), and Penelope purpurascens (c).....................................................................................100 2-5. Phalanges 1 of digitus segundus (bottom row) and digitus tertius (top row) in lateral aspect from OGSE-80 specimens (a), Rallus longirostris (b), Aramides cajanea (c), Porphyrula martinica (d), and Gallinula chloropus (e).........................................109 2-6. The lower mandible in dorsal aspect from the La Chimba specimen (a), Pionus fuscus (b), Pionopsitta haematotis (c), Pionites melanocephala (d), and Aratinga erythrogenys ...........................................................................................................137 2-7. Plantar aspect of the tarsometatarsus in Amazona autumnalis (a), the La Chimba specimen (b), and Ara severa (c)...........................................................................147 2-8. Carpometacarpus from the La Chimba specimen (a), Aulacorhynchus prasinus (b), A. derbianus (c), and A. haematopygus (d)............................................................167 2-9. Skull in dorsal aspect from La Chimba specimen (a) and Buthraupis montana (b).168 2-10. Dorsal aspect of the mandible fro m the El Azúcar specimen (a) and Spiza americana ...............................................................................................................171


xiii 2-11. A principal components analysis on five variables of the coraco id in large oscine passerines ( Psarocolius s. l. includes Gymnostinops montezuma , Psarocolius viridis , P. decumanus , P. angustifrons , and Zarhynchus wagleri ; Cyanocorax s. l. includes Psilorhinus morio , Cyanocorax affinis , C. cayanus , C. cyanomelas , C. dickeyi , C. sanblasianus ), supporting the notion that the archaeological specimens come from oropendolas (Family Icteridae)............................................................176 2-12. Variation in the second pneumatic fossa in the passerine humerus from absent (a) to highly developed (g, j). Species are Grallaria przewalskii (a), Megarhynchus pitangua (b), Tityra semifasciata (c), Cyanolyca viridicyana (d), Turdus serranus (e), Mimus gundlachii (f), Anisognathus lachrymosus (g), Saltator maximus (h), Cacicus haemorrhous (i), and Sturnella militaris (j).............................................179 3-1. A map of the Santa Elena Peninsula, Ec uador, showing the locations of Site 80 ( ), several towns ( ) as well as permanent and temporal rivers.................................205 3-2. A map showing different localities menti oned in the text. The gray line represents the 200 m contour level of the Cordillera Chongón-Colonche..............................206 3-3. A map of Ecuador showing the 21 provinces...........................................................211 3-4. The locations of several Peruvian arch aeological sites menti oned in the text in relation to Site 80...................................................................................................224 4-1. Map showing major towns ( ) as well as the El Azúcar ( ), La Ponga ( ), Loma Alta ( ), Valdivia Village ( ), and Real Alto ( ) archaeological sites on the Santa Elena Peninsula, Ecuador.......................................................................................231 4-2. A. Percentage minimum number of indi viduals (MNI) and B. percentage number of individual specimens (NISP) of passerines (songbirds; ), columbids (ground doves; ), and other birds ( ) for each lot at site 47, El Azúcar..........................239 4-3. A. Percentage minimum number of indi viduals (MNI) and B. percentage number of individual specimens (NISP) of passerines (songbirds; ), columbids (ground doves; ), and other birds ( ) for each excavation level at site 30, El Azúcar....240 4-4. A. The valley of El Azúcar taken from the road about 2 km from the modern town. B. A side valley of the Río Valdivia near La Ponga..............................................264 4-5. The mudflats of the Río Chanduy estuary with several Rhizophora sp. saplings....267 5-1. A map of the study area with contour lines at 400 meter intervals; the dark square in the inset shows the location of the study area within Ecuador...............................276 5-2. Relationship between sediment volume a nd NISP (number of individuals specimens) for A) TP-5 and B) TP-7........................................................................................283


xiv 5-3. Change of percentages of A) MNI (min imum number of individuals) and B) NISP (number of individuals specimens) of bird bones for the three periods of La Chimba classified by habitat associ ation: páramo vegetation ( ), dry/open habitat ( ), humid forest/forest edge ( ), and wetlands ( )....................................................291 5-4. A side valley of the La Chimba valley, showing the dark green remnants of humid forest as indicated by the arrows............................................................................301 6-1. Map showing the location of all archaeological sites ( ) and major cities and towns ( ) mentioned in the text. Inset show s enlarged Santa Elena Peninsula................308 6-2. Fisher’s alpha biodiversity index for the Ecuadorian sites of OGSE-80, El Azucar, and La Chimba, as well as the Mexican site of Paso de la Amada........................316 6-3. Species accumulation curves based on the number of identified specimens (NISP) for A) OGSE-80, El Azúcar, La Chimba, a nd Paso de la Amada; and B) for OGSE80, El Azúcar, La Chimba, and Paso de la Amada truncated at 40 specimens, as well as Valdivia Village, La Ponga, and Loma Alta..............................................317


xv Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy PREHISTORIC EXPLOITATION AND BIOGEOGRAPHY OF BIRDS IN COASTAL AND ANDEAN ECUADOR By Markus Patricio Tellkamp December 2005 Chair: David W. Steadman Major Department: Zoology The Holocene (the last ca. 10,000 yr BP) is usually thought of as a climatically stable period. Bird species dist ributions in the Neotropics are assumed to have been stable throughout this time as well. A review of recen t literature, however, shows that climate was more variable. Although a picture about changes in plant distributions is now emerging, little is known about how animals mi ght have responded to climatic change. Through a detailed study of comparative osteol ogy, I analyzed the bird assemblages from five archaeological sites in coastal Ecuador (ca. 10,000–1670 yr BP) and one site from the northern Andes (ca. 2640–1700 yr BP) to establish how biogeographic patterns could be related to past climate or other natural as well as anthropogeni c factors. This work suggests several broad patterns in bird expl oitation and avian biogeography. Birds are only a minor component of archaeological si tes in Ecuador, and prehistoric people in Ecuador consumed mostly common species from the immediate surroundings of their settlements. The dominant species are sma ll to medium-sized birds (30–300 g) with


xvi terrestrial foraging habits that make them easy to hunt. Ease of procurement has probably influenced the decision by Early Holocene hunters of coastal Ecuador to focus on terrestrial and mangrove-associated animals, rath er than marine birds and fish as in Peru. Biogeographically, this study show s that 20% of the species from archaeological sites no longer occur there in the present. Range cha nges can be attributed to human and natural changes in habitat. Deforestation may have been severe locally as at El Azúcar (2370– 2030 yr BP) where most species recovered archaeologically are open-habitat birds associated with brushy vegetation. Climateinduced changes are probably related to fluctuations in rainfall associated with the El Niño Southern Oscillation (ENSO). Tectonic uplift has significantl y altered the distribution of mangroves and their associated avifaunas. Despite these changes most spec ies have been able to rebound after severe habitat change and hunting in the past. I therefore suggest that in some areas habitat restoration projects could be more successful in the long term than implied by the current focus on areas of “pristine” habitat.


1 CHAPTER 1 INTRODUCTION: THE HOLOCENE ENVIRONMENTS OF ECUADOR Introduction The present distribution of speci es is a result of a variet y of factors that span both space and time. Biogeography, the study of the factors that determine the geographical distribution of species, can be divided into two main areas of inquiry based on two broad approaches. The ecological approach studies the present interactions that shape geographic boundaries, dispersal, and persistence. The histori cal approach examines past events that contributed to the present distri bution of species by tr ying to reconstruct how historical contingencies have le d to processes such as expans ion or contraction of ranges, speciation, local extirpation, and global extinction. The great diversity of tropical ecosystems is of special interest both because of current concerns about massive extinction from human activity a nd the existence of complex distributional patterns. Although ecological interacti ons are often important in maintaining these patterns, the ultimate cause for the spatial arrangement of species is often historical in nature. Extinction and sp eciation dynamics are se t against the backdrop of large-scale, density-independent factor s, such as global climate change, regional climatic oscillations, and various localized disturbance regimes that ultimately set the stage for ecological interactions. For at least the last 11,000+ radi ocarbon years before present (14C yr BP) people have been present in the Americas. There is substantial evidence that prehistoric humans have altered landscapes and, th erefore, may have shaped th e distribution of many species


2 (Martin and Steadman 1999, Gnecco 2003). Presen t species distributions therefore have to be interpreted as the re sult of a possible combination of natural and anthropogenic factors. Floras and faunas of different bi omes may react very differently to the environmental changes caused by human ac tivities (Gade 1999, Sarmiento and Frolich 2002). Given that people have been hunting w ildlife for thousands of years, it would be of great interest to determine the ecological correlates that are associated with long term persistence of species. Zooa rchaeological material sheds light on this issue. The distribution of plants and animals in Ecuador (Figure 1-1), especially with regard to the location of centers of endemi sm, depends to a large degree on the way in which different taxonomic groups have res ponded to environmental change during the Quaternary. The recent biodiversity crisis caused by the interplay among the growing agricultural frontier, burni ng, logging of hardwood tree sp ecies, petroleum and mineral exploitation, and fossil fuel consumption has renewed the interest in determining how the plant and animal communities may have res ponded to recent clima tic changes and early human populations in South America during the Holocene. Studies on the historical development of plant and animal life in Ecuador may shed light on the ability of the biota to resist change or to recover after major ch ange has occurred. In this review I summarize our current understanding of climate change, paleoecological conditions, and biogeographical patterns in Ec uador during the Holocene. I also present a brief summary of the human footprint during this time a nd the dynamics between early cultures and nature. It is not my intention to present an exhaustive literature re view since many of the aspects described here have recently been re viewed extensively, but to set the stage for further studies in the Ecuadorian Holocene.


3 Figure 1-1. Map of Ecuador showing the lo cations mentioned in the text. (1) Lago Yaguarcocha; (2) Lago San Pablo; (3) Laguna San Marcos; (4) Maxus Core 4; (5) Maxus Core 5; (6) Lago Yambo; (7) Mera; (8) Laguna Chorrera; (9) Laguna Pallcacocha; (10) Laguna Llaviu cu; (11) Laguna Ayauchi; (12) San Juan Bosco; (13) Deep Sea Sediment Co re Tri 163-31B; (14) La Chimba; (15) La Ponga; (16) Loma Alta; (17) Valdivia ; (18) OGSE-80; (19) Real Alto; (20) La Libertad; (21) Quebrada Cuesaca; (22) La Calera; (23) Punín. Open squares represent paleontol ogical sites, gray squares represent archaeological sites, and gray circles represen t location of sediment cores. Even though the Pleistocene-Holocene tran sition is geologically speaking a recent event, the exact climatic conditions and vegetation cover for large areas of South America have yet to be determined for this period of change. A preliminary picture can be painted from evidence gathered from palynology, sedimentology, zooarchaeology, and paleontology.


4 Palynology Pollen studies from Ecuador are scant compar ed to those from other countries such as Colombia (Hooghiemstra and van der Ha mmen 1998). Within Ecuador cores have been obtained from Yasuni National Park (M axus cores, Athens and Ward 1999, Weng et al. 2002), Mera (Liu and Colinvaux 1985), Sa n Juan Bosco (Bush et al. 1990), Lago Surucucho (Colinvaux et al. 1997), Lago San Pablo (Athens 1990), Lago Cunro, Lago Yaguarcocha, Lago Yambo (Colinvaux et al. 1988), Lago Chorreras and Lago Pallcacocha (Hansen et al. 2003). Additional da ta were obtained from deep sea cores off the coast of Ecuador (Heusser and Shacklet on 1994). Pollen records from the Galápagos Islands (Colinvaux and Schofield 1976) are also relevant to understanding continental climate. The pollen record in Ecuador reveals relatively little climatic change after the onset of the Holocene compared to the Pleistocene. Variation in the pe rcentanges of various pollen types cannot be explained easily b ecause of the complex interactions among temperature, humidity and wind regimes (Han sen et al. 2003, Bush et al. 2004). The most important feature of Holocene climate suggest ed by most cores is a noticeably drier and locally colder period between ca. 9000–5800 14C yr BP which has many implications for cultural and natural ev ents in northwestern South America (Paulsen 1976, deMenocal 2001). Eastern Lowlands In the eastern lowlands two major trends emerge. In the lower tropical zone the Holocene is characterized by seemingly conflic ting trends as determined from two cores that are located within 25 km. Both cores sugge st drier conditions dur ing the first half of the Holocene. In Maxus 5 core (Athens a nd Ward 1999) drier condi tions are inferred


5 from an increased charcoal concentration and the increased presence of Asteraceae, Cyperaceae, and Poaceae. Athens and Ward (1999) suggest that increased charcoal levels are due to humans taking a dvantage of prolonged seasonal dr y seasons to cut and burn the forest for incipient agriculture. The lack of Cecropia pollen in this core is surprising given that this pioneer tree is commonly associated with natu ral and anthropogenic disturbances. In contra st, Weng et al. (2002) fi nd a period of abundant Cecropia pollen in Maxus 4 core, about 25 km SE from Maxus 5. Constituting 40-80% of the pollen for ca. 2600 years (8300–5700 14C yr BP) during the Early to Mid Holocene, the high abundance of Cecropia pollen was not likely the result of human activity as there were no indications of Zea mays (maize), and extremely low levels of other Poaceae and charcoal. High levels of Cecropia were attributed to high tree mo rtality due to long periods of drought. The droughts are not thoug ht to have been cyclical , but rather unpredictable (Weng et al. 2002). At about 5800 14C yr BP the climate became wetter and pollen of Cecropia decreases. The sequential peaks of diffe rent species may be interpreted as successional replacements or stochastic ch anges from short-term flooding. Increased precipitation is suggested by the formation of a Mauritia swamp at 3700 14C yr BP which is replaced by a hydrarch succession at 1000 14C yr BP resulting in the modern vegetation found in this area of Ecuador today (see We ng et al. [2002] and references therein). The two sites located in the upper foothi lls, Mera (Liu and Colinvaux 1985) and San Juan Bosco (Bush et al. 1990), were comp letely forested over the entire Holocene. The presence of many Andean taxa at these si tes resulted from an estimated temperature decrease of 7°C during the Last Glacial Ma ximum (LGM). Andean species disappeared as temperatures increased during the Holocene.


6 Although there was a signal of drier conditi ons during the early Holocene from the Maxus cores, no such signal was found from Mera or San Juan Bosco (Liu and Colinvaux 1985, Colinvaux et al. 1997). During the LGM the cloud layer of condensing moisture near the Andean mountain chain probably form ed at lower elevati ons than during the interglacials resulting in a constant local climate that did not experience the same conditions of drought as the Maxus sites (Bush et al. 2004). It is im portant to note that increased humidity in the foothills as well as a blending of highland and lowland plants appear to be two general consequences of lowered temperatures (Colinvaux et al. 1997, Bush et al. 2004) regardless of the dating unc ertainties associated with the Mera and San Juan Bosco sites in particular (Heine 2000). High Andes Several lakes in the Ecuadorian highlands have been cored. Unfortunately, many studies still remain unpublishe d. Generally, conditions became warmer at the onset of the Holocene followed by a slight decrease in temperature. The late Holocene shows indications of human habitation. The clima tic conditions therefore are difficult to reconstruct for the last 2,000 yr. At about 10,000 14C yr BP Colinvaux et al. (1997) postulate a warming trend at Laguna Llaviucu (3120 m) due to the increased presence of Cecropia , other Moraceae, Anacardiaceae, Sterculiaceae, Meliaceae, and Sapindaceae in the pollen record. This increase in taxa more typical of warmer conditions (lower elevations) follows a spike in Polylepis at the Pleistocene/ Holocene transition. As Cecropia decreased again ca. 6000 14C yr BP, Alnus appeared to replace it as a major disturbance species. Podocarpus maintained constant levels during the Holocene until ca. 2000 14C yr BP, when its pollen percentages drop. The pollen record of Lake Ch orrera (3700 meters), 15 km to the west at


7 Cajas National Park (Hansen et al. 2 003), does not track these changes as Polylepis peaks three times: at the beginning of the Holocen e, during the Mid Holocene, and during the Late Holocene. Podocarpus is most common during the first 4000 years of the Holocene. The difference between Chorrera and Llaviucu is related to the elevation of the sites relative to tree line. As Llaviucu lies below the average Holocene tree line, only one peak is observed marking the upslope movement dur ing the Early Holocene. For the remainder of the epoch, Polylepis is located well above Llaviucu. On the other hand, Chorrera is located near Polylepis vegetation for most of the Holocene. Thus, any fluctuation in elevation due to changes in temperature or humidity, or changes in wind direction should be reflected in relatively minor changes compared to the peaks observed at Lluviacu (Hansen et al. 2002). The data from Lago Cho rrera therefore suggest that since about 10,000 14C yr BP species composition in high elev ation habitats have remained fairly constant. Species more indicative of higher temperatures (Colinvaux et al. 1997) were well below their tolerance limit to the cold climate near the tree line and did not reach 3700 m. From the northern Andes several lakes were studied at lower elevations. Unlike the southern cores, these do not represent tree-line habitat. Lago Yaguarcocha was surrounded in the early Holocene by Weinmannia forests with Podocarpus and Cecropia also present. At around 5000 14C yr BP the two humid forest taxa ( Podocarpus , Weinmannia ) decreased considerably and pe rsist at only very low levels. Miconia also decreases apart from a small peak at about 1800 14C yr BP. Cecropia and Alnus , two disturbance species from di fferent elevational belts, we re present throughout. Other plants associated with drier habitats, such as Mimosaceae, were present for most of the


8 Holocene (Colinvaux et al. 1988). These data su ggest that the early Holocene tended to be wetter than the second half. The presence of pollen from different types of vegetation may also be due to changes in wind directi on. Lake Yaguarcocha is located within an ecotone between humid montane forests to the s outh and dry forests, characteristic of the arid Chota valley, just to the north. Evidence for considerable dry conditions in the Andes, at le ast for the last 2000 14C yr BP, come from Lago Yambo. In the inter-And ean valley many areas are located in rain shadows that lead to locally very low levels of rainfall. However, the sediment load also suggests that higher in the mountains rains must have been frequent (Colivaux et al. 1988). Lago Yambo is also interesting because of different pollen types that come from the lowlands. These may have been disperse d by wind over the cordilleras. On the other hand, a second hypothesis that involves humans as dispersal agents should be evaluated as well. People traveling to and from the lowlands may have transported pollen on their clothing and/or trade goods. This hypothesis would be strengthened if Lago Yambo were known to be located along an established tr ade route. Western Lowlands No pollen studies have been conducted in the western lowlands of Ecuador. The absence of such research, despite the existe nce of wetlands and lakes, is surprising and possibly reflects a bias against working in greatly degraded and populated areas. The most pertinant pollen data to this re gion come from deep sea core Tri 163-31B 300 km WSW of the Río Guayas delta (Heusser and Shackleton 1994). Ironically, Heusser and Shackleton (1994) focused mostly on Andean taxa in their study. Since the core undoubtedly must contain lowland taxa as well, their data plot s of arboreal species could be taken tentatively as an indication fo r changes in the lowlands as well. Lowland


9 species must be represented to a non-trivial degree since the Guayas watershed covers a large expanse of tropical lowl ands from Santo Domingo in th e north to Guayaquil in the south. In addition, pollen in the core may stem from several rivers that flow through as much lowland as mountainous terrain, dropping from the southern highlands into the Gulf of Guayaquil. Based on this assump tion, highland and lowland forests (at least within the greater watershed of the Río Guayas ) contracted and were replaced to a large degree by grasslands or savannas. Arboreal po llen was at its lowest levels at ca. 15,000 14C yr BP. After that point the Late Pleist ocene was characterized by an increase in arboreal pollen which closely follows the 18O curve for the core except for the Early Holocene at which the oxygen curve changes more gradually. The amount of arboreal pollen in turn reaches a stable plateau rapidly (Heusser and Shackleton 1994, Fig. 2). However, since the contribution of lowland ta xa to the total arborea l pollen count is not indicated, we do not know how the lowland fl ora may have tracked climatic changes during this period. Mangrove ( Rhizophora ) pollen peaked at the Pleist ocene/Holocene transition and may indicate increased precipitation (runoff) and temperature as well as increased sea levels. Levels of Rhizophora pollen were much reduced by 5000 14C yr BP (the endpoint of the core). This drop may represent a re duction in precipitation during the Holocene or a drastic change in the shoreline along the Ec uadorian coast. As arboreal pollen remained high during the Holocene (compared to Poaceae), increased aridity does not seem likely. An estimate of the changes in shoreline lo cation by Stothert et al. (2003) supports the possibility that increased sea levels proba bly reduced the amount of shallow coastal waters in which Rhizophora could grow.


10 Some limited evidence about the possibl e vegetation cover in northwestern Ecuador comes from a pollen record obtaine d in the southern Chocó of Colombia. Spanning the last 7000 14C years, the core from Lake Piusbi shows very little change during the mid to late Holocene (Behling et al. 1998). Throughout this time the entire area was covered by a diverse tropical rainforest . This stands in star k contrast to the high level of variation suggested for the arid z one of southwestern Ecuador (see Paulsen 1976). Evidence of human activity at La ke Piusbi was found beginning at 3500 14C yr BP, although human populations appear to have been low even when maize ( Zea mays ) was cultivated about 1700 14C yr BP. Sedimentology Changes in temperature and precipitation as reflected in glacial, periglacial, and fluvial records have recently been review ed by Heine (2000). Despite an increase in temperature, the Pleistocene/Holocene transi tion was marked in the Andes by an advance of glaciers in at least two areas of the northern Andes in Ecuador: Papallacta pass and Pichincha volcano. This advance was probably the result of increased humidity after a generally dry LGM (Heine 2000). This findi ng is in agreement with studies from southern Ecuador that also detected a hum id transition (Hansen et al. 2003). Glaciers advanced two more times in the Late Holocen e as a result of both decreased temperatures and increased humidity (Heine 2000; Fig. 4). For the humid tropics, the nature of clim atic change varies geographically. Changes in sea surface temperature (SST ) and the intertropical converg ence zone (ITCZ) varied in different areas of the world (Thomas 2000). Par ticularly revealing to climatic change in Ecuador is the finding that Amazonia receiv ed high amounts of rain fall during the Late Pleistocene and Early Holocene with an in terruption by th e Younger Dryas. In Amazonia


11 rainfall may have peaked at ca. 13,500 yr BP as suggested by the alluvial record. This record also shows that after 8000 14C yr BP rainfall decreases to reach current levels (Thomas 2000). In western Ecuador rainfall reach ed its lowest levels at 15,000 14C yr BP as suggested by deep sea core Tri 163-31B 300. Oxygen isotope data from the site show a gradual decrease throughout the Holocene punctuated by higher levels at ca. 8000–9000 14C yr BP. These higher levels indicate of a short period of cooli ng. Lower temperatures for this time period have also been record ed from other studies on oxygen isotopes from several sources, glacial adva nces, lake cores, and polle n abundances among others (compiled by Mayewski et al. 2004). Mayewski et al. (2004) identify this period as one of six periods of “rapid climate change” (RCC) for the Holocene. They assign four of the other five periods to a “cool poles, dry tropi cs” and one to a “cool poles, warm tropic” RCC pattern, respectively. For equatorial nor thwestern South America the RCCs were characterized by dry climate with more humid conditions suggested only for the periods between 3500–2500 14C yr BP and 600 yr BP (M ayewski et al.2004; Fig. 5). Present climate in the wester n lowlands of Ecuador is gove rned to a large degree by the Peru Current and the El Niño Southern Oscillation. The upwelling of the ocean current along the South American coast is re sponsible for dry conditions from Chile to southern Ecuador. The ITCZ is typica lly located further north (between 3 and 10 N) here than in other areas. Just south of the ITCZ a band of westerli es brings humid air toward the continent resulting in the high amount of rainfall in the Colombian Chocó. Farther south the cold SST of the Peru Curre nt leads to dry conditi ons for most of the year. From January to March, as the ITCZ attains its most sout herly position, western


12 Ecuador receives most of the yearly rainfall despite the fact that the ITCZ is weakened compared to the rest of the year. How the precise locations of the ITCZ, SST, and/or ENSO differed in the past and shaped the cl imate is suggested by a few recent studies. As the ITCZ, SST, and ENSO are interrelated in th e eastern Pacific, I tr eat them as a set of interrelated phenomena and will not attempt to separate their individual contributions to climate in Ecuador. A reconstruction of SST using magnesium/cal cium ratios in foraminifera from the Galápagos Islands suggests that during the LGM the temperature was 1.2 °C lower than in the present (Koutavas et al. 2002). Koutavas et al. ( 2002) concluded that tropical temperature gradients were less steep, Hadley and Walker circulations were weakened, the ITCZ was located more toward the south, and an El Niño-like pattern characterized the tropical Pacific. Maximum SSTs we re found for the Bø lling-Allerød (13,000–14,800 yr BP) and Younger Dryas (11,500–13,000 yr BP) periods , during which SSTs also were highly variable. During the Early to Mi ddle Holocene variability was low with temperatures intermediate between LGM and present levels. Low variability was associated with an infrequent or weakened ENSO during this time as evidenced by the low frequency of sediments with a low conten t in organic material in Lake Chorreras (Rodbell et al. 1999, Moy et al. 2002, Hansen et al. 2003). During the period of low ENSO activity several tropical marine taxa appear on the central Peruvian coast (Andrus et al. 2002, 2003). Among these is the Peruvian sea catfish ( Galeichthys peruvianus ), whose southern limit is normally off the coast of Ecuador, despite a broad tolerance fo r water temperatures. Analyses of 18 O from the otoliths suggest that SSTs were 3 °C higher th an at present. These data seemingly conflict


13 with the suggestion of a prolonged La Niña off the Ecuadorian coast. Throughout the Pleistocene subtropical incursions into the coas tal area of Peru have occurred three times during the last 400,000 years duri ng interglacials. During thes e times the currents were weakened and/or winds were diminished (Molina-Cruz and Herguera 2002). During interglacials that coincide with a northern position of the eastern tradewinds, the Chile Current was weakened (Molina-Cruz a nd Herguera 2002), but tradewinds may nonetheless remain strong (Clement et al. 2001, Moy et al. 2002). Strong upwelling of the Peru Current and a weakened Chile Current would allow the Peru Countercurrent to wedge itself between these two currents and reach the Peruvian coast (Molina-Cruz and Herguera 2002). I hypothesize that tropical mari ne species could have reached the central coast of Peru by migrating during strong El Niño years during th e early Holocene and finding temporal refuge in coastal embaym ents. The Peruvian Countercurrent between 8000–5800 yr BP would have allowed tropical and subtropical species to survive at these low latitudes for several millennia. In summary, the first half of the Ho locene may have been marked by strong easterlies that led to a shallow thermocline in the eastern Pacific and therefore increased upwelling off the Ecuadorian coast. The incr eased upwelling (Koutav as et al. 2002) and dry conditions in the western Amazon (Weng et al. 2003) and the Andes (Paduano et al. 2003, Weng et al. 2004) both suggest that the ITCZ was located further to the north than presently between 8000–5800 14C yr BP and confirm the da ta of Haug et al. (2001) on titanium and iron concentrations from the anoxic Cariaco Basin. Apart from these long term changes in climat e, solar forcing appears to have led to large-scale changes in atmospheric circul ation patterns (Bush et al. 2002, Schimmelmann


14 et al. 2003). Increased solar activity produced increased leve ls of insolation which, in turn, strengthened trade winds and broadened th e latitudinal extent of the Hadley cell. On the other hand, decreased in solation would weaken trad e winds and constrict the latitudinal extent of the Hadley cell. These cyclical changes have resulted in periods of prolonged drought and humid conditions, respec tively, independently of ENSO activity (Schimmelmann et al. 2003). Paleontology Little paleontological res earch has been conducted in Ecuador. Most of the work has focused on Late Pleistocene megafauna fr om both highland and we stern lowland sites (Edmund 1965, Hoffstetter 1986, Colt orti et al. 1998, Ficcarelli et al. 2003). The analysis of the bird fauna in tar pits at Carolina in Ecuador and Talara in Peru document Late Pleistocene faunas in the dry coastal ar eas (Campbell 1976, 1979, 1982). These studies will be discussed in terms of their relevance to the Holocene. As elsewhere in the Late Pleistocene wo rld, the megafauna of Ecuador disappeared in the Late Pleistocene (Hoffstetter 1 986, Coltorti et al. 1998). The exact timing, however, has not been well established. In the northern Andes of Ecuador a bone fragment presumably be longing to the mastodont Haplomastodon chimborazi (= Cuvierionus hyodon ; Prado et al. 2005) has been dated to be about 16,670 ± 80 14C yr BP (Coltorti et al. 1998). Fragments of the mylodont Glossotherium wegeneri were dated to 12,350 ± 70 14C yr BP. Further south in the A ndes of Chimborazo province, where glaciation was more pronounced, a bone of H. chimborazi was dated to 20,980 ± 530 14C yr BP (Coltorti et al. 1998). There is some confusion about the exact identity of the H. chimborazi bones as Coltorti et al . (1998, p. 584) list most of their bone material as “Miscellaneous bone” and only one as Glossotherium in a summary table of the


15 radiocarbon dates. In the remainder of the text, however, some of the same “Miscellaneous bone” fragme nts are then treated as H. chimborazi . Despite the confusion about the identity of the bones, these dates fall within the range of mass extinctions of large mammals in South America. Carbon isot ope ratios obtained from bone and teeth (enamel and/or dentine) found at the same or nearby sites suggest that Cuvierionus (Haplomastodon) hyodon had a mixed diet of C3 (trees, bushes, shrubs, forbs, and high elevation and high latitude grasses) and C4 (tropical grasses and sedges) plants (Sánchez et al. 2004). Therefore, in northern Ecuador mastodonts had access to tropical grasses as well as broad leaved vegetation/high elev ation grasses. Sinc e the dentition of C. hyodon suggests grazing habits (MacFadden and Shocke y 1997), I conclude that at nearly 3000 m elevation this species lived in a grass do minated habitat with possibly some brushy vegetation. In the western lowlands the megafauna ma y have persisted until the early Holocene as some bones were associated with mollusks which were dated to 8680 ± 80 14C BP (Ficcarelli et al. 2003). Given that most large mammals became extinct before 11,000 14C yr BP, this date is surprisingly late. More information on the context and number and species of specimens sampled, as well as a discussion on the likeli hood that the mollusks used for radiocarbon dating had not been rewo rked, would have been desirable to assess this potentially important finding. The assembla ge of vertebrates from tar pits in La Libertad is unusual and includes species from open savanna, forested habitats, and wetlands (Edmund 1965, Campbell 1982, Ficcarelli et al. 2003), such as crocodilians, Geochelone cf. gallardoi , Geomyda sp., Haplomastodon chimborazi (= Stegomastodon waringi in Sánchez et al. 2004, Prado et al. 2005), Eremotherium laurillardi or E.


16 rusconii , Scelidotherium reyesi , Glossotherium tropicorum , Chlamytherium occidentale , Equus santa elenae , Palaeolama aequatorialis , Odocoileus salinae , Neochoerus sirasakae , Echimyidae, Dusicyon sechurae , Protocyon orcesi , Puma sp., Smilodon neogaeus . Ficcarelli et al. (2003, p. 840) assume th at the mastodon from the site is of “forest origin,” but Sánchez et al. (2004) find clear evidence based on carbon isotope ratios that Stegomastodon waringi fed mostly on C4 plants suggesting that it foraged in open vegetation dominated by tropical grasses. Tomiati and Abazzi (2002) analyzed the bones of Odocoileus deer from the La Libertad site (Early Holocene according to Fi ccarelli et al. 2003) and suggest that the morphological differences with recent comparative specimens of O. virginianus ustus warrant distinct species status. In additi on, differences in the occlusal profile of O. salinae , similar to modern deer from the coast, we re interpreted as an indication of poorly abrasive vegetation more typical of forested than grass dominated habitats (Tomiati and Abazzi 2002). Thus, forested habitats may have been present since the Late Pleistocene to Early Holocene. According to Campbell (1976, 1982), paleon tological data on birds suggest savanna-like vegetation for the western lowla nds during the Late Pl eistocene as birds indicative of both grasslands and more hum id climate were found in tar pits dating 14,000 yr BP. A closer examination of the species lists, however, suggests that most of the species associated with a humid habitat ar e wetland species such as herons and egrets (Ardeidae), ibises (Threskiornithidae), flamingoes (Phoenicopteridae), and ducks (Anatidae). Many of these species prefer marshy or lake-shore habitats rather than rivers surrounded by gallery forest. Thus, the bird as semblage suggests a mixture of savanna


17 with open wetlands and possibly gallery forest. The existence of wetlands is puzzling, but may be explained by a higher water table and/ or the presence of large mangrove flats exposed by lower sea levels during the both th e Late Pleistocene and the Early Holocene (Ferdon 1981). With regard to the wetland birds, raptor s, and raptorial birds the discovery of several morphologically distinct species and even genera from the Talara and La Carolina tar seeps (Campbell 1976, 1979) have yielded several new species of ducks, shorebirds, raptors, and vultures. The bird assemblage suggests the presence of wetlands as well as open areas where microand mega-mammals mu st have been abundant to sustain seven species of hawks and six species of vultures. In the present the arid zone of Santa Elena Peninsula supports only one common hawk ( Buteo polyosoma ) and two vultures ( Coragyps atratus , Cathartes aura ). A taxonomic revision of some of the birds described by Campbell (1976, 1979) may be necessary in light of the greater availability of modern, comparative specimens. It is imperative to dete rmine the status of these birds in order to assess the biogeographical signif icance of the arid landscapes of Ecuador and Peru for the evolution of the arid equatorial fauna (sensu Chapman 1926). The Talara tar pits have reve aled a vertebrate fauna that is richer in species than that of the Santa Elena Peninsula. Apart fr om the same mammals found at la Libertad, it includes the large predators Felix atrox , Canis dirus as well as bats, skunks, and armadillos (Churcher 1959, Campbell 1982). These tar pits have also yielded large amounts of insect remains (Churcher 1966). Most are aquatic insects, suggesting that the tar pit was covered by a film of water during the time of deposition. Most animals perished as they inadverten tly got caught in the tar conc ealed under the film of water


18 (Churcher 1966). It is also possible that the tar pools were te mporarily covered by a shallow body of water such as a coastal lagoon or marsh. Such short-lived (years to decades) wetlands could have formed as a re sult of changes in sea level and geological uplifting along the coast. For discussions on the formation of the low lying table lands called tablazos, the most common geological feat ure resulting from uplift see Ficcarelli et al. (2003), Stothert et al. (2003), a nd Cantalamessa and DiCelmo (2004). Archaeology Recent research in tropical zooarcha eology and archaeobotany have greatly improved our understanding of past climatic changes, biogeographical patterns, and human adaptations to local conditions. Methodo logical innovations, such as the use of starch grains and phytoliths, have led to major leaps in our interpretations from archaeological sites (Piperno et al. 2000a,b; Piperno and Stothert 2003). The detailed analysis of zooarchaeological remains also provides a powerfu l tool for inferences about human exploitation of wildlife and interpre ting the habitats surr ounding settlements (Kent et al. 1999, Steadman et al. 2003). Cultural pe riods mentioned in this section are summarized in Figure 1-2. Apart from the Cueva Negra de Chobshi site in southern Ecuador, which may have been occupied as early as 8000 14C yr BP, zooarchaeological remains dating to the Early Holocene are virtually absent in Ecuador (review ed by Stahl 2004). Located at 2400 meters, Chobchi Cave has an unusual fauna that includes highland as well as lowland taxa (contra Stahl 2004; Table 1-1). Presently these taxa do not overlap and the highland species are not obser ved at this elevation. There are various competing hypotheses regarding the species assemblage and it may be impossible to distinguish among these (Stahl 2003). 1) The vegetation cons isted of montane forest similar to what


19 we would expect to find at this elevation today. The highland species were brought to the site through trade with other pre-ceramic groups living high er in the Andes. 2) The vegetation was modern in character and th e animals were obtained through long-range hunting outings. 3) Chobshi Cave was located still near the tree-line and lowland species were obtained through trade. 4) The plant co mmunities during the glacials had no modern analogues and were characterized by a comb ination of lowland and highland elements (see Colinvaux et al. 1997). At 8000 14C yr BP the faunal and floral communities may still have been in transition. Extensive studi es from Colombia about 1000 km to the north have uncovered human remains in association with mega-mammals dating to 11,740 14C yr BP, followed by the Abra Stadial during wh ich the Sabana de B ogotá consisted of subpáramo, and a warmer Early Holocene during which most of the Sabana was probably covered by Andean forest (reviewed by Co rreal 1999). Natural communities appear to track climatic changes rapidly. Thus, if the last hypothesis were true, human occupation would be expected to have occurred before 8000 14C yr BP. By implication, the 14C dates obtained for this site would be in error. Most of the other site s in the Andes belong culturally to the Late Formative. Only a few of these sites, however, have been analyz ed in any detail. In the north, La Chimba has yielded the largest amount of bone of any site in Ecuador (Athens 1990). The fauna is dominated by species characteri stic of tree-line habitat (Stahl and Athens 2001). In the south, zooarchaeological analyses have been more limited. Studies on a series of sites have yielded faunal lists that lack the taxonomic rigor necessary to make good inference about the paleoecological conditions. Most of the taxa are typical of the modern highlands, which I consider as the area above 2000 meters (reviewed by Stahl 2004).


20 Year (BC) Period Guayas Highlands 500 Inka 750 Cochasquí 1000 INTEGRATION Manteño 1250 Socabamba 1500 --------------------------------------Late La Chimba 1750 2000 REGIONAL DEVELOPMENT Guangala Middle La Chimba 2250 Cerro Narrío 4 2500 --------------------------------------Early La Chimba 2750 Cerro Narrío 3 LATE FORMATIVE Chorrera 3000 Pirincay 3250 ··················································· Cotocollao, Cerro Narrío 2 3500 Machalilla Challuabamba 3750 8 4000 7 Cerro Narrío 1 4250 6 4500 EARLY FORMATIVE 5 Valdivia 4750 4 5000 3 5250 2 5500 --------------------------------------1 6000 Late Vegas 8000 PRECERAMIC Chobshi Early Vegas 10,000 Figure 1-2. Chronology for major cu ltural periods mentioned in the text (simplified after Stahl 2004).


21 The peninsula of Santa Elena was inhab ited by humans at least since the early Holocene (Stothert et al. 2003). It is an ar ea of great importance for the origin of agriculture as attested by the early domestication of Cucurbita and maize (Piperno and Stothert 2003). Preliminary studies on site OGSE-80 (hereafter, Site 80) show that the Vegas people relied on resources associated with mangroves, inshore and offshore fish, and many species of terrestrial snakes, frogs , mammals, and birds (Stothert et al. 2003). Many of these terrestrial vertebrates, such as Bufo sp. (Bufonidae), are associated with freshwater streams or swamps which no longer exist in this arid region of Ecuador. Most of the mammals and birds, however, are more typical of mangroves and/or dry forests (Stothert et al. 2003), hinting at conditions sim ilar to those presently found in dry forests. Thus, rainfall may have been highly seasonal. Compared to the paleontological data from the Late Pleistocene (Campbell 1976), the Holocene archaeofauna suggests an increase in forest cover and a continue d presence of wetlands. A high water table and seasonal rainfalls could have maintained such a habitat. Phytoliths from Site 80 suggest that grasses and shrubs were the dominant vegetation. The absence of palm phytoliths le d Piperno and Pearsall (1998) to conclude that overall conditions must have been dry. During the occu pation of Real Alto (ca. 6200–3800 14C yr BP, Zeidler 2003; Fi gure 1-1) climate was generally dry, but phytoliths and macrobotanical data from this site suggest that conditions were at least temporarily and/or locally (around rivers, for example) humid enough to sustain humid forest trees (Pearsall 1988). In the Jama Valley about 240 km to the north, the vegetation was considerably more humid. The ca. 4000 yr old pre-occupational sediment strata reveal the presence of intact humid forest (Veintimilla 2000).


22 Table 1-1. Fauna from Chobshi Cave* w ith modern habitat associations†. Common Name Scientific Name Habitat Association White-eared Opposum Didelphis albiventris Arid and temperate (including humid forest) regions cf. Domesticated Dog Canis cf. familiaris Human habitation Spectacled Bear Tremarctos ornatus Forest and paramo above 2000 meters Unidentified procyonid Procyonidae sp. Variable Unidentified mustelid Mustelidae sp. Variable cf. Mountain Tapir Tapirus cf. pinchaque Forest and paramo above 2000 meters White-tailed Deer Odocoileus virginianus Usually open habitats and secondary vegetation; not a true forest species Northern Pudu Pudu mephistopheles Forest and paramo above 2000 meters cf. Mountain paca Paca cf. taczanowski High montane forest Bicolor-spined Porcupine Coendou bicolor Lowland to premontane forest up to 2500 meters Brazilian Cottontail Sylvilagus brasiliensis Lowland forest, cloud forest, into paramo Unidentified tinamou Tinamidae sp. Variable * From Lynch and Pollock (1981, p. 98) as cited by Stahl (2004). † Sources: Eisenberg and Redford (1999), Emmons and Feer (1997). During the Middle to Late Holocene, the archaeofauna is derived from communities indistinguishable from modern ones (Stahl 2003). Excavations from the Valdivia Valley have recovered many marine species of birds as well as terrestrial vertebrates that still occur there today. Howe ver, these archaeologica l sites are from the vicinity of the coastal cordillera and may not reflect overall paleoclim atic patterns due to the special biogeographic significance of thes e mountains (see belo w). In addition, the inherent difficulties associated with correla ting changes in the stra tigraphy to changes in the environment preclude the ability to de tect fine-scale temporal changes in


23 paleoecological conditions (Stahl 1991). A pr eliminary conclusion is that environmental conditions in western Ecuador ha ve been sufficiently constant to maintain the same types of habitat over the last six millennia. Apart from the fauna and flora associated with archaeological sites, patterns in settlement and abandonment of upland and coas tal areas, respectively, can be interpreted as the result of sever drought. Paulsen (1976) found that on the Santa Elena Peninsula site occupation varied since the Guangala phase of regional development. Through times of abundant rainfall, Guangala people used sma ll, open catchment basins, called albarradas, to sustain agriculture during the dry mont hs. During periods of severe droughts these people abandonded their settlements, moved to coastal areas, and may have lost contact with adjacent cultural groups. This pattern was repeated several times throughout the Late Holocene (Paulsen 1976). Schimmelma nn et al. (2003) have correlated flooding events in California and other paleoclimatic records with periods of cultural transitions across much of Central and South Am erica over the last 2000 years. Biogeographical Patterns Although most biogeographical pa tterns are the result of l ong term processes, many may also reflect recent events related to the Pleistocene/Holocene transition. The characteristics of the bird faunas associated w ith coastal cloud forest s (garúa forests sensu Becker and López 1997) and the Tumbesian Cent er of endemism both contribute to our understanding of long-term climate and vege tation change. A center of endemism is “defined by the distributional congruence of constituent taxa” (Cracraft 1985a, p. 50). The “constituent taxa” are those species that share a common geographic realm that can be defined based on vegetation cover, geograp hic elements (such as mountain tops or island), or useful abstracti ons of territorial or landscap e organization (for example,


24 country limits or areas bounded by rivers, re spectively). The concept is therefore dependent on the scale at which distribu tions are analyzed (Cracraft 1985a). The Tumbesian Center of endemism, located in southwestern Ecuador and northwestern Peru, is particularly important for suggesting past climatic conditions in the arid coastal region of Ecuador. The low clouds that form over the cool waters of the Peru Current shroud the mountain tops of the coastal cordillera and the lower slope s of the southern Andes in mist. The vegetation that grows in these ar eas is characterized by a high diversity of epiphytic orchids and bromeliads reminiscen t of the mid-elevation cloud forests (Best and Kessler 1995). Many Andean bird species can be found at unusually low elevations within these forests. In th e coastal Cordillera Chongón-Colonche, for example, 21 species are Andean (Becker and López 1997). Many of these, especially the hummingbirds, appear to be seasonal visito rs, although many understory spec ies are likely residents. These birds are mostly poor dispersers that are quite tolerant of edge or secondary vegetation (Tellkamp 1999). The presence of Andean bird species in the coastal Cordillera Chongón-Colonche, about 130 kilometers from the main mountain chain, suggests that su itable conditions for dispersal must have existed between the tw o mountainous systems at some time during the (geologically) recent past. Many of these Andean species are poor fliers that are not likely to be long-distance dispersers. They would require suitab le cool and humid conditions in the Guayas watershed between the coastal mounta ins and the Andean backbone to overcome the “lowland barrier.”


25 The lowest elevational distribution of thes e species at present is about 1400 meters (Ridgely and Greenfield 2001b). Given an adia batic temperature grad ient of 0.5 °C per 100 meter elevation, the temperature must have dropped by at least 5–6 °C to result in lowland temperatures low enough to be tolera ted by these Andean birds. During the LGM temperatures were probably low enough (Co linvaux et al. 1997), but the extremely dry conditions during the LGM have probably result ed in reducted forest cover in the Andes and lowlands (Heusser and Shackleton 1994). Th erefore, I suggest that the dispersal event may have occurred during a previous, more humid glacial. Morphological and genetic studies on the phylogeographic affinities of these populations are needed to determine their source populations and potential time since divergence. Assuming that thes e populations dispersed befo re the last LGM, it seems plausible that the Cordillera Chongón-Colonche has been a local, moist refugium for humid forest species in an otherwise arid en vironment. Thus, the coastal cordillera may have played an important role in the reestab lishment of bird populations requiring at least seasonally humid habitats north of the Sant a Elena Peninsula duri ng the Early Holocene. Many endemic birds of the Tumbesian Center show a close affiliation with sister taxa inhabiting the more humid north (C hapman 1926, Best and Kessler 1995). These species are usually found in or associated with riparian th ickets or the low cloud forests of the southwest, and are morphologically di stinct enough to warrant species or genus status. The distribution of distinct Tumbesia n birds suggests that cloud forests may have persisted in southwestern Ecuador and/or northwestern Peru for a long time (Chapman 1926), possibly throughout the Pleistocene as suggested by a molecular analysis of the chat-tyrants (genera Tumbezia , Ochthoeca , Silvicultrix ; García-Moreno et al. 1998).


26 Analyses predicting vegetational changes sin ce the LGM also support the idea of a dry to moist forest refugium in southwestern Ecua dor and northwestern Peru (Best and Kessler 1995). Micro-climatic conditions as illustrated may have modulated the general trend in a reduction of the arboreal vegeta tion during the Late Pleistocen e and had an important role in the secondary dispersal of many species. Historical Accounts At the time the Spaniards arrived in nor thwestern South America, native peoples were undergoing major changes in terms of popul ation structure, trade, and warfare from the devastating effects of Old World diseases. By the time that Amerindians saw the first conquistadores, diseases had already d ecimated many human populations (Newson 1995). Natural ecosystems were thus on the reb ound; the writings of the early chroniclers describe a natural environment that is not repr esentative of the latest phase of Ecuadorian prehistory. Nonetheless, they suggest the pos sible vegetation types that many areas were able to sustain in many cases in the absence of human settlements. The following discussion is primaril y based on Hidalgo Nistri (1998), who reconstructed the distribution of the coasta l and Andean forests. Although the habitat classifications used by Hidalgo Nistri are very simplified and he frequently extrapolated from much later descriptions by 19th century naturalists, his study is quite informative about many aspects of prehisto ric/historic vegetation cover. Major sources for his work were the writings of Cieza de León and othe r chroniclers as well as documents found in 11 archives in Ecuador and Spain. The northern Andes consisted of a comple x mix of both humid and arid vegetation types that I will describe from north to south. The northern province of Carchi was mostly covered by montane forest. The xer ophytic vegetation of the Chota valley, which


27 intersects the western cordillera, was traversed by lush vegetation bordering the major rivers. The upper Chota watershed, which incl udes the city of Ibarra, was vegetated by xeric forests as well. The Otavalo valleys in turn consisted of agricultural lands with “abundant” patches of humid forest. The Mo janda Massive was surrounded by montane forests that formed an east-west corridor connecting the two cordilleras. The Cayambe valley still had many forest remnants (the te rm montaña is interpreted as forest). The upper Guayllabamba valley was a dry forest island separated from the dry Río Patate watershed by another, much wider east-west corridor that stretched from Ilaló volcano south to the Iliniza mountains. The Río Pata te watershed in turn was separated by a tongue of humid forest and the dry Río Chambo watershed from the valley surrounding the city of Riobamba. The eastern slopes of the western cordil lera included some discontinuous patches of humid montane forests that disappeared in areas with páramo at high elevations and xeric vege tation in the dry inter-Andean valleys. A similar picture was reconstructed for the western slopes of the eastern cordillera. The western slopes of the western Andes were clad in premontane and montane forest from Ca rchi south at least to the Río Chimbo valley, which currentl y is almost entirely deforested. Toward the coast, a savanna stretched from the lower Río Daule watershed north of Guayaquil along the coast south to El Oro province. Dry forests existed in western Manabí, the Santa Elena Peninsula, and most of El Oro province, interspersed with humid tropical lowland forests. The shores off the Colonche Mountains are described as humid tropical lowland forest as well. According to Hidalgo Nistri (1998) the Punta de Santa Elena was forested as well, suffering major deforestation only after oil exploration began in the early 1920s.


28 Synthesis The studies from different disciplines l ead to similar conclusions. The Holocene began with moist conditions and temperatur es similar to modern ones. At around 8000 14C yr BP the climate became drier in the east, and the El Niño phenomenon was considerably weakened. A warming trend occurre d in the Andes. In the east the location of the ITCZ may have been more variable as the earth’s inclination on its axis was near 24.5° rather than 23.5° today (Jackson and Broccoli 2003, Fig. 1). The temporal absence of easterlies may have resulted in lower input of moisture fro m the Atlantic. At the same time, the cold Peru Current became cold er and increased upwelling caused a steep thermocline in the eastern Pacific. Warm wa ters, however, penetrat ed the north central Peruvian coast due to a weakened Chile Cu rrent. Under this s cenario, southwestern Ecuador should have received moisture mostly in the form of garúa mist. Little empirical evidence exists to reflect the environmental c onditions of this period for the coast (Table 1-2). After ca. 5800 14C yr BP temperatures dropped and humidity increased again. During this time the climatic conditions were very similar to the pr esent ones punctuated by periods of abundant rain and severe drought s as ENSO activity in creased. Cycles of prolonged droughts during the Holocene probably were related to solar activity with cyclic large-scale changes in globa l circulation patte rns (Table 1-2). Environmental Change as the Result of th e Interaction between Culture and Nature The South American Holocene is often inte rpreted as a time of fairly constant climate punctuated by periods of severe drough t and rains, respectively, driven by ENSO and orbital forcing. Humans responded to these climate changes by adapting to the environment or by migrating out of aff ected areas (Paulsen 1976, deMenocal 2001).


29 Humans and nature alike also were at the mercy of vol canic eruptions throughout the Holocene. These eruptions have caused the abandonment of many settlements due to heavy ashfall, especially in western Ecuador (Zeidler and Isaacson 2003) . It is likely that thick layers of tephra made many stretches of land uninhabitable for fauna and flora as well. Table 1-2. Summary of climatic changes duri ng the Holocene as suggested for the three major regions in Ecuador. Region/sub region Dates Eastern lowlands Andes Western lowlands 10,000–9000 humid/warm humid/cold dry/warm 9000–8000 seasonally dry/warm dry/warm dry/warm 8000–7000 seasonally dry/warm dry/warm very dry/cool* 7000–6000 seasonally dry/warm dry/warm very dry/cool* 6000–5000 humid/warm dry/cold very dry/cool* 5000–4000 humid/warm dry/cold dry/warm 4000–3000 humid/warm humid/cold dry/warm 3000–2000 humid/warm humid/cold dry/warm 2000–1000 humid/warm dry (N), humid (S) dry/warm 1000–Present humid/warm humid dry/warm *: Inferred from present climatic conditions during northern position of ITCZ when SST are lowest. Humans not only respond to changes in the environment but also have the ability to induce local and even regional changes in vegetation cover and micro-climates. Even archaic peoples had the necessary tools to clear large expanses of forest by setting fires or hunting keystone species, such as large ma stodons, in the Late Pl eistocene (Martin and


30 Steadman 1999). With the advent of agriculture , forest was cleared to grow crops. Thus, humans were both agents of change as we ll as sufferers of change (Stahl 1996). The pollen record and phytoliths shed some light on early anthropogenic change. Humans may have cleared tropical rainforest as early as 8000 14C yr BP in the eastern lowlands (Athens and Ward 1999) where ev idence for maize cultivation suggests agriculture around 6000 14C yr BP (Bush et al. 1989). In the Andes corn and other domesticates appear after 4000 14C yr BP. In the western lowlands domestication of plants may have started at the onset of th e Holocene (Piperno and Stothert 2003). Despite the evidence for agriculture, large-scale vegetation change is not evident before 2000 yr 14C BP. It is likely that some heavily populated areas, such as the Jama Valley in Manabí province or Santa Elena Peni nsula, regional change wa s considerable. Even though humans have had the ability to change th e environmental setting around settlements for more than 10 millennia, change has been most drastic within the last 500 years, especially during the 20th Century with the growth of popula tions and improvement of technology (Dodson and Gentry 1991, Gade 1999, Hidal go Nistri 1998, Newson 1998, Wille et al. 2002). Ecuador has been throughout the Holocene a patchwork of diff erent natural and anthropogenic ecosystems that varied consider ably spatially and temporally.With current levels of deforestation this c onclusion allows for a little bi t of hope. Because most species were able to withstand great changes in the pa st, they may be fairly resistant to changes in vegetation cover. On the other hand, the sc ale of ecosystem conversion to agricultural lands today probably is of a much larger magnitude and may achieve what ENSO,


31 glacials, and volcanoes were not able to do— cause massive extincti ons in one of the world’s biodiversity hotspots.


32 CHAPTER 2 COMPARATIVE OSTEOLOGY OF ECUADORIAN BIRDS FROM ARCHAEOLOGICAL DEPOSITS Introduction Recent studies of botanical and faunal re mains from archaeological sites have produced promising results regarding the orig ins of and agriculture and domesticated animals in South America (P iperno and Pearsall 1998, Stahl 2004). Despite their potential for paleoenvironmental reconstructions with biogeographic implications, many of these studies lack the taxonomic rigor needed to accu rately describe the flora and fauna from archaeological contexts. The dearth of species -level, and even genus-level, identifications is especially notorious in birds. With a few exceptions, this taxon is usually a minor component of archaeofaunas (Chapter 6). Give n the emphasis on patterns in resource use as they relate to nutritional contributions to prehistoric people, zooarchaeologists have focused mostly on abundant fish and mammals (e.g., Byrd 1976). Birds pose an especially difficult challenge for zooarchaeo logists. By selecting for aerodynamic and light bodies, the ability to fly imposes lim itations on the evolution of novel structures. High levels of conservation in the general body plan and bone structur e thus characterize most birds, and identificati ons require in-depth studies of avian osteology. Given the large time investment required for the identi fication of birds and their generally minor contribution to archaeofaunas, zooarchaeologist s often forgo a detailed analysis of avian remains.


33 The fairly recent attempts to use archaeofaunas for paleoenvironmental reconstructions in the species-r ich tropics (Stahl 2000) highli ght the necessity of specieslevel identifications as many genera and families are highly heterogeneous in terms of habitat preferences. The great body of literature on the distribution and habitat selection in birds make this taxon especially useful in relating biogeographic patterns to past climatic/environmental variables. The pale ontological contributi ons by Campbell (1976, 1979, 1982) are good examples of the usefulne ss of birds for paleoenvironmental reconstruction in the Neotropics. Two zooa rchaeological studies with species-level identifications from Mexico and Bolivia, re spectively, have provided a great wealth of data for the reconstruction of past (anthr opogenic) environments, suggesting that most birds were obtained within a 20 km radius of human settlements, but that other species were traded over considerable distance s (Kent et al. 1999, Steadman et al. 2003). Stahl (2003, Table 1) compiled an impressive list of all vertebrates identified to date from archaeological sites spanning th e Formative Periods of Ecuador (ca. 5500– 2500 yr BP). Among the 22 identified bird taxa , only three have been assigned to species and an additional 10 to genera. Faunal lists for fish and mammals are much more extensive. My study of avian bones from se ven archaeological sites: OGSE-80 (Chapter 3), El Azúcar (Chapter 4), Loma Alta (Cha pter 4), La Ponga (Cha pter 4), Real Alto (Chapter 4), La Chimba (Chapter 5) is a first step to fill the void for prehistoric birds in Ecuador. I provide detailed descriptions and measurements as part of a comprehensive osteological analysis of all sp ecimens recovered from these site s. It is my hope that these descriptions will be helpful for zooarchaeo logists and paleontologists working in the hyper-diverse Neotropics.


34 Methods Details on the archaeologi cal sites and methods of excavation can be found in chapters 3, 4, and 5. Each of these chapters al so has maps of each site. The identification of the bird bones is based mo stly on a qualitative approach. Where possible, I identified traits unique to the different families, genera, and/or species. The traits were selected based on their utility to discriminate be tween taxa and do not necessarily reflect phylogenetic relationships. For each family I describe the general osteology followed by a section on comparative osteology for local spec ies or genera. Because of high levels of conservation in the passerine osteological bauplan, I treat the general and comparative osteology for the Passeriformes at the orderlevel rather than the family-level. Unless otherwise noted, osteological terminology follows that of Baumel et al. (1993). The extirpation of two migratory species, Anas cyanoptera (Cinnamon Teal) and Fulica americana (American Coot), in historic ti mes (Ridgely and Greenfield 2001a) suggests that other species might have ha d a past non-resident winter population in Ecuador. I thus studied materi al from species currently not found in Ecuador, including several migratory ducks (Anatidae) and songbirds (Passeriformes). For some taxa I took morphometric measur ements of the bones (after von den Driesch 1976, Steadman 1980). I used principl e components analysis (PCA) to look for patterns in overall morphology as a means to separate difficult groups. Analyses were performed on XLStat add-in for Microsoft Excell software. Bones from archaeological contexts were compared to modern museum specimens from the Florida Museum of Natural Hist ory (FLMNH). The archaeological specimens are permanently housed at the Environmen tal Archaeology Division of the Florida Museum of Natural History.


35 Results The organization of this section reflects my interest to make this study in comparative osteology more accesible to zo oarchaeologists and paleontologists. Following Woolfenden (1961), I trea t each skeletal element separa tely within each family (order in the case of the Passeriformes). Th is approach makes it easier to find specific information regarding the element under st udy and should decrease the amount of time otherwise spent on searching for relevant in formation scattered th roughout the entire document. I divide the section for each el ement into two parts on general osteology and comparative osteology. For each element I provi de the name of each species identified, the name(s) of the archaeological site(s) from which it was recovered, and the number of specimens from each site. Annotated faunal lists and details on individual bone specimens are found in Chapters 3–6. System atic sequence and nomenclature follow Ridgely and Greenfield (2001a,b). In accord ance with recent molecular evidence (Chesser 2004), I place the genus Tityra in the Cotingidae. Identifications Order TINAMIFORMES Family TINAMIDAE Only three of the genera (four specime ns) found in Ecuador were available for comparison, including Nothocercus nigricapillus , Crypturellus soui , C. cinnamomeus , and Nothoprocta cinerascens . The genus Tinamus was not available, but is readily identifiable by its large size.


36 BASIS CRANII EXTERNA General Osteology Basis cranii externa of the Tinamidae is ch aracterized by a broad os supraoccipitale and a ventrad orientation of a broad processu s paroccipitalis. Foramina venae occipitalis externae are located dorsad. A deep sulcus orig inates from the foramina and runs ventrad, ending as a short canal just above forame n magnum. Fossa subcondylaris is deep and lamina parasphenoidalis is sl ightly longer than wide, endi ng rostrally in a triangularly shaped area. Four foramina, tw o at the rostral and two at th e caudal end of lamina, are the corner points of an almost perfect imagin ary rectangle. Rostrum parasphenoidale bears two processi basipterygoideae nearly at a right angle to axis rostrocaudalis. Comparative Osteology Only the width of foramen magnum, width of lamina parasphenoidalis, width of basis rostris parashenoidales, and shape of processus basipterygoideus are needed to separate the three available genera. Nothoprocta cf. curvirostris – La Chimba (1): The archaeological specimen is slightly smaller than that of N. cinarescens and similar in size to that of N. nigricapillus . Although none of the reference skeletons coincide s well, five traits suggest that the La Chimba specimen belongs to the genus Nothoprocta . Foramen magnum is wide (narrow in N. nigricapillus , very wide in C. cinnamomeus ). Canalis nervorum hypoglossi is large (large in N. nigricapillus and small in C. cinnamomeus ). The rostral end of basis rostri parasphenoidali has the shape of a narrow triangle (wide in C. cinnamomeus ). Ala parasphenoidalis is expanded rostrally (a ventral proj ection is found in N. nigricapillus whereas no enlarged area exists in C. cinnmomeus ). Processus basipterygoideus


37 originates dorsad (dorsad in N. nigricapillus and ventrad in C. cinnamomeus ). The identification to species is tentat ive and is based on the evidence for N. curvirostris by the rostrum maxillae (see below). OSSA MAXILLAE ET PALATI General Osteology The bill of the Tinamidae has a stout, but variably long rostrum maxillae, a very flat ventral surface with low crista tomialis , and a long and caudally flattened processus frontalis of os premaxillare. Comparative Osteology The length of rostrum maxillare, width of the ventral surface of os premaxillare, and degree of curvature allow for a good separa tion of the genera and species examined. Nothoprocta cf . curvirostris – La Chimba (1): Similar to N. cinerascens , the rostrum maxillae with a partial processus frontalis of os prem axillare, is longer and wider than that found in N. nigricapillus , C. cinnamomea , Nothura maculosa , and Eudromia elegans . This long bone is also notably decu rved. In Ecuador only two species of Nothoprocta are known, N. curvirostris and N. pentlandii (Ridgely and Greenfield 2001a,b). In the north only N. curvirostris has been observed. Ossa maxillae et palati of this fairly long decurved bill is suggestive of N. curvirostris (bill decurved, but shorter and slender in N. pentlandii ). In absence of comparative ma terial, I assume that most of the specimens assignable to medium-sized Nothoprocta belong to this species. Distinctly smaller and qualitatively different specimens belonging to the genus Nothoprocta are probably N. pentlandii (see bellow)


38 STERNUM General Osteology The sternum of the Tinamidae has several sa lient characteristics. It is long and slender with deep incisurae lateralis that come to within 10-15% of the length of the bone to the labrum internum. Trabecula latera lis is long, thin, and rod-like. Processus craniolateralis is long and gives the bone a horned app earance. Sulcus articularis coracoideus is broad and bordered laterally by a deep foramen. Spina interna is of variable width and spina extern a is absent. Sulcus carinae is often deep and apex carinae is set back, being more caudal than spina externa in lateral aspect. Comparative Osteology Most of the differences in bone morphology can be found in the cranial end of the bone, specifically in spina inte rna and sulcus carinae. Diffe rences in general poportions are also useful. Nothoprocta small (cf. N. pentlandii ) – La Chimba (2): Nothoprocta has a broad spina interna with three ventral ridges that fo rm two triangles. Due to these ridges, the central part of spina inte rna protrudes ventrad a l ittle further than the rest of this structure. Sulcus carinae is very deep dorsa lly where it lacks a medial ridge. Crypturellus has a narrow spina interna with a single ridge in th e middle. The spina interna appears notched from a dorsal view. Sulcus carinae is shallow and a medial ridge comes close to the labrum externum. Nothocercus is similar to Crypturellus in that spina interna has a notched appearance and sulcus carinae is sh allow with the medial ridge coming close to the labrum externum, but differs in that th e spina interna is broad and lacks a ventral ridge. Instead of the ridge a flat tuberos ity can be found. Although the cranial fragments


39 of sternum share all the aforem entioned characteristics with Nothocercus nigrocapillus and Nothoprocta cinerescens , overall proportions are sli ghtly different. Specifically, sulcus articularis coracoideus and labrum in ternum are narrow, and the distance between the most caudad point of labrum internum and the most craniad margin of incisura lateralis is reduced. Sulcus carinae is not deep dorsally, forming no shallow depression between labra externa. The distance between th e labra externa is larg e and sulcus carinae therefore does not have the semi-closed app earance (although present in one specimen) of N. nigrocapillus and N. cinerescens . The La Chimba specimens share the following traits with Nothoprocta , but not Nothocercus . 1) Spina interna has a pr ominent central ridge in on ventral aspect. 2) Labrum internum is narrow (wide in Nothocercus ). 3) Apex carinae is narrow, clearly separated from pila carinae (apex carinae vitually absent in Nothocercus in which the cranial margin of pila carinae forms the crania l margin of carna sterni). 4) Although broken, a cr anio-lateral orientation of pr ocessus craniolateralis at a ca. 45 û angle to axis rostrocaudalis is suggest ed by the base of the process. In Nothocercus , processus craniolateralis is almost at a 90 û angle to axis rostrocaudalis. Nothoprocta medium-sized (cf. N. curvirostris ) – La Chimba (7): The seven cranial fragments are larger than those of the small tinamou as evidenced in particular by a lower degree of curvature of labrum externum. SCAPULA General Osteology The Tinamidae have a characteristic sca pula with a broad, half-circular facies articularis humeralis (at a 30 û angle to axis proximodorsalis in dorsal aspect), a strong acromion with a small, stout or no medial pro cess, a thick collum scapulae, and in most


40 species one or two small to large pneumatic foramina on the lateral aspect of the proximal end at the height of facies articularis humeralis. Comparative Osteology Size, the shape of acromion, and the size of pneumatic formamina are good traits to separate the Ecuadorian genera. Crypturellus cf. transfasciatus – El Azúcar (1): The proximal end compares to Crypturellus cinnamomeus by a narrow and stubby acromion, resulting in a convex surface between facies articularis humeralis and acromion. Large formanina on the lateral aspect can also be found in Crypturellus (as well as Nothoprocta ). Nothoprocta small (cf. N. pentlandii ) – La Chimba (3): All three proximal ends have the acromion broken to some extent. A broad acromion, however, is suggested by a concave surface dorsal to facies articularis humeralis leading up to acromion. The three specimens have small pneumatic foramina (very small in N. nigricapillus and large in N. cinarescens ) and a narrow medial projection of facies articularis humeralis (most noticeable in medial aspect; narrow in N. cinarescens and Crypturellus , and broad in N. nigrocapillus ). Nothoprocta medium-sized (cf. N. curvirostris ) – La Chimba (4): The two complete and two proximal ends of scapulae are similar to the small species of tinamou. Because these specimens differ from the small ones not only in overall size, but also in the size of the pneumatic foramina (being mu ch larger) and in latero-medial thickness (being proportionately thinner), I assume that they represent a different species rather than members of the smaller sex.


41 CORACOID General Osteology The coracoid of the Tinamidae is eas ily identified by a large proximo-dorsal foramen just proximal to facies articular is scapularis, a very short processus procoracoideus, and a squared extremitas omalis coracoidei with a nearly perpendicular processus acrocoracoideus in relation to axis proximodistalis. Comparative Osteology Phylogenetically informative characters have been identified by Bertelli and Chiappe (2005), including the degree of deve lopment of processus lateralis, the position (and depth) of impressio ligamenti acrocoracohu merale in relation to facies articularis clavicularis, presence/absence of impressio musculi biceps brachii on ventral surface of acrocoracoid process, development of processus acrocoracoideus, presence/absence of “medial expansion” of processus acrocoracoideu s, and the shape of the medial edge of processus procoracoideus. Crypturellus cf. transfasciatus – El Azúcar (3): Two distal ends of coracoids are from Crypturellus based on size ( Tinamus much larger), a wide dorsal foramen (narrower in Nothocercus and Nothoprocta ), a large dorsal foramen locat ed on the distal surface of facies articularis scapularis (small or absent in Nothoprocta ), a deep impressio ligamenti acrocoracohumeralis (shallow in Nothocercus and Nothoprocta ), and an overarching, ventrad oriented tuberculum ventra le (only slightly overarching in Nothocercus and intermediate levels of overaching in Nothoprocta ). As comparative material was not available, I assign these specimens tent atively to the locally fairly common C. transfasciatus .


42 Nothoprocta small (cf. N. pentlandii ) – La Chimba (12): The archaeological specimens have a long proce ssus lateralis (short in Crypturellus and Nothocercus ), presence of foramina on impressio musculi st ernocoracoidei, small (or absent) foramina on the distal side of facies ar ticularis scapular is (large in Crypturellus and Nothocercus ; as this trait is not visibl e on reference specimens of Nothocercus , I rely on Bertelli and Chiappe [2005] for a description of this ch aracter), a well developed medial edge of processus procoracoideus pr ojecting as a short and stout crest (absent in Nothocercus ), and a shallow impressio ligamentum acrocoracohumeralis (deep in Crypturellus ). These smaller coracoids differ from the larger ones in having processus lateralis at the same level as incisura nervosa supracoracoidei (whe re visible in the larger specimens it tends to be more proximal on the bone as incisura nervosa supracoracoid ei; see Bertelli and Chiappe 2005). In addition, the shaft tends to be slender and the dorsal foramen wide, bearing one or two small foramina (narrow and with one large foramen in the larger specimens). Nothoprocta medium-sized (cf. N. curvirostris ) – La Chimba (12): By the above reasoning I assign the 12 medium-s ized specimens tentatively to N. curvirostris . Three specimens are unusual, however, in havi ng a broader processus acrocoracoideus. Processus acrocoracoideus shows sexual differen ces in some species, most notably within the Anatidae (personal observat ion); I therefore assu me that these three specimens reflect sexual dimorphism in N. curvirostris .


43 HUMERUS General Osteology The humerus of the Tinamidae has a thic k head, a well developed pneumatic fossa with a small fossa, a broad impression coracobr achiale, usually a slender or crescent-like impressio musculi brachiale, a long condylus ve ntralis (as long or longer than condylus dorsale when comparing the length of each along their respective main axes) and a thick, nearly round condylus vent ralis (personal observation; Bertelli and Chiappe 2005). Despite a superficial resemblance with that in the Galliformes (for example, Penelope montagnii ), the Tinamidae differ noticeably in havi ng a pointy crista bicipitale (due to a stubby tuberosity serving as attachment site of musculus scapulohumeralis caudalis), a small pneumatic foramen, a broad and flat fo ssa pneumotricipitalis (small and deep in Penelope ), a shallow fossa musculi brachialis, and a wide and raised tuberculum supracondylare ventrale. Comparative Osteology Phylogenetically informative traits in the tinamid humerus are a tuberosity obstructing the proximal end of incisura capitis, the extent of opening of pneumatic fossa, shape of impressio musculi brachialis, number of fossae on the dorsal aspect of the distal end between processus supr acondylaris dorsalis and epic ondylus dorsalis, and distal extent of processus flexorius (bas ed on Bertelli and Chiappe 2005). Crypturellus cf. transfasciatus – El Azúcar (2): A complete hu merus and distal end of humerus are too small for Tinamus . Both specimens can be assigned to Crypturellus based on the presence of a tuberosity at th e distal end incisura capitis (absent in Nothocercus and Nothoprocta ), a small pneumatic fossa (large in Nothocercus and


44 Nothoprocta ), a narrow, crescent-shaped impre ssio musculi brachia lis (broad and crescent-shaped in Nothocercus and oval in Nothoprocta ), the presence of three fossae on the dorsal aspect of the distal end (two in Nothoprocta ), and a short processus flexorius that does not extend as much distad as condylus ventralis. Nothoprocta small (cf. N. pentlandii ) – La Chimba (3): The three fragments belong to Nothoprocta based on the above characteristics. Compared to the other La Chimba humeri, they are considerably smaller, mo re slender, and have a narrow impressio musculi brachialis. Only in one specimen is condylus ventralis preserved. It is narrower than that found in the medium-sized Nothoprocta from La Chimba. More modern skeletal material is needed to see if this trai t occurs consistently in this species. Nothoprocta medium-sized (cf. N. curvirostris ) – La Chimba (7): Nothoprocta by the above traits, an oval impressio musculi br achialis and/or large size suggest that the seven specimens belong to a medium-sized Nothoprocta . Unfortunately, Bertelli and Chiappe (2005, p. 20) were unable to define a character state for N. perdicaria , possibly the sister taxon of N. curvirostris (Bertelli and Porzecanski 2004), which could have improved the resolution of this identification. ULNA General Osteology The ulna is characterized by a deep and dorso-ventrally flattened olecranon, a well developed attachment of musculus anconeus (George and Berger 1966), a curved and flat shaft, a wide distal end with a bulky tube rculum ventrale, and a low, poorly defined condylus ventralis et dorsalis.


45 Comparative Osteology The shape of olecranon, development of tuberculum ligamentosum collateralis ventralis, degree of curvature of the shaft, shape of tubercul um carpale, and the length of condylus dorsalis ulnaris in cran ial aspect are helpful traits. Crypturellus cf. transfasciatus – El Azúcar (1): The most distinctive character of this proximal end is the shape of olecranon in dorsal aspect, extending more caudad than that of the other genera and gently curving distad. In Nothocercus the olecranon is shorter, having two caudal angular bends. Nothoprocta also has a short olecranon, but only one distal angular bend. Nothoprocta small (cf. N. pentlandii ) – La Chimba (5): Four complete ulnae and distal end of ulna share several traits with Nothoprocta . Olecranon is angular in dorsal view (with only one distinct angle; se e above). Tuberculum ligamenti collateralis ventralis is raised (low in Nothocercus and Crypturellus ). The dorsal margin of depressio radialis is a very high ridge (lower in the other genera). Depressio radialis is small (intermediate in Nothocercus and large in Crypturellus ). Tuberculum carpalis is long with its tip oriented distad, thus leaving a deep furrow between tuberculum carpalis and condylus ventralis ulnaris (similar in Nothocercus , rounder and oriented more craniad in Crypturellus ). Condylus dorsalis ulnaris is s hort in ventral view (short in Nothocercus and long in Crypturellus ). Nothoprocta medium-size (cf. N. curvirostris ) – La Chimba (10): Four complete ulnae and two proximal ends share the same traits as the small Nothoprocta . They differ in being longer, having a larger depressi o radialis, and possessing a proportionately thicker shaft.


46 RADIUS General Osteology The radius has an oval cotyla humeralis, a shallow tuberculum bi cipitalis radialis that is partially located on a triangular tuberosity (in caudal aspect; I will refer to this tuberosity as “proximo-caudal tuberosity” he nceforth), a round sli ghtly distad curving capital tuberosity (Howard 1929), a triangular cr oss-section of mid-sh aft, and a slightly expanded distal end with a nearly straight facies ar ticularis radiocarpalis. Comparative Osteology The capital tuberosity, the pr oximo-caudal tuberosity, the degree of curvature of the shaft, a small depression on the ventral aspe ct of the distal end, and the degree of curvature of facies articularis radiocar palis are good traits to distinguish among individuals from three out of four Ecuadorian genera. Nothoprocta small (cf. N. pentlandii ) – La Chimba (2): As in the reference specimen of N. cinerascens (UF 38951), the two proximal ends have a broad capital tuberosity (reduced to small tuberosities in Nothocercus and Crypturellus ) and a short, rounded proximo-caudal tuberosity (pointy in Crypturellus ). Nothoprocta medium-sized (cf. N. curvirostris ) – La Chimba (3): The three specimens are slightly larger than the above and the two proximal ends agree well with Nothoprocta in all characters. The specimen of a dist al end reveals that the shaft is only slightly curved (strongly curved in Crypturellus ), a small depression can be found on the ventral aspect of the distal radius (not found in Nothocercus and Crypturellus ), and facies articularis radiocarpalis has a small caudal bend (straight in Crypturellus ).


47 CARPOMETACARPUS General Osteology The carpometacarpus is a stout bone with a very deep attachment for ligamentum ulnocarpometacarpalis ventralis and a proximad th ickened ventral rim of trochlea carpalis diagnostic for the Tinamidae. Processus extensorius is small and knobby, os metacarpi minus is slender curving dorsad, facies articu laris digitalis major and facies articularis minor are separated distally by a deep notch that leaves only a sl ender ventral bony ridge connecting the two articular surfaces. Comparative Osteology Although characteristic for the Tinamidae, within this family, carpometacarpus provides few distinguishing features. The degr ee of thickening on the proximal portion of the ventral rim of trochlea carpalis, the dept h of the attachment for ligamentum ulnocarpometacarpalis, the loca tion of a small proximal tube rosity on the insertion of musculus interosseus dorsalis, and the developm ent of facies articula ris digitalis majus et minus are uselful to some extent. Crypturellus cf. transfasciatus – El Azúcar (4), Loma Alta (2): In all complete carpometacarpi and proximal ends (four specime ns), a proximal thickening of the ventral rim of trochlea carpalisa is absent as in Crypturellus , whereas it is present in Nothocercus and Nothoprocta . The attachment for ligamentum ulno carpometacarpalis is deep in both Crypturellus and Nothoprocta , but shallow in Nothocercus . Similarly, Nothoprocta and Crypturellus share nearly the same location for the small proximal tuberosity (just distal of symphysis proximalis). In Nothocercus this tuberosity is locat ed more distad. The two


48 distal ends and the complete carpometacarpi ha ve narrow facies articularis digitalis major et minor, as found in Crypturellus . Nothoprocta small (cf. N. pentlandii ). – La Chimba (3): By the above characteristics the two sm all carpometacarpi belong to Nothoprocta . Because of their small size I place them in N. pentlandii . Nothoprocta medium-sized (cf. N. curvirostris ) – La Chimba (6): Nothoprocta by the traits mentioned above, it is slightly larger than the previous three elements. I tentatively identify these specimens as N. curvirostris . OS COXAE ET SYNSACRUM General Osteology The very characteristic pelvic girdle a nd synsacrum stand out by the absence of a lamina infracristalis ilii, a broad and distally widening ala ischii, a large foramen obturatum, a flat and long tuberculum preacet abulare, a non-perforate foramen acetabuli, and a broad synsacrum with a pair of long proc essus transversus that reinforces the pelvic girdle by attaching to the ischium. Comparative Osteology The shape of ala ischii, the shape and orientation of tuberculum preacetabulare, and the degree of fusion between ischium a nd pubis are taxonomically informative. Nothoprocta sp. – La Chimba (2): The round ala aschii (squared in Nothocercus and Crypturellus ) and a closed angle between th e long tuberculum preacetabulare (broken in one specimen) and ala preacetabular is ilii (short in at a larger angle in Nothocercus and Crypturellus ) are the only two traits obser vable in the archaeological specimens. They are sufficient, however, to identify them as Nothoprocta . As the two


49 fragments are of similar size and as I cannot co mpare them to modern skeletal material of N. pentlandii and N. curvirostris , I cannot assess any differen ces in size. Only a generic identification is therefore possible. FEMUR General Osteology The femur is a highly curved bone with high, cranially protruding crista trochanterica, a highly curved facies articu laris antitrochanterica th at protrudes caudally to form an overhanging ledge, a flat condylus medialis, a wide, circular impressio ansae musculus iliofibularis, and a triangular caudal facet of condylus medialis. Comparative Osteology Apart from differences in general proporti ons and size, a few traits are of limited utility for the identification of the archaeol ogical specimens. They are the shape and size of femoral head, including the degree of curvat ure of facies articula ris antitrochanterica, orientation of the caudal ledge formed by faci es articularis antitrochanterica, the shape and location of impressio ansae musculus ilio fibularis, and the shape and proximal extent of condylus medialis et lateralis in caudal aspect. Crypturellus cf. transfasciatus – El Azúcar (1), Loma Alta (1): Having proportionately more slender and longer heads (as found in Crypturellus ) than the femora of Nothocercus and Nothoprocta , I assign the two proximal ends tentatively to C. transfasciatus . Nothoprocta small (cf. N. pentlandii ) – La Chimba (8): Although the head of femur is more slender than that of N. cinarescens , the proximal ends agree well with those in Nothoprocta in having a more steeply sloped facies articularis antitrochanterica in cranial


50 aspect (intermediate in Nothocercus , low in Crypturellus ), and a long crista trochanterica (short in Nothocercus and Crypturellus ). The distal end coin cides with that of Nothoprocta in a concave oval impressio ansae musc ulus iliofibularis (round and concave in Nothocercus and oval and flat in Crypturellus ). The specimens assigned to this species are smaller and more slender compared to the below. Nothoprocta medium-sized (cf. N. curvirostris ) – La Chimba (6): The archaeological specimens are similar, but larger than the above. TIBIOTARSUS General Osteology The long tibiotarsus of the Tinamidae has several salient features. On the proximal end, tuberositas popliteus is a high and strong ridge, and cr ista cnemialis cranialis is reduced. On the distal end, sulcus extensoriu s runs parallel and close to the medial margin. It maintains its direction when passi ng through canalis extens orius such that the canal opens distad right above the main axis of condylus medialis. Comparative Osteology Bertelli and Chiappe (2005) have iden tified 15 phylogenetically informative characters of which six (Characters 49, 51, 53, 54, 58, and 61) are useful for the present study (Table 1). They are the proximal extent of crista cnemialis cranialis (0: greatly extended beyond articular surface, 1: slightly beyond articular surface; TiTi1), relative widths of condylus lateralis and condylus medi alis (0: lateral condyale broader, 1: nearly equal; TiTi2), relative lengths of condylus lateralis and condylus medialis (0: nearly equal, 1: slightly shorter, 2: much shorte r; TiTi3), shape and de pth of medial condylar depression (0: reduced to poorly developed, 1: deep pit on crania l margin, 2: groove


51 along caudal margin bordered by a crest; TiTi4), degree of de velopment of a ridge along the proximal end of incisura intercotylar is (0: not developed, 1: incomplete, not continuous medially, 2: complete and conti nuous across supratendina l bridge; TiTi5), and relative position of pons suprat endineus with respect to the distal end of condylus lateralis in cranial aspect (0 : distal, 1: proximal; TiTi6). Crypturellus cf. transfasciatus – El Azúcar (1): Table 1 suggests that the distal end of tibiotarsus is from a species of the genus Crypturellus , although the rela tive lengths of condylus medialis and condylus lateralis are variable within this genus (Bertelli and Chiappe 2005, Appendix 3). Given the lack of comparative material I assign this specimen tentatively to C. transfasciatus . Table 2-1. Character matrix for the tibiotarsus in the Family Tinamidae. The characters in this table correspond to characters 49, 51, 53, 54, 58, and 61, respectively, of Bertelli and Chiappe (2005). See text for definition of characters. Characters (TiTi) Species 1 2 3 4 5 6 Nothocercus nigrocapillus 1 0 2 1 2 0 Crypturellus cinnamomeus 1 0 1 1 1 1 Nothoprocta cinarescens 0 0 1 2 0 1 Nothoprocta pentlandii* 0 0 1 2 0 1 cf. Nothoprocta pentlandii (LC) 0 0 1 2 0 1 cf. Nothoprocta curvirostris (LC) 0 0 1 2 0 1 Crypturellus cf. transfasciatus (EA) 0 2 1 1 1 * Character scores for N. pentlandii are from Bertelli and Chiappe (2005).


52 Nothoprocta small (cf. N. pentlandii ) – La Chimba (5): Small size and a low degree of curvature of the medio-distal margin just proximal to condylus medialis distinguish this species from N . cf. curvirostris (high degree of curv ature; Table 2-1). Nothoprocta medium-sized (cf. N. curvirostris ) – La Chimba (10): Regarding the relative lengths of the tibiotarsal condyles, th is species appears to differ from its possible sister taxon N. perdicaria (Bertelli and Porzecanski 2004, Bertelli and Chiappe 2005). TARSOMETATARSUS General Osteology In the Tinamidae this bone is characterized by an eminentia intercotylaris that extends proximally beyond the hypotarsus (B ertelli and Chiappe 2005), a single, deep sulcus hypotarsi formed by a large crista medi alis hypotarsi and a small crista lateralis hypotarsi, a slightly concave sulcus extensorius, a distally flattening shaft, and trochlea metatarsi II extending more distad than trochl ea metacarpi IV. (This la tter trait is shared by several orders including th e Procellariiformes, Cicon iiformes, Anseriformes, and Gruiformes which differ in the shape of troc hlea metacarpi and thic kness of the shaft). Comparative Osteology The main difference in the tinamid tarsometatarsus lie in overall proportions, the shape and distal extension of crista lateralis et medialis hypotarsi relative to foramina vascularia proximalia (Bertelli and Chiappe 2005 ), and the depth of sulcus extensorius. Tinamus cf. major – OGSE-80 (1): The general osteological traits identify this specimen as tinamou. Although superficially simi lar, Procellariiformes and Gruiformes have a thicker distal shaft. The large size suggests that the specimen comes from a member of the genus Tinamus .


53 Nothoprocta small (cf. N. pentlandii ) – La Chimba (5): The complete tarsometatarsi and one distal end approach the short and stout tarsometatarsus of N. cinarescens in overall proportions. Nothocercus and Crypturellus have long and slender tarsometatarsi. The acuminate crista lateralis et medialis hypotarsi (i n plantar aspect) end distal to foramina vascularia proximalia, a condition common to the Tinamidae except in Nothocercus . Sulcus extensorius is deep (shallow only in Nothocercus ). Nothoprocta medium-sized (cf. N. curvirostris ) – La Chimba (8): All specimens are very similar to cf. N. pentlandii and differ mainly in being larger. Order PROCELLARIIFORMES Family PROCELLARIIDAE CORACOID General Osteology The procellariid coraco id is a short bone with a wide and flat sternal end, presence of fenestra nervus supracoracoidei, and a fl at, angular processus acr ocoracoideus. Facies articularis sternalis is long and processus lateralis stout and pointy. Sulcus musculi supracoracoidei often bears a longitudinal groove. Processu s acrocoracoideus projects ventrad at an angle of ca. 30 û in relation to the main shaft. Comparative Osteology Size and overall proportions are important ch aracters in this family. The shaft is relatively wide due to a wide proximal extens ion of processus procoracoideus. The shape of processus acrocoracoideus, as well as th e location and depth of the groove in sulcus musculi supracoracoidei, also help to distinguish species.


54 Puffinus griseus – Valdivia Village (2): The incomplete right and proximal end of left coracoids are from a medium-sized pr ocellariid. It is similar to that in P. griseus in overall size and width of the shaft. The sh aft is proportionately wider in the small Daption capense and large Procellaria aequinoctialis . Being slightly smaller, Puffinus bulleri also has a slender shaft. Pterodroma inexpectata has a narrow shaft proximally that becomes wide distally. As in the modern specimens of P . buller i and P. griseus , the narrow groove in sulcus musculi supracoracoideus originates centrally on the shaft at its proximal end and runs obliquely to end dista lly on the dorsal margin of sulcus musculi supracoracoideus. Whereas in Daption the groove is absent, it is wide and shallow, running along the ventral margin of sulcus, in Procellaria , and wide and deep, and entirely located centrally in Pterodroma . HUMERUS General Osteology The humerus in the Procellariiformes is superficially similar to that of the Charadriiformes with a deep fossa pneumotricipitalis and a long, flat processus supracondylaris dorsalis. A long tuberculum ve ntralis, a round fossa pneumotricipitalis, and a flat distal end with a pronounced epic ondylus ventralis readil y distinguish the two orders. Comparative Osteology On the proximal end, informative charac ters include caput humeri (bulky in Puffinus and Procellaria ), crista deltopectoralis (knob-like tip in Puffinus , pointy tip in Pterofroma and Procellaria ), extent of fossa pneumotricip italis (undercuts caput humeris in P. bulleri , Pterodroma inexpectata , and Daption capense ), and impressio


55 coracobrachialis (deep in Puffinus ). The distal end can be di fferentiated by the shape of fossa musculi brachialis, shape of condylus vent ralis et dorsalis, thickness of epicondylus ventralis. Puffinus griseus – La Ponga (1): The distal end of humerus can be assigned to P. griseus . 1) Fossa musculi brachialis is round a nd shallow (oval to oblong in the other genera). 2) Condylus dorsalis is flattened cr anially in distal aspect (more rounded in Puffinos bulleri and Procellaria aequinoctialis , more pointy in Pterodroma inexpectata and Daption capense ). 3) Epicondylus ventralis is flat with the attachment sites for musculus flexor carpi ulnari s being oriented distally (m ore bulky with a more ventral orientation in the other genera). PHALANX 1 DIGITUS MAJOR General Osteology Phalanx 1 of digitus major is a long a nd slender bone in th e Procellariiformes similar to that of other long-winged orde rs. It differs by having a broad and round concave caudal surface on the proximal pila cranialis which serves as the passage of musculus interosseus dorsalis (sensu Zusi and Bentz 1978). The blade is located ventrally on the same plane as the vent ral margin of the shaft. Comparative Osteology Size, overall proportions, the shape of the caudo-proximal margin of pila cranialis, and location and thickness of an oblique, thic kened bar on the flat blade can be used to distinguish among genera. Some genera may also show fenestra on the flat blade. Puffinus griseus – Valdivia Village (1): By size and overall proportions ( P. inexpectata is smaller and shorter, P. bulleri smaller, D. capense is smaller with a


56 tapering proximal end due to the narrowing of the caudal blade, P. aequinoctialis is larger) the proximal end of this phalanx corresponds to P. griseus . In addition, the concave dorsal surface of the proximal pila cranialis is wide and deep (shallow and narrower in P. inexpectata and D. capense , shallower in P. bulleri , and similar in P. aequanoctialis ) and the oblique bar on the distal blad e is more thickened distally than proximally (more evenly thickened in the other genera). Order PELECANIFORMES Family PHALACROCORACIDAE OS PREMAXILLARE (+ OS NASALE) General Osteology The os premaxillare of the Pelecaniforme s is characterized by impervious nari (Sibley and Ahlquist 1990). In the Phalacrocoracidae the nari are very small, and located just rostrally to zona flexoria craniofascia les. Septum nasi osseum and conchae nasales are heavily ossified (Baumel et al. 1993). Comparative Osteology Within the Pelecaniformes, the Phaethon tidae, Pelecanidae, and Anhingidae can be readily distinguished by size. In addition, the overall shape of the bill differs considerably among these families. Phalacrocorax bougainvillii – Valdivia Village (1). Os nasale lies centrally and dorsally to the rami of the premaxillary. In the Pelecanidae, Phalacrocoracidae, and Fregatidae these two bones are easily delineated as the lateral margins of os nasale meet os premaxillare at a steep angle. In the Ph aethontidae, Sulidae, and A nhingidae, there is a smooth transition between these two bones, resulti ng in an inverted Uor Vshape of the


57 upper jaw. In the Phaethontidae these two bones bones can be separated by pneumatized line caudally to the nari and at the nari. In the Sulidae the divisi on between the bones is marked by a long, slender groove. In the Anhingi dae a poorly ossified line of ankylosis marks the transition between the bones from the caudal end to the rostrum maxillae. By the size and shape of the os nasale, th is caudal fragment of the ossa maxillae belongs to the Phalacrocorax . Only in this genus is os na sale very high compared to the height of the rami. Whereas os nasale is broa d relative to the rami of os premaxillare in P. brasilianus , it is slender in P. bougainvillii . In addition, os nasale of P. bougainvilli is characterized by an area of cons triction rostrally to the nari just before widening toward zona flexoria craniofacialis. HUMERUS General Osteology The humerus of the Phalacrocoracidae is a long bone with a deep incisura capitis that undercuts caput humeri (found in all Pelecan iformes except Pelecanidae; also absent in Procellariiformes), a gr eatly reduced crista deltop ectoralis (shared with the Pelecanidae, Sulidae, and Anhingidae), and a high and bladelike epicondylus ventralis (limited to Phalacrocoracidae and Anhingida e; Cracraft 1985b). On the distal end the condyla are strongly rotated cr anially, processus supracondyl aris dorsalis is lacking (present in Procellariiformes and Charadr iiformes), and fossa musculi brachialis is shallow (deep in Procellariiformes and Charadriiformes). Comparative Osteology The Phalacrocoracidae is a fairly homoge nous family in terms of the humerus, although a variety of traits can be identifie d when comparing individual species (see


58 below). Only two species are currently known from western Ecuador, P. brasilianus and P. bougainvillii , which differ considerably in size. A third species, P. gaimardii , is considered here as well because of its relative proximity to Ecuador (ca. 660 km to the south). Because P. gaimardii was not available for comparison, I used the osteologically similar P. punctatus (Siegel-Causey 1988). Phalacrocorax bougainvillii – Valdivia Village (1): A series of traits suggests that the specimen belongs to P. bougainvillii . The specimen of P. punctatus is similar (it differs only in character 4), but smaller (J ohnsgard 1993). 1) The ventral margin of sulcus ligamenti transversus forms a little ledge in cranial aspect (not in P. brasilianum ). 2) The distal extreme of crista deltopectoralis is thickened forming a short tuberosity. 3) Fossa musculi brachialis broadens proximally whereas in P. brasilianum it narrows. 4) Processus supracondylaris dorsali s is less pronounced than in P. brasilianus . 5) Sulcus scapulotricipitalis is relativel y wide (in proportion to the width of the distal end). 6) The ligamental attachment site of tuberculum supracondylare dorsale is long and shallow in P. bougainvilli as opposed to circular and deep in P. brasilianus . RADIUS General Osteology The cormorant radius is a long, roughly S-sh aped bone in dorsal aspect. The distal margin of tuberculum bicipitale radii partially overhangs the sh aft, a trait I have only seen in Phalacrocorax . The capital tuberosity (Howard 1929) is very prominent. The crosssection of the central shaft is roughly tria ngular. The distal end has a long facies articularis radiocarpalis that ends distally in a well defined, pointy tuberosity.


59 Comparative Osteology As with the humerus, there is a c onsiderable size difference between P. bougainvillii and P. brasilianus . Tuberculum bicipitale radii, the capital tuberosity, and facies articularis ulnare are good traits to distinguis h between the two species. Phalacrocorax bougainvillii – Valdivia Village (1): The proximal end with half a shaft is P. bougainvillii by size. In addition, the overhang of tuberculum bi cipitalis radii ends distally in a pointy tuberosi ty (tuberosity more proximal in P. brasilianum ), the capital tuberosity is rounder in dorsal view (more square in P. brasilianum ), and facies articularis ulnare is longer along th e axis proximodistalis than in P. brasilianum . These three thraits are shared between P. bougainvillii and P. punctatus . CARPOMETACARPUS General Osteology The carpometacarpus of the Phalacrocoracid ae superficially resembles that of the Anhingidae, Ardeidae, and Anatidae. It is a long bone with a round trochlea carpalis, a deep fovea carpalis cranialis, a deep fovea carpalis dorsalis, and slightly curved os metacarpale minus. The procellariid carpomet acarpus is also a long bone, but differs by having a broad processus extensiorius and less excavated proximal foveas. Comparative Osteology The shape of processus extensorius and fovi a carpalis cranialis allows for a reliable separation of P. brasilianus from P. bougainvilli and P. gaimardii . Phalacrocorax brasilianus – Loma Alta (1): The proximal end of a carpometacarpus is smaller than that of P. bougainvilli from which it also differs by a narrower processus extensorius and a deeper and narrower fovea carpalis cranialis. In


60 similarly sized P. punctata it has a narrow processus extensorius, but with a more proximad orientation. Only P. brasilianus shares these traits with the specimen from the archaeological site. Order CICONIIFORMES Family ARDEIDAE SCAPULA General Osteology Superficially similar to the Threskiornithid ae and Anatidae, the ardeid scapula has a flat cranial end with a flat, slightly late rally bent acromion (in the other two families bulkier) and a slight medial curvature. A t uberculum coracoideum is present. Facies articularis humeralis is square-shaped. Comparative Osteology Variation is limited to the shape of the acromion, presence or absence of crista ligamenti acrocoracoacromali, a nd shape of facies articulari s humeralis (see below). Size limits the potential species to three genera. Tigrisoma lineatum – OGSE-80 (2): The two cranial ends from the archaeological site are heavily abrade d, but the three distinguishing traits are still visible. Acromion is slightly curved late rally, but to a much lesser degree than in Nycticorax , Egretta , and Ardea . Crista ligamenti acrocoracoacromali is most pronounced in Ardea , to a lesser degree Nycticorax and Egretta , and least in Tigrisoma . Facies articularis humeralis has a rounded medial margin in Ardea , an almost straight margin in Nycticorax , and a squarelooking margin in Tigrisoma and Egretta . Differences between T. lineatum and T. fasciatum are more subtle. Despite being nearly identical in size, the scapula of T.


61 lineatum has a bulkier acromion as seen in cranial view. The transition from the medial margin to the caudal margin of facies articularis humeralis is abrupt in T. fasciatum , giving the cranial end of the medial margin of the scapula a tria ngular look in medial view. In T. lineatum this feature is more rounded. CARPOMETACARPUS General Osteology The carpometacarpus is roughly similar to that of the Phalacrocoracidae, Threskiornithidae, and Anatidae. Pro cessus extensorius is stouter in the Phalacrocoracidae, Threskiornithidae, and Anatid ae, symphysis distalis is shorter than in the Phalacrocoracidae, the dorsal margin of trochlea carpalis meets the shaft at a steep angle distally (it converges smoothly onto the shaft in the Phalacrocoracidae and Anatidae), the bone is overall shorter (relative to width)in the Anatidae and Threskiornithidae, and the distal end of os di gitalis major is deeply excavated ventrally as os digitalis major flar es out cranially. Comparative Osteology The only heron in this size-class in Ecuador is Ardea alba , with Nycticorax , Nyctinassa , Botaurus , Tigrisoma , and Egretta being smaller and A. cocoi and A. herodias being much larger. There are also subtle differences in the shape of the processus extensorius. Ardea alba – El Azúcar (1), Loma Alta (1): The proximal ends of carpometacarpi are close in size to that in A. alba . Although the El Azúcar specimen is very large, exceeding in size a carpometacarpus from a 1200 g individual (UF 39323), it is considerably smaller than that of A. cocoi and A. herodias .


62 TIBIOTARSUS General Osteology The tibiotarsus is a long bone with short crista patellari s et cnemialis cranialis, a long crista fibularis, a vaguely S-shaped sh aft in medial view, and a medially sloping distal shaft in cranial aspect. The cranial side of the distal shaft is flat, giving it a squared appearance. Sulcus extensorius and pons supr atendineus lie on the medial half of the shaft. Condylus medialis is narrower and l onger than condylus latera lis in distal view. Comparative Osteology Significant differences in the tibiotarsus can be found in the proximal and distal ends. Tigrisoma cf. fasciatum – OGSE-80 (1): This specimen belongs to Tigrisoma based on two traits. 1) The distal shaft is not square, but rounded laterally (in Botaurus shaft is square distally, but unlike most herons gradually becomes rounded proximally). 2) Judging from the distance between crista fibul aris and the medial tuberositas retinaculi extensorius, the shaft is relatively shorter th an that of any other genus in the Ardeidae. Identification to species of this abraded shaft is tentative. The shaft is short and stout as in T. fasciatum . Despite having similar weights (8 50 g vs. 823-897 g; Dunning 1993), the distance between the foramen just distal to the crista fibularis and lateral tuberositas retinaculi extensorius is shorter in T. fasciatum than that in T. lineatum . By this measure, the archaeological specimen is slightly longer than the modern specimen of T. fasciatum (UF 43434, which at 760 g is ca. 100 g lighter than published weights for this species; Dunning 1993). In addition, the archaeologi cal specimen agrees with that in T. fasciatum in that the foramen distal to the crista fibularis is a long, slit-like opening (short in T.


63 lineatum ). The condition of the tuberositas re tinaculi extensorius, which is long in T. lineatum and short in T. fasciatum , cannot be determined because the bone is too fragmentary. TARSOMETATARSUS General Osteology The distal end of the tarsometatarsus is hi ghly diagnostic in Ardeidae. Trochlea are large and round in medial view, being delinea ted sharply from the shaft by a different texture. In the large species the foramen dist ale lies between trochlea metatarsi III and IV. With decreasing size foramen is located more proximally. Figure 2-1. The distal end of the tarsometatarsus of Ixobrychus exilis (UF 42030) in dorsal (a) and plantar (b) asp ect. Reference bar is 1 cm. Comparative Osteology Small size sets the genus Ixobrychus apart from the other genera of this family. Ixobrychus exilis – El Azúcar (1): Despite the sm all size of the distal fragment, several key features allow me to identify this specimen as I. exilis . Trochlea metacarpale III is round in medial aspect and at an angle to the shaft in cranial aspect. The distal foramen is clearly located between the proxi mal ends of trochlea metatarsi II and III (Figure 2-1). a. b.


64 ANATIDAE STERNUM General Osteology Carina sterni is less deep in the Anatidae th an in other taxa with strong flight such as the Columbidae or Apodiformes. Apex carin ae extends craniad beyond spina externa. The ventral margin of carina st erni is straight, giving it a tr iangular shape. Spina externa is a pointy and slender projecti on or absent and the spina in terna is reduced to a small ridge or absent. Labrum internum is rounde d and usually curves ventrally toward the center of the body. Processus craniolateralis is short, broad, and somewhat squared in appearance. Foramen pneumaticum is large a nd variable in shape and position. Caudally, trabecula medialis extends as far back as margo caudalis, often forming a fenestra medialis. Comparative Osteology Anas sp. – Real Alto (1): Only a small cranial fr agment was found in the sample from Real Alto. The fragment includes sulcus articularis coracoideu s, portions of labra externa and interna,and as the cranial margin of foramen pneumaticum. Spina externa is missing. This fragment is an anatid based on a reduced, ridge-like labrum internum, a slender spina interna as judge d by the breakage area, a shar ply bending labrum externum which is reduced laterally to a mere ridge , and a round foramen pneumaticum. The shape of labrum internum suggests that this bone fragment comes from a species in the genus Anas : labrum internum slopes ventrad toward the medial plane and is replaced on the cranio-dorsal margin of the sternum by a sm all ridge which may represent a vestigial spina interna. In Netta foramen pneumaticum lies in a sm all pit-like dorsal surface and


65 spina interna is replaced by a notch. Th e latter characteristic is shared by Aythya . In Oxyura labrum internum and a vestigial spina in terna are not well defined as the labrum becomes a sharp ridge medially. Sulcus articularis coracoideus is much narrower in Oxyura than in any other duck. SCAPULA General Osteology The scapula in the Anatidae is a slender a nd highly curved bone. Corpus scapulae is narrow throughout, widening only marginally towa rd the caudal end before decreasing in width again. The curvature is most pronoun ced just caudally to collum scapulae. The proximal end is flat with an oblong facies articularis humeralis, a short and round tuberculum coracoideum, and a pointy facies articularis clavicularis being the most prominent features. The scapula in the Ana tidae resembles in many ways the scapula of the Ardeidae which have a slightly bigger facies articularis hu meralis and a broader acromium. Comparative Osteology The scapula is fairly uniform in this fa mily. Despite large differences in body mass of the species examined, the size of the sca pula does not vary mu ch unless the species considered are at the small or large size extr emes. Based on present distributions (Ridgely and Greenfield 2001a,b) eight species of ducks can be found in or near the peninsula of Santa Elena. Of these, Anas discors (small) and Sarkidiornis melanotus (large) are outside of the possible size range. Sarkidiornis melanotus is also different osteologically by having a very broad facies articularis scapularis.


66 The proximal end exhibits a few traits that can be used to distinguish among different genera of this family, including the shape of acromion, and the shape of the surface between acromion and facies ar ticularis humeralis (Woolfenden 1961). Anas bahamensis – El Azúcar (1), OGSE-80 (1): The proximal end of a mediumsized scapula with a partia lly broken acromion is similar in size to that in Dendrocygna , Anas , Aythya , and Netta . Both Dendrocygna and Netta can be excluded based on a very long facies articularis humeralis. Aythya has a wide and bulky the acromion in medial aspect. Although partially broken in the speci men, the acromion has a narrower base in the archaeological specimen. In addition, the area between acromion and facies articularis humeralis is very deep in Aythya , whereas in Anas it is shallow. Aythya is also unlikely given its current distribution, having been observed only once in the Ecuadorian Andes (Ridgely and Greenfield 2001a). Within the genus Anas , A. clypeata has a long facies articularis humeralis and a small acromion. F acies articularis hume ralis is short and acromion bulky in A. bahamensis and A. georgica . Anas bahamensis can be distinguished by a more protruding medial process of acr omion (sensu Höfling and Alvarenga 2001; I will use this name interchangeably with the new term processus acromii medialis), a deep central area between acromion a nd facies articularis humeral is (almost undercut), and a more rectangular (rather than squa re) facies articularis humeralis. Anas georgica – Real Alto (1): The proximal end can be assigned to A. georgica based on a short medial process of acromi on, a shallow central area between acromion and facies articularis humeralis, and a square facies articularis humeralis.


67 CORACOID General Osteology The anatid coracoid is very characteristic with a flat and wide proximal end, a deep cotyla scapularis, and no inci sura nervum supracoracoidei. In addition, extremitas omalis coracoidei is short with a medially flatte ning neck. Tuberculum brachialis is flat and short. Comparative Osteology Characteristics of the proximo-ventral surf ace of the shaft, the shape of membrana sternocoracoclavicularis, tuberculum brachiale, and impressio ligamenti acrocoracohumeralis can be used to distinguish between sp ecies (personal observation; Woolfenden 1961). The most distinctive genus is Oxyura with a long shaft in re lation to the extremitas omalis coracoidei, a short processus acrocoraco ideus, and a practically absent tuberculum brachiale. Netta erythrophthalma and Dendrocygna bicolor are also quite distinct with a sharp attachment line for the membrana sternocoracoclavicularis that gives the processus procoracoideus a broader and flatter appearan ce. On the ventral side both genera show a shallow depression not found in any other of th e genera examined. Th e medial margin of the neck is very flat only in Anas whereas in Sarkidiornis , Netta , and Aythya this margin is wider in decreasi ng order. Curiously, A. georgica has a somewhat intermediate thickness reminiscent of Aythya rather than other Anas . Dendrocygna sp. – Loma Alta (1): The proximal end has a proximal depression on the ventral side, near facies articularis sternalis. I have seen this depression only in Dendrocygna and to a lesser degree in Netta . As D. autumnalis and D. bicolor have a


68 coracoid of roughly the same size with sim ilar proximal ends, I cannot identify this specimen to species. Figure 2-2. Dorso-medial aspect of coracoi d, showing the width of impressio ligamenti acrocoracohumerale of Anas georgica (a), archaeological specimen (cf. A. georgica ) from OGSE-80 (b), and A. bahamensis (c). Reference bar is 1 cm. Anas georgica – OGSE-80 (5), Real Alto (5): Identification of the highly fragmentary material is difficult. The proximal ends share with A. georgica a wide impressio ligamenti acrocoracohumeralis of moderate length and a marked distal curvature (Figure 2-2). In A. bahamensis and A. clypeata the impressio is narrow of moderate and great lengths, respectively. Anas acuta has a proportionately narrower and shorter impressio ligamenti acrocoracohumer alis. A principal components analysis on four traits places the archaeo logical specimens close to A. georgica (Figure 2-3). The proximal end of the coracoid is wider and deeper and has a wider facies articularis humeralis (PC1) than in A. bahamensis and A. clypeata , having a longer facies articularis humeralis (PC2) than in A. clypeata (Table 2-2). The broad, slightly curved impressio ligamentum acrocoracohumeralis of moderate length also fits the description of A. amotape Campbell 1979, an extinct species of northwestern Peru (Campbell 1979). Further studies are needed to determine if th e specimens from the archaeological sites are A. amotape , or if A. amotape is in fact synonymous with A. georgica .


69 Anas bahamensis – Real Alto (1): The distal end overlaps with both A. bahamensis in the PCA (Figure 2-3), but the width and curvature of impressio ligamenti coracohumeralis is of the A. bahamensis type, becoming wider distally and having a marked curvature. In A. clypeata impressio ligamenti corac ohumeralis is narrower and straight. -1.0 0.0 1.0 -2.0- PC1PC2 Archaeological Specimens Anas bahamensis Anas georgica Anas clypeata Figure 2-3. A principal components analysis on four variables of the distal end of coracoid, suggesting that most archaeogi cal specimens tend to be stout. Oval shapes are only drawn for visualizing pa tterns They do not re present statistical confidence intervals. Anas cf. discors – OGSE-80 (2), Real Alto (1): By size the proximal ends of coracoids could be assigned to either to A. discors or A. cyanoptera . Given mid-twentieth century reports of breeding A. discors on the Santa Elena Peninsula (Marchant 1958), I consider this species perhap s to be more likely, although A. cyanoptera has a breeding population in northwestern Peru. Netta erythrophthalma – Loma Alta (1): The subfossil is smaller than several reference specimens of N. erythrophthalma (UF 38914, UF 23524, UF 23525), but agrees well in terms of the broad lateral margin of the medial neck, dorsally oriented facies articularis humeralis, a sharp and extended attachment line for the membrana


70 sternocoracoclavicularis, and a broad and flattened lateral shaft seen in one of the specimens. Although most of these traits are also shared with Dendrocygna bicolor , the shaft of N. erythrophthalma is proportionately longer and more slender. Table 2-2. Unrotated factor loadings for f our measurements in a principal components analysis on the distal end of the coracoid in several species of Anas (see Figure 2-3). Trait* PC1 PC2 PC3 PC4 1 0.786 0.604 -0.131 0.014 2 0.915 -0.242 -0.066 0.317 3 0.900 -0.010 0.427 -0.089 4 0.897 -0.273 -0.247 -0.246 * 1: length of facies articularis humeralis along axis proximodorsalis 2: dorso-ventral width of shaft, measured from the do rsal border of f acies articularis humeralis to the ventral depression just mediad of facies articularis humeralis 3: width of facies articularis humeralis 4: depth of distal shaft (measurement taken on dorso-ventral axis) HUMERUS General Osteology The anatid humerus is a long but stout bone with a bulky and round caput humeri, a single pneumatic fossa, a strong epicondylus ventralis, a pronounced tuberculum supracondylare ventrale, and a small, sh allow impressio musculus brachiale. Comparative Osteology The humeri from the archaeological sites all belong to the genus Anas which is characterized by the absence of a marked margo caudalis (very prominent in Dendrocygna ), an open pneumatic fossa (closed in Aythya ), and an equal caudal extent of epicondylus ventralis et dorsalis (when laying the bone on a flat surface with the cranial side faced down epicondylus ventralis longer in Netta ; Woolfenden 1961). Within Anas ,


71 the size, thickness of caput humeri, depth of impressio coracobrachiali s, width of shaft, and relative position of impressio musculi br achialis are all useful characters to distinguish among especies. Anas andium – La Chimba (1): Although slightly larger, the proximal humerus and shaft share many traits with that in A. flavirostris , its closest relative. Caputi humeri is thin in caudal aspect (thick in A. bahamensis and A. georgica ) and is undercut by the broad surface occupied by musculus triceps brachii (not undercut in A. bahamensis , A. georgica , A. clypeata ). Crista bicipitalis ends distally in an angular bent (straight in the other species), and impressio musculi brachiale is short (reaching onl y the center of the shaft; longer in other genera). Anas georgica – OGSE-80 (1), Real Alto (1): A distal end can be assigned to A. georgica based on a short and broad fossa mu sculi brachialis (long and narrow in A. bahamensis ). The proximal end has a deep and narrow incisura capitis (wider and more open in A. clypeata ), a wider tuberculum vent rale (not as wide in A. clypeata ), and an oval pneumatic fossa (round in A. bahamensis ). ULNA General Osteology The ulna is a strong bone with a stou t olecranon, a deeply excavated trochlea humeralis ulnaris, a long condyl us dorsalis ulnaris, a shallow to deep depressio radialis, and a bulky and large tuberculum carpale. Comparative Osteology The ulna is a uniform bone with only mi nor differences among genera, especially Anas , Aythya , and Netta . In both Dendrocygna and Oxyura this element is long and


72 slender. Only Oxyura can be distinguished readily by si ze. The strength of olecranon, the shape of cotyla dorsalis, the location, size, and shape of tuberculum carpale, and the shape of cotyla dorsalis ulnaris are th e taxonomically most useful characters. Anas georgica – OGSE-80 (1), Real Alto (1): The proximal end has a stout and rounded (but slightly asymmetric ) olecranon in cranial aspect and a broad cotyla dorsalis. It shares these two traits with A. georgica . In A. bahamensis olecranon is more slender and pointy and in A. clypeata it is very bulky and rounde d (symmetric). In addition, cotyla dorsalis is wide in A. clypeata and narrow in Netta erythrophthalma . The distal end agrees with A. georgica in having a stout tuberculum carpale and a long condylus dorsalis ulnaris in ventral aspect. Anas bahamensis differs in having a shorter condylus dorsalis ulnaris, and A. clypeata and N. erythrophthalma in having a more slender tuberculum carpale and a less extensive condyl us dorsalis ulnaris (b roader, flaring out caudally in A. georgica ). RADIUS General Osteology Compared to those in other orders, the radius of the Anatidae is a stout bone with a bulky proximal end and a slightly widened di stal end. The capital tuberosity (sensu Howard 1929) is round and tuberc ulum bicipitale ra diale is a low structure with a round depression in its center. Comparative Osteology Apart from size and a less pronounced capital tuberosity in Dendrocygna , there are no traits that allow me to di stinguish among the proximal ends of different genera in this family.


73 Anas sp. – OGSE-80 (1): Because of the abundance of A. georgica and A. bahamensis at site OGSE-80, I tentatively assign this proximal end to the genus Anas . CARPOMETACARPUS General Osteology The anatid carpometacarpus is a long bone with a stout processus extensorius, a deep fossa infratrochlearis, a deep fovea carpa lis caudalis, and a ventrally curving distal end of os metacarpale majus. Superficiall y it resembles that of the Ardeidae. Comparative Osteology The osteology of this family has been described in detail by Woolfenden (1961). Since only two proximal ends were recovered from the Ecuadorian archaeological sites, I will not discuss the distal end. Taxa can be distinguished based on the dorsal rim of trochlea carpalis, the location of attach ment for ligamentum ulnocarpometacarpalis ventralis, and the shape, size, and or ientation of processus extensorius. Dendrocygna autumnalis – Loma Alta (1): The absence of a notch in the dorsal rim of trochlea carpalis, char acteristic of the genus Dendrocygna (Woolfenden 1961), makes this proximal end similar to that in Egretta , which can be distinguished by a rounder, shorter ventral rim of trochlea carpalis, a round fossa infratrochlearis (oval in Dendrocygna ), and a shorter proximal symphysis. The other genera of Anatidae all have a distinct notch in the dorsal rim of trochlea carpalis, a br oad processus extensorius, and an oblong attachment for ligamentum ulnocar pometacarpalis (rather than round as in Dendrocygna ). The two possible species of Dendrocygna can be distinguished by size ( D. autumnalis is smaller), the shape of fossa infratrochlearis (round and deeply


74 excavated in D. bicolor ), and the attachment area on pr ocessus extensorius (located proximally in D. autumnalis and extending more distad in D. bicolor ). Anas bahamensis – OGSE-80 (1): The proximal end can be assigned to A. bahamensis because of a broad processus extensorius (narrower in Dendrocygna , broader in Aythya and A. georgica , shorter in N. erythrophthalma ), a round and deep fossa infratrochlearis (shallow in N. erythrophthalma ), and a large and low ligamentum ulnocarpometacarpalis dorsalis (small and raised in Aythya ). FEMUR General Osteology The femur of the Anatidae is a variably long but stout bone. The proximal shaft is broad and medio-laterally flattened. Proxi mally, facies dorsalis is at a ca. 45 û angle to plana dorsalia. Toward the distal end, facies dorsalis is parallel to plana dorsalia giving the shaft a twisted appearance. The wide di stal end is character ized by a deep fossa poplitea and broad trochlea fibularis. Comparative Osteology The comparative osteology of the femur in ducks has been discussed in detail by Woolfenden (1961). The Ecuadoria n archaeological sites yielde d only one specimen of a shaft. Merganetta armata – La Chimba (1): The shaft of the La Chimba specimen is short and thick. Fossa poplitea is deeper than in Anas . Attachment for musculus femorotibialis lateralis is long and prominent, stretching fr om the proximo-lateral margin of trochlea fibularis proximally for about one third of th e length of the shaft. I have not found such a


75 long and pronounced attachment for musculus fe morotibialis lateralis in any other anatid taxon. TIBIOTARSUS General Osteology The tibiotarsus of the Anatid ae is a slender bone with a large facies gastrocnemialis and a characteristic bulky, mediad curving distal end. Condylus me dialis and condylus lateralis are similar in width and length. Comparative Osteology The tibiotarsus of the Anatidae does not have many distinguishing characterisitcs, making even the identificati on of genera difficult (Wool fenden 1961). Although several traits can be identified on the distal end upon closer exam ination, the proximal end is very uniform. Anas bahamensis – OGSE-80 (1): The proximal end of tibiotarsus is smaller than in Dendrocygna and larger than in Oxyura . Aythya is also slightly smaller and distinguishable by a deep fossa flexoria. The tibiotarsus of N. erythrophthalma is similar in size, but has a proportionately shorter crista fibularis. The three species of Anas considered here ( A. bahamensis , A. georgica , and A. clypeata ) are almost indistinguishable, with A. bahamensis tending to have a more slender shaft. My identification is based on the stoutness of the shaft.


76 TARSOMETATARSUS General Osteology The tarsometatarsus of the Anatidae is a compact bone with three hypotarsal sulci/canals, a slightly concave facies dorsali s, and a plantad rotated trochlea metatarsi II that is more proximally located than trochlea metatarsi IV. Comparative Osteology The only tarsometatarsi from the Ecuadoria n archaeological sites are two complete bones from the highland site of La Chimba . Overall proportions can be used for a preliminary sorting of this element. Other useful traits includ e the proximo-medial margin of the shaft, the position of forame n distale, and width and distal extent of trochlea metatarsus II (Woolfenden 1961, M. P. Tellkamp personal observation, ). Merganetta armata – La Chimba (2): The tarsometatarsus of M. armata is very distinctive because of its sl ender shaft (most slender shaf t in the dabblers, perching ducks, and divers conbined), a plantad sl oping medial margin of the proximal shaft (squared and more dorsal in dabblers), s hort trochlea metatarsi II et III (longer in Anas and Aythya ), a broad trochlea metatarsi II (narrow in Anas ), and a proximal trochlea metatarsi II that shows little plantar rotation (distal rim of trochlea metatarsi II does not reach proximal end of trochlea metatarsi III as it does in Anas and Aythya ; Woolfenden 1961). Order ACCIPITRIFORMES Family ACCIPITRIDAE The genera found in the dry pacifi c zoogeographic area of Ecuador are Elanoides , Gampsonyx , Elanus , Rhostramus , Harpagus , Ictinia , Accipiter , Geranospiza ,


77 Leucopternis , Buteogallus , Heterospiza , Parabuteo , Buteo , and Spizaetus . The diversity of raptors is smaller in the high Andes wh ere some of the same genera can be found. Genera that occupy diffe rent types of habitats in the highlands are Elanoides , Circus , Accipiter , Parabuteo , Geranoaetus , Buteo , and Oroaetus . SCAPULA General Osteology The scapula of the Accipitr idae is an unusually strong bone, with a broad proximal end (due to a greatly expanded medial pr ocess of acromion [Höfling and Alvarenga 2001]), prominent crista ligamenti acrocoraco acromiali, distal pneumatic fossa of acromion, small tuberculum coracoideum, long facies articularis humeralis, narrow and thick proximal corpus scapulae, and a greatly expanded and thin distal corpus scapulae. Comparative Osteology A single scapula of a mediumto large-si zed hawk was recovered from the coastal site of Validivia Village. I therefor e limit this account to the genera Leucopternis , Buteogallus , Parabuteo , and Buteo . Leucopternis was not available for comparison. Distal width, shape of acromion, location and size of pneumatic fossa, shape of facies articularis in lateral aspect, outline of facies articularis hum eralis in medial aspect, and relative width of proximal corpus scapulae ar e useful characters to distinguish among these genera. cf. Leucopternis sp. – Valdivia Village (1): The proximal end with a slightly broken medial process of acromion does not ma tch any of the species examined. The combination of a small pneumatic fossa (large in Buteogallus anthracinus , B. meridionalis , and Buteo nitidus ), very broad distal end (narrow in B. aequanoctialis ,


78 Parabuteo unicinctus , and B. brachyura ), a distally overarc hing medial process of acromion (found in no species examined), a broa d and short facies articularis humeralis (found in no species examined), a highly angled outline of facies articularis humeralis in medial aspect (found in no species examined ), and a narrow proximal end of corpus scapulae (approximated only in P. uncinctus and Buteo ), is unique. The only other genus of hawk that might show these traits is Leucopternis . Until comparative specimens of Leucopternis become available the identification remains tentative. CORACOID General Osteology The coracoid of the Accipitridae is a st out bone with a short shaft, large dorsoventrally rotated extremitas om alis coracoidei, pronounced faci es articularis clavicularis, and sharp processus procoracoid eus with an incisura nervi supracoracoidei. Most groups have a pneumatic fossa in sulcus musculi supracoracoidei. Comparative Osteology Two coracoids were recovered from El Az úcar and Loma Alta, respectively. The El Azúcar specimen belongs to a small to medium -sized raptor and the Loma Alta specimen to a medium-sized hawk. Most of the differen ces in the coracoid of this family can be found in facies articularis st ernalis, the margins of impr essio musculi sternocoracoidei, the relative width of the shaft, the shape of a tuberosity just distal to impressio musculi sternocoracoidei, the shape and presence of a pneumatic fossa within the sulcus musculi supracoracoidei, the shape of f acies articularis clavicularis, and the shape of the processus acrocoracohumeralis. Both coracoids are dama ged to such a degree that many of these traits are not available, making the identification difficult.


79 Buteo cf. nitidus – El Azúcar (1): The size of the bone reduces the number of possible genera (smaller in Gampsonyx , Harpagus , and neotropical Accipiter , larger in Buteogallus , Heterospiza , Parabuteo , and Spizaetus ). Of the remaining genera, Elanoides differs by being strikingly flat, and Ictinia lacks a pneumatic fossa in sulcus musculi supracoracoidei, leaving only Leucopernis and Buteo sensu lato (including Buteo [ Asturina ] nitidus and Buteo [ Rupornis ] magnirostris [Riesing et al. 2003]). Unfortunately, no specimens of Leucopternis were available for comparison. The fossil shares several traits with Buteo sensu lato: a prominent tuberosity just distal to impressio musculi sternocoracoidei, a somewhat more ve ntrally oriented pro cessus procoracoideus, and the unusual shape of the latera l half of cotyla scapularis which is not raised, gradually merging proximally with the shaft. Of all the species of Buteo examined, B. nitidus compares most favorably with the specimen fr om El Azúcar because of 1) a low degree of curvature of the ventral margin of extremitas omalis coracoidei (more curved in B. platypterus , and B. brachyurus , but similar in B. magnirostris ), 2) a slender shaft (proportionately wider in B. platypterus and B. brachyurus , more slender in B. magnirostris ), 3) a more mediad orientation of coty la scapularis (dorsa d in the other three species), and 4) a thick pro cessus procoracoideus that bears a broad facies articularis scapularis (thin in the other species). Buteo swainsonii – Loma Alta (1): The only genera that approximate the size of the fossil are Buteogallus , Heterospiza , Parabuteo , and Buteo . Of these, Buteogallus and Heterospiza have a very pronounced tuberosity distal to impressio musculi sternocoracoidei which is lack ing in the subfossil. Although Parabuteo lacks this tuberosity, it differs by having a long and slender shaf t. In all three genera the coracoid is


80 a fairly straight bone in which the most distal end of facies articularis humera lis is located dorsally in relation to facies ar ticularis sternalis in lateral or medial aspect. On the other hand, Buteo is characterized by a more curved coraco id such that the distal end of facies articularis humeralis lies on the same plane as facies articularis sternalis in medial or lateral aspect. Based on these observations , the subfossil belongs to the genus Buteo and compares closely in size and shape to that of a female B. swainsoni . HUMERUS General Osteology The humerus is a thick bone with a broa d proximal end that has a bulky tuberculum supracondylare ventrale, a strong and knobby pr ocessus supracondylaris dorsalis, and a long crista deltopectoralis. Fossa pneumotricipitalis is lacking. Fossa musculi brachialis is large, and raised muscle attachment lines can be found near margo caudalis. Crista deltopectoralis has thickened margins. The hum eri of the Accipitridae and Falconidae are fairly similar; lthough in the Accipitridae it te nds to be longer in re lation to the width of the distal end and to have a narrower pr oximal end. The distance between tuberculum dorsalis and tuberculum ventralis is sma ller than the distance between processus supracondylaris dorsalis and epicondylus ventralis. In the Falconidae the inverse relationship exists. Comparative Osteology The humerus is represented by only two fo ssils, a proximal end from Loma Alta and a distal end from La Ponga. The humerus from Loma Alta has the same provenience as the coracoid described earlier and likew ise belongs to a medium-sized raptor. The bone from La Ponga is much smaller.


81 Within the Ecuadorian members of this family, the main differences on the proximal end can be found in the relative thic kness of caput humeri , the proximal end of crista deltopectoralis, and the shape and dept h of impressio coracobrachialis. The distal end differs mainly in the strength and curv ature of processus flexorius, the distance between epicondylus dorsalis and processus supracondylaris dorsalis, and the thickness of the ventral margin of fossa musc uli brachialis in cranial aspect. Buteo swainsonii – Loma Alta (1): The proximal fragment is similar in size to the genera Buteogallus , Heterospiza , Parabuteo , and Buteo . The specimen can be ascribed to the genus Buteo based on the following characteristics. 1) Caput humeri is less bulky than in Buteogallus , Heterospiza , and Parabuteo , especially ventrad. 2) Tuberculum dorsale is narrow, with cranial and caudal marg ins being nearly parallel. Whereas Parabuteo shares this trait with Buteo , in Buteogallus and Heterospiza the tuberculum dorsale gradually tapers toward the dorsal end and is not clea rly delimited from caput humeri. 3) Crista deltopectoralis is thickened proximally in Buteo and to a lesser degree in Heterospiza , whereas in Buteogallus and Parabuteo it is flat. 4) The distal e nd of crista deltopectoralis is thickened in Buteo but not in any of the other genera. 5) The ventral margin of crista deltopectoralis sits on a “ledge” extending ventrad for about a third of the width of the shaft in Buteo and most other genera exept Parabuteo (the ledge is absent and the ventral margin of crista is close to the dorsal margin of the shaft). 6) The caudal muscle attachment line of the musculus deltoideus ma jor runs just dorsally to margo caudalis, a condition not found in Heterospiza . The specimen compares very favorably to Buteo swainsonii , although B. albonotatus cannot be ruled out. In addition, Campbell (1976) described an extinct late


82 Pleistocene species of Buteo , which may have persisted in coastal Ecuador to the Mid Holocene. Based on the very good match with B. swainsonii , I tentatively identify the specimen as that species. An associated cor acoid supports that iden tification (see above). Ictinea plumbea – La Ponga (1): The medium-sized accipitrid is in the size range of Ictinia , Rostrhamus , Chondrohierax , and a small Buteo sensu lato. The specimen most closely resembles Ictinia , however, in three characters. 1) Processus flexorius does not extend distad beyond cotyla vent ralis in cranial aspect. In Buteo sensu lato, processus flexorius is longer and extends beyond cotyla ventralis. The short pr ocessus flexorius in Chondrohierax is similar to that in Ictinia ; in Rostrhamus processus flexorius is very short and compressed cranially, unlike in any of the other taxa. 2) The dorsal ridge bordering sulcus scapulotricipita lis in Ictinia and the other kites is knob-like rather than sharp as in Buteo . 3) The ventral margin of fossa musc uli brachialis is a sharp ridge close to the ventral margin of the shaft as found in Ictinia and Buteo magnirostris (possibly not a Buteo sensu stricto; Riesing et al. 2003), and to a lesser degree in Rostrhamus and Chondrohierax . In more typical Buteo , such as B. platypterus and B. nitidus the ridge is rounded and located more centrally in cranial aspect. All traits described occur only in Ictinia . The only species of this genus pr esently found in western Ecuador is I. plumbea . The closely related I. mississippiensis is virtually indistingui shable osteologically; however, since I. mississippiensis has rarely been recorded in Ecuador (Ridgely and Greenfield 2001a), I consider it to be very unlikely.


83 ULNA General Osteology The acciptrid ulna is a long but strong bone with a short and knob-like olecranon, a flat and lobe-shaped processus cotylaris dorsalis, deep impressio brachialis, bulky tuberculum ligamenti collateralis ventralis , large and round condylus dorsalis ulnaris, proximally extending condylus dorsalis ulnaris that is raised above the surface of the shaft, bulky condylus ventralis ulnaris, short tuberculum carpale, and shallow depressio radialis. It is somewhat similar to the ul nae of the Falconidae a nd Strigidae, differing mostly in having a deep pit caudal of crista intercotylaris (interme diate in the Falconidae and shallow in the Strigidae), bulky olecranon that is round in proximal view (oblong in the Falconidae) and thick in dorsal view (fla ttened in the Strigidae), short and raised tuberculum ligamenti collateralis ventralis, de ep depressio radialis (absent in Strigidae and shallow in Falconidae), and flat condylus ventralis ulnaris (rais ed in the Falconidae and pointy in Strigidae). Comparative Osteology On the proximal end, differences lie mos tly in the shape of olecranon, the shape and orientation of cotyla dorsalis, the lo cation of impressio scapulotricipitalis, the strength and position of tuberculum ligamen ti collateralis ventralis, the width of impressio brachialis, the attach ment for musculus bicipitalis brachii, and the location of the line of attachment for anchoneus. On th e distal end, the shape of condylus ventralis ulnaris and the shape and relativ e position (with regard to co ndylus ventralis ulnaris) of tuberculum carpale are mo st useful (Table 2-3).

PAGE 100

84 cf. Buteogallus anthracinus – OGSE-80 (1): This distal shaf t is difficult to place due to the poor quality of the specimen. The surface layer of the bone is mostly abraded, suggesting that this was a juve nile individual, and covered by mineral deposits. It is a hawk because of the long condylus dorsalis ulnaris in ventral aspect, a deep depressio radialis, and the position and shape of the su rface leading up to tuberculum carpale. The most similar species examined is Buteogallus anthracinus . It agrees well in having a long condylus dorsalis ulnaris in ventral view, sa ddle-shaped tuberculum ventralis in distal aspect, broad surface just proximal to tuberculum ventralis in cranial aspect, and deep depressio radialis. The identification is very tentative. Table 2-3. Characters and character states fo r the ulna of medium -sized Accipitridae. Character Character States AcU1 Olecranon 0: small 1: large AcU2 Impressio scapulotricipitalis 0: round 1: oblong AcU3 Tuberculum ligamenti collateralis ventralis 0: narrow, flat 1: narrow, raised 2: broad, flat 3: broad, raised AcU4 Impressio brachialis 0: short, wide 1: short, narrow 2: long, wide 3: long, narrow AcU5 Attachment for musculus bicipitalis brachii 0: located dorsad 1: located centrad or slightly dorsad AcU6 Attachment of anchoneus 0: dorsad in re lation to impressio musculi brachialis 1: originating from impressio musculi brachialis AcU7 Condylus ventralis ulnaris, ventral aspect 0: round and bulky 1: low, but sharp 2: raised and sharp AcU8 Tuberculum carpale, ventral aspect 0: round 1: pointy Parabuteo unicinctus – Loma Alta (1): The proximal end of an ulna is slightly damaged, with some tuberosities and other ra ised features being h eavily worn. In terms of size, it compares to that in Parabuteo , Buteo , and a male Buteogallus anthracinus .

PAGE 101

85 Other examined species that are slig htly smaller or larger in size are Elanoides forficatus , Circus cyaneus ( C. cinereum was not available), Rostrhamus sociabilis , Buteogallus urubitinga , and Heterospiza meridionalis . I identified the specimen from Loma Alta as P. unicinctus as the muscle attachment sites for mu sculus bicipitalis brachii do not coincide in Circus , Buteogallus urubitinga , or Heterospiza . Circus also has a flatter tuberculum ligamenti collateralis ventralis. In B. urubitinga tuberculum ligamenti collateralis ventralis is also located more proximad. Table 2-4. Character matrix for the ulna of me dium-sized Accipitridae . For definitions of character states see Table 2-3. Characters (AcU) Species 1 2 3 4 5 6 7 8 Elanoides forficatus 0 1 1 2 1 0 0 0 Rostrhamus sociabilis 0 0 1 3 1 0 2 0 Elanus leucurus 0 1 1 2 1 0 1 1 Ictinea plumbea 0 0 0 0 0 0 1 0 Circus cyaneus 1 1 1 1 0 0 2 0 Chondrohierax uncinatus 1 0 2 1 1 0 1 0-1 Accipiter cooperi 1 0 2 3 0 1 1 0 Parabuteo unicinctus 1 1 2-3 2 0 1 2 0 Buteo platypterus 1 1 3 3 0 0 1 1 Buteo brachyura 1 1 2-3 0 0 1 1 0-1 La Chimba Specimen 1 1 3 3 0 0 1 0-1 Buteo cf. platypterus – La Chimba (1): The complete ulna corresponds in size to that of a small Buteo . It agrees in character states with the ulna of B. platypterus (Table 24). Because a specimen of B. albigula was not available, the identification is tentative.

PAGE 102

86 CARPOMETACARPUS General Osteology The carpometacarpus of the Accipitridae is superficially similar to that of the Falconidae and the Strigiformes. A strong pro cessus extensorius, a ventrally offset os metacarpi minus, and a deep sulcus tendineus are shared by the Accipitriformes and the Strigiformes. The Strigiformes can be dist inguished by the presence of a ridge extending proximally from the ventral rim of os metacar pi minus and reaching the base of processus pisiformes, a rounder and more club-like pro cessus extensorius, and the knob-like shape of the cranial rim of facies articularis digitali s major. In the Accipitriformes the distal end flares out allowing for a broa d facies articularis. The Falconidae differ in having a proportionately shorter carpometacarpus, a less ventrad oriented (also found in some kites) and pointy (rather than rounded) proces sus extensorius, and in an extreme ventral position of os metacarpi minus. Comparative Osteology Apart from size and overall porportions, se veral subtle differences exist between species of similar size with regard to the shape of processus extensorius, the crosssectional shape of os metacarp i majus, and the strength of various tuberosities as well as the depths of fossas and foveas. Elanus leucurus – La Chimba (1): Several traits of th e broken specimen coincide with the carpometacarpus of E. leucurus . 1) Fossa supratrochlearis is concave and less steep than in Buteo . 2) Processus extensorius in na rrower and more squared than in Buteo or Circus . 3) Processus extensorius does not show the extreme dorsad curvature found in Circus and Buteo . 4) Just distal to the dorsal aspect of processus alularis a little concave

PAGE 103

87 depression can be found which is absent in Elanoides and large and elongated in Circus . 5) Fovea carpalis cranialis is broad and round in Elanus , but long and round in all other genera. 6) Although the ventral rim of trochlea carpalis is broken, fovea carpalis caudalis evident (absent in Parabuteo and Buteo ). 7) Sulcus tendineus shif ts from cranial to dorsal facies within the last fout h of os metacarpi majus. 8) A concave surface between symphysis distalis and the ventro -cranial rim of facies articu laris digitalis major is broad and shallow (deep and oblong in Circus , Buteo , and Parabuteo ). 8) Facies articularis digitalis major is slightly angled in distal as pect, ending dorsally in a small tuberosity that points into a dorso-caudad direction. In Circus facies articularis is rounded, whereas in Buteo it is highly angled. The dorsal tuberosity points into a more caudad direction in Circus and Buteo . FEMUR General Osteology The accipitrid femur is a highly characterist ic bone with a low crista trochanterica, a deep fovea ligamenti capitis, strong im pressiones iliotrochantericae, a proximal pneumatic fossa in cranial aspect located between impressiones il iotrochantericae and linea intermuscularis cranialis, a proximally cen trad linea intermuscularis cranialis, and a strong distal end with deep impressio ansa musculi iliofibularis pars caudalis, an expanded proximal end of the lateral rim of trochlea fibularis, and a wide sulcus intercotylaris. Comparative Osteology The femora from the Ecuadorian archaeol ogical sites are from mediumto largesized hawks. Based on size, I limited the taxa examined to members of the genera

PAGE 104

88 Buteogallus , Parabuteo , and Buteo . Leucopternis was unavailable. The fragmentary nature of the archaeological material limits the available traits to those found on the extreme proximal end as well as available on mo st of the distal end. I have found several useful traits, however, including the relati ve size of the hea d, location of linea intermuscularis cranialis, width of condylus lateralis, shape of sulcus intercondylaris in caudal aspect, location and shape of medial muscle attachment site just proximad of cotyla medialis, depth of impressio ansa mu sculi iliofibularis, and the orientation of a proximal tuberosity on the lateral rim of trochlea fibularis in caudal aspect. Buteogallus meridionalis – La Ponga (1): The proximal fragment of femur has a very large head. I have seen such a proportionately large head only in B. meridionalis . Parabuteo unicinctus – El Azúcar (1): Identificat ion of the fragmentary proximal end is problematic. One of the most distin ctive features of th e proximal end of the accipitrid femur is the location and shape of the proximal pneumatic fossa in cranial aspect, which is absent in the archaeological specimen. As in Parabuteo and Buteogallus , fovea ligamenti capitis is deep; it is sha llow in the femora of similarly sized Buteo . Three species of Buteogallus , B. anthracinus , B. urubitinga , and B. aequatorialis , are very large. In the smaller B. meridionalis , linea intermuscularis cranialis touches the proximal pneumatic fossa (it does not in Parabuteo ). Although I consider it likely that this specimen is from B. unicinctus , the identification is te ntative because of the incompleteness of the bone. Buteo cf. nitidus – El Azúcar (2): The two proximal ends can be assigned tentatively to B. nitidus based on a narrow muscle attach ment site just proximal to condylus medialis (stronger in Parabuteo and Buteogallus ), a highly asymmetric sulcus

PAGE 105

89 intercondylaris in cranial vi ew (nearly symmetrical in Buteogallus and Parabuteo ), and a straight and large proximal tube rosity on the lateral rim of tr ochlea fibularis (at an angle to the rim in Parabuteo and Buteo brachyura , much reduced in Buteogallus ). TIBIOTARSUS General Osteology The tibiotarsus of the Accipitridae is a hi ghly characteristic bone with a small and deep facies gastrocnemialis and an oblique sulcus tendineus at a ca. 45 û angle to axis proximodistalis. A wide distal end, due to the outward flaring of condylus medialis, supports an expanded trochlea cartilaginis tibialis and wide incisura intercondylaris. Comparative Osteology As distinctive as the tibiotarsus is am ong avian families, it is rather homogenous within the buteonine Accipitridae. Size a nd overall proportions, and a few other small differences, result in tentative identifications. Geranoaeutus melanoleucus – La Chimba (1): The distal end of a large tibiotarsus is of the buteonine type and is identical to the modern comparative specimen of G. melanoleucus . The great resemblance to Buteo provides additional evidence that Geranoaetus is closely related to that genus (Riesling et al. 2003). Buteo cf. brachyura – El Azúcar (1): The archaeological specimen does not agree with any of the species examined. The be st match is with the tibiotarsus of Buteo brachyura , which shows the same location of cana lis transversus with regard to the laterad extent of the proximo-lateral aperture of the canal (more medial in most species except P. unicinctus ), and the large size of a shelf fo rmed at the proximo-medial end of pons supratendineus (smaller in P. unicinctus ).

PAGE 106

90 Buteo nitidus – Loma Alta (1): The distal end with a sharp medial margin (broad in Buteogallus urubitinga , B. anthracinus , and Buteo swainsonii ), an intermediate level of distal widening of condylus me dialis (little widening in Buteogallus anthracinus and B. urubitinga , more widening in B. meridionalis , P. unicinctus , and B. brachyura ), and a thick shaft fits the characters of B. nitidus . Buteogallus aequinoctialis is similar, but has a small shelf at the proximo-medial end of pons supratendineus. TARSOMETATARSUS General Osteology As was the case for the tibiotarsus, the ta rsometatarsus of the Accipitridae is highly distinctive. Two cristae hypotar sorum, a medially flattened shaft, and strong tuberositas musculi tibialis cranialis charac terize the proximal end. Facies plantaris has a deep sulcus flexorius, and facies subcutanea lateralis is deep. The shaft has a roughly triangular shape in cross-section. On the dist al end, trochlea metatarsi II, III, and IV have roughly the same distal extent and form a half circle in distal aspect . The lateral rim of trochlea metatarsi IV and the medial rim of tr ochlea metatarsi II are greatly expanded. Comparative Osteology As with the tibiotarsus, the tarsometatar sus is a uniform bone among the buteonine hawks. Subtle differences in the shape of cristae hypotarsorum, the size of tuberositas musculi tibialis cranialis, the length and width of trochleae metatarsorum, and the shape of the wing-like medial rim of trochlea meta tarsi IV are useful when considered in conjunction. Geranoaetus melanoleucus – La Chimba (2): The two pitted proximal ends of carpometacarpi are very close in size and shap e to a reference specimen of a juvenile G.

PAGE 107

91 melanoleucus (UF 33557). No comparative material was available for other large-sized raptors, such as Oroaetus isidori or Morphnus guyanensis . Because of the excellent match, I identify the archaeological specimens as G. melanoleuca . Buteo nitidus – El Azúcar (2): A proximal and a distal fragment agree best with that of B. nitidus . On the proximal end, the relative position of foramen vasculare distale and fossa metatarsi I coincide only with specimens of B. nitidus . In addition, the fragment comes from a species with a broad tarsometatarsus. OSSA DIGITORUM PEDIS General Osteology The pedal phalanges of the Accipitridae ar e characteristic for raptors with a deep trochlea articularis and deep fovea ligamenti collateralis. Phalanges ungualis have a large tuberculum extensorium and t uberculum flexorium, and lack a sulcus neurovascularis. Basis phalangis of phalanges proximales et intermediae are intermediate in width between the wide Falconidae and narrow Stri gidae. Phalanges ungualis have flat lateral and medial margins in plantar aspect in cont rast to the round margins in the Strigidae. Comparative Osteology Size and overall proportions are the best tr ait to distinguish among taxa in this family. Parabuteo unicinctus – El Azúcar (3): In terms of length to width ratio, the species with the proportionately longest and most slender phalanx 2 of hallux is P. unicinctus . I assign two associated phalanges ungualis to P. unicinctus as well.

PAGE 108

92 cf. Buteogallus anthracinus – OGSE-80 (1): Buteogallus anthracinus has the thickest and shortest phalanx 2 of hallux that I have examined. I therefore identify this bone tentatively as B. anthracinus . Geranoaetus melanoleuca – La Chimba (3): By size and the same reasoning as for the tarsometatarsus, I refer these phalanges (phalanges 1 and 2 of hallux, phalanx 1 of digitus quartus) to Geranoaetus melanoleuca . cf. Buteo nitidus – OGSE-80 (1): This small phalanx 4 of digitus quartus corresponds in length an d width to that of B. nitidus . The identification is tentative, however, as other species of the ge nus are similar osteologically (e.g., B. magnirostris , which tends to have longer elements). Order ACCIPITRIFORMES Family FALCONIDAE SCAPULA General Osteology Raptors of the orders Accipitriformes a nd Strigiformes have scapulae with bulky, pneumatized proximal ends. The acromion and medial process of acromion (Höfling and Alvarenga 2001) together form a broad dorsal ha lf in cranial aspect, often with a cranial pneumatic fossa. Crista ligamenti acroco racoacromiali is well pronounced. In small hawks (e.g., Gampsonyx swainsonii ) the acromion is longer and less pneumatized whereas in small owls (e.g., Glaucidium spp.) it is more trian gular shaped in cranial aspect.

PAGE 109

93 Comparative Osteology Large differences in size allow for an easy separation of large and small species. The shape of medial process of acromion (pro cessus acromialis medialis) is characteristic for different genera. Falco sparverius – La Chimba (1): Only extant species of Micrastur and Falco have scapulae similar in size to the La Chimba specimen. Micrastur differs by having a large and flat medial process of acromion. Falco rufigularis and F. columbarius are similar in size to F. sparverius , but differ in a broad facies articularis humeralis (oblong in F. sparverius ) and a shorter processus acromii medialis in cranial view. HUMERUS General Osteology The humeri of the Falconidae are superficially similar to those of the Accipitridae (see above). Comparative Osteology In addition to size and overall proportions , the shape of crista deltopectoralis, tuberculum ventrale, tuberculum suprac ondylare ventrale, and impressio musculi brachialis are useful traits to distinguish among species. Phalcoboenus carunculatus – La Chimba (1): Differing in several traits from the similarly sized Polyborus plancus and Falco peregrinus , the extremely well preserved complete humerus can be assigned to P. carunculatus . 1) The bone is longer and more slender than in F. peregrinus . 2) Tuberculum ventrale is less bulky than in P. plancus and F. pereginus . 3) Crista deltopectoralis is more triangular than in P. plancus . 4) Fossa musculi brachialis is oval shaped as in P. carunculatus , but more oblong and squared in

PAGE 110

94 P. plancus and F. peregrinus . 5) Tuberculum supracondylare ventrale is less raised than in P. plancus and less circular than in F. peregrinus . At this point it is impossible to differentiate between P. carunculatus and P. megalopterus , species considered conspecifics by some authors (Ridgely and Greenfield 2001a). The two species have allopatric distributions, with northern Ecuador being part of the range of P. carunculatus . Because of its large size, the spec imen probably came from a female. CARPOMETACARPUS General Osteology The carpometacarpus of the Falconidae is superficially similar to that of the Psittacidae. A flat and extended processus extensorius, proximally sloping processus alularis, and a squared and flat shaft in cross-section are ch aracteristics of the falcons. A distal depression on os metacarpale major (f or passage of musculus ulnometacarpalis dorsalis) separates symphysis metacarpalis di stalis from a cranially expanded facies articularis digitalis major. Comparative Osteology The Falconinae can be distinguished from the Caracarinae by os metacarpale minus inserting ventrally between processus pisiform es and the caudo-distal end of the ventral rim of trochlea carpalis, and by a longer proce ss bearing facies articu laris digitalis minor. The more distal insertion of os metacarpa lis minus suggests that the specimen from OGSE-80 came from a species of Caracarinae. Overall size, shape and size of processus extensorius, and the shape of os metacarpale major in cr oss-section provide diagnostic traits.

PAGE 111

95 Herpetotheres cachinnans – OGSE-80 (1): The proximal end of carpometacarpus has a rounded articular surface on pro cessus extensorius (less rounded in Daptrius and Micrastus ), a shallow fovia carpal is cranialis (deep in Micrastur ), and a cranio-caudally flat os metacarpale major with a sharp vent ro-caudal margin (somewhat flattened with no sharp ventro-caudal margin in Micrastur and triangular in Daptrius ). These traits are distinctive for Herpetotheres . Falco sparverius – El Azúcar (1): With a ventral depressi on on the distal end of os metacarpale major and a broad facies articu laris digitalis major (causing distal os metacarpale major to flare out dorso-cranially ) the distal end of carpometacarpus is from a small falcon (the distal end does not flare out as much in small hawks and kites). The long process bearing facies ar ticularis digitalis minus sugge sts that the bone is from a species of Falco . The size of the specimen agrees best with that of F. sparverius (smaller than in F. rufigularis and F. columbarius ). Order GALLIFORMES Family CRACIDAE FEMUR General Osteology The femur of the Cracidae ranges from slender in Ortalis to stout in Crax . The bone appears to be rotated around the main axis su ch that the cranial si de of the distal end is oriented cranio-laterall y. Proximally, impressiones obturat oriae are very prominent, giving the bone a U-shaped form in proximal aspect. In many species the proximal end is further characterized by a fossa on the cranial side just distal to facies articularis antitrochanteris. Crista trochanteris is slig htly curved medially and facies articularis

PAGE 112

96 antitrochanteris is at a 30 û angle with respect to axis proximodistalis. Caudally, facies articularis antitrochanteris forms a little ledge . On the cranial side of the shaft, linea intermuscularis cranialis extends from the crista trochanteris across the bone almost reaching the proximal end of the condylus medi alis. On the distal end, condylus lateralis reaches more distad than the condylus media lis. Fossa poplitea usually bears two to three foramina and is otherwise a variably textur ed, poorly structured depression. Impressio ansae musculus iliofibularis is always well pronounced, forming a shallow depression. Impressio ligamenti cruciati cranialis is shallow and always accompanied by another, smaller and more proximal impressio ligam enti. A well pronounced fovea tendineus musculi tibialis cranialis can be found on condylus lateralis. Comparative Osteology The femora of the large species within this family are much more textured than those of the smaller species with muscle att achment sites protruding to a larger degree and linea intermuscularis forming well defi ned ridges. Tuberosities are less pronounced in large Penelope than in the even larger Crax . The femur in Ortalis is relatively more slender than in other genera. The dorsal ledge of facies articularis antitrochanteris is highly pronounced in Crax and Ortalis , less so in Penelope , and almost absent in Aburria , where broad impressiones obturatori ae reduce the ledge to a small medial “ridge.” The distal end of the femur differs in overall shape among the different genera examined. In Crax the distance between the medial and lateral extremes of the condyla is more than twice as long as the condylus fi bularis from a caudal aspect (from impressio ansae musculus iliofibularis to its most distal point). The ratio between these two

PAGE 113

97 measures is smallest in Aburria and intermediate in Penelope and Ortalis . Condyla fibiolaris and lateralis are oriented more caudally in Aburria whereas in Ortalis , Penelope and Crax they are rotated outward pointing in a more lateral direction. The foramina in the fossa poplitea differ among the specimens ex amined as well. The large species have a row of at least three foramina at the proxima l margin of the fossa. In smaller species only one fossa may be present. In Aburria and Ortalis the fossa is located near the proximal margin of the fossa poplitea, whereas in Penelope montagnii it is located more distally toward the center of the fossa. Crista supr acondylaris medialis is short and stout in Penelope and Aburria , but narrow and long in Ortalis and Crax . I have found seven traits that are useful in identifying the specimens from the archaeological sites (Table 2-5). Table 2-5. Characters and character states for the femur in the family Cracidae. Character Character States CrF1 Stoutness 0: very slender 1: slender 2: stout CrF2 Fossa distal to facies articularis antitrochanteris 0: absent 1: present CrF3 Dorsal ledge of facies articularis antitrochanteris 0: small 1: pronounced CrF4 Width of distal end in caudal aspect 0: condyla fibularis < ½ th e distance between the condyla lateralis and medialis 1: condyla fibularis > ½ the distance between the condyla lateralis and medialis CrF5 Orientation of condyla fibularis and lateralis 0: caudal 1: caudolateral CrF6 Location of major foramina within the fossa poplitea 0: distal 1: proximal CrF7 Shape of the crista supracondylaris medialis 0: bulky 1: narrow, sharp-edged cf. Chamaepetes goudotii – La Chimba (1): Based on the traits in Table 2-6, I can exclude Penelope as the likely genus for the fossil. Penelope has a more slender, longer femur with a less massive proximal shaft. Th e foramen of the fossa poplitea is located proximally as in Aburria and not distally as in Penelope montagii . On the distal end, most

PAGE 114

98 of which is destroyed in the fossil femur, cr ista supracondylaris medialis appears to be more bulky, extending more proximad than in Penelope . This can also be seen in Aburria . The femur from La Chimba is too small to be either species of Aburria present in Ecuador. Given the present distribution of cracids in Ecuador, the most likely species is Chamaepetes goudotii , which has an Andean distri bution similar to that of Penelope montagnii , and at least in Bolivia it has been r ecorded at high elevations (Remsen and Cardiff 1990). Chamaepetes has been placed near Aburria in some classifications (see Vaurie 1968), although recent molecular work wa s inconclusive, suggesting that together with Penelopina it is in a sister clade to (( Aburria , Pipile ), Penelope ) (Pereira et al. 2002). Table 2-6. Character matrix for seven traits found in the femur of th e Cracidae. See Table 2-5 for definitions of the character states. Characters (CrF) Species 1 2 3 4 5 6 7 Penelope purpurascens 2 1 0 1 1 1 0 Penelope montagnii 1 1 0 1 1 0 0 Ortalis leucogastra 0 0 1 0 1 1 1 Aburria pipile 2 1 0 0 0 1 0 Crax rubra 2 0 1 1 1 1 1 Crax globulosa 2 0 1 1 1 1 1 “Cracid” 2 2 1 0 1 0 Ortalis cf. erythroptera – Loma Alta (1): This large and slender proximal end with shaft is clearly Ortalis . In addition to overall proportions , the lack of a cranial pneumatic foramen is distinctive for this genus (as well as the large currasows).

PAGE 115

99 TIBIOTARSUS General Osteology The cracid tibiotarsus is a la rge bone with nearly parallel crista cnemialis cranialis et cnemialis lateralis, almost parallel trochlea in cranial view, a wide pons supratendineus, and a very prominent distal tuberculum retinaculi musculi extensorius. Table 2-7. Characters and character states fo r the distal end of the tibiotarsus in the Cracidae. Character Character States CrT1 Relative position of the margins of the cotyla medialis and lateralis in cranial aspect 0: parallel 1: at an angle CrT2 Shape of the distal exit of the canalis extensorius 0: oblong 1: oval 2: almost round CrT3 Foramen on the pons supratenineus 0: absent 1: present CrT4 Incisura intercondylaris 0: shallow 1: deep Comparative Osteology The only members of the Cracidae that are large enough to be clos e in size to the subfossil tibiotarsus are Aburria aburri , Penelope jacquacu , P. purpurascens , Nothocrax urumutum , Mitu salvini , and various species of Crax . Only specimens of P. purpurascens , A. pipile , C. globulosa , and C. rubra were available for comparison. I identified four characters on the distal end of the tibiotarsus to distinguish between five genera of cracids (Table 2-7). Aburria aburri – La Chimba (1): The La Chimba specimen is the fragment of a left tibiotarsus. It includes most of the condyles , a groove for the peroneus profundus, and the distal portion of the pons supratendineus on the cranial side. The caudal side is badly damaged. The bone is slightly larger than that of P. purpurascens and smaller than that of C. rubra . It compares best in size with C. globulosa . In terms of the four characters on

PAGE 116

100 the distal end of the tibiotars us (Tables 2-7 and 2-8, Figure 24), the fossil is most similar to that of Aburria pipile . I hypothesize that the bone is from the closely related A. aburri (Vaurie 1968, Grau et al. 2005). Figure 2-4. Cranial aspect of the three specimens of cracids, representing Aburria pipile (a), a large specimen from La Chimba probably belonging to Aburria aburri (b), and Penelope purpurascens (c). Reference bar is 2 cm. TARSOMETATARSUS General Osteology The cracid tarsometatarsus is a flat bone with a shallow sulcus flexorius and a very thin proximo-medial margin, just distal of co tyla medialis. Facies subcutanea lateralis is flat and at a sharp angle to th e sulcus flexorius. A flat ta rsometatarsus of a medium to large sized bird can only be found in the Craci dae. Other traits char acteristic of this family include a round and low eminentia in tercotylaris, bordered caudally by a cup shaped area intercotylaris, and a single sulcus hypotarsi flanked by a thin crista lateralis hypotarsi and a thick crista me dialis hypotarsi. Trochlea me tatarsi III is the largest trochlea and extends more distad than troc hlea metatarsi II and trochlea metatarsi IV, which are at the same level. Trochlea metatarsi II has a stubby and pointy wing.

PAGE 117

101 Table 2-8. Character matrix for the distal e nd of the tibiotarsus in four species (three genera) of Cracidae. See Table 2-7 for definitions of the character states. Character (CrT) Species 1 2 3 4 Penelope purpurascens 1 0 0 1 Penelope montagnii 0 1 0 1 Penelopina nigra 0 2 0 0 Ortalis leucogastra 0 0 0 1 Aburria pipile 0 1 1 1 Crax rubra 1 2 1 0 Crax globulosa 1 1 1 1 “Cracid” 1 0 1 1 1 Comparative Osteology Based on size, the tarsometatarsus from La Chimba cannot belong to any of the large-bodied cracids. Therefore, I will focus on the smaller species: Ortalis vetula , Penelope montagnii , and Aburria pipile . Overall proportions as we ll as the attachment for ligamentum rectum hallucis are informative taxonomically. Penelope montagnii – La Chimba (1): General proportions agree the most with P. montagnii . The tarsometatarsus is thicker in Aburria and more slender in Ortalis . In Ortalis the shaft is also narrower, reducing the flat appearance of the bone. The attachment for ligamenti rectum hallucis on the caudo-medial margin leading up to trochlea metatarsi II is well pronounced in both Penelope and Aburria , but not in Ortalis . Based on the few characters visible on the shaft and overall size and proportions, this tarsometatarsus belongs to Penelope montagnii . Comparison with a specimen from

PAGE 118

102 Chamaepetes goudotii is needed, but based on the relativ e proportions of the La Chimba femur and the tarsometatarsus of Aburria pipile , I believe that this species is unlikely. Order GRUIFORMES Family RALLIDAE CORACOID General Osteology The coracoid of the Rallidae is a highly distinctive bone. The sterno-dorsal portion is concave. Facies articularis sternalis is located on the dorsal side, with the medial and lateral ends of the proximal ma rgin being highly angular. Pro cessus lateralis is small and impressio musculi sternocoracoidei deep (lyi ng in concave proximal end). A fenestrate incisura nervi supracoracoidei is locate d at the proximal margin of processus procoracoideus, which is long and craniad proj ecting in this famil y. The distal end is characterized by a stubby, s hort extremitas omalis cora coidei and a nearly round impressio ligamenti acrocoracohumeralis. Comparative Osteology Only two species can be found in the size class (ca. 75 g) of the archaeological specimens from OGSE-80 and La Chimba. Th ey differ in overall proportions, shape of the processus procoracoideus, and the width of processu s acrocoracoideus. Porzana carolina – OGSE-80 (3): The three specimens are short bones with a wide shaft as P. Carolina . (In Rallus aequatorialis it is longer, with a mo re slender shaft.) In addition, the dorsal end of processus acroc oracoideus is thin (more pronounced and thicker in R. aequatorialis ). Processus procoracoideus, which is wide and blunt in R. aequatorialis and narrow and pointy in P. carolina , is missing from the three specimens.

PAGE 119

103 Rallus aequatorialis – La Chimba (1): A long, slender, nearly complete coracoid with a stubby dorsal tip of processus acr ocoracoideus, this specimen belongs to R . aequatorialis based on the characters just given above. HUMERUS General Osteology The rail humerus has a narrow and pointy cap ut humeri, a deep incisura capitis, a low, crista deltopectoralis, a stout tuberulu m dorsale, and a square-s haped distal end in cranial aspect (due to a flat epicondylus dor salis and tuberculum supracondylare ventrale being on the same level, parallel to axis proximodorsalis) with a shallow fossa musculi brachiale. Comparative Osteology Size distinguishes many species in this family. Once placed into a proper size class, subtle differences in shape of various structures can be used to differentiate even further, as discussed below. Porzana carolina – OGSE-80 (3): The three distal e nds of humeri share several characters with that of P. carolina . Fossa musculi brachialis is 1) shallow and 2) at a steep angle to axis proximodistalis (deep and at small angle in Rallus aequatorialis ). 3) Epicondylus ventralis does not protrude as much as in R. aequatorialis . Rallidae sp. – OGSE-80 (1): The proximal end of hu merus is heavily abraded and damaged, but can be identified as a rallid by th e distinctive pointy caput humeri and stout tuberculum dorsalis. In term s of size, it corresponds to Aramides , large Rallus , Gallinula , and a small Fulica . A round caput humeri and a flat margo caudalis (due to the caudal position of impressio musculi humerotricipitalis dorsalis [new term: a long depression

PAGE 120

104 just ventral of margo caudalis is the site of att achment of musculus triceps brachii]) exclude Aramides from further consideration. A short cr ista bicipitalis th at is at a steep angle to the shaft and a pointier caput humeri can be found in Rallus and Pardirallus , but not in Gallinula or Fulica . Three additional traits suggest that this specimen belongs to a species close to Pardirallus although other species within the Tribe Rallini (Livezey 1998) cannot be excluded: 1) a poin ty caput humeri (more pointy in Pardirallus than in any other rallid examined); 2) a less rais ed tuberculum dorsale (on pedestal in Rallus ); 3) a more rounded intumescentia humeralis; and 4) a proximally broader impressio tricipitalis. The identification is somewhat tentative because the abraded nature of the bone. ULNA General Osteology The Rallidae have a distinct ulna with a short olecranon, proximally strongly curving shaft, distally flatte ned shaft, shallow depressio radialis, and deep incisura tendinosa. Comparative Osteology Size, overall proportions, shape of impr essio brachialis, width of olecranon, location and size of foramen on the cranial si de of shaft, and shape and length of the proximal end of condylus dorsalis allow differe ntiating among species of similar sizes. Neocrex sp. – OGSE-80 (1): The distal end with half of the shaft is from a small species, larger than in large Laterallus , but smaller than in Porzana carolina or Rallus aequatorialis . Because of size and a bulky condylus ve ntralis ulnaris, which agrees with the closely related P. carolina (Livezey 1998), but not R. aequatorialis , I tentatively

PAGE 121

105 identify this specimen as Neocrex sp. Neocrex erythrops is the most likely species in western Ecuador based on size and distribution. Porzana carolina – OGSE-80 (1): This nearly complete ulna is slightly longer than that of modern P. carolina , but smaller than that of R. aequatorialis . Apart from length, overall proportions and curvatur e of the shaft in dorsal aspect agree well with that of P. carolina . In both the archaeological and modern specimen of P. carolina , the ventral margin of cotyla ventralis is at a greater angle to axis proximodistalis than in R. aequatorialis . All specimens I have examined are from Florida. Overwintering Florida birds migrate over smaller distances than those that reach northerwestern Peru. Difference in ulna length may reflect the tre nd in increased wing leng th for long-distance migrants (Marchetti et al. 1995). Porphyrula martinica – La Chimba (1): The distal end from the archaeological site corresponds in size to specimens of P. martinica , G. chloropus , and some Aramides . Fulica is considerably larger. Aramides and Gallinula have a more ventrally oriented proximal end of condylus dorsali s in cranial aspect than P. martinica . Aramides differs from Gallinula and Porphyrula by a flatter ventral margin of condylus ventralis in cranial aspect (sharper and more massive in Gallinula ). The two species can be separated on the basis of size (individuals of P. martinica with nearly the same weight as that of G. chloropus have ulnae with shorter and more slende r shafts) and the dist al extent of the condylus dorsalis ulnari s (short and blunt in P. martinica , long and pointy in C. chloropus ).

PAGE 122

106 CARPOMETACARPUS General Osteology The carpometacarpus in the Ra llidae has a characteristic shape due to a nearly symmetric curvature of os metacarpale minus (notably asymmetric in other families). Trochlea carpalis is round vent rally, but oblong dorsally. Proce ssus extensorius is slender and and at a steep angle to axis proximodist alis. Distally, the bone tapers due to the curvature of os metacapalis minus. Facies arti cularis digitalis minus et digitalis major lie on the same plane, perpendicular to axis proximodistalis. Comparative Osteology The distal end of a rallid carpometacarpus represents a small (ca. 75 g) bird. In western Ecuador, only Rallus aequatorialis and P. carolina are in this size class. Porzana carolina – OGSE-80 (1): The distal end (including most of os metacarpalis major) agrees with P. carolina in being proportionately more robust and having the caudal margin of the distal symphysis at a greater angle to axis proximodistalis (a small, proxi mal notch is also present in P. carolina , whereas in R. aequatorialis the symphysis is slightly concave proximally) than R. aequatorialis . TIBIOTARSUS General Osteology The proximal end of the tibiotarsus in the Rallidae is characterized by a proximally oriented crista cnemialis cranialis, larg e and expanded facies gastrocnemialis, and centrally thickened crista patellaris. The distal end is offset medially in relation to the shaft, with the lateral margin of condylus lateralis being at the same level, or more medial, than the lateral margin of the shaft. The distal end of sulcus extensorius and pons

PAGE 123

107 supratendineus are located on the medial half of the bone, being almost parallel to axis proximodistalis (and not oblique as in many other families). Condylus medialis is much thicker than condylus lateralis in cranial aspect, and the distal margin of condylus lateralis is flattened (not round). Comparative Osteology Size, stoutness, and the shape of crista cnemialis cranialis and sulcus extensorius are useful characters to separate most species. Porzana carolina – OGSE-80 (5): By size, the five distal ends (some quite fragmentary) belong to P. carolina or R. aequatorialis . All specimens have a broad sulcus extensorius as in P. carolina . In contrast, R. aequatorialis has a narrow sulcus extensorius, resulting in a flat surface just la terally to sulcus muscu li fibularis. None of the archaeological specimens possesses this feature. Fulica ardesiaca – La Chimba (1): A bulky crista patell aris also agrees with Fulica . In terms of size, the La Chimba specimen is only slightly smaller than the modern specimen of a female F. ardesiaca (UF 39760), but is much larger than a male F. americana (UF 16940). Size and a flat cranial surface suggest that the proximal end of tibiotarsus is from the genus Fulica (smaller in Gallinula , cranial surface rounded in Rallus and Aramides ). TARSOMETATARSUS General Osteology Cotyla lateralis is slightly distal to co tyla medialis. The dorsomedial passage of musculus flexor digiti longus c onsists of closed canals, a nd the dorsolateral passage of musculus flexor hallicus longus consists of open sulci (inst ead of closed canals; Livezey

PAGE 124

108 1998). Facies dorsalis is flat, a nd the distal margin of trochl ea metatarsi II is located proximad of trochlea metatarsi IV (roughly at same level as foramen distale). Comparative Osteology Based on size, the two specimens from the archaeological site are either Rallus aequatorialis or Porzana carolina . Porzana carolina – OGSE-80 (2): The two proximal ends agree with P. carolina by having an excavated and broad area in tercotylaris and a narrow hypotarsus in proximal view. Area intercotylaris is na rrow and not deeply excavated, and the hypotarsus is broad in proximal aspect in R. aequatorialis . OSSA DIGITORUM PEDIS General Osteology Pedal phalanges in the Rallidae are relativel y longer than those of small Ardeidae, but a little shorte r than those of Jacana . Small herons have more highly expanded proximal ends of phalanx 1 of all digits. Jacana has a proximal groove on facies plantaris that divides the proximal end into two nearly symmetrical halves. Comparative Osteology Overall size and proportions can be used to distinguish among most genera. In addition, the degree of ossification and size of a pair of tuberosities serving as muscle attachment sites for musculus abductor digi ti II, musculus adductor digiti II, musculus extensor proprius III, and musculus extens or brevis digiti III, respectively, on the proximal end are useful characters.

PAGE 125

109 Table 2-5. Phalanges 1 of digitus segundus (bo ttom row) and digitus tertius (top row) in lateral aspect from OGSE-80 specimens (a), Rallus longirostris (b), Aramides cajanea (c), Porphyrula martinica (d), and Gallinula chloropus (e). Reference bar is 1 cm. Rallus cf. longirostris – OGSE-80 (7): The seven specimens of ossa digitorum pedis include whole a nd partial elements. I base the identification mostly on the best preserved elements, phalanges 1 of digitus segundus and digitus tertius, respectively (Figure 2-5). Only the phalanges of R. longirostris agree in terms of overall size and proportions, being slender, intermediate in total length, and interm ediate in proximal length (as judged by the distan ce between the proximo-dorsal margin and the plantar tuberosities). Phalanges in Aramides cajanea are more robust, with a wider proximal end. In contrast, Gallinula chloropus and Porphyrula martinica have long and strong pedal

PAGE 126

110 phalanges. Whereas phalanges of Gallinula have highly excavated proximal grooves on facies plantaris, a ve ry short proximal end characterizes those of Porphyrula martinica . Order CHARADRIIFORMES Family SCOLOPACIDAE HUMERUS General Osteology The humerus of the Charadriiformes is highly characteric wi th a wide and long fossa pneumotricipitalis, low crista deltopectoralis, promin ent tuberculum dorsale et ventrale, and long and expanded processus su pracondylaris dorsalis (a similar condition is only seen in the Pr ocellariiformes; Baumel et al. 1993, Figure 4.12). Comparative Osteology Albeit similar, the Stercorariidae, La ridae, and Sternidae tend to have proportionately longer wing elements with a relatively longer proximal end and a more pronounced processus supracondylaris dorsalis than the other charadriiform families. Assigning archaeological or pale ontological specimens to a pa rticular size-class and then comparing the specimens within size-classes greatly facilitates identification in this diverse order (see Campbell 1979). Characters that have proven to be useful are the shape of caput humeri, depth and proximal extent of fossa pneumotricipitalis, width of incisura capitis, shape and size of tuberculum dorsali s et ventralis, shape of area between margo caudalis and crista deltopectoralis in caudal view, areal extent of impressio coracobrachialis, size, shape, and position of processus supracondylar is dorsalis, depth and location of fossa musculi brachialis, relative sizes of c ondylus dorsalis and condylus ventralis, the shape and thic kness of epicondylus ventralis , the orientation of the

PAGE 127

111 attachment for ligamentum collaterale dorsalis and musculus flexor ca rpalis ulnaris, and shape and size of processus flexorius among ot hers (Table 2-9; see Brodkorb [1955] and Campbell [1979] for more traits of taxa not treated here). Tringa melanoleuca – La Chimba (1): The distal end of a humerus is in the size range of that in T. melanoleuca and Bartramia longicauda . For comparison I considered other smaller or larger species as well, including Catoptrophorus semipalmatus , Heteroscelus incanus , and Limnodromus griseus among others. Given the combination of traits in Table 2-10, I conclude that the specimen belongs to T. melanoleuca . Table 2-9. Characters and character states fo r the humerus in the family Scolopacidae. Character Character States ChH1 Width of proximal end 0: narrow 1: wide ChH2 Depth and extent of fossa pneumotricipitalis 0: shallow and proximal of caput humeri 1: deep, often undercutting caput humeri ChH3 Prominence and orientation of tuberculum dorsale 0: not prominent, at angle to axis proximodistalis 1: prominent, parallel to axis proximodistalis ChH4 Crus dorsale (thickening in pneumatic fossa, often ventral to crus dorsale sensu Baumel et al. [1993]) 0: absent 1: present ChH5 Fossa musculi brachialis 0: deep dorso-ventrally 1: deep throughout ChH6 Area between proximal extent of epicondylus ventralis and attachment of musculus flexor carpalis ulnaris 0: long and slightly concave 1: short and deeply concave ChH7 Relative size of condylus ventralis to condylus dorsalis 0: small 1: large ChH8 Size and orientation of attachment of musculus flexor carpalis ulnaris 0: small, oriented ventrally 1: large, oriented cranio-ventrally ChH9 Processus flexorius 0: straight or round bulge on ventral margin 1: sharp bulge on ventral margin Arenaria interpres – Valdivia Village (2): A proximal and a distal end, probably from the same bone, agree best with A. interpres (Table 2-10). Gallinago sp. – La Chimba (1): The distal end of humerus from La Chimba is from medium-sized scolopacid that agrees well with that in Gallinago (Table 2-10). Three

PAGE 128

112 species of large Gallinago occur in the northern Andes of Ecuador, G. nobilis , G. jamesoni , and G. imperialis . Without additional compara tive material, this specimen cannot be identified to the species level. Table 2-10. Character matrix for the humerus in 8 species (8 gene ra) of Scolopacidae. Character (ChH) Species 1 2 3 4 5 6 7 8 9 Tringa melanoleuca 1 1 1 0 0-1 1 1 1 0 Catoptrophorus semipalmatus 1 1 1 0 1 1 1 1 1 Heteroscelus incanus 1 0-1 1 0 0 0 0 1 0 Bartramia longicauda 0 1 0 0 1 0 0 0 1 Arenaria interpres 1 1 1 1 0 1 0 1 0 Calidris canutus 1 1 0 0 1 1 1 0 0 Limnodromus griseus 1 0 1 0 1 0 0 0 0 Gallinago gallinago 1 1 0 0 1 1 0 0 1 La Chimba 1 0-1 1 1 1 0 La Chimba 2 1 1 0 1 Valdivia Village 1 1 1 1 0 0-1 0 1 Family THINOCORIDAE HUMERUS General Osteology The Thinocoridae have a humerus with wide proximal and distal end, narrow fossa pneumotricipitalis, low tuberculum dorsale, narrow and deep incisura capitis that undercuts caput humeri, shallow fossa musc uli brachialis (Campbe ll 1979), and a short processus supracondylaris dorsalis.

PAGE 129

113 Comparative Osteology Only two genera, each represented by one species, occur in Ecuador. They can be distinguished by size alone ( Attagis much longer than Thinocorus ). Attagis gayi – La Chimba (1): Despite the lack of comparative material, this stout, complete humerus is a thinocorid which by its large size can only be A. gayi . An unusual charadriiform sternum (fused spina externa et interna, labrum internum curving up dorsally toward the center of the bone) may al so belong to this species. Because of the distinctive features characteriz ing this bone, I only can spec ulate on its identity based on the available comparative material. Family LARIDAE CORACOID General Osteology The larid coracoid is typical for the Ch aradriiformes in having a concave proximal end (in dorsal aspect) and a ve ntrally rotated processus acro coracoideus. Gulls and terns differ from other Charadriiformes in having longer and more slender coracoids with a fenestrate processus procoracoideus. Comparative Osteology The comparative osteology of northern Pe ruvian gulls was decribed in detail by Campbell (1979), who pointed out that th ree of the nine Peruvian species ( Larus belcheri , L. dominicanus , and Creagrus furcatus ) can be separated by large size (in Ecuador L. argentatus , L. fuscus , and L. californicus overlap in size with L. belcheri and C. furcatus ) and one by small size ( Xema sabini ). The specimens from Va ldivia Village and Loma Alta are in a size class containing L. modestus , L. delawerensis , L. cirrocephalus , L.

PAGE 130

114 serranus , L. atricilla , and L. pipixcan . To distinguish among these species Campbell (1979) used five characters (Table 2-11). Ba sed on comparisons with modern specimens, L. delawarensis and L. serranus are probably too large for th e archaeological specimens. Table 2-11. Characters and character states for the coracoid of simila rly-sized species of Larus , summarized from Campbell (1979). Character Character States ChC1 Processus acrocoracoideus (= “anterior projection of head”) 0: short 1: long ChC2 Facies articularis humeralis (= “glenoid facet”) 0: elliptical 1: wide and round ChC3 Concavity where musculus biceps attaches 0: shallow 1: deep ChC4 Cotyla scapularis (= “furcular facet ”) 0: elliptical of moderate length 1: long and broad ChC5 Impressio ligamenti acrocoracohumeralis (= “attachment of ligamen tum humero-coracoideum anterius superius”) 0: elliptical, deep, and narrow 1: small, round to long oval Larus pipixcan – Loma Alta (1), Valdivia Village (1): A high degree of intraspecific variation makes identification of these complete coracoids problematic (Campbell 1979). Among the smaller species of Larus ( L. cirrocephalus , L. atricilla , L. pipixcan ), L. pipixcan is unique in having a long pro cessus acrocoracoideus and round impressio ligamenti acrocoracohumeralis (long and oval in L. cirrocephalus and L. atricilla ). Unfortunately, a modern specimen of L. cirrocephalus was not available and I rely on Campbell (1979) for the ch aracter matrix (Table 2-12). HUMERUS General Osteology Compared to other Charadriiformes, the Laridae have a longer humerus with a long crista deltopectoralis. Pro cessus supracondylaris dorsalis is large and fossa musculi brachialis is wide.

PAGE 131

115 Table 2-12. Character matrix for the coracoid in six species of Larus . See Table 2-11 for definition of character states. Character ChC Species 1 2 3 4 5 Larus modestus 1 1 1 0 1 Larus delawarensis* 1 0 0 1 1 Larus cirrocephalus 1 0 0 0 0 Larus serranus 0 1 1 1 1 Larus atricilla 0 0 0-1 0 0 Larus pipixcan 0-1 0-1 0-1 0 1 Loma Alta specimen 1 0 0 0 1 Valdivia Village specimen 1 1 1 0 * Not studied by Campbell (1979). Characters scored by author. Comparative Osteology Campbell (1979) has identified four traits on the distal end that separate the five Peruvian species (Table 2-13). Table 2-13. Characters and character states for the humerus of similarly-sized species of Larus , summarized from Campbell (1979). Character Character States ChH10 Distal end 0: laterally compressed with small condyles 1: laterally compressed with condyles of moderate size 2: laterally compressed with large condyles 3: broad with wide condyles ChH11 Incisura intercondylaris (= “intercondylar groove”) 0: moderately deep 1: deep 2: very deep ChH12 Fossa musculi brachialis (= “brachial depression”) 0: moderately deep 1: deep 2: very deep ChH13 Epicondylus dorsalis (= “ectepicondylar prominence”) 0: large, projecting palmarly almost perpendicular to palmar surface 1: as in 0, but broader and shorter or larger 2: moderate size, projecting at an angle to palmar surface

PAGE 132

116 Larus pipixcan – Valdivia Village (1): Although I find certain characters given by Campbell (1979) difficult to use (especially with regard to epicondylus dorsalis), I suggest that the distal end of humerus from Vald ivia Village belongs to L. pipixcan (Table 2-14). The small condyles and deep inci sura intercondylaris appear to distinguish this specimen well from the humeri of L. modestus and L. atricilla . Table 2-14. Character state matrix for the humerus in six species of Larus . See Table 213 for definitions of the charcter states. Character ChH Species 10 11 12 13 Larus modestus 3 2 2 1 Larus delawarensis* 3 1 0 1 Larus cirrocephalus 3 1 2 1 Larus serranus 2 1 2 1 Larus atricilla 1 0 1 0 Larus pipixcan 0 1 0 2 Valdivia Village specimen 0 1 0 * Not studied by Campbell (1979). Characters scored by author. ULNA General Osteology In ulnae of the Laridae, the olecrano n is short and stubby, tuberculum ligamenti collateralis ventralis long and pronounced, inci sura radialis deep, and impressio brachialis wide. The distal end has a na rrow trochlea ulnaris and a pr oximally located tuberculum carpale.

PAGE 133

117 Comparative Osteology Campbell (1979) provided no characters for th is element. The three small species of Larus that compare best in size to the archaeological specimen are L. modestus , L. atricilla , and L. pipixcan . Of these, L. modestus is not available, but given the large differences found in the other elements (Cam pbell 1979), I assume that it is distinctive. The depth of impressio brachialis, the sh ape of the proximal surface of tuberculum ligamenti collateralis ventralis, and the spaci ng between papilla remigialis are useful traits. Larus pipixcan – Valdivia Village (1): As with the associated bones, the proximal end of ulna appears to be that of L. pipixcan . It differs from that in L. atricilla by having a deep impressio brachia lis (shallower in L. atricilla ), short and round proximal surface of tuberculum ligamenti collateralis ventralis (longer and more oval in L. atricilla ), and short spacing between papilla re migialis (spacing greater in L. atricilla ). CARPOMETACARPUS General Osteology The carpometacarpus of the Laridae has n early parallel os metacarpi major et minus. Processus extensorius is broad and flat . Trochlea infracondylaris tends to be deep and fovea carpalis cranialis sh allow. The dorsal and ventral rims of trochlea carpalis are round to oval. Facies articularis digitalis major is very broad. Comparative Osteology Diagnostic characters have been suggested by Campbell (1979; Table 2-15). Larus pipixcan – Valdivia Village (1): The proximal end of carpometacarpus does not agree in all characters with that in L. pipixcan , especially the groove in trochlea

PAGE 134

118 carpalis. This trait may, however, depend to a la rge degree on the leve l of ossification of the rims of trochlea carpalis and thus be qu ite variable. Consideri ng its association with other bones that have been identified as L. pipixcan and agreement in most traits, I classify this specimen as L. pipixcan . Table 2-15. Characters and character states for the carpometacarpus of similarly-sized species of Larus , summarized from Campbell (1979). Character Character States ChCMC1 Fossa infratrochlearis (= “internal ligamental fossa”) 0: shallow 1: deep ChCMC2 Groove in trochlea carpalis 0: moderately deep 1: deep ChCMC3 Dorsal slant of ventral rim of trochlea carpalis 0: little 1: moderate 2: great ChCMC4 Shaft depth immediately distal to processus alularis (= “pollical facet”) 0: small 1: moderate 2: great TARSOMETATARSUS General Osteology The tarsometatarsus of the Laridae is slende r and short with a str ong shaft that is as wide as it is deep. Cotyla la teralis et medialis are wide a nd hypotarsals form a series of canals and sulci. Foramen vasculare distale is wide and trochlea metatarsi II does not project distally as much as trochlea metata ri IV (trochlea metata rsi III is longest). Superficially, the tarsometatarsi of other charadriiform families, the Rallidae, and Procellariiformes look similar. Comparative Osteology Size alone allows placing the archaeological specimen into a group of five species of terns, including Sterna (= Gelochelidon ) nilotica , S. sandvicensis , S. elegans , S. hirundinacea , and Larosterna inca . Overall dimensions and pr oportions are sufficient to

PAGE 135

119 tell most of these species apart. Also, in so me of these species facies dorsalis is deeply excavated proximally, and trochlea metatarsi III shows a slight dorsad rotation. These traits are especially useful when only th e proximal or distal end is available. Table 2-16. Character matrix for th e carpometacarpus in six species of Larus . See table 215 for definitions of character states. Character ChCMC Species 1 2 3 4 Larus modestus 1 1 1 2 Larus delawarensis* 1 0 2 2 Larus cirrocephalus 0 0 0 1 Larus serranus 1 1 2 1 Larus atricilla 1 0 1 2 Larus pipixcan 1 0 2 0 Valdivia Village specimen 1 1 1-2 0 * Not studied by Campbell (1979). Characters scored by author. Sterna cf. hirundinacea – Valdivia Village (1): The distal end of a tarsometatarsus has a thicker shaft than in S. nilotica , but thinner than that of L. inca . Trochlea metatarsi III is located more dorsally than in S. nilotica and L. inca . Although the shaft of the archaeological specimen is slender, the di stance between foramen vasculare distale and the distal extent of trochlea metatarsi III in this specimen is similar to that of S. maxima . None of the other species has such a long dist al end. The combination of a fine shaft and a long distal end suggests that this species comes from a longlegged species. I tentatively assign this specimen to S. hirundinacea , which, among terns in Ecuador (Ridgely and Greenfield 2001a,b) has the pr oportionately longest legs.

PAGE 136

120 Order COLUMBIFORMES Family COLUMBIDAE STERNUM General Osteology Large pectoral muscle masses have resulted in very deep carina sterni in this family. Spina interna is broad, stout, roughly triangular in outline, and much wider than spina externa. Sulcus articu laris coracoideus is broad. A large foramen pneumaticum and short processi craniolaterales characterize pars cardiaca, and long trabeculae laterales and a pair of fenestra mediales are distinctive char acters for pars hepatica (see Baumel et al. 1993, Fig 4.11). Comparative Osteology Only minor differences exist among the st erna of Ecuadorian Columbidae. These include the lengths of proce ssus craniolateralis and trabecula lateralis as well as the degree of ossification of margo caudalis (resulting in a closed fenestra medialis in some species and open incisura medialis in othe rs). In archaeological middens, the most commonly preserved part of the bone is th e most cranial portion of pars cardiaca, including spina interna et externa (often bor ken), sulcus articularis coracoideus, pila carinae, and foramen pneumaticum. The best di stinguishing characters are size, width of sulcus articularis coracoideus, and shape of foramen pneumaticum. Columba speciosa – La Chimba (1): The cranial fragment of sternum has a wide sulcus articularis coracoideus typical for the genus Columba. Compared to specimens from C. fasciata monilis Vigors 1939 (UF 32305, UF 32306, UF 32307), the archaeological specimen and two museum specimens of C. speciosa (UF 13618, UF

PAGE 137

121 38772) have a narrow and short spina inte rna rostri (making spina interna look proportionately broader than in C. fasciata ), a more curved sulcus articularis coracoideus, a more anteriad curved pila carina, a nd a narrower foramen pneumaticum. Although the foramen pneumaticum is wide in a small specimen of C. speciosa (UF 25412) and narrow in a specimen of C. f. fasciata Say 1823 (UF 32308), the spina interna rostri is consistently wider in C. fasciata than in C. speciosa . Columba plumbea can be ruled out based on a significantly smaller and propor tionately wider spina interna rostri. Zenaida auriculata – La Chimba (2), El Azúcar (3), OGSE-80 (5): Among Columbidae of overlapping size ( Zenaida , Leptotila , Geotrygon , small Columba ), Zenaida auriculata stands out in having a narrow and long spina interna ro stri relative to the length of complete spina interna. Spina interna rostri is intermediate in length and width in Z. asiatica (specimen UF 40008 of a large Z. asiatica was used, as no specimen of Z. meloda was available), broad and short in L. verreauxi , very broad and short in G. montana , and intermediate in length but broad in C. cayannensis . Zenaida meloda – El Azúcar (1): By the shape of spina interna rostri (see above), this cranial fragment can be assigned to Z. meloda . Columbina sp. – OGSE-80 (1): The very small cranial fragment is referred to Columbina by size. The sterna in Columbina picui (closest relative of C. cruziana ) and C. buckleyi are very similar; I have not been able to find consistent differences. SCAPULA General Osteology This bone is characterized by a nearly r ound facies articularis humeralis, a dorsally broadening proximal end in proximal aspect with a flat acromion, a pronounced crista

PAGE 138

122 ligamenti acrocoracoacromiali, and an initially widening corpus scapul ae that tapers in the distal third of the bone to end in a pointy tip. Comparative Osteology The proximal end in proximal view has two useful characteristics, the shape of the midsection of the proximal scapula and the outli ne of acromion (which is related to the position of crista ligamenti acrocoracoacrom iali). The presence of pneumatic foramina just ventral to acromion is a less reliable trait. Zenaida auriculata – OGSE-80 (3), El Azúcar (9), La Chimba (1): Having a narrow midsection of the proximal end, the acromion in Z. auriculata expands medialy. The dorsal margin appears flat as crista lig amenti acrocoracoacromiali is situated on the medial half of acromion. Geotrygon montana and Leptotila verreauxi are similar in outline except for the dorsal margin which is slightly more pointed in Leptotila due to a central position of crista ligamenti acrocora coacromiali. These two species also differ by a pneumatic foramen which is present only in G. montana (not present in the larger G. frenata from the Andes). Zenaida meloda – El Azúcar (7): Similar to the above but larger, this dove differs from Columba cayennensis in having a constr icted midsection. In C. cayennensis the lateral and medial margins are at a small angle causing the proximal end to become progessively wider dorsad. Acromion bears a pneumatic foramen in Columba as well. Columbina minuta – El Azúcar (2): Very small size is di agnostic for this species. Columbina cruziana – El Azúcar (6): This small grounddove has a stout proximal end with a thick midsection in proximal vi ew and a short medial process of acromion

PAGE 139

123 (sensu Höfling and Alvarenga 2001). Columbina buckleyi has a narrow midsection and a large medial process of acromion. CORACOID General Osteology The coracoid of the Columbidae is highly distinctive from that of other families, but difficult to distinguish w ithin the family. The proximal end of the bone has a saddleshaped facies articularis sternalis, long a nd pointy processus latera lis, and highly textured impressio musculi sternocoracoidei. Small fo ramina can be found in most specimens at the base of the impressio musculi sternocoraco idei. On the distal end extremitas omalis coracoidei is stout and compact. The flattened tuberculum brachiale is close to the main body of the bone. Processus procoracoideus is flattened and expanded towards its tip. Impressio ligamenti acrocoracohumeralis is broad and at a ca. 45 angle to axis proximodistale. The line of attachment for me mbrana sternocoracoclavicularis is slightly raised just proximad of pr ocessus procoracoideus. Comparative Osteology The coracoid of the Columbidae is very homogeneous in morphology, and similarly-sized species are often difficult to identify. Thus, apart from size, which only allows for identification of the largest ( Columba ) and smallest ( Columbina ) genera, only a few traits are consisten tly useful (Table 2-17). Zenaida auriculata – OGSE-80 (15), El Azúcar (17), La Ponga (1), Loma Alta (2), La Chimba (2): A fairly thick shaft, processus acrocoracoideus bearing two dorsal apices in distal aspect, and a wide and pronounced attachment of membrana sternocoracoclavicularis proximal to processu s procoracoideus, distinguish this genus

PAGE 140

124 from other genera of similar size. The proximal end in Z. auricularis differs from that in Z. meloda in having linea intermuscularis join the lateral end of facies articularis sternalis. Table 2-17. Diagnostic characters and char acter states for the coracoid in the Columbidae. Character Character State CoC1 Area of proximal shaft laterad of linea intermuscularis ventralis in relation to the width of the shaft 0: Narrow 1: Broad CoC2 Proximal end of linea intermuscularis ventralis 0: lateral of lateral edge of facies articularis sternalis 1: coinciding with lateral edge of facies articularis sternalis CoC3 Membrana sternocoracoclavicularis proximal of processus procoracoideus 0: Wide and pronounced 1: Narrow and only slightly raised CoC4 Dorsal side processus acrocoracoideus in distal aspect 0: Moderately thin with one dorsal apex 1: Thick with two dorsal apices CoC5 Thickness of shaft* 0: Slender 1: Stout * Difficult to see in incomplete specimens. Zenaida meloda – OGSE-80 (1), El Azúcar (8), La Ponga (1): In addition to the above mentioned qualitative traits, Z. meloda is considerably larger than Z. auriculata , often reaching the size of C. cayennensis from which it differs considerably (Table 2-18). Columbina minuta – El Azúcar (1): Of the five genera found in Ecuador, Columbina is the most distinct because of its small size. Within the genus, C. minuta is significantly smaller than the similarly-sized C. buckleyi and C. cruziana . Columbina buckleyi – El Azúcar (3): Despite being similar in size to C. picui , the coracoid of C. buckleyi is much more slender and has a shallower processus acrocoracoideus due to a more slender impr essio acrocoracohumera lis. Despite having a narrower shaft, the neck (sensu Howard 1929) is as wide as or wider than that of C. picui . Columbina cruziana – OGSE-80 (2), El Azúcar (14): This species has an overall stronger coracoid than C. buckleyi (see above).

PAGE 141

125 Table 2-18. Character matrix for the coracoid in nine species of Co lumbidae. Characters are defined in Table 2-17. Character CoC Species 1 2 3 4 5 Columba cayennensis 0 0 0 0 1 Zenaida auriculata 1 1 0 1 1 Zenaida meloda 1 0 0 1 1 Columbina minuta 1 1 0 0 0 Columbina buckleyi 1 1 0-1 0-1 0 Columba cruziana 1 0 0 0-1 1 Claravis pretiosa 1 1 1 1 0 Leptotila verreauxi 1 1 1 0 0-1 Geotrygon montana 0 1 1 0 0 HUMERUS General Osteology The columbid humerus is a stout bone with several salient features. The proximal end has a large, triangular deltoid crest that is almost perpendicular to the intumescencia humeri. Incisura capitis is wide and tubercul um ventralis pronounced. At the distal end, processus supracondylaris dorsalis is located proximad of the proximal end of the brachial depression. Comparative Osteology As is the case with other elements, the humerus is a very uniform bone. Size differences are enough to classify some genera without furthe r analysis of other traits. General proportions regarding the thickness of the shaft and caput humeri are also useful. On the distal end the location of impressio coracobrachiale, the width of intumescentia

PAGE 142

126 humeri, and the orientation of the proxima l margin of crista deltopectoralis are taxonomically informative. The distal end has on ly one consistent character, the thickness of the bony ridge between condylus ve ntralis and processus flexorius. Columba cf. plumbea – La Chimba (1): The large proximal end corresponds in size to a small Columba . Leptotila and Zenaida asiatica are not as large and have a shorter caput humeri, giving the distal end proxi mo-distally compressed look compared to Columba . In addition, impressio coracobrachialis lies more centrad on the shaft in these two genera, whereas in Columba this impressio is located more dorsad, mostly on crista deltopectoralis. Geotrygon frenata compares well in size, but has a very deep concavity proximad of impressio coracobrachialis and a centrad impressio co racobrachialis. Among Columba , C. fasciata and C. speciosa are very large. Columba cayannensis and C. plumbea are within the size range of the archaeological specimen, but C. plumbea has fine and narrow caput humeri for this genus. Zenaida auriculata – OGSE-80 (15), El Azúcar (17), La Chimba (6): This dove has a humerus that overlaps in size with that of Leptotila verreauxi and Geotrygon montana . Proximally, it has a rounder condylus ventrali s, and thin caput humeri (thicker in L. verreauxi and intermediate in G. montana ), intumescencia humeri is narrow (also narrow in L. verrauxi , thick in G. montana ), impressio coracobrachiale is located mostly on crista deltopectoralis (located mostly on shaft in L. verreauxi and G. montana ), and the proximal margin of crista deltope ctoralis is vertical to axis proximodistale or at a small angle pointing proximad (point ing slightly distad in L. verreauxi and G. montana ). Zenaida meloda – OGSE-80 (4), El Azúcar (5): Larger than Z. auriculata , this closely related dove has a thick ridge between condylus ventralis and processus flexorius,

PAGE 143

127 which is even proportionatel y thicker than that of C. cayennensis . A thin caput humeri in Z. meloda further differentiates this species from C. cayennensis . The same traits that distinguish Z. auriculata from L. verreauxi also apply to the comparison of Z. meloda and L. verreauxi . Columbina minuta – El Azúcar (2): Tiny size is diagnos tic for this species. Columbina buckleyi – El Azúcar (2): As for other elements, the humerus of C. buckleyi is significantly more slender a nd less massive than that of C. picui (and presumably its sister species C. cruziana ). Columbina cruziana – El Azúcar (6): This species has an overall stronger humerus than C. buckleyi (see above). ULNA General Osteology The columbid ulna is a strong, highly curved bone similar in overall shape to that of the Psittacidae. Tuberculum ligamenti collate ralis ventralis is a slightly raised and broad surface, trochlea humeralis ulnaris is deep and well developed, and condylus ventralis and condylus dorsalis are nearly parallel in vent ral view, being separated by a narrow sulcus intercondylaris. Comparative Osteology This element has only few taxonomically in formative characters, including size, general proportions, the shape of tuberculum bicipitale, and th e shape and relative size of tuberculum carpale. Zenaida auriculata – OGSE-80 (17), El Azúcar (13), Loma Alta (2), La Chimba (3): Zenaida auriculata differs from Leptotila verreauxi by having a short and broad

PAGE 144

128 tuberculum bicipitalis ul naris (long and slender in L. verreauxi ) and a deep incisura tuberculi carpale. Geotrygon has a porportionately longer and more slender ulna than Z. auriculata , but when only proximal and distal en ds are preserved, these species are almost impossible to sepa rate osteologically. The distal end of ulna in Z. auriculata has a deeper incisura tuberculi carpale than that in G. montana . Interestingly, G. frenata has an exceptionally short ulna. Zenaida meloda – El Azúcar (4): Except for size and a shallower incisura tuberculi carpale, the ulna in Z. asiatica is very similar to that of Z. auriculata . It also is much larger than in L. verreauxi and G. montana . Columbina minuta – El Azúcar (2): Small size is dia gnostic for this species. Columbina buckleyi – El Azúcar (1), Valdivia Village (1): This ground dove is similar to C. picui , but more slender. Tuberculum carpale is noticeably thinner. Columbina cruziana – El Azúcar (15): This species has a more robust bone than C. buckleyi . RADIUS General Osteology The radius is a stout bone in the Columbidae compared to most other families. Being located slightly caudally on the proximal shaft, tuberculum bicipitalis radialis is a low circular ridge enclosing a shallow and roughly circular concave depression. Capital tuberosity (sensu Howard 1929) is delim ited caudally by a deep incision formed by meniscus radioulnaris. Just di stal to this incision lies the attachment for ligamentum transversum radioulnare. The shaft has severa l raised lineae intermusculares, giving the

PAGE 145

129 shaft an angular aspect. The distal end is cu rved ventro-caudad. The cranial margin of the distal end flares out dorso-cr anially. Tuberculum aponeurosis ventralis is thin and long. Comparative Osteology Size and overall proportions are the taxonom ically most informative characters. The distal end also shows variation in the degr ee of curvature of the shaft and the location and depth of a ventral depression just proxima d of facies articularis radiocarpalis. Zenaida auriculata – OGSE-80 (1), El Azúcar (3): The radius in this dove differs from that of G. montana in being stouter overall and by having a much broader distal end with a deep ventral depression, which is absent in G. montana . Despite being very similar, the radius of Leptotila is more slender and has more widely separated lineae intermusculares. Zenaida meloda – El Azúcar (4): The radius of this species is unusually short and stout. The proximal end approaches that of Columba cayennenis in width, but differs in having less excavated attachment sites fo r meniscus radiouln aris and ligamentum transversum radioulnare. CARPOMETACARPUS General Osteology In the carpometacarpus of the Columbidae, processus extensorius is about half as long as the proximal symphysis. The processus alularis is narrow in the middle, giving the distal end of processus extensorius a squared appearance. All columbids have a pronounced and deep fovea carpalis craniali s. The attachment for the ligamentum ulnocarpometacarpalis ventralis extends to the proximal end of os metacarpale minus. Processus pisiformes is well pronounced. Sulc us tendineus extends from a dorso-cranial

PAGE 146

130 position on the proximal end to a dorso-ca udal position on the distal end, where two tuberosities flanking the sulcus almost form a canal. The distal end is characterized by facies articularis digitalis minus extending slig htly more distad than facies articularis digitalis major. Comparative Osteology Size, orientation of processus extensoriu s, and the shape of facies articularis digitalis majus are the only re liable traits I have found for similarly sized columbids in the genera Zenaida , Leptotila , and Geotrygon . For Columbina only the attachment for ligamentum radiocarpometacarpal e dorsale is consistent. Zenaida auriculata – OGSE-80 (2), El Azúcar (15), La Chimba (1): Compared to Leptotila verreauxi , Z. auriculata has a less ventrad oriented processus extensorius. The main trait distinguishin g the carpometacarpi in Z. auriculata from that of G. montana is a narrower cranial facies articularis digitalis ma jus (the facies being divided into a wide cranial facet and a round to oval caudal facet). In addition, the carpometacarpus of G. montana is slightly more slender a nd has a narrower proximal end. Zenaida meloda –El Azúcar (2): Apart from size, the carpometacarpus in this species is similar in all aspects to that of Z. auriculata . It differs in C. cayennensis in having a more oval (less rounded) ventral rim of trochlea car palis and a slightly thinner shaft in ventral aspect. Columbina minuta – El Azúcar (3): Small size is dia gnostic for this species. Columbina cruziana – El Azúcar (7): Columbina picui (in lieu of C. cruziana ) differs from C. buckleyi by having a thicker and stouter bone in all aspects. In addition,

PAGE 147

131 the attachment for ligamentum radiocarpometaca rpale dorsale is located on a flat surface whereas in C. buckleyi the attachment sits on the caudal si de of a slightly raised surface. FEMUR General Osteology Usually a long and slender bone in the Co lumbidae, the femur has a high crista trochanteris, angled facies articularis antitrochanterica with respect to axis proximodistalis, concave and slightly pneumatized fossa poplitea, long crista tibiofibularis compared to the lateral rim of trochlea fibularis, and acuminate crista supracondylaris caudalis. Condylus medialis an d condylus lateralis ha ve the same distal extent in cranial view. Comparative Osteology Apart from differences in size and gene ral proportion, I have found only the shape of impressiones obturae, the proximal thicke ning of the shaft, and a concave ligamental attachment (sensu Howard 1929) on the medial si de of condylus medialis to be consistent characters. High levels of intr aspecific variation obscure mo st interspecific differences. Columba cf. cayennensis – La Ponga (1): The almost complete femur is a long and stout bone with highly raised impressiones obturatoriae (giving th e proximal shaft a nearly triangular cro ss-section) as found in C. cayennensis and other species of Columba . Although the femur of Leptotila verreauxii is almost as long, the bone is much more slender. Zenaida auriculata – OGSE-80 (7), El Azúcar (11), La Ponga (1): This dove has a short and slender femur. This element is considerably longer in all other species ( C.

PAGE 148

132 cayennensis , L. verreauxi , G. montana ) or larger, but identical in overall proportions ( Z. asiatica ). Zenaida meloda – El Azúcar (2), Loma Alta (1): The femur in this species is very similar to that in Z. auriculata but is larger. Columbina minuta – El Azúcar (1): Small size is dia gnostic for this species. Columbina buckleyi – El Azúcar (1): As with all other skeletal elements, the femur of C. buckleyi has a much more slender build. Columbina cf. cruziana – El Azúcar (3): The femur in this species is stout compared to that in C. buckleyi . Leptotila cf. verreauxi – El Azúcar (1): In the size range of L. verreauxi or a small Z. meloda , I assign this archaeological specimen to L. verrauxi for two reasons. 1) Facies articularis antitrochanterica is almost at a right angle to axis proximodistalis (at a steeper angle in Z. meloda ), and 2) linea intermuscularis crania lis has a more laterad position in L. verreauxi . TIBIOTARSUS General Osteology The columbid tibiotarsus has very reduced cristae cnemialis cranialis et lateralis. Crista patellaris is located only slightly pr oximad of area interarticularis. The shaft is slightly curved crania lly. The distal end of sulcus ex tensorius and pons supratendineus are located on the medial half of the shaft, leaving a broad area on the lateral half which bears the attachment for retinaculum exte nsorium tibiotarsi and retinaculum musculi fibularis. Sulcus musculi fibular is is hardly visible, its lo cation only being indicated by two very low ridges. Condylus medialis and con dylus lateralis are almost equal in length

PAGE 149

133 in cranial aspect. The distan ce between the medial margin of condylus medialis and the lateral margin of condylus lateralis decreases fr om the proximal to the distal extremes of the condyla. Comparative Osteology Size and overall proportions are the only reliable criteria fo r separating species based on the proximal end. The relative le ngth of condylus medialis and condylus lateralis in cranial aspect, the shape of condylus lateralis in lateral aspect, the position of pons supratendineus, and the extent of the c oncave area laterad of pons tendineus are all taxonomically useful traits. Zenaida auriculata – OGSE-80 (2), El Azúcar (12), La Chimba (2): Being comparatively shorter than in L. verreauxi and G. montana , the tibiotarsus of Z. auriculata also tends to have a more slender proximal end. On the distal end, tuberculum retinaculi lateralis is a small tuberos ity and the concave area laterad of pons supratendineus is very wide (wid est among the species examined). Zenaida meloda – El Azúcar (1): Having a larger tibiotarsus than in Z. auriculata and G. montana , Z. meloda is similar in size to Leptotila . Zenaida meloda has a more developed tuberculum retinaculi musculus fibularis, a more proximad tuberculum retinaculi extensorium tibiotarsi medialis, and less separated and dist ally more tapering condyla medialis et lateralis. Columbina cruziana – El Azúcar (7): The tibiotarsus if this species is stronger than that of the similarly sized C. buckleyi . Geotrygon montana – El Azúcar (1): With a narrow sh aft, an oblique canalis extensorius, a small concave area latera d of pons supratendineus, and long condyla

PAGE 150

134 lateralis et medialis in cranial aspect (as co mpared to the width of the distal end), this specimen agrees very well with the tibiotarsus of G. montana . TARSOMETATARSUS General Osteology The tarsometatarsus varies widely in lengt h and width. Cotylae me dialis et lateralis are separated by a pointy eminentia intercotylar is. Crista hypotarsi medialis is long and has a dorso-laterad orientati on. Usually the hypotarsal pattern consists of one canal and one or two sulci hypotarsorum. Impressio re tinaculi extensorii ar e well developed. The medial margin of the proximal shaft is thick a nd the lateral margin is very thin. The distal end is concave in plantar aspect as trochleae metatarsorum II et IV are oriented medially and laterally, respectively. The distal extent of trochlea metatarsi II is slighty greater than that of trochlea metatarsi IV. Comparative Osteology Zenaida auriculata – El Azúcar (11), La Chimba (1): The tarsometatarsus of Z. auriculata is unique in its overall dimensions and size. It is short and slender ( Columba short and stout, Leptotila long and strong, Geotrygon long and slender) and significantly smaller than that of the very similar Z. meloda . Zenaida meloda – El Azúcar (3): Similar in overall pr oportions, the tarsometatarsus is larger and stronger than that of Z. auriculata . Columbina minuta – El Azúcar (1): Small size is dia gnostic for this species. Columbina buckleyi – El Azúcar (2): The slender tarsom etatarsus of this grounddove is longer than that of C. cruziana.

PAGE 151

135 Columbina cruziana – El Azúcar (5): A stouter build and shorter length distinguish this species from C. buckleyi . Claravis pretiosa – El Azúcar (1): The complete tarsometatarsus is intermediate in size between Z. auriculata and C. buckleyi , two species with similar overall proportions. The only species in this size class in western Ecuador is C. pretiosa . Order PSITTACIFORMES Family PSITTACIDAE OSSA CRANII General Osteology The rostro-dorsal cranium of the Psittaci dae has a flat os frontale, thin or syndesmotic/synovial zona flexoria craniofaci alis, and squared head of os lacrimale (sensu Cracraft 1968). A symmetrical set of fora mina perforates os la crimale just laterally to a small os mesethmoidale. Comparative Osteology Within the parrot family, size is a good tr ait to limit the pool of species to be considered. The genera with species approxi mately the size of the La Chimba specimen are Aratinga , Leptosittaca , Pionites , Pionus , Amazona , and Deroptyus . Good traits to distinguish among these genera are the shape of the head of os lacrimale (PsOL1; 0: square and narrow, 1: tetraor pentahedrallike narrowing ventrally); the lateral margin of os frontale just dorsad of orbita (PsOL2; 0: well below cen tral part of os frontale in lateral aspect, 1: nearly same or same leve l as central part of os frontale); and the arrangement of foramina passi ng through the rostral end of os lacrimale in caudal view (PsOL3; 0: a pair of large foramina on each side with division by a rostral bar, 1: a pair of

PAGE 152

136 foramina with medial foramen usually sma ller and division by bony sheets starting more caudad; Table 2-19). Table 2-19. Character state matrix for ossa cr anii of the Psittacidae. Character states explained in text. Character (PsOL) Species 1 2 3 Aratinga finschi 0 0 0 Aratinga erythrogenys 0 0 0 Pyrrhura picta 0 0 0 Deroptyus accipitrinus 1 0-1 0* Pionites melanocephalus 1 0-1 0 Pionus fuscus 1 1 1 Amazona autumnalis 1 1 1 * A single long foramen in which dividing bar probably has been lost. cf. Leptosittaca branickii – La Chimba (1): The rostro-dor sal portion of cranium from the La Chimba specimen shares all traits with Aratinga (Table 2-19). The head of os lacrimale is narrow and square, the lateral margin of os frontale is well below the central region of os frontale in lateral aspect, and a pair of large fo ramina perforates the rostral wall of os lacrimale on each side of the skull. It differs from that of Aratinga by a comparably narrower and longer os frontale, indicative of a long and slender head. No comparative material was available and no information on the cranial osteology of Leptosittaca has been published, but because of its close affiliation with the genus Aratinga and high Andean habitat, I tent atively assign this specimen to L. branickii . The also related Nandayus nenday is unlikely due to its south temperate distribution.

PAGE 153

137 OS DENTALE PARS SYMPHASIALIS General Osteology Os dentale in the parrots is a uniquely br oad and short bone with a U-shaped rostral margin and small fenestra. Comparative Osteology Williams (1999) has pointed out that the Ne otropical parrots can be divided into two major clades (his tribes “Arini” and “Pionini”) by th e shape of os dentale pars symphasialis (the symphysis). As the La Chimba specimen has a squared caudal margin of the symphysis, great dorso-ventral depth in lateral aspect, and a short rostro-dorsal process of the os dentale (in addition, a si ngle large foramen on th e dorsal surface of the symphysis is absent in Amazona and Pionus ; Figure 2-6), I restrict my analysis to the Arini (sensu Williams 1999), including Ara , Aratinga , Pyrrhura , Deroptius and Pionites . The caudal margin of the symphysis, size, and arrangement of several fossae and foramina on the dorso-medial surface of the symphysis ar e informative characters. Figure 2-6. The lower mandible in dorsal as pect from the La Chimba specimen (a), Pionus fuscus (b), Pionopsitta haematotis (c), Pionites melanocephala (d), and Aratinga erythrogenys . Reference bar is 1 cm.

PAGE 154

138 Pionites melanocephala – La Chimba (1): Having a round, but wide and slightly squared caudal margin, the symphysis from the La Chimba archaeological deposits is very different from that in Ara (square and much larger), Aratinga (very squared with central notch and larger), and Deroptyus accipitrinus (somewhat square and larger; Figure 2-6). Whereas Pyrrhura and Pionites are similar in shape, no species of Pyrrhura is large enough to be considere d. In addition, a pair of foramina at the distal end of each ramus is arranged rostro-caudally in Pyrrhura (as well as in Aratinga ) whereas in the La Chimba specimen it is arranged dorso-ventrally as in Pionites . Although slightly larger and apparently more heavily ossified than the mandible from modern specimen UF 26068, Pionites agrees well in terms of the caudal margin of the mandibular symphysis and the dorso-medial foramina of the proximal symphysis. A specimen of L. branickii is sorely needed, but because of its close relatedness to Aratinga (and Nandayus , both genera sharing the same condition of the sym physal foramina), I think it is unlikely that the archaeological specimen be longs to this species. CORACOID General Osteology Superficially the psittacid coracoid resembles that of the Columbidae (doves and pigeons). It is a stout bone with a bulky processus acrocoracoideus and a sheet-like tuberculum brachiale. Facies articularis sternalis is thin, impressio ligamenti acrocoracohumeralis is long and more dorsall y oriented than in the Columbidae, and tuberculum brachialis has a more extensive dorso-medial surface than in the Columbidae.

PAGE 155

139 Comparative Osteology The two major tribes of New World parro ts can be separated on the basis of tuberculum brachiale, which is closer to the shaft in the Arini (including Pionites and Deroptyus ) than in the Pionini (sensu Williams 1999). In addition, the corcacoids of the Arini are generally more slender than those of the Pionini. Size is also a useful trait. The two very small coracoids from El Azúcar are Forpus -sized. Pionites melanocephalus – La Chimba (2): The two complete coracoids are very similar to those of Pionites melanocephalus in size and shape. Whereas some modern specimens of captive Pionites (UF 20083, UF 26067, UF 26068) , differ by lacking the heavy pneumatization of the distal sulcus mu sculi supracoracoidei, having a more slender shaft, and having a finer pro cessus procoracoideus, the cor acoid of a wild-caught bird (UF 19482) approaches the specimen in all tr aits. Given their association with the mandible of P. melanocephala , I assign the coracoids to this species as well. I do not think these specimens are from an Aratinga as facies articularis sternalis is not bordered distally by an acuminate ridge, and the distal end of sulcus musculis supracoracoidei is not as pneumatized. Forpus coelestis – El Azúcar (3): Identification of thes e specimens is based on the small size. The distal ends of all three cor acoids have the typical Pionini condition of a more open, more caudally proj ecting tuberculum brachiale. Amazona farinosa – OGSE-80 (1): The heavily fragmented specimen is a stout bone with a narrow proximal end, small proces sus lateralis, deep cotyla scapularis, and narrow processus procoracoideus. This combin ation of traits can only be found in the Psittacidae (proximal end wider in the Anatidae, shorter and broader in the

PAGE 156

140 Accipitriformes; processus procoracoideus fene strate and wider in the Strigiformes). By the width of the shaft it is an Amazona and by sizeit isreferred to A. farinosa (considerably larger than in A. autumnalis ). Psittacidae sp. – La Ponga (1): Given the comparative material available, I am unable to identify this complete coracoid wh ich is similar in size to a specimen of Pionites melanocephala (UF 26068). Because of the more caudally oriented tuberculum brachiale, I assume that it belongs to a memb er of the Pionini (sensu Williams 1999). The only species of parrot in western Ecuador that is in the size range of the archaeological specimen is Pionopsitta pulchra . HUMERUS General Osteology The psittacid humerus is a stout bone with strong proximal and distal ends. Crista deltopectoralis is very thick. Superficially, the humerus of the Psittacidae resembles that of the Columbidae, but it differs in havi ng a more pointy caput humeri and a shallow fossa pneumotricipitalis (completely absent in the doves and pigeons). Comparative Osteology Extreme differences in body mass found in this family make size a good trait to distinguish among many genera. Within size classes, however, no single characters may allow distinguishing between genera and/or species. Ara sensu lato and Amazona farinosa differ markedly in the shape of fo ssa musculi brachialis. The humerus of Aratinga tends to have a long impressio coracobr achialis, extending distally as far as the apex of crista deltopectoralis . The lack of good characters makes the humerus difficult to identify in certain size classes.

PAGE 157

141 Forpus coelestis – El Azúcar (2): Based on size, the complete humerus and proximal end of humerus can be identified as F. coelestis . Pionites melanocephala – La Chimba (1): This humerus is very close in size and shape to that in P. melanocephala . In addition to an associ ation with other bones from this species, a broad fossa pneumotricipitali s that undercuts caput humeri, and an angled tuberculum supracondylare venralis in ventral aspect (straight in Aratinga ) support the identification. Amazona farinosa – La Chimba (1): The specimen of a distal end and most of the shaft of a large humerus is cl ose is in the size range of Amazona farinosa or a small Ara . Fossa musculi brachialis tapers distally and is bordered along its distal margin by a smooth, featureless plateau th at gently slopes cranially toward the condyla. This characteristic condition of A. farinosa cannot be found in Ara . Psittacidae sp. – La Ponga (1): By size possibly Pionopsitta pulchra . ULNA General Osteology Like the humerus, the strong ulna shows a strong resemblence to that of the Columbidae. A prominent tuberculum ligamenti collateralis ventralis and a noticeable depressio radialis distinguish ps ittacid from columbid ulnae. Comparative Osteology Size is again an important character. Furt hermore, medium-sized Arini and Pionini can be distinguished on the basis of olecr anon (which tends to be more slanted proximally in the Arini), the orientation of a ttachment site of musculus extensor carpi ulnaris that emanates from the dorsal base of processus cotylaris dorsalis (line at smaller

PAGE 158

142 angle to axis proximodistalis in Arini), and the width of de pressio radialis (broader in Pionini). Large species, such as Ara sensu latu and Amazona farinosa differ also by a prominent muscle attachment line in Ara (absent in Amazona ), width of depressio radialis, and degree of separation of condyla ulnare (both of which are greater in the Arini). Aratinga erythrogenys – OGSE-80 (1): By size, shape of olecranon, and the attachment of musculus extensor carpi ulnaris, the proxima l end of ulna is A. erythrogenys . Forpus coelestis – El Azúcar (2): Size of the two proximal ends suggests Forpus . Pionites melanocephala – La Chimba (1): Although not shared by the ulnae of captive birds of this species, a steep angle of the attachment site of musculus extensor carpi ulnaris in relation to ax is proximodistalis and a wide depressio radialis of this complete ulna were found in the specimen of a wild-caught bird. W ith the exception of some Pionites , these two traits are mainly found in the Pionini. Given the association with other bone elements (see abov e); however, I refer this ulna in P. melanocephala . Amazona autumnalis – Loma Alta (1): In terms of size and shape the most likely genus for this distal end is Amazona . Distributional patterns a nd near perfect agreement with comparative skeletons suggest that this specimen belonged to A. autumnalis . Amazona farinosa – Loma Alta (2): The proximal end compares in size and shape to A. farinosa . A small species of Ara can be ruled out by the lack of a raised attachment of musculus extensor car pi ulnaris (see above).

PAGE 159

143 CARPOMETACARPUS General Osteology The carpometacarpus of the Psittacidae is characterized by near ly parallel ossa metacarpale majus et minus, erect processus ex tensorius, a V-shaped proximal symphysis pars centralis (due to a rais ed ridge representing the proxim al continuation of the ventral margin of os metacarpale minus to the base of processus pisiformes), and a tendinal canal (or pair of canals) on distal os metacarpale majus in caudal aspect. Comparative Osteology The lack of informative traits within the Psittacidae (Williams 1999) makes size the best character to begin identification of ar chaeological specimens. Within any given sizeclass, differences have to be evaluated on a case to case basis. For instance, similarly sized Aratinga and Pionus differ in the depth of fovea cran ialis carpalis and a short, distal ridge projecting cranially from the cranial marg in of facies articularis digitalis major to the proximal margin of the distal symphysis. Aratinga cf. wagleri – La Chimba (1): This specimen is a nearly complete carpometacarpus (lacking os metacarpalis minus) with a deep fovea carpalis cranialis and a high distal ridge as in Aratinga (see above). Smaller than that of A. mitrata , but larger than that of A. erythrogenys , this specimen compares favorably in terms of overall proportions to these two species (in A. leucophthalmus the distal symphysis is proportionately narrower cranio-caudally). Th e La Chimba specimen may also represent an individual of Leptosittaca branickii , but without further comp arative material I cannot confirm nor reject that possibilit y. Because of the good agreement with A. mitrata and A. erythrogenys , I believe that L. branickii is unlikely.

PAGE 160

144 Forpus coelestis – El Azúcar (1): The proximal end of carpometacarpus corresponds in size to that of Forpus . Pionites melanocephala – La Chimba (1): This complete carpometacarpus is associated with other specimens of Pionites melanocephala . It is slightly stronger and more heavily build than modern specimens, but shares a deep and round attachment for ligamentum ulnocarpo-metacarpal e ventralis (shallower in Aratinga ) and a highly raised attachment for ligamentum ulnocarpo-metacarpale dorsalis (also found in Aratinga ). Amazona cf. mercenaria – La Chimba (1): This complete carpometacarpus is slightly smaller than that of a modern specimen of A. autumnalis (UF 25771). Because the tendinal passage of the distal end of os metacarpale majus in caudal aspect tends to be closed in Ara , forming a canal, and open in Amazona , forming a sulcus, I suggest that this large carpometacarpus is from a species of Amazona . Being smaller than that of A. autumnalis , I tentatively refer this specimen to A. mercenaria , a common highland parrot. Specimens of A. mercenaria need to be examined next to Amazonian species of the same genus to confirm the identification. Psittacidae sp. – La Ponga (1): This bone may be a ssociated with the humerus above and may therefore re present the same species, Pionopsitta pulchra . FEMUR General Osteology The psittacid femur has a large head, with a large fovea ligamenti capitis that opens dorsad. Crista trochanteris is broad and heavy, but trochanter femoris is absent. Condylus medialis and condylus lateralis broaden distally, resulting in a broad sulcus intercondylaris and a bulky-looki ng distal end. Impressio ansae musculus iliofibularis is

PAGE 161

145 wide and deep. In medium-sized parrots, condylus lateralis does not extend further distally than condylus lateralis, a cond ition I have not seen in any other taxon. Comparative Osteology The osteological separation of Ara sensu lato and Amazona is hampered by the large degree of intra-specific and intra-ge neric variation within both genera. There are several trends, however, that are more prevalen t in one than the other genus. The distal end tends to be broader in Ara due to a stronger medial curv ature of condylus medialis. The proximo-caudal articular su rface of condylus medialis also tends to be narrower in Ara . In Amazona the distal medial margin of condylus medialis is almost parallel to axis proximodistalis, rapidly slanting toward the sh aft (giving the proximal medial margin of condylus medialis a highly angled look, whereas in Ara the margin is rounder). The proximal margin of the articu lar surface of condylus medialis comes closer to condylus lateralis in Amazona than in Ara . Sulcus intracondylaris therefore appears to narrow proximally in caudal aspect. Impressio ligam enti cruciaticranialis tends to be more excavated and closer to the distal end of sulcus intracondylaris in Amazona than in Ara . Finally, the shaft just proximal to condyla medialis tends to be thicker in medial aspect in Amazona . Medium-sized species of Aratinga and Pionus differ in the thickness of the shaft and width of trochlea fibularis. Size and pneumatization of fo ssa poplitea are less reliable but useful traits. Aratinga cf. wagleri – La Chimba (1): This nearly complete femur has a narrow shaft, a wide distal end, and a wide and pne umatized fossa poplitea, conditions associated

PAGE 162

146 with Aratinga . Close agreement with that of A. mitrata and A. erythrogenys in terms of overall proportions makes Leptosittaca branickii unlikely. Amazona farinosa – La Chimba (1): Despite of its large size, the La Chimba specimen is slightly smaller than the smallest available specimen of Ara . In addition, the distal end is in all traits more similar to that of A. farinosa than any species of Ara examined (see above). TARSOMETATARUS General Osteology The psittacid tarsometatarsus is a very shor t, flat (in dorsal-plantar dimension), and compact bone. The proximal end is wide and cristae hypotarsorum join to produce a variably shaped hypotarsal canal . The extremely wide distal end has a highly rotated trochlea metatarsi IV, typi cal of a zygodactyl foot. Comparative Osteology Williams (1999) identified the presence of a well-developed crista hypotarsi medialis and sulcus hypotarsi as a diagnostic traits for Tribe Pionini. Crista lateralis hypotarsi is also pronounced, but much broader. In the Arini, the pl antar margin of the fused cristae hypotarsorum is a flat surface. On the dorsal surface of the proximal end, impressio ligamenti collateralis lateralis is deeper and wider in the Arini, causing the proximo-medial margin to flare out. On the dist al end, the shape of trochlea metatarsi can be used to separate the two tribes. The late ral rim of trochlea metatarsi IV has a flat plantar surface in the Arini and a rounded surface in the Pionini (making this condyle look more rod-like; Figure 2-7).

PAGE 163

147 Figure 2-7. Plantar aspect of the tarsometatarsus in Amazona autumnalis (a), the La Chimba specimen (b), and Ara severa (c). Reference bar is 1 cm. Ara cf. ambiguus – El Azúcar (1): Based on size and th e flat plantar surface of condyla lateralis of trochlea metatarsi IV, th e distal end of a tarsometatarsus can be assigned to Ara . Within the genus identification is difficult due to great intra-specific variation. The bone is small, possibly bel onging to a female, and has bulky rims of trochlea metatarsi III, similar to that in A. ambiguus . In A. ararauna , A. macao , and A. chloroptera the rims are finer, resulting in a br oader sulcus trochlea ris (“narrow groove between lateral and medial rims of trochlea;” Livezey 1998, p. 2122). In A. ararauna the rims of trochlea metatarsi III are also at a great er angle to axis proxim odistalis than in the other species. Because of these traits and current distributional pa tterns of macaws in northwestern South America (only A. ambiguus occurs in western Ecuador, with recent records of A. ararauna probably representing escaped pets; Ridgely and Greenfield 2001a), I tentatively refer this specimen to A. ambiguus . Ara cf. chloroptera – Cotocollao (1): Villalba (1988, Lámina 60) presents a photo of a complete psittacid tarsometatarsus in plantar aspect. Large si ze (scale provided), a shallow crista hypotarsi medialis and a flat pl antar surface of the lateral rim of trochlea

PAGE 164

148 metatarsi IV can be seen on the picture, sugge sting that the specimen is from a species of Ara . Six characters (Table 2-20) suggest this bone belongs to A. chloroptera (Table 221). Table 2-20. Characters for the tarsom etatarsus of large species of Ara (Psittacidae). Character Character States PsTMT1 Crista hypotarsi medialis 0: low 1: high PsTMT 2 Sulcus hypotarsi medialis 0: shallow 1: deep PsTMT 3 Plantar surface of fused cristae hypotarsorum 0: narrow 1: widening laterad or centrad PsTMT 4 Shaft 0: slender 1: stout 2: very stout PsTMT 5 Amount of lateral margin protruding laterally from cotyla lateralis in proximal view 0: < ½ the width of proximal articular surfaces 1: about ½ the width of pr oximal articular surfaces PsTMT 6 Facies articularis phalangis trochlea metatarsi II 0: straigth in plantar aspect 1: curved in plantar aspect Ara severa – La Chimba (1): The proximal end has partially damaged hypotarsi, but the typical Arini condition of a late rally narrowing hypotarsal surface in plantar aspect is clearly visible. Being larger than any available modern specimen of Aratinga , this bone is even slightly la rger than a 328 g specimen of A. severa from western Ecuador (UF 22234). The modern and archaeological spec imens also share a hypotarsus with two (rather than one) hypotarsal canal s in proximal aspect. Based on size and tendinal canals, I tentatively identify this specimen as A. severa , a small macaw that has been observed up to 1800 m in the eastern Andes of Ecuador (R idgely and Greenfield 2001a). Even though I did not have a comparative specimen, I believe that Orthopsittaca (= Ara ) manilata is less likely because of its present distribut ion in the Amazonian lowlands beneath 400 m (Ridgely and Greenfield 2001a,b).

PAGE 165

149 Table 2-21. Character matrix for six traits found in the tarsom etatarsus of the Psittacidae. See Table 2-20 for definitions of the character states. Characters (PsTMT) Species 1 2 3 4 5 6 Ara ambiguus 0 0 0 2 1 0 Ara ararauna 1 1 1 0 1 0 Ara macao 0-1 0-1 1 0 0 1 Ara chloroptera 0-1 0-1 0 1 0 1 Forpus coelestis – El Azúcar (1): By size the complete tarsometatarsus comes from a F. coelestis , the only species of this genus to o ccur in western Ecuador (Ridgely and Greenfield 2001a,b). OSSA DIGITORUM PEDIS General Osteology The zygodactyl foot of the Psittacidae is ve ry compact and short, a characteristic reflected in stout phalanges 1 of digits 2, 3, and 4 with broad proximal articular surfaces. Phalanges 2, 3, and 4 of digits 2, 3, and 4, re spectively are slender and long, comparable in overall shape to those in the Accipitriformes. Comparative Osteology Within Ara intraspecific variation in qualitati ve osteology is great; often the only characters to distinguish among species is size and overall proportions. Ara cf. ambiguus – El Azúcar (8): The eight phalangeal specimens and associated tarsometatarsus are from small species of Ara . In overall proportion, phalanx 1 of digit 3 is similar to that of A. ambiguus , A. macao , and A. chloroptera . Only A. ararauna differs by greater dorso-ventral depth and proportiona tely longer and more slender bones (also

PAGE 166

150 true for other phalanges). The sulcus between the tuberosities servi ng as attachment sites of musculus extensor proprius III and musculus extensor brev is digiti III is more open in A. chloroptera than in A. ambiguus and A. macao . Since an associated tarsometatarsus (see above) appears to be from A. ambiguus , I also tentatively identify the phalanges as this species. Order STRIGIFORMES Family TYTONIDAE OSSA CRANII General Osteology The ability to rotate the skull nearly 360 û requires a ventral position of condylus occipitalis in the Strigiformes. The cranium of Tyto alba , the only extant member of this family in South America (König et al. 1999), is long with a high a nd round calvaria. Os lacrimale is large and expanded rostro-cauda lly. The orbital proce ss of os frontale is greatly reduced. Apertura nasi are narrower than in the Strigi dae. Basis cranii externa is characterized by an extensive basis rostri parasphenoidali. Except for Asio , this structure is greatly reduced in the Strigidae. Comparative Osteology The tytonid cranium is very different from that of the Strigidae. The only strigid genus showing some simlilarity with Tyto is Asio which differs in having a less furrowed condylus occipitalis, a more ventral forame n magnum, and shallower fossa parabasalis. Tyto alba – La Chimba (1): The above menioned traits suggest this basis cranii is from T. alba .

PAGE 167

151 HUMERUS General Osteology The humerus in Tyto is similar to that of othe r raptors with a long crista deltopectoralis and a st rong distal end. Tuberculum vent rale and the pneumatic fossa are small, and impressio musculi brachiale is broad and long compared to the Accipitriiformes. Comparative Osteology The humerus of the Tytonidae is similar to that of the Strigidae, although caput humeri is less massive, impressio coracobrach ialis is less extensive, and pneumatic fossa opens more dorsad. The distal end is not as wide, but has a bulkier processus supracondylaris dorsalis, a more raised t uberculum supracondylare ventrale, and a round pit-like attachment for muscul us flexor carpi ulnaris. Most importantly, the caudal attachment of musculus deltoid eus major merges or nearly me rges with the distal end of crista deltopectoralis in the Strigidae, but runs along the caudal margin of the shaft, nearly parallel to the crista, in Tyto . Tyto alba – La Chimba (1): The caudal line formed by musculus deltoideus major suggests that this highly frag mented specimen belongs to Tyto . Crista deltopectoralis is much shorter than in the modern comparat ive skeletons, however, and comparison with the Andean T. a. contempta would be desirable to confirm the identification. ULNA General Osteology The ulna of the Tytonidae is typical for the Strigiformes with a strong tuberculum bicipitale, pronounced tuberc ulum ligamenti collateralis ventralis, long impressio

PAGE 168

152 brachialis, pointed tuberculum carpale, a nd ventrally projecting condylus ventralis ulnaris. Comparative Osteology Tyto alba – La Chimba (1), Loma Alta (2): A craniad orientation of both cotyla, proximally highly angled and generally bulky tuberculum ligamenti collateralis ventrale, and a greatly raised tuberculum bicip itale distinguish the proximal end of Tyto from that of similarly sized strigid owls . On the distal end, tuberculum radiale is short and round, and incisura tuberculi carpalis is shallo w containing a single, medium-sized foramen (deeper with one or several sma ll foramina in the Strigidae). RADIUS General Osteology In the Tytonidae tuberculum bicipitale radiale is rounder, having a more ventrad orientation and a reduced cranial margin than in the Strigidae. The distal end also is more expanded, with the caudal ridge of sulcus tendineus raised and located centrad (rather than caudally). Comparative Osteology Tyto alba – La Chimba (1): The location of the caudal ridge of sulcus tendineus and a small tuberculum aponeurosis ventralis (large in many Strigidae) suggest that the distal end from La Chimba belongs to T. alba . CARPOMETACARPUS General Osteology The cranio-ventral margin of os metacarpi minus continues proximad beyond the distal margin of the proximal symphysis as an elevated ridge, reaching, or almost

PAGE 169

153 reaching, processus pisiformes. It differs in being proportionately longer, having a broad processus extensorius, a less proximad projecting dorsal rim of trochlea carpalis, a more ventrad position of os metacarpi minus, a bul ky cranial process on the distal end just proximal to facies articularis digitalis major (flat in Psittaciformes; see above), lacking two distal tendinal canals, and having a distally expanded sulcus tendineus. Comparative Osteology Tyto alba – La Chimba (1): The proximal end of car pometacarpus has the V-shaped symphysis due to the ridge formed by the cran ial margin of os metacarpi minus in the Psittaciformes and Strigiformes. Processus extensorius is broad, the dorsal rim of trochlea carpalis extends only slightly beyond the ventral rim, and the proximal end of os metacarpi minus has a ventral position (such th at it is visible in cranial view; in the Psittaciformes os metacarpi minus is completely hidden behind os metacarpi majus in cranial aspect). These traits suggest that the La Chimba sp ecimen belongs to an owl. A short processus intermetacarpalis that does not fuse proximally with the distal symphysis, as present on the archaeological specimen, is found only in Tyto . OSSA DIGITORUM PEDIS General Osteology The pedal phalanges of tytonid owls bear a superficial resemblance to those of the Accipitriformes and Psittaciformes. Phalanges pr oximales et intermediae of digits 2 and 3 are long, and those of digit 4 (with the exception of the last phalanx intermedia) are short. Facies dorsalis is characterized by a flat proximal surface, giving the proximal end a squared appearance in cross-section. Fov ea ligamenti collateralis is deep. Phalanges intermediae bear a long proximal process on faci es dorsalis that partia lly covers trochlea

PAGE 170

154 articularis of the more proximal phalanx. Th is dorsal process is shorter in the Accipitriformes and Psittaciformes. Phalanx ung ualis is distinct in having a bony sheet originating from the margin of cotyla arti cularis and the dorsal ma rgin of tuberculum flexorium. Also unusual is the ab sence of sulcus neurovascularis. Comparative Osteology Tyto alba – La Chimba (3): The flat surface on facies dorsalis is diagnostic for Tyto alba . Family STRIGIDAE OSSA CRANII General Osteology As for Tyto alba , the triangular cranium (in dorsa l aspect) of the Strigidae is characterized by a ventrad location of c ondylus occipitalis and formamen magnum. Condylus occipitalis is large a nd slightly oval in shape. Fo ssa temporalis is wide and processus paroccipitalis great ly expanded. The orbit is large and opens rostrad. Os frontale is thick and bears a stout orbital process. Apertura nasi are holorhinal and located rostrad just distal to rostrum maxillaris. Conchae nasales are extensively ossified. Comparative Osteology Apart from size, which distinguishes th e largest and smallest genera (large Bubo vs. small Otus and Glaucidium ), the shape of lamina paraspheno idales as well as the location of several canales nervi and/or foramina va scularia can be used to distinguish among different genera of owls. Bubo virginianus – La Chimba (1): The size of condylus occipitalis and foramen magnum suggest that this ventro-ca udal fragment of skull is from B. virginianus . In

PAGE 171

155 addition, lamina parasphenoidales is approxima tely triangular and of intermediate width, a condition shared by Bubo and the much smaller Strix virgata . In Pulsatrix perspicillata the lamina is wide and short; in Asio it is narrow and long. ULNA General Osteology The proximal end is characterized by a short olecranon, deep trochlea humeralis ulnaris, and prominent tuberculum ligamenti co llateralis ventralis that forms a descending ridge distally, bordering impressio brachiali s. This ridge makes the proximal end look club-like as only processus coty laris dorsalis protrudes dorsa d (at a right angle to the shaft) in cranial view. Incisura radialis is well defined and attachment of tuberculum bicipitale ulnare is bulky tuberosity that forms the di stal border of incisura radialis. Tuberculum bicipitale ulnare is the point of origin of raised lineae intermuscularis. The nine to eleven papillae remigalis ventralis are not very pronounced and in some cases barely visible. The proximal half of the bone is curved whereas the distal half is straight. Incisura tendinosa is short a nd runs obliquely along a caudallateral to a cranial-medial axis. There is no obvious depressio radial is as found in the Accipitriformes. Comparative Osteology I examined seven species in five genera. They differ greatly in overall size. For most of the species that can be expected to o ccur at high elevations, size of the ulna is a good trait. Bubo virginianus is the largest species and doe s not show any overlap. As for most large species, muscle attachment si tes are very pronounced and tubersities are bulky. The smallest genera, Glaucidium and Otus , are alone in their respective size classes. For similarly sized species, the ratio between the length and width of the bone is

PAGE 172

156 useful. Separation of taxa can also be achie ved based on size and shape of olecranon, the location and prominence of tuberculum bici pitale, shape and orientation of cotyla dorsalis, shape of tuberculum ligamenti colla teralis ventralis, the proximal end of impressio brachialis, shape of tuberculum br achiale, and the proximal extent of ventral condylus dorsalis. Asio cf. stygius – La Chimba (2): The complete ulna a nd proximal end of ulna are very similar to those of Asio flammeus , although the bone is proportionately slightly longer. Strix differs in having a longer olec ranon, an elongate, curved, and low tuberculum bicipitale, and a na rrow tuberculum ligamenti collateralis ventralis. The distal end differs in a pointier tuberculum radiale, a more proximately slanted attachment of ligamentum ulno-ulnocarpale pr oximale, and a more proxima d ventral condylus dorsalis. A specimen of A. stygius was not available, but the La Chimba specimen differs in having a bulkier tuberculum ligamenti collatera lis ventralis, a bulkier tuberculum carpale, and a proportionately longer shaft. The differe nces in strength of ligamental attachment sites and total length are in agreement with larger body size (ca. 675 g vs. 206-450 g in males and 284-505 g in females) and longe r wings (ca. 292-349 mm vs. ca. 281-335 mm) for A. stygius (König et al. 1999). Size differences between A. flammeus and A. stygius are probably accentuated in the Ecuador ian Andes as the local subspecies A. flammeus bogotensis is even smaller than nominate race av ailable for comparison. Identification is tentative until more comparative material becomes available.

PAGE 173

157 RADIUS General Osteology The radius of the Strigidae is superficiall y similar in appearance to that of the Accipitriformes. The proximal aspect of cotyla humeralis is oval in shape due to a large capital tuberosity (sensu Howard 1929), tuberc ulum bicipitale radi i is large and deep (lying more ventrad than in the Accipitriformes), and ligamental papilla is prominent. The area comprised of tuberculum bicipitale radii and the ligamental papilla is at an angle to axis proximodistalis giving the proximal end a slanted appearance. The shaft is sshaped in ventral aspect. Lineae intermuscula re on the ventral side are prominent and form a small arch in the proximal third of the bone. On the distal end, sulcus tendineus lies medially and tuberculum aponeurosis ventralis points dorsad. Comparative Osteology The radius is very uniform within the Strigidae. The best characteristic to distinguish the differe nt genera is size. Bubo virginianus and Glaucidium are at the large and small extremes, respectively, showing no ove rlap in size with those of any of the other genera. Otus is also in a size class by itself. The radius of Strix ( Ciccaba ) virgata is considerably smaller than in Pulsatrix perspicillata , Asio flammeus , and Strix varia . Most of the differences among genera of similar size are found in the distal end of the radius. Proximal to depressio ligamenti, a small t uberosity is found in most species. It is especially prominent in Bubo , less so in Asio , Strix , and Otus , and almost absent in Glaucidium and Pulsatrix . Caudad to that tuberosity is another tuberosity that is especially pronounced in Pulsatrix and barely noticeable is Bubo . It is intermediate in the other genera except in Glaucidium where I have not found it at all. Dorsal facies

PAGE 174

158 articularis ulnaris is separate d by a deep impression of meni scus radioulnaris from the remainder of facies articularis ul naris. This canal is narrow in Bubo , Glaucidium , Otus , and Asio , and wide in Strix and Pulsatrix . The capital t uberosity (sensu Howard 1929) that bears facies articularis ulnaris is long and pointy in Pulsatrix whereas it is round in all the other genera. On the dorsal side the ridges of sulcus tendineus are variable in shape and precise location. Bubo virginianus – La Chimba (1): By size and above described traits, the complete radius belongs to B. virginianus . Asio cf. stygius – La Chimba (1): This complete radius is from an Asio based on overall size, size and position of the capita l tuberosity, width of the impression of meniscus radioulnaris, and shape and position of sulcus tendineus. This element was associated with a complete ulna identified tentatively as A. stygius . I therefore also assign this radius tentatively to A. stygius . CARPOMETACARPUS General Osteology The carpometacarpus of the Strigidae is ve ry similar to that of the Tytonidae described above. It differs in that processus intermetacarpalis is fused with the distal end of proximal symphysis, and in many species os metacarpi minus is more rounded and more separated from os metacarpi majus than in Tyto . Asio resembles Tyto in having the proximal portion of os metacarpi minus highly angled in ventral aspect, and the distal portion nearly parallel to os metacarpi majus.

PAGE 175

159 Comparative Osteology Size clearly separates most of the genera. The shape of os metacarpi minus is also characteristic for some taxa. Asio cf. stygius – La Chimba (1): The nearly complete carpometacarpus, lacking only processus extensorius, is smaller than that of B. virginianus , but larger than in Otus and Glaucidium . Among medium-sized owls , it is distinct from Strix virgata and Pulsatrix perspicillata in having a proximally angled and di stally straight (parallel to os metacarpi majus) os metacarpi minus. Although similar to that of T. alba , the archaeological specimen has a long symphysis (short in T. alba ), processus intermetacarpalis is fused to proximal symphys is, the distal, expanded area of sulcus tendineus is shallow (deep in T. alba ), and facies articularis digitalis majus is narrow (broad in T. alba ). As with the associated ulna an d radius, the La Chimba specimen is proportionately longer than modern specimens of A. flammeus , and I tentatively refer it to the longer-winged A. stygius (König et al. 1999). FEMUR General Osteology The strigid femur is a variable bone with a steep facies articularis antitrochanterica, low and strong trochanter femoris, strong a nd raised impressiones obturatoriae, deep fovea ligamenti capitis, strong lineae interm uscularis cranalis et caudalis, sharp and pronounced crista supracondylaris caudalis, and deep fossa pop litea. The proximal end of linea intermuscularis cranialis originates on th e medial margin of trochanter femoris or from the distal end of trochanter femoris. In dorsal view, condyla medialis et lateralis become wider distally.

PAGE 176

160 Comparative Osteology Size and relative proportions allow for a pr eliminary separation of large and small genera. The angle of facies articularis antitr ochanterica, depth of a ventral fovea just distal to facies articularis antitrochanteric a, origin of linea intermuscularis cranialis, presence of a trochanteric foramen, width of fossa popl itea, and width of crista tibiofibularis are also useful characters for similarly sized species. Asio cf. stygius La Chimba (2): These slender complete and an incomplete (lacking the distal end) femora are smaller than those of Bubo and larger than those of Otus and Glaucidium . Among intermediate-sized owls, the stout and short femur of Pulsatrix is unique in having a prox imal trochanteric foramen. Strix virgata has a proportionately shorter femur in which linea intermuscularis cranialis originates on the medial margin of trochanter femoris. The steeply angled facies articularis antotrochanterica, distal origin of linea intermuscularis crania lis at the base of trochanter femoris, slender shaft, and narrow crista tibiofibularis found in the archaeological specimen all coincide with Asio . The femora are longer than those from reference specimens of A. flammeus (as was the case for the ulna, radius, and carpometacarpus described previously) and mo re slender than those of A. clamator ; I therefore tentatively place these specimens into A. stygius which is intermediate in robustness between A. flammeus and A. clamator . Because of overall osteologi cal similarities between Asio and Tyto , I consider it important pointing out the differences be tween the femora of these two genera. Tyto has a deep fovea just distal to f acies articularis antitrochanter ica in the caudal aspect, linea intermuscularis cranialis originates proxima lly along the medial margin of trochanter

PAGE 177

161 femoris, the distal end is more massive, th e distal origin of cr ista supracondylaris medialis is thicker, crista tibiofibularis is wider, and fovea tendineus musculus tibialis cranialis is larger and located more caudad. TARSOMETATARSUS General Osteology The strigid tarsometatarsus is very distinct from that of other families. Crista medialis hypotarsi is large compared to crista lateralis hypotarsi, wh ich is reduced to a small process. On the ventral side, a tendinal bridge can be found on the medial half just proximal to foramina vascularia proximalis. Bo th sulcus flexorius and sulcus extensorius are very deep and the shaft is curved dorsa lly in a medial aspect. Distally, trochleae metatarsorum appear to form a half-circle in a distal aspect due to a more plantar location of trochleae metatarsorum II et IV (compared to trochlea metatarsi III) and enlarged medial and lateral wings on trochleae meta tarsorum II et IV, respectively. Trochlea metatarsi IV is strongly rotated laterally, its articular surface being perpendicular to facies plantaris. Comparative Osteology One tarsometatarsus was recovered from La Chimba. Given its size, it must belong to Pulsatrix , Asio , or Strix . Superficially the tarsometatarsi of Strix and Asio are very similar. Pulsatrix perspicillata (in lieu of P. melanota ) has a much stouter tarsometatarsus that is only slightly longer, but about twice as thick as the archaeological specimen. Tuberositas musculi tibialis crania lis is located more proximally than in Strix and Asio . Two small, ridge-like tuberosities or iginate from cotyla lateralis. These tuberosities are not found in Asio or the fossil. Strix bears only one ridge . In strigids, a

PAGE 178

162 muscle attachment line originates just distad of foramina vascularia proximalia lateralia. This line is more distad in Asio , Strix , and the archaeological specimen, than in P. perspicillata . On its more massive distal end, Pulsatrix differs mostly in the shape and position of cotyla metatarsi II. The medial margin of the bone leading up to cotyla metatarsi is at a steeper angle to axis proxim odistalis in cranial aspect than in the other two genera, and the wing of the cotyla meta tarsi II is stouter, projecting caudo-medially rather than caudally. On the basis of these traits P. melanota seems highly unlikely. Asio cf. stygius – La Chimba (1): The La Chimba specimen does not coincide completely with either Asio flammeus or Strix virgata , although it shares more traits with the former. Tuberositas musculi tibialis cran ialis is bulky and located more distad; no tuberosities emanate from cotyla lateralis on th e cranial side (a single tuberosity found in S. virgata ); and trochlea metatarsi IV is rounde d in shape (rather than oblong as in S. virgata ). It differs from that of A. flammeus by being considerably longer and more slender, though. As with other elements, I tentatively identify this specimen as A. stygius . OSSA DIGITORUM PEDIS General Osteology Ossa digitorum pedis of the Strigidae are very similar to those of the Tytonidae. They do lack, however, the flat proximal surface of facies dorsalis. Comparative Osteology Size alone allows separating large an d small genera. Among similarly sized species, general proportions regarding length, wi dth, and depth of phalanges help resolve some identifications. The widt h of the proximo-dorsal process is a reliable trait for

PAGE 179

163 phalanges intermediate, and the length and sh ape of tuberculum flexorium are good traits also. Bubo virginianus – La Chimba (4): The two phalanges in termediae (phalanx 2 of digitus secundus, and phalanx 3 of digitus ter tius) and two phalanges ungualis (hallux and digitus tertius) overlap in size with those of B. virginianus and P. perspicillata . Phalanx 2 of digitus secundus a more robus t and more deeply excavated fa cies plantaris than that of P. perspicillata . Phalanx 3 of digitus tertius is small for B. virginianus , but agrees well by having a similarly slender midsection and a more excavated facies plantaris (compared to P. persipicillata ). Small size of the archaeologica l specimens compared to modern reference skeletons reflects a general trend in decreasing body size toward the equator for this and other species (König et al. 1999). Asio cf. clamator – OGSE-80 (1): I assign this nearly complete phalanx 2 of digitus secundus to Asio sensu lato based on the narrow proximo-dorsal process and overall proportions as Asio sensu lato has longer and more sl ender phalanges than other genera. The archaeological specimen is significantly la rger and stouter than the phalanges from A. flammeus . I therefore assign it to Asio (= Rhinoptynx ) clamator , a species with reportedly powerful talons (König et al. 1999). Asio cf. stygius – La Chimba (7): Two complete phalanges 2 and complete phalanx ungualis of digitus secundus, complete pha lanx 2 and complete phalanx ungualis of digitus tertius, complete phalanx 3 and complete phalanx ungualis of digitus quartus correspond in slender shape and na rrow proximo-dorsal process to Asio . Strix virgata has shorter and thicker elements. Because they are larger than those of the reference

PAGE 180

164 specimens, and in accordance with the identifications of an ulna, radius, carpometacarpus, and femur (see above), I refer these phalanges to A. stygius . Order CUCULIFORMES Family CUCULIDAE SCAPULA General Osteology The ventromedial process of acromion (sen su Höfling and Alvarenga 2001) in the scapula of the Cuculidae is large and often grooved on its cranial margin. Acromion and the cranial margin of facies lateralis side form a raised ridge. Facies lateralis of corpus scapulae descends caudally in a concave sl ope from that ridge. Facies articularis humeralis lies in a plain parallel to facies lateralis of corpus scapulae. Table 2-22. Diagnostic characters and character states for th e proximal end of the scapula in the Cuculidae. Character Character State CSc1 Shape of facies articularis humeralis 0: round 1: oval CSc2 Size of ventromedial process of acromion 0: short 1: long CSc3 Relative size 0: small 1: large CSc4 Thickness of cranial margin 0: thin 1: thick Comparative Osteology Most of the differences within the Cuculidae are in the shape of facies articularis humeralis, the size of the ventromedial proce ss of acromion, and the na ture of the cranial groove on ventromedial process of acromion. Size is also an informative character (Table 2-22).

PAGE 181

165 Table 2-23. Character state matrix for diagnostic characters of the cuculid scapula in five species. Characters and character st ates are explained in Table 2-22. Character (CSc) Species 1 2 3 4 Coccyzus erythrophthalmus 0 0 0 0 Coccyzus americanus 0 0 0-1 0 Piaya cayana 1 1 1 1 Crotophaga ani 1 1 0-1 1 Crotophaga sulcirostris 1 1 0-1 0 Tapera naevia 1 0 0 1 Crotophaga sulcirostris – El Azúcar (1): Facies articula ris humeralis is roughly oval-shaped. The cranial end of scapula has a long and slender vent romedial process of acromion. This scapula of a medium-sized cu culid also has a thin cranial margin. The combination of these four traits is only found in Crotophaga sulcirostris (Table 2-23). Its congener Crotophaga ani is similar, but has a wide cranial margin. Piaya cayana also strongly resembels the El Azúcar specimen, but is larger (especially corpus scapulae) and has a thicker cranial margin. Order PICIFORMES Family RAMPHASTIDAE CARPOMETACARPUS General Osteology The carpometacarpus of the Ramphastidae is similar to that of other Piciformes of similar size. The ventral rim of trochlea carpalis is broad (slightly angled caudally in Rhamphastidae and Captitonidae and round in Bucconidae and Picidae), and the dorsal

PAGE 182

166 rim of trochlea carpalis is narrow, giving ex tremitas proximalis carpometacarpi a slanted appearance. Processus pisiformes is flat (m ore pointy in the Picida e and Galbulidae) and contiguous with the attachment of ulnare (sensu Hamon 1964). Located midway between the proximal and distal ends, processus in termetacarpalis is long (extending past os metacarpale minus). Os metacarpale minus is curved, bow-like (straight in Picidae). The distal end of distal symphysis, bearing faci es articularis digita lis minus, is short. Superficially the piciform carpometacarpus rese mbles that of the Passeriformes. In the passerines it tends to be longer, has a smaller extremitas proximalis carpometacarpi in relation to the total leng th of the bone, and has a longer distal symphysis. Comparative Osteology The ramphastid genera are quite distinct in terms of size which limits the La Chimba specimen to the genera Selenidera and Aulacorhynchus ( Andigena , Pteroglossus , and Ramphastus are larger). The shapes of proc essus pisiformes and processus extensorius as well as the lengt h of the distal symphysis can be used to distinguish among species and genera. Aulacorhynchus prasinus – La Chimba (1): The complete carpometacarpus can be assigned to Aulacorhynchus as Selenidera has a stubbier processus pisiformis and a more excavated attachment of ulnare. The most distinctive species of Aulacorhynchus is A. haematopygus with a very stout processus extens orius and short distal symphysis. Aulacorhynchus prasinus differs from A. derbianus by a more slender processus extensorius, a less oval, nearly round attachment ulnare, a nd a longer symphysis distalis (Figure 2-8).

PAGE 183

167 Figure 2-8. Carpometacarpus from the La Chimba specimen (a), Aulacorhynchus prasinus (b), A. derbianus (c), and A. haematopygus (d). For each carpometacarpus both the dorsa l (left) and ventral (righ t) aspects are shown in this composite image. Reference bar is 1 cm. Order PASSERIFORMES OSSA MAXILLAE (OS NASALE, OS PREMAXILLARE ET ROSTRUM MAXILLAE) General Osteology Ossa maxillae (upper jaw bones) are usua lly a stout set of bones with small holorhinal apertura nasi oss ea. Only in the Furnariidae are aperturae nasarum ossea pseudo-schizorhinal (the opening is not confined to ossa max illae, but extends into zona flexoria craniofascialis; Feduccia 1973). The nares are delimited rostrad by os premaxillare and caudad by os nasale. Crista to mialis is sharp and the ventral surface is bisected by a pair of canales neurovascularia maxillae. At the Ecuadorian sites, rostrum

PAGE 184

168 maxillae is the most commonly preserved part of the upper jaw, ofte n lacking the rostral end and os nasale. Comparative Osteology There is great inter-familiar variation in the passerine rostrum. The most useful characters to match archaeological specimens to modern reference specimens are size and shape of nares, the strength of processus na sale of os premaxillar e, the thickness of the lateral struts of os nasale, the degree of tapering of os premaxillare toward the rostral end, and the width and orientation of canales neurovascularia. Family Thraupidae, Buthraupis montana – La Chimba (1): The wide upper jaw includes complete os premaxillaris, os nasale, os lacrimale, and the cranial end of os frontalis. Nares are oval and notched ventrall y. Processus nasale of os premaxillare is very stout and the lateral struts of os nasale are flat. Canale s neurovascularia are wide and at a small angle to plana dorsalis, ankylosi ng along a slightly raised central line. Os lacrimale is bulbous and prominent (Figure 29). I have not seen this combination of traits in any other species. Figure 2-9. Skull in dors al aspect from La Chimba specimen (a) and Buthraupis montana (b).

PAGE 185

169 Family Emberizidae, Volatina jacarina – El Azúcar (1): With large nares, flat os nasale, and a cone-shaped os premaxillaris, this small os maxillae belongs to a small emberizid. Sporophila has a broader upper jaw with a st ronger processus nasale of os premaxillare, and Tiaris has concave tomia in vent ral view. In addition, only V. jacarina agrees in having fairly straight culmen w ith a low angle to axis rostrocaudalis (more angled in Tiaris , arched in lateral aspect in Sporophila ; Steadman 1982, Steadman and Morgan 1985) and slightly curved pr ocessus jugales (strongly curved in Tiaris , Sicalis , and Sporophila ). Family Emberizidae, cf. Sporophila peruviana – El Azúcar (1): A short, rounded, and stout os maxillae with a very thick proce ssus of os nasale, suggest this specimen is from a species of Sporophila . This combination of traits o ccurs in no other genus except Melopyrrha from the West Indies (Steadman a nd Morgan 1985, Fig. 2). The rostrum in Tiaris , Volatina , Oryzoborus angolensis , and Phrygilus is longer and finer; in O. funereus (in lieu of O. crassirostris ) it is much larger and has a high ly curved os maxillare; it is also much larger in Sicalis . Family Emberizidae, cf. Sicalis taczanowskii – El Azúcar (4): Identification of these four rostra is problematic due to th e lack of comparative material. Having the typical emberizid-type os maxillae (parallel medi al bars of os nasale, flat os nasale, the wide os premaxillare, highly curved culmen in lateral aspect, and ventrally deflected os maxillaris) suggests that this specimen comes from a seed-eating bird with a deep bill. Processus medialis of os nasale and os max illare are not as thick and curved, respectively, as in O. funereus . The flat and large os nasale, cone -like shape of os premaxillare in dorsal aspect, as well as ove rall size, are suggestive of Sicalis or Aimophila . The

PAGE 186

170 specimens share with Sicalis stout lateral struts of os nasale (finer in Aimophila ), a highly curved os maxillare (straighter in Aimophila ), centrally constricted processus of os nasale (equally thick along its entire length in Aimophila ), a triangular and deep processus medialis of os nasale in cross-section (in Aimophila , ventral vertex of processus medialis not as deep), and round nares (oval in Aimophila ). A specimen of S. taczanowskii is needed to confirm this identification. Family Icteridae, Sturnella bellicosa – OGSE-80 (1), El Azúcar (2): This os premaxillare, lacking most of the rostral portion of rostrum maxillae, comes from a narrow and long beak. The rostral portion of na res as well as parts of processus medialis of os nasale and parts of os maxillare ar e preserved in the specimen. The round rostrum in rostral aspect, and pointed shape are only found in icteri ds (in large dendrocolaptids the culmen is much more triangular). I have found the strong processus nasale and short, ventrad curving os maxillare only in S. militaris . I assume that the closely related S. bellicosa differs mostly in size and not in shape). OSSA MANDIBULAE General Osteology Os dentale is usually thick rostrally a nd thin caudally, the transition marked by angulus mandibulae. At angulus mandibulae the ramus bends ventrally. Fenestra may be present in the caudal intramandibular flexi on zone (Baumel et al. 1993, p.74). Processus mandibulae medialis is large and flat, a nd processus retroarticularis is thick. Comparative Osteology As with the ossa maxillae, the ossa mandibulae are highly variable in the Passeriformes. Good comparative features are its shape of in caudal or ventral aspect,

PAGE 187

171 length of the mandibular symphysis in relation to total length of the mandible, degree of curvature at angulus mandibulae, number of fenestra, and si ze and shape of processus mandibularum medialis et retroarticularis. Family Cardinalidae, Spiza americana – El Azúcar (1): This partial symphysis with a portion of the left ramus is unusual in severa l regards. The thin section of os dentale rapidly increases in depth just caudal to angulus mandibulae, a character found in one museum specimen of S. americana (UF 41291). The transition from the thick rostrum mandibulae to the thin ramus mandibulae pars intermedia begins rostrad on the dorsal side and ends caudad on the ventral side, a cond ition I have only found consistently in the Cardinalidae. The degree of curvature of angul us mandibulae in medial view is very low, lower than that of any other specimens examined. The caudal end of mandibular symphysis is broad and round (Figure 2-10). Th e combination of traits appears to be restricted to the Cardinalidae, wherein Pheucticus spp. and Passerina are similar to Spiza , but much larger and much more massive. Figure 2-10. Dorsal aspect of the mandibl e from the El Azúcar specimen (a) and Spiza americana .

PAGE 188

172 Family Icteridae, Icterus cf. graceannae – El Azúcar (1): The orioles, blackbirds, grackles, caciques, and oropendolas are characterized by a pronounced processus retroarticularis and an elongate but strong cotyla lateralis. This process is very long in Sturnella , Psarocolius , Cacicus , and some Icterus , short and wide in other Icterus , and short and knobby in Quiscalus and Dives . The caudal fragment from the archaeological specimen has the short and flat condition of Icterus mesomelas , but is smaller overall. I therefore tentatively assign this fragment to the smaller I. graceannae . Family Icteridae, Sturnella bellicosa – El Azúcar (5): By the same traits as the ones described for the previous species, the two caudal ends correspond to an icterid. Size and length of the processus retroartic ularis suggest they are from S. bellicosa . The three rostral ends can be ascribed to this species because of size, concave rami in dorsal view, a thin symphysis, and low degree of thickening of rostrum mandibulae, all characteristics of long slender bills. STERNUM General Osteology The passerine sternum is a very uniform element with an enlarged and cranially bifurcating spina externa. Spina interna is absent. Most incisurae intercostales are located a flat, expanded, and cranially oriented proce ssus craniolateralis. The traberculae laterales expand caudally. Fenestra are absent. Comparative Osteology The sternum of passerines is difficult to identify except in certain large species. Large suboscines such as Grallaria przewalskii , Querula purpurata , and Rupicola peruviana have a ventral, sheet-like projection on spina externa, which is absent in the

PAGE 189

173 oscines. In Xiphocolaptes promeropirhynchus , this projection is also absent. This species also differs in that spina externa is much reduced, probably one of many adaptations in the shoulder girdle to a scansorial lif e-style (M. P. Tellkamp unpublished data). Family Icteridae, Psarocolius cf. angustifrons – La Chimba (1): Lack of a ventral sheet on the spina exte rna excludes suboscines for this cranial ster nal fragment. Within the oscines, corvids have a thin spina externa in which the vent ral margin is as long as the dorsal margin in lateral aspect. In Turdus and Psarocolius the ventral margin is longer than the dorsal margin. Turdus differs by a thin ventral spina (heavy and triangular shaped in cross-section in Psarocolius ) SCAPULA General Osteology Variation in the passerine scapula is larg e. Suboscines, having a long acromion and short/absent ventromedial process of acr omion (sensu Höfling and Alvarenga 2001), resembling that in the Cuculiformes and Galbulidae (Order Piciformes), whereas the oscines, usually bearing a pronounced ve ntromedial process of acromion, show resemblance to that of the Coraciiformes and Piciformes (except Gal bulidae). In all cases is the passeriform scapula less textured (in term s of muscle attachment sites) and of finer build. In addition, the oscine sc apula differs from that of ot her orders in having a flat ventromedial process of acromion. Comparative Osteology As with other elements, there is a clear sp lit between the subosc ines and the oscines in the scapula. Suboscines have a much re duced ventromedial process of acromion or

PAGE 190

174 lack it completely. Oscines in turn have a very pronounced ventromedial process of acromion. I used presence/absence of the ventromedi al process of acromion (simply “process” for the remainder of this paragraph) and its sh ape (when present), as well as the structure of acromion proper to assign specimens to gene ral size classes (Chapt er 4). Corvids have a medially truncated process and a stout, pointy acromion. Turdus thrushes have a rounded process in cranial aspect and a stout , pointy acromion bearing a small ventral tuberosity. The process is laterally truncated in Campylorhynchus and dorsally truncated in Mimus , with acromion being otherw ise similar to that of Turdus . Saltator and other cardinalids have a pointy process and a fla ttened acromion. A foramen can often be found at the cranial origin of the process in these bi rds. In the Icteridae, the process is dorsally truncated and acromion is knob-like. In Ps arocolius a foramen may be found at the ventral-medial base of the acromion. CORACOID General Osteology The passerine coracoid is highly distinc tive, being a slender bone with a narrow proximal end, a short, triangular processu s procoracoideus, and a long, decurved tuberculum brachiale. Facies articularis ster nalis is located dorsad and impressio musculi sternocoracoidei is delimite d by sharp ridges. Processus acrocoracoideus has a pointy appearance due to a steep impressio ligamen ti acrocoracohumeralis and the decurved tuberculum brachiale. The coracoid of the Piciformes is similar in having an overall slender shape (Höfling and Alva renga 2001), but differs as faci es articularis sternalis is

PAGE 191

175 more raised, rounder and deeper, processus ac rocoracoideus is rounder, and tuberculum brachiale is absent. Comparative Osteology The suboscine/oscine dichotomy is reflect ed clearly in the distal end of the coracoid. Whereas oscines have a highly decurved, pointy tuberculum brachiale (curvature nearly 180°), in suboscines the often flat, rounde d tuberculum brachiale is almost at a right angle to axis proximodista lis (curvature ca. 100-120°; trait PaC1). Other traits distinguishing among different taxa are th e shape of the facies articularis sternalis (trait PaC2), the length of impressio musculi sternocoracoidei in rela tion to the length of the shaft (trait PaC3), the shape of the pro cessus procoracoideus (trait PaC4), and the shape and relative position of the insertion for aponeurosis of musculus biceps brachii (trait PaC5). Family Turdidae, Turdus fuscater – La Chimba (2): The two complete coracoids are from a medium-sized oscine, larger than in Butraupis montana , but smaller than in Psarocolius angustifrons . Apart from size, Turdus fits the archaeological specimen very well, as it shares with the Turdidae a broad upper shaft in dorsal aspect, a long impressio ligamenti acrocoracohumeralis, and a short pr ocessus acrocoracoideus. Together, these three traits give the distal end a broad and blunt, rather than poin ty appearance as found in most other oscines. Many suboscines al so have a broad distal end, but differ substantially by the l ong, mediad oriented tuberulu m brachiale (PaC1). Processus procoracoideus (PaC4) has a low, asymmetric , triangular shape in most Turdidae as well as in the La Chimba specimens.

PAGE 192

176 Family Turdidae, Turdus cf. maculirostris – El Azúcar (1): This bone is very similar in shape and size to T. serranu s. The similarly-sized Turdus maculirostris is the only large thrush in the dry western lowlands. -1.500 0.000 1.500 -5.0000.0005.000 PC1PC2 Archaeological Specimens Psarocolius Cyanocorax Figure 2-11. A principal components analysis on five variables of the coracoid in large oscine passerines ( Psarocolius s. l. includes Gymnostinops montezuma , Psarocolius viridis , P. decumanus , P. angustifrons , and Zarhynchus wagleri ; Cyanocorax s. l. includes Psilorhinus morio , Cyanocorax affinis , C. cayanus , C. cyanomelas , C. dickeyi , C. sanblasianus ), supporting the notion that the archaeological specimens come from oropendolas (Family Icteridae). Family Icteridae, Psarocolius cf. angustifrons – La Chimba (2): The two large specimens can be differentiated from the co racoids of other large Passeriformes as follows: 1) a caudad pointing tuberculum brach iale (PaC1; medially pointing in most suboscines, tuberculum brachiale almost absent in Family Dendrocolaptidae), 2) a pointy, almost symmetrical processus procoracoideus (PaC4; flat, asymmetric in Corvidae and Turdidae), and 3) the oblong in sertion for aponeurosis of musculus biceps brachii (trait PaC5) being located medially, having a mediolateral orientation (more distal and with cranio-caudal orientation in the Corvidae). Co rvids and oropendolas ove rlap very little in

PAGE 193

177 morphometric space (Figure 2-11). Both pr incipal component 1 (PC1) and principal component 2 (PC2) place th e zooarchaeological specimens with the oropendolas. Family ICTERIDAE, cf. Molothrus ( Scaphidura ) oryzivora – La Chimba (2): Identification of this pair of nearly complete, large coracoids is tentative as a modern specimens for M. oryzivora were not available. Despite havi ng an icterid distal end, they have a short impressio musculi sternocoracoidei in relation to the shaft (PaC3; short in Molothrus bonariensis , but longer in Psarocolius ), and a flat and asymmetric processus procoracoidus (PaC4; pointier in Psarocolius ). The only species of icterid large enough to be considered for the two specimens is M. oryzivora . This cowbird is a parasite of oropendolas and could have been captured t ogether with oropendolas (possibly as a nestling). This hypothesis is weakly supported by the fact that the two bones were found in the same quadrate an d level as those from Psarocolius . Table 2-24. Unrotated factor loadings for f our measurement in a principal components analysis on the distal end of the coraco id in several species of Corvidae and Icteridae (see Figure 2-6). Traits* F1 F2 F3 F4 F5 1 0.965 0.161 -0.183 -0.019 0.098 2 0.951 0.280 0.052 0.078 -0.094 3 0.966 -0.008 0.203 -0.154 0.044 4 0.958 -0.221 -0.131 -0.094 -0.088 * 1: Distance between processus procoracoi deus and extremitas omalis coracoidei. 2: Distance between processus procoracoideus and the base of processus glenoidales. 3: Length of impressio ligamenti acrocoracohumeralis. 4: Length of facies articularis humeralis. Family Icteridae, Sturnella bellicosa – El Azúcar (1): I assigned this complete coracoid to Sturnella based on size, long impressio mu sculi sternocoracoidei (PaC3), shape of insertion for aponeurosis of muscul us biceps brachii (tra it PaC5), and a pointy

PAGE 194

178 distal end. All three traits ar e found in the icterids. Despite being of similar size, the coracoid of Icterus is proportionately shorter. HUMERUS General Osteology The passerine humerus is easily distinguished from that of other avian orders. Most obvious is the distal end with a long, stout, and distad proj ecting processus flexorius and a prominent processus supracondylaris dorsalis. On the ventro-medial side the attachment site for the ligamentum craniale cubiti is a conical tuberosity. Fossa musculi brachiale is small, somewhat triangular, and restrict ed to the ventral side of the bone. Comparative Osteology Most informative traits of the humerus are found in the proximal end, with the condition of the pneumatic fossa (PaH1) ta king a central role (Bock 1962; Table 2-25, Figure 2-12). The ancestral condition is a single pneumatic fossa found in suboscines. The derived condition is the double pneumatic fo ssa developed to vary ing degrees in the oscines. I identified six character states for the pneumatic fossa, ranging from completely absent to very deep (Table 2-25). Based on th ese character states, the suboscines can be readily distinguished from the oscines, as no suboscines have a state higher than 1 and no oscines score below 2. Other traits that appear to vary among the families are thickness of the bicipital crest (PaH2) and the s ite of attachment of supraspina tus (or musculus scapulohumeralis anterior sensu Bock [1962]; PaH3). These tw o traits are coupled in all the specimens examined (Table 2-26): a thickened bicipita l crest and internal location of musculus

PAGE 195

179 scapulohumeralis anterior can be found in all oscines wi th the exception of Progne chalybea (Hirundinidae). Figure 2-12. Variation in the second pneuma tic fossa in the passerine humerus from absent (a) to highly deve loped (g, j). Species are Grallaria przewalskii (a), Megarhynchus pitangua (b), Tityra semifasciata (c), Cyanolyca viridicyana (d), Turdus serranus (e), Mimus gundlachii (f), Anisognathus lachrymosus (g), Saltator maximus (h), Cacicus haemorrhous (i), and Sturnella militaris (j). Note that S. maximus is unusual for Emberizidae by having a strong medial bar, and C. haemorrhous is unusual among icteri ds in having a poorly developed secondary pneumatic fo ssa. White bar represents 1 cm. The distal end does not have the same le vel of variability as the proximal end. However, several traits distinguish the most s uboscines from the oscines: fossa olecrani is shallow, sulcus scapulotricipitalis is shallow, and the fossa musculi brachialis is not as deep. Finally, processus supracondylaris dor salis is usually well developed in both suboscines and oscines, but in some furnar iids it is reduced to a mere tuberosity.

PAGE 196

180 Family Cotingidae, cf. Tityra semifasciata – La Ponga (3): The fragments are of the type 2 second pneumatic fossa, which seems to be characteristic of the Tyranni. Specieslevel identification is based on an as sociated complete ulna (see below). Table 2-25. Diagnostic characters and charac ter states for the proximal end of the humerus in the Passeriformes. Character State PaH1 Double pneumatic fossa 0: Second pneumatic fossa is absent and capital shaft ridge slightly raised resulting in a concave surface betwee n the ridge and the pneumatic fossa; concave surface lies distal to cauda l margin of the pneumatic fossa. 1: As previous state, but the concave surface extends more proximally to the posterior surface of the pneumatic fossa. 2: A second pneumatic fossa is a slight depression distal to the head of the humerus and posterior to the capital-shaft ridge. 3: Second pneumatic fossa extends under the head of the humerus. 4: The second fossa is large and deep but the medial bar is strong; the capital-shaft ridge well developed with a well defined posterior wall; the bicipital crest meets the shaf t at almost a right angle. 5: Similar to three, but the medial bar is virtually absent PaH2 Crista bicipitalis 0: Thickened dorsally, but very narrow ventrally making fossa less defined. 1: Thickened throughout raising the crest above the level of the interior wall of the fossa. PaH3 Location of m. scapulohumeralis anterior 0: Distal to the pneumatic fossa, not well defined. 1: Inside pneumatic fossa in well defined depression. Family Turdidae, Turdus fuscater – La Chimba (5): In the three specimens that include at least a portion of the proximal end, the type 4 double pneumatic fossa is clearly visible. Since this type is associated with 10-primarie d oscines as well as some Cardinalidae (Table 2-26), Thraupidae (t anagers) and Icterida e (oropendolas, and caciques) can be excluded. The only highland 10 -primaried oscine to be large enough to meet the conditions of both large size and a type 4 double pneumatic fossa is T. fuscater . The distal end can be assigned to T. fuscater based on size and a long fossa olacrani (round, or long with round depr ession, in Icteridae). One of the bones belongs to a juvenile and the identification is therefore tentative. The second pneumatic fossa appears to be of type 3, but the lack of depth might be due to

PAGE 197

181 Table 2-26. Character state matrix for the pa sserine humerus regard ing three diagnostic characters in 18 species of Passeriformes . Characters and character states are explained in Table 2-25. Characters PaH Species 1 2 3 Furnarius leucopus 0 0 0 Syndactyla subalaris 0 0 0 Grallaria przewalskii 0 0 0 Megarhynchus pitangua 1 0 0 Lipaugus vociferans 1 0 0 Querula purpurata 1 0 0 Rupicola peruviana 1 0 0 Corvus imparatus 1 0 0 Cyanocorax cyanomelas 2 0 0 Progne chalybea 3 0 0 Riparia riparia 3 0 0 Turdus serranus 4 1 1 Mimus gundlachii 4 1 1 Buthraupis montana 5 1 1 Sicalis flaveola 5 1 1 Saltator maximus 4 1 1 Leistes militaris 5 1 1 Psarocolius angustifrons 3-4 1 1 incomplete development of the bone. This specimen is not Cacicus , the only oscine of this size class with a type 3 secondary pneumatic fossa, because of the low curvature of caput humeri (rounder in Cacicus ), flat distal margin of the caput humeri (forms a

PAGE 198

182 “widow peak” in Cacicus ), a more dorsal position of th e margo caudalis, and a more vertical orientation of tuberculum ve ntralis with regard to the shaft. Family Mimidae, Mimus cf. longicaudatus – El Azúcar (1): Having a type 4 second pneumatic fossa, this specimen differs from that of Turdus , Campylorhynchus , and Saltator maximus by a compressed caput humeri and a sh arply angled crista bicipitalis that makes the head look proportionately wider. The second pneumatic fossa undercuts caput humeri. I have not found this condition in any other oscine. It is approached to some degree in Pheucticus chrysogaster , which has an overa ll stouter humerus. Family Hirundinidae, Riparia riparia – El Azúcar (2): The two complete humeri are swallows by their type 3 second pneuma tic fossa, high degree of rotation, and their short, stocky built. Within the Hirundinidae, Stelgidopterix is larger and has a more angled crista bicipitalis, Petrochelidon is larger and has a lo nger tuberculum ventrale, Notiochelidon is smaller and has a less angled crista bicipitalis, Tachycineta is stouter having tuberculum ventrale closer to the shaft (more narrow head overall), and Hirundo is larger with tuberculum ventrale closer to the shaft (as in Tachycineta ), lacks a thickening of the margo caudalis, and has a longer processus flexorius. Family Icteridae, Psarocolius cf. angustifrons – La Chimba (7): All of the specimens are oropendolas by size and a type 3 second pneumatic fossa and/or a long fossa olecrani with a central r ound depression. Due to the large degree of overlap in size and high intraspecific variability, identifica tion to species is difficult. I assign the specimens to P. angustifrons based on the shape of the ca put humeri (more massive in Zarhynchus wagleri ), the length of the crista deltopectoralis (longer in Z. wagleri ), the shape of the crista deltopectoralis in dorsal aspect (apex more distad in P. decumanus ),

PAGE 199

183 and the shape of fossa olecrani (not as well delimited along condylus ventralis in Z. wagleri and P. decumanus ) Family Icteridae, Quiscalus mexicanus – El Azúcar (1): This is a large proximal end with a type 5 second pneumatic fossa. Although smaller than the specimens examined, Quiscalus is the only large genus with this type of secondary pneumatic fossa in western Ecuador. Family Icteridae, cf. Dives warszewiczi – El Azúcar (2): The archaeological specimen is similar to that of Quiscalus and Sturnella , but it is intermediate in size and has noticeable thicker medial bar (almost approaching Turdus in thickness). Family Icteridae, Sturnella bellicosa – OGSE-80 (3), El Azúcar (5): In terms of shape and proportions the specimen matches this species perfectly. No other species of this size class has this wide type of secondary pneumatic fossa. ULNA General Osteology In the passerine ulna, the olecranon is usually pointy, cotyla dors alis “ear-like”, the shaft dorso-ventrally compressed, condylus vent ralis ulnaris wide a nd flat ventrally, and tuberculum carpale flat and round. The long and narrow impressi o brachialis is located on ventral half of the ve ntral side, hardly reaching the cran ial half of the ventral side. The attachment of ligamentum radioulna re interosseum is very deep. Comparative Osteology As in the humerus, oscines have a pneu matized fossa on the proximal end. In the ulna it is located on the dorsal half of the ventral side of the shaft, just dorsad of the proximal end of impressio brachiale. Species may differ in terms of the exact position of

PAGE 200

184 the fossa. In suboscines this fossa is absent . The olecranon is longer in oscines than in suboscines; in ventral aspect, it forms a straight line to the proximal end of tuberculum brachiale from whence the shaft bends craniall y. The bend is smooth in most oscines, but pronounced in most suboscines, especially the Furnariidae. Table 2-27. Character states of the suboscine ulna. Character Character State PaSU1 Shape of caudal margin of olecranon and proximal shaft in dorsal aspect 0: Strongly S-shaped 1: Weakly S-shaped PaSU2 Shape of the proximal ventral margin of the impressio brachialis leading up to the tuberculum ligamenti collateralis ventralis 0: Strongly curved (ca. 80 û angle to the shaft) 1: Weakly curved (ca. 60 û angle to the shaft) PaSU3 Relative length of ulna relative to femur (given than species were selected on approximate weight) 1: Shorter 2: Equal 3: Longer Family Cotingidae, Tityra semifasciata – La Ponga (4): Being from a suboscine by the lack of a proximal pneumatic fossa, thes e ulnae belonged to a medium-sized bird. I thus compared this specimen to the ulnae of Furnarius , Thripadectes , Xiphorhynchus , Taraba , Grallaria , Lipaugus , Tityra , Megarhynchus , and Myiodynastes (Tables 2-27 and 2-28). Only that of Tityra semifasciata agrees in terms of the shape of olecranon (PaSU1) and the shape of the ventral margin of im pressio brachiale (PaS U2; Table 2-28). In addition, even though species in the character matrix were selected based on body mass, the ulnae of Taraba major and Grallaria przewalskii are too small to be considered. Family Mimidae, Mimus cf. longicauda – El Azúcar (2): The two proximal ends possess a single pneumatic fossa. Within the oscines, I only found the genus Mimus to have a reduced tuberculum ligamenti collate ralis ventralis that does not extend caudodistally beyond the proxim al margin of impressio brachialis. In addition, Mimus has a shallow impressio brachialis than in other oscines.

PAGE 201

185 Family Turdidae, Turdus cf. fuscater – La Chimba (1): The large and slender complete ulna has a more dorsal pneumatic fossa than typically found in the stout ulna of the Thraupidae. The shape of trochlea carpalis in ventral view is also indicative of Turdus . The angle formed between condyla dorsalis et ventralis is sharper than that found in Buthraupis . Condylus ventralis is rounder and bulkier in Buthraupis as well. Table 2-28. Character states for the ulna in nine suboscine genera. See Table 2-27 for definition of the characters. Characters (PaSU) Species 1 2 3 Furnarius cinnamomeus 1 0 2 Thripadectes rufobrunneus 0 1 1 Xiphorhynchus triangularis 0 1 2 Taraba major 0 0 0 Grallaria przewalskii 0 1 0 Lipaugus vociferans 0 1 2 Tityra semifasciata 1 1 2 Megarhynchus pitangua 0 0 2 Myiodynastes maculatus 0 0 2 Family Turdidae, Turdus serranus – La Chimba (1): Being very similar to the ulna of T. fuscater qualitatively, this complete ulna is about 20% smaller, which matches the proportional difference in wing measurements between T. fuscater and T. serranus given by Clement (2000). This bone only is slightly larger than in a 73 g museum specimen of Turdus serranus (UF 43409). Family Thraupidae, Sericossypha albocristata – La Chimba (1): The proximal end of ulna differs from that of Buthraupis and Turdus by a flat olecranon, proximally deeply

PAGE 202

186 excavated impressio brachialis, and deep pr oximo-caudal depression (area bordered by olecranon, cotyla dorsalis, and trochlea hum eralis ulnaris). I have not found tis combination of traits in any other large oscine. Family Thraupidae, Buthraupis montana – La Chimba (1): A large ulna belonging to Buthraupis by the characteristics mentioned above in the description for Turdus fuscater . The large size makes the smaller B. eximia unlikely (Isler and Isler 1987). Family Icteridae, Psarocolius cf. angustifrons – La Chimba (1): This very large ulna with a pneumatic fossa comes from a la rge oscine. The only oscine in this size category is Psarocolius sensu lato. The archaeological specimen agrees with that of P. angustifrons rather than P. decumanus in having a less bulky tuberculum ligamenti collateralis ventralis and a mo re slender shaft, and from Zarhynchus wagleri by being larger, having a more excavated impressio brachialis, and a more slender shaft. Family Icteridae, Sturnella bellicosa – El Azúcar (1): This medium-sized, complete ulna has a pneumatic fossa that extends cranially ben eath the tuberculum ligamenti collaterialis ventralis (typical for the Icteridae) as S. bellicosa . I have not seen such an expanded pneumatic fossa in any other species. CARPOMETACARPUS General Osteology The carpometacarpus of the Passeriform es is a very uniform bone. Trochlea carpalis is highly asymmetric in proximal view , with the dorsal rim of the trochlea being located more cranially and proximally than the ventral rim. Processus pisiformes is reduced to a small, flat protuberance. The lig amental attachment of ulnare (sensu Hamon 1964), a shallow fossa infratrochlearis, a nd elongate and concave attachment for

PAGE 203

187 musculus flexor digiti minoris, divide the ve ntral surface into three distinct areas. The attachment for musculus flexor digiti minoris extends distad well onto metacarpal III. The dorsal surface of trochlea carpalis is na rrow and elongate. Fovea carpalis caudalis is usually deep. A processus intermetacarpalis is found in all birds of this order. Facies articularis digitalis minor is located on a long protuberance that extends well beyond facies articularis digitalis major. This char acter readily distinguishes the Passeriformes from the superficially similar Piciformes. Comparative Osteology There are few characters that reliably distinguish between different taxonomic groups in this order (Table 2-29). Protubera ntia metacarpalis (PaCMC3) can only be found in the nine-primaried oscines. The cran ial portion of sulcus te ndineus (PaCMC4) is consistently longer in oscines. Facies articularis digitalis minor (PaCMC6) is at a right angle to axis proxiodistalis in the oscines a nd at a more closed a ngle in the suboscines. Other characters are less cons istent, but can help distinguis hing between different species of similar sizes within the suboscines and oscines, respectively. The limited number of large passerines for both the Andean and the coastal sites makes it possible to identify large specimens to genus and even species. Family Cotingidae, Tityra semifasciata – La Ponga (1): The distal end of this carpometacarpus is a suboscine based on shape of facies articularis digitalis minor (trait PaCMC6). The identification of this specime n is based on its association with other bones, including a complete ulna (see above).

PAGE 204

188 Table 2-29. Diagnostic characters and charac ter states for the carpometacarpus (Order Passeriformes). Character Character State PaCMC1 Attachment of ulnare 0: shallow, oblong 1: round, deep PaCMC2 Fovea carpalis caudalis 0: shallow 1: deep PaCMC3 Protuberantia metacarpalis 0: absent 1: present PaCMC4 Sulcus tendineus Fraction of metacarpal II for which sulcus tendineus can be seen in cranial view 0: 1/2 1: > 1/2 PaCMC5 Depression on ventral symphysis metacarpalis distalis 0: absent 1: shallow 2: deep PaCMC6 Facies articularis digitalis minor 0: narrow ventral articular su rface quickly at a steep angle to axis proximodistalis giving distal end a pointy appearance in dorsal aspect 1: narrow ventral articular su rface at right angle to axis proximodistalis, giving the dist al end a squared appearance in dorsal aspect Family Turdidae, cf. Turdus fuscater – La Chimba (1): Only T. serranus and smaller Turdus spp. were available for comparison. Having the oscine condition of facies articularis digitalis minor (PaCMC6), I refer this complete carpometacarpus to Turdus based on the combination of a shallow attachme nt of ulnare (PaCMC1) and a moderately deep depression on the ventral si de of symphysis metacarpalis distalis (PaCMC5) that is not as deeply excavated as in some Icterid ae. In addition, the atta chment of musculus flexor digiti minoris begins more distad (d istad of processus piciformes) than in the Icteridae (proximad to processus piciformes). Family Thraupidae, Buthraupis montana – La Chimba (1): Being smaller than that of Cacicus and intermediate in size between those of Turdus fuscater and Turdus serranus , this complete carpometacarpus belongs to a mediumto large-sized oscine (see traits PaCMC3, PaCMC5, PaCMC6; Tabl e 2-30). It differs from that of Turdus in having a deep attachment of ulnaris (PaCMC1), and from that of Cacicus in having a shallow

PAGE 205

189 depression on the ventral symphysis metacar palis distalis (PaCMC5). The specimen is also different from the carpometacarpus of Sericossypha albicristata by having a smaller processus piciformes and a more slender at tachment area for musculus flexor digiti minoris. Family Icteridae, Psarocolius angustifrons – La Chimba (2): This massive, complete carpometacarpus from a large os cine (PaCMC3, PaCMC5, PaCMC6; Table 230) belongs to an oropendola. No species of oropendola presently occurs above 2000 m making it necessary to compare several species of Psarocolius as well as Zarhynchus wagleri . The large degree of variation in body mass results in great overlap between P. decumanus , P. angustifrons , P. viridis , and Z. wagleri (Dunning 1993), making size an inadequate character to distinguish among species. The large size of the La Chimba specimen does suggest, however, that it co mes from a male. The carpometacarpus of Psarocolius angustifrons and P. decumanus are very similar overall, with the main difference found in the shape of the processu s extesorius in proximal aspect. Whereas that of P. angustifrons has a narrow and sharp cranial marg in that widens at the base of processus alularis, in P. decumanus there is a broad and round proximal margin. In Z. wagleri this condition is intermediate. Zarhynchus differs from P. angustifrons further in having a slightly deeper fossa in fratrochlearis with a distinct foramen in its center, and a broader and shorter distal end (p ast facies articularis major). Family ICTERIDAE, Quiscalus mexicanus – El Azúcar (1): The large distal end of an oscine carpometacarpus (PaCMC6) comes from Q. mexicanus based on size and the depth of the depression in th e distal symphysis (PaCMC5; Table 2-30). These traits are not found in the other large passeri ne on the Santa Elena Peninsula, Cyanocorax

PAGE 206

190 mystacalis . The genus Cacicus can be ruled out because of a shorter facies articularis digitalis major, more extensive depression in the distal symphysis, and the cross-sectional shape of the dorso-distal margin of os metacarpale minus (flat in Quiscalus and round in Cacicus ), and an almost flat caudal surface of os metacarpale majus (more concave in Quiscalus ). FEMUR General Osteology The passerine femur is characterized by a small, dorso-ventrally compressed head and the lack of trochanter femoris. Fovea ligamenti capitis is deep and wide, covering most of the dorsal surface of the head. The di stal end has a curved appearance in cranial view, mostly due to the medial orientati on of condylus medialis and the longer and bulkier condylus lateralis. A less curved distal end and a flattened distal shaft differentiate the femur of most Picifo rmes from that of the Passeriformes. Comparative Osteology The femur is a conservative bone with few distinguishing features and great intraspecific variation. Reliable species id entifications are difficult and often only possible in a few taxa with distinct osteol ogical features. In ge neral, the shaft of saltatorial species and woodcr eepers is highly curved in la teral view. Corvids can be distinguished by a long concave surface just distal to the condylus medialis that serves as the attachment site of musculus femorotibiali s internus (after Berger 1968). The medial margin of the distal shaft is sharp an d ridge-like in corvids. The Turdidae are characterized by a foramen just distad of the head on the proximal end of the femur. This condition can also be found in Sturnella bellicosa .

PAGE 207

191 Table 2-30. Character state matrix for the carpometacarpus in the Passeriformes. See Table 2-29 for definitions of the character states. Characters (PaCMC) Species 1 2 3 4 5 6 Furnarius leucopus 1 0 0 0 0 0 Syndactyla subalaris 0-1 1 0 0 0 0 Grallaria przewalskii 0 0-1 0 0 0 0 Phlegopsis nigromaculata 0 0 0 0 0 0 Megarhynchus pitangua 0 0 0 0 1 0 Tyrannus melancholicus 0 0 0 1 0 0 Querula purpurata 0 1 0 1 0 0 Rupicola peruviana 0 1 0 0 0 0 Corvus imparatus 0 1 0 1 1 1 Cyanocorax cyanomelas 0-1 0-1 0 1 1 1 Progne chalybea 1 1 1 1 2 1 Turdus grayi 0 1 1 1 1 1 Turdus serranus 0-1 1 1 1 2 1 Mimus gundlachii 1 1 1 1 1 1 Campylorhynchus zonatus 1 1 1 1 1 1 Saltator maximus 1 1 1 1 1 1 Tangara vitriolina 1 1 1 1 1 1 Buthraupis montana 1 1 1 1 1 1 Thraupis episcopus 1 0-1 1 1 1 1 Ramphocelus icteronotus 1 1 1 1 1 1 Sicalis flaveola 1 1 1 1 1-2 1 Arremon aurantiirostris 1 1 1 1 1 1

PAGE 208

192 Table 30. Continued. Characters (PaCMC) Species 1 2 3 4 5 6 Arremonops conirostris 1 1 1 1 1 1 Cacicus haemorrhous 1 1 1 1 2 1 Psarocolius angustifrons 1 1 1 1 2 1 Molothrus bonariensis 1 1 1 1 1-2 1 Dives dives 1 1 1 1 1-2 1 Icterus mesomelas 1 1 1 1 1 1 Sturnella militaris 1 1 1 1 2 1 Family Thamnophilidae, cf. Taraba major – El Azúcar (1): The proximal end of this femur is highly curved and coinci des in size and shape with that of Taraba major . I have not found another passerine of this size with an equall y decurved shaft. Family Corvidae, Cyanolyca sp. – La Chimba (1): The concave attachment area for musculus femorotibialis internus suggests this bone belongs to a corvid. Intermediate in size between Cyanocorax yncas and C. violaceus / C. mystacalis , it probably belonged to a Cyanolyca . Family Corvidae, Cyanocorax cf. mystacalis – El Azúcar (1): Qualitatively similar to the previous specimen, this distal end of a femur is much larger. Cyanocorax mystacalis is the only large corv id in western Ecuador. Family Turdidae, Turdus cf. fuscater – La Chimba (1): The proximal end with a strong shaft is referred to Turdus by the proximal formamen on the dorsal side just distad of the head.

PAGE 209

193 Family Turdidae, Turdus cf. maculirostris – El Azúcar (1): This proximal end is referred to Turdus by the proximal formamen on the dorsal side just beneath the head. Family Icteridae, cf. Zarhynchus wagleri – La Chimba (1): The distal end of a very large passerine humerus, this specimen can only be Psarocolius sensu lato or Rupicola peruviana by size. Based on the position a nd shape of tuberculum musculi gastrocnemialis lateralis (more laterad and oval in Rupicola , more caudad and round in Psarocolius sensu lato), the width of tr ochlea fibularis (narrow in Rupicola , wide in Psarocolius sensu lato), and the width of the distal end relative to the length of the condyla (wide and short, respectively, in Rupicola , and narrow and long in Psarocolius sensu lato), R. peruviana is highly unlikely. Within Psarocolius sensu lato the femur is very similar; although in Zarhynchus it differs from that of Psarocolius in the shape of fossa poplitea and the location of two popliteal foramina. There is only one foramen in the round, deep popliteal fossa of Psarocolius , but two in the wide, shallow fossa of Z. wagleri . The condition of the popliteal fossa is similar in P. viridis , but the popliteal foramina are not as separated and the proxima l margin of the condylus medialis not as wide as in Zarhynchus . Other species of large Icte ridae were not available for comparison. Thus, this identification is tentative. Family Icteridae, Psarocolius cf. angustifrons – La Chimba (1): This large distal end is Psarocolius by size as well as the size and sh ape of fossa poplitea (see above). As with most large specimens from La Chimba, I assign this femur tentatively to P. angustifrons .

PAGE 210

194 Family Icteridae, Sturnella bellicosa – El Azúcar (1): This highly curved, complete femur has a proximal foramen just beneath the head and a highly curved shaft. In terms of general proportions and shape, th is specimen agrees very well with S. bellicosa . TIBIOTARSUS General Osteology The proximal end of the tibiotarsus of th e Passeriformes has a very thin crista cnemialis and a deep fossa retropatellaris. The distal end has a latera l groove for perineus profundus, a deep tendinal gr oove, and more or less pronoun ced tuberositas retinaculi extensoris. The lateral tuberositas retinaculi ex tensoris is located, at least partially, on pons supratendineus. The medial tuberosity is located proximad of pons supratendineus. Trochlea cartilaginis tibialis is bisected by a raised ridge. Comparative Osteology The passeriform tibiotarsus is a fairly uni form bone with few diagnostic characters to distinguish among the many passerine families. The lack of distinguishing features is reflected in the low number of specimens iden tified from this element. Nevertheless, the following observations still prove useful. 1) Impressio ligamenti collateralis medialis is raised and gives the proximal end a somewhat flattened appearance. This can be seen in the ovenbirds (Family Furn ariidae; especially Syndactyla subalaris ) and antbirds (Family Formicariidae). 2) The emberizid specimens all show a slightly bent distal end with the condyles being at an angle to the main axis of the shaft. 3) In oscines, the proximal end of crista cnemialis cranialis bends laterally, ar ching partially over sulcus intercnemialis. Whereas the crista cnemialis cranialis is slight ly bent in the Tyranni, it is straight in the Furnarii. 4) In the swallo ws (Family Hirundinidae), the distance between the medial rim

PAGE 211

195 of condyla medialis and the lateral rim of condyl us lateralis is about 2-3 times as large as the thickness of the shaft in cranial aspect. Incisura intercondylaris is broad and condyla medialis et lateralis are short (along axis proxim odistale) in this family as well. The distal end thus looks almost club-like. 5) Woodcreep ers (Family Dendrocolaptidae) have a very broad, flattened sulcus intercnemialis. Family Formicaridae, Grallaria sp. – El Azúcar (2): Because of the narrow shaft, the two distal ends superficially resemble the tibiotarsi of Grallaria antpittas, Troglodytidae (wrens), Mimi dae (mockingbirds), and Turdus (thrushes). Wrens can be excluded based on size. Mockingbirds have a narrower pons supratendi neus and a thicker condylus medialis. The tibiotarsi of the rema ining two genera can be separated on the basis of two traits. 1) Latera l tuberositas retinaculi extensoris is located at the proximal margin of the pons supratendineus in Turdus and distal to the margin in Grallaria . 2) In Turdus condylus lateralis is about as long as cond ylus medialis in di stal aspect, whereas in Grallaria condylus medialis is noticeably longer. Other families, such as the Icteridae and Thraupidae, have thicker shaf ts that are flattened cranially. Family Corvidae, Cyanolyca sp. – La Chimba (3): Probably belonging to the same bone, the proximal and distal ends are from a medium-sized to large passerine (ca. 120 g). This size-class is limited to species in the genera Grallaria , Agriornis , Querula , Cyanolyca , Turdus , and Cacicus . In terms of the distal end, Grallaria differs in that condylus medialis is longer than condylus lateralis, Agriornis by posessing more separated condyla and a U-shaped shaft in cross-section ( Agriornis was not available for comparison, but I assume that it is similar to Muscisaxicola ), and Querula by being smaller and appearing symmetrical in cran ial view (as the margins of condylae are

PAGE 212

196 roughly equidistant from the shaft). Cacicus differs in having a wide ned distal shaft. In Turdus and Cyanolyca it is very similar. Both share a narrow shaft, equally long lateral and medial condylae in distal view, and a pr oximally highly asymmetric shaft in crosssection due to a narrowing of the lateral marg in. They differ, however, by the thickness of the shaft (narrower in Turdus ), degree of separation of condylae (larger in Cyanolyca ), and the shape of the condylus me dialis (flaring out proximally in Cyanolyca ). The archaeological specimen is larger than that of a 111 g specimen of C. viridicyana (UF 43499) from Peru, and may belong to either C. armillata or C. turcosa . More comparative material is needed to attempt a species-lev el identification. Family Turdidae, Turdus fuscater – La Chimba (1): The distal end and slender shaft of this medium-sized tibiotarsus come from a long-legged bird, such as Turdus and Grallaria . These two species can be separated on th e basis of the two criteria described above. The La Chimba specimen is considerably larger than a modern specimen of a 73 g T. serranus (UF 43409) and thus probably belongs to T. fuscater . Family Troglodytidae, cf. Henicorhina leucophrys – La Chimba (1): The distal end with shaft comes from a small bird with th e body mass of 15-20 g. In this size range, I compared the archaeological specimen to that of Synallaxis , Henicorhina , Thryothorus , Zonotrichia , Diglossa , Hemispingus . Most of these genera can be discarted on the basis of the distance from tibiotarsal condylae to crista fibularis (refelcting overall length of the bone; shorter in Synallaxis , Thryothorus , and Hemispingus ), the thickness of the shaft (wider in Diglossa , Hemispingus ), and the distance between the cotylae (smaller in Synallaxis , larger in Hemispingus ). Zonotrichia is similar in proportions, but has the typical emberizid distal end (which is at a sm all angle to the shaft). The only species that

PAGE 213

197 agrees in all regards is Henicorhina . The La Chimba specimen is slightly smaller than a 15.9 g specimen of H. leucoptera and probably belonged to a small individual of H. leucophrys (see Dunning 1993). Family Icteridae, Psarocolius cf. angustifrons – La Chimba (2): This large tibiotarsus compares best to Psarocolius sensu lato in size and overall shape, but I cannot identify it to species. As most other large Psarocolius bones appear to be from P. angustifrons , I tentatively assign this bone to this species as well. TARSOMETATARSUS General Osteology The tarsometatarsus of the Passeriformes has a shallow sulcus extensorius. The proximal end is characterized by distinctive hypotarsi with a six-can al pattern (Manegold et al. 2004). Only a minority of unrelated sp ecies deviate from that pattern. The shaft has a broad, flat lateral surface and a narrow, often sharp medial surface. Trochlea metatarsi are roughly of the same length, and are lo cated on the same dorso-plantar plane. Throchlea metatarsi IV is usually much sm aller, being reduced to a small disk-like structure with indistinguishable condyla. Comparative Osteology The best characters for distinguishing be tween taxa are the arrangement of the hypotarsal canals and the degree of separa tion and size of trochleae metatarsorum. Suboscines tend to have long trochlea metata rsi II with a strong medial inflection. A distinctive group are the Dendrocolaptidae (Woo dcreepers) in which a scansorial habit has selected for a high degree of separati on of trochleae metatarsorum and a deeply excavated trochlea metatarsi III (Feduccia 1973).

PAGE 214

198 Family Turdidae, Turdus fuscater – La Chimba (3): The long and slender, but robust shaft is typical of Turdus . Other genera with larg e species have either proportionately shorter ( Rupicola , Cyanocorax , Psarocolius ) or longer ( Grallaria ) tarsometatarsi. Rupicola also differs by a highly a ngled trochlea metacarpi IV. Cyanocorax has a different hypotarsal formula and a more slender trochlea metacarpi II. Psarocolius exhibits a different hypotarsal formula, a longer fossa metatarsi I, and a caudally bend distal end (at the height of fossa metatarsi I). Grallaria has a broad trochlea metatarsi IV with two distinct coty la. In this genus trochlea metatarsi I is significantly shorter than trochlea metatarsi III (traits re garding os hypotarsi and trochlea metatarsi II could not be determined on the damaged modern museum specimen). Family Mimidae, Mimus logicaudata – El Azúcar (1): Identification of the distal end is based on the deeply excavated and th e distally located fo ssa metatarsi I, two characters that I have not f ound in other passerine examined. Family Icteridae, Dives warszewizci – El Azúcar (1): The distal metatarsal end of a medium-sized oscine is an icterid by the ca udo-distal bend of the sh aft, and a notch lying cranially on the medial rim of trochlea metata rsi II. A robust trochlea metatarsi II and a strong medial inflection of trochlea metatarsi IV further distinguish this bone from that in the similarly sized Saltator (which has slender trochlea metacarpi II and straight trochlea metacarpi IV) and Turdus (characterized by straight trochlea metacarpi IV). Family Icteridae, Psarocolius cf. angustifrons – La Chimba (2): Size and a strong distal bend of the shaft are found only in large icterids such as Psarocolius . The hypotarsal formula (with only five canals, all of which are closed) also agrees with that of

PAGE 215

199 Psarocolius . Identification as P. angustifrons is tentative and based on the presence of other elements at La Chimba that appear to be from this species (see above). Discussion Throught careful osteological analysis I was able to id entify 95 taxa, most to the species level. The identification of zooarch aeological material was only made possible by the availability of a good reference collection. The taxonomic completeness of such a collection will limit to some extent the amount of possible species-leve l identifications. Taxonomic incompleteness can be offset to some degree by the availability of closel y related taxa, assuming that these are very similar osteologically. Similarity in turn can be assessed to by cons idering size, weight, and habits of a species and its cl ose relative. For instance, I used Columbina talpacoti in addition to a single specimen of C. buckleyi to look at possible intr aspecific variation in C. buckleyi . I assumed that C. buckleyi and C. talpacoti are very similar as they are once considered to be sister species and they are often treated as consp ecific. The two species differ markedly, however, such that the comparison of C. cruziana and C. talpacoti in Campbell (1979) lacks much validity in assessing differences between C. cruziana and C. buckleyi . Columbina talpacoti has a much stouter build than C. buckleyi and a much shorter tarsometatarsus. On the othe r hand, morphological descriptions of Asio stygius , in addition to examining modern skeletons of A. flammeus , helped in the tentative identification of several bones from La Ch imba. Reference to proportionately longer wings and larger size in A. stygius (König et al. 1999) and inc ongruence of archaeological specimens with modern skeletons of A. flammeus , made A. stygius a more likely identification. Species-level iden tifications will remain in so me flux as long as even the

PAGE 216

200 largest museum collections are so incomplete. This problem will surely afflict zooarchaeological studies of bi rds for some time to come. For groups that are taxingly difficult to identify, such as the Passeriformes, new molecular tools may prove to be helpful. Not only will these techniques increase the certainty of identifications, but they will al so allow phylogeographic studies of recent and prehistoric birds. For instance, in this study the identity of Psarocolius would allow for testing alternative hypotheses regarding the biogeography of this genus and patterns of resource utilization in Andean Ecuador (Cha pter 5). Molecular tec hniques are costly and at time problematic, however, and in the near future will not be able to replace careful studies of comparative osteology. Both approa ches will require the continued work of museums in collecting modern specimens.

PAGE 217

201 CHAPTER 3 EARLY HOLOCENE MANGROVE AND WE TLAND BIRDS ON THE ARID SANTA ELENA PENINSULA, ECUADOR Introduction Climatic conditions in western Ecuador ar e strongly influenced by the cold Peru Current that moves north along the South Amer ican coast. Upon reaching the Santa Elena Peninsula of southern Ecuador, the trade winds push the current to the west. The dry air above the Peru Current produces arid condi tions in southern Ecuador. In northern Ecuador, the moisture-laden air associated w ith the Northern Equatorial Countercurrent condenses upon rising in the Andes to support lu sh tropical rain fore sts. Western Ecuador is therefore characterized by a latitudinal humidity gradient. The Santa Elena Peninsula is located at the center of the gr adient, projecting into the Pa cific Ocean like a giant probe recording climatic conditions. The record is ke pt in tar seeps and ar chaeological sites that contain plant and animal microand macro-fo ssils from different time periods. The tar seeps at La Libertad and Talara (Peru) ha ve yielded fossils of Late Pleistocene age (Lemon and Churcher 1961, Edmund 1965, Campbell 1976, 1979). Archaeological sites provide the only evidence so far for Holocen e paleoenvironmental conditions. Most of the sites date to the Mid to Late Holocene. In western Ecuador, a series of sites on the Santa Elena Peninsula belonging to the pre-ceramic Vegas cu lture are Early Holocene in age (Stothert 1985). Of these, sites OGSE-80 (henceforth Site 80) and M5 A4-67 (Site 67) are the most important (Piper no and Stothert 2003) as they appear to have been ritual centers occupied for two to four mille nnia (Raymond 2003, Stothert et al. 2003).

PAGE 218

202 The archaeological excavations at Site 80 have revealed th e details on burial practices and settlement pa ttern of the Vegas people (Stothert 1983, 1985). Site 80 probably was a central locati on for a nomadic or semi-nom adic people who also used many other satellite sites throughout th e peninsula (Stothert 1985, Raymond 2003, Stothert et al. 2003). A disc ontinuous shell layer marks a stratigraphic break in the midden of Site 80, allowing the Vegas occupati on to be divided into two phases. The early Las Vegas phase spanned ca. 2000 years between 10,000 and 8000 14C yr BP. The late phase lasted from 8000 to 6600 14C yr BP (Stothert 1985). Vegas people subsisted on a variety of re sources from mangroves and terrestrial habitats supplemented by early domesticated fo rms of squash, and at a later stage, corn (Piperno and Stothert 2003, Stothert et al. 2003). Resource use changed over time. During the early phase, fish was less importa nt than during the la te phase. Terrestrial sources of protein, in turn, decrea sed in significance. In addition, Anadara tuberculosa , a clam common in mangroves, were extensively used during the early phase, but later was replaced by other mollusks. The shift ma y reflect modifications of the social organization, increased importance of domes ticated plants, as well as environmental changes (Stothert et al. 2003). The faunal remains from Site 80 suggests overall arid environmental conditions similar to present ones (Stothert et al . 2003). The absence of palm and bamboo phytoliths, and certain mammals such as monke ys, have been interp reted as evidence for a dry climate (Piperno and Pearsall 1998). A lthough the absence of these taxa in an archaeological context is suggestive, archaeo logical assemblages of plants and animals are the result of taphonomic processes that are strongly biased by a “cultural selection

PAGE 219

203 filter.” Human food preferences are not only di ctated by mere protein requirements, but also by taboos and other social rules govern ing the consumption of plants and animals (Stahl 2003). A lack of palm and bamboo phytoliths may also be indicative of human preference (bamboo may not have been used, yet) or local habitat alte ration rather than a complete absence from the Santa Elena Penins ula. It is noteworthy that in the more humid Jama Valley about 240 km to the north, bamboo is absent from the ca. 4000 yr old lower, pre-occupational sediment strata, a nd highly abundant only during the Valdivia 8 (Early Piquigua Ceramics) phase of occupa tion when people had cleared much of the forest (Stahl 2000, Veintimilla 2000, Pearsall 2004). Paleoecological interpretations are further hampered by the lack of clear stratigr aphies that do not allow detecting significant short-term (years, decades) or long-term (cen turies or millennia) change (Stahl 1991). A low resolution in the identification of floral and faunal material makes paleoenvironmental reconstructi ons very difficult. There is a great need for species-level identifications to extrapolate present patterns in habitat selection to past distributions, assuming that habitat preferences are indica tive of vegetation stru cture and composition as well as climatic conditions. The object of this study is to identify and interpret the bird re mains from Site 80. Species-level identification of bird bones will improve our understanding of paleoecological conditions in Early Holocen e of southwestern Ecuador. Only one bone from Site 80 had been identified previously as an unidentified psittacid (parrot family; Stothert et al. 2003). Living birds are a relatively well st udied taxonomic group for which much of the current distribu tions and broad patterns in habitat selection are well known. This knowledge strengthens infe rences on paleoenvironmental conditions. Little is known

PAGE 220

204 about trade in pre-ceramic cultures of we stern Ecuador which may have been very limited (Stothert 1988). I therefore assume that most birds found in the midden of Site 80 come from within a local hunting radius of 20 km (S tothert et al. 2003, Fig. 3). Methods Site 80 is located on the Santa Elena Peni nsula near the seasonal Río Las Vegas on a small hill about 33 meters above sea level (Figures 3-1 and 3-2). The bay of Santa Elena is located about 3.5 km to the north, and the beaches to the south are about 9 km away. Presently, the site cove rs an area of about 0.225 ha, but its original size may have been 1.3 ha (Stothert 1988). In the early 1970s , when the excavations began, the area was surrounded by desert scrub (Stothert 1988, Fi g. 2.5). The area has become urbanized since, and the site is preserved as part of an archaeological museum. Rainfall and temperature are highly seasonal. A dry a nd cool but cloudy season lasts up to nine months. For the remaining three to four months (usually December/January through March) a generally clear sky and abundant s unshine generate warm days punctuated by strong rains. The El Niño Southe rn Oscillation (ENSO) is part icularly important in this area as it creates years of abundant rain alte rnating with relativel y dry years (Stothert 1988). All the faunal material was recovered fr om excavations done by Karen Stothert in the early and late seventies (S tothert 1985, 1988). Approximately 300 m2 of the deepest deposits were excavated as a series of tr enches to depths from 150 to 450 cm below datum (50 to 320 cm below ground surface; Stothert 1988, Figs. 2.7-2.20). Despite lack of visible stratigraphy in the cultural deposits, a discontinuous shell layer at ca. 100 cm below datum marks a transition between two o ccupational phases that are characterized by different patterns in res ource utilization and have been dated to 10,000 to 8000 14C yr

PAGE 221

205 BP (henceforth referred to as the early period) and 8,000 to 6,600 14C yr BP (henceforth the late period), respectively (Stothert 1985). Figure 3-1. A map of the Santa Elena Peninsula, Ecuador, show ing the locations of Site 80 ( ), several towns ( ) as well as permanent and temporal rivers. All sediments were screened using a 5 mm mesh. A subsample was wet-screened using a 2 mm mesh and some samples were s ubjected to flotation (Stothert 1988). Except for bones on display at the Museo Amantes de Sumpa, Santa Elena, Ecuador, all the faunal material is curated at the Florida Museum of Natural History I determined minimum numbers of indivi duals (MNI) and number of individual specimens (NISP) for the combined data se t and three depth categories, respectively. These two measures provide a range for th e number of individuals found, with MNI being the most conservative estimate and NISP representing the maximum number of

PAGE 222

206 Figure 3-2. A map showing different localitie s mentioned in the text. The gray line represents the 200 m contour level of the Cordillera Chongón-Colonche. individuals found. To determine an asso ciation between excavation depth (roughly corresponding to the early and late Vegas peri ods) and bird habitat, I conducted a Wilk’s G-test on XLSTAT software for both MNI a nd NISP. The three depth categories were 20-90 cm, 90-100 cm, and 100-160 cm below datum. In contrast to Stot hert et al. (2003), I treated the layer between 90 and 100 cm as a di stinct category because it is associated with the Anadara tuberculosa stratum, contains a large pr oportion of the total bone, and appears to represent a di stinct depositional event.

PAGE 223

207 Results Much of the bird bone from Site 80 is hi ghly abraded and/or fragmentary, making identification difficult. In total, 151 specime ns belonging to at least 20 species in 10 families were identified (Table 3-1). The most common group is the Columbidae (pigeons and doves) which make up more th an 50% of all the specimens. The Eared Dove ( Zenaida auriculata ) is the most common species of bird in the faunal sample of Site 80. Its congener, Z. meloda , is much rarer with only five elements (minimum of three individuals) assigned to this species. The Rallidae (rails, gallinules, coots) is th e second most abundant family. With four species it is also the most dive rse. The migratory Sora Rail ( Porzana carolina ) is represented by 18 specimens. The Spotted Rail ( Pardirallus cf. maculata ) and a Neocrex rail are each represented by a single element. Seven pedal phalanges were assigned to the genus Rallus , which is clearly distinguishable from Aramides , Gallinula , Porphyrio , and Fulica , the four genera in the same size cl ass (Chapter 2). Based on the present distribution and on the comparativ e material examined, I classify the elements tentatively as R. longirostris . Other wetland species correspond to one species of tiger-heron and at least three species of duck. Two species of hawk, one falcon, and one owl are the only predators. A single species of icterid (blackbirds, grackles, orio les, meadowlarks) is the only passerine. Two elements belong to the parrot family, Psittacidae. Of these Amazona farinosa is notable as it does not presently occur anywhere near th e study area. Equally unusual for the Santa Elena Peninsula is a large tinamou, Tinamus cf. major .

PAGE 224

208 Table 3-1. Minimum number of individuals (M NI) and number of individual specimens (NISP) for all species at the OGSE80 archaeological site, Santa Elena, Ecuador. Current distributi on is shown as absence () or presence (+) within 20 km and 100 km radii from site 80. Modern Records Family/Species MNI NISP Within 20 km Within 100 km TINAMIDAE Tinamus cf . major 1 1 ARDEIDAE Tigrisoma fasciatum 1 1 x Tigrisoma lineatum 1 2 ANATIDAE Dendrocygna sp.a 1 1 x x Anas cf. discors 2 2 x x Anas bahamensis 2 3 x x Anas georgica 5 6 ACCIPITRIDAE Buteo cf. nitidus 1 1 x x Buteogallus anthracinus 2 2 x FALCONIDAE Herpetotheres cachinnans 1 1 x x RALLIDAE Rallus cf. longirostrisb 1 7 x cf. Pardirallus sp.c 1 1 x Porzana carolina 9 18 x x Neocrex sp.d 1 1 x

PAGE 225

209 Table 3-1. Continued. Modern Records Family/Species MNI NISP Within 20 km Within 100 km RALLIDAE Rallidae sp. 1 1 ----COLUMBIDAE Columbina sp.e 2 4 x x Zenaida auriculata 17 66 x x Zenaida meloda 3 5 x x Zenaida sp. 2 2 ----Columbidae sp. 9 10 ----PSITTACIDAE Aratinga erythrogenys 1 1 Amazona farinosa 1 1 STRIGIDAE Asio cf. clamator 1 1 x ICTERIDAE Sturnella bellicosa 2 3 x x Oscine sp. ( Sturnella size) 2 3 ----Passeriformes sp. 1 4 ----a D. bicolor and D. autumnalis , both within 20 km. b No recent records within range c P. maculatus within 100 km, P. sanguinolentus more than 100 km away d N. erythrops within 100 km, N. columbianus more than 100 km away e C. buckleyi and C. cruziana Early and late Vegas occupations show ma rkedly different bird faunas. During the early occupation, mangrove/freshwater wetland species make up about 50% of the bird remains, but during the late period only 10-20% (Table 3-2). The differences are nearly

PAGE 226

210 significant statistically for MNI (Wilk’s G2 = 5.552, df = 2, p = 0.062) and significant for NISP (Wilk’s G2 = 16.450, df = 2, p = 0.0003). As in other studies, MNI and NISP are closely correlated (Grayson 1984, Chapters 4 and 5). Systematic List This section summarizes the material rec overed from each species and places each species in a biogeographic context. The se quence of species nomenclature, and the current distributional data, unless otherwis e noted, are from Ridgely and Greenfield (2001a,b). The comparative osteology and justif ications for the identifications are found in Chapter 2. Order TINAMIFORMES Family TINAMIDAE Tinamus cf. major Material.Distal end of left tarsometatarsus. Remarks.The only tinamous to be large e nough to be considered given the size of the tarsometatarsus belong to the genus Tinamus . In western Ecuador the only species belonging to this genus is the rare to locally fairly common T. major . Although presently restricted to the humid lowland forests of northwestern Ecuador, there are several old records from Manabí and Guayas provinces (Figure 3-3). This species has also been observed at the Reserva Ecológica Manglares Churute, suggesting that its distribution may have been more extensive in the past . In a straight line, the distance to the mangroves of Churute is ca. 130 km (Figure 3-2).

PAGE 227

211 Figure 3-3. A map of Ecuador showing the 21 provinces. 1. Esmeraldas, 2. Carchi, 3. Sucumbios, 4. Imbabura, 5. Manabí, 6. Pichincha, 7. Napo, 8. Orellana, 9. Guayas, 10. Los Rios, 11. Cotopaxi, 12. Tungurahua, 13. Pastaza, 14. Bolívar, 15. Chimborazo, 16. Morona-Santiago, 17. Cañar, 18. Azuay, 19. El Oro, 20. Loja, and 21. Zamora-Chinchipe. Order CICONIIFORMES Family ARDEIDAE Tigrisoma fasciata , Fasciated Tiger-Heron Material.Shaft of right tibiotarsus. Remarks.This rare heron usually occurs along rapidly flowing and rocky rivers in the Andean foothills. This species has also b een observed at Loma Alta in the Cordillera Chongón-Colonche (ca. 47 km from site 80), just east of Valdivia as well as in the humid lowlands of Esmeraldas provi nce (Figures 3-2 and 3-3).

PAGE 228

212 Tigrisoma lineatum (Boddaert 1783), Rufescent Tiger-Heron Material.Proximal ends of right and left scapula. Remarks.The rare Rufescent Tiger-Heron can be found on slow-flowing streams, freshwater swamps and marshes, as well as mangroves. In western Ecuador this heron is known only from several isolated observations, mostly within the Río Daule basin more than 100 km away from Site 80 (Figure 32). One record come s from the Reserva Ecológica Manglares Churute so uth of Guayaquil although no details of that observation are provided. Order ANSERIFORMES Family ANATIDAE cf. Dendrocygna sp., Whistling Ducks Material.Proximal end of left coracoid. Remarks.Dendrocygna bicolor and D. autumnalis occur in western Ecuador where they are locally and seasonally comm on. Both species make local movements, with D. bicolor following changing water levels a nd sometimes congregating in large numbers (up to 30,000 have been observed at the Reserva Ecológica Manglares Churute; Figure 3-2). Although either species inha bits freshwater ponds and marshes, D. bicolor prefers open areas with tall grass whereas D. autumnalis , known to roost in mangroves, prefers to perch on tree limbs. Both species o ccur on the Santa Elena Peninsula, but given present habitat preferences, D. autumnalis would have been more likely in the mangroves surrounding the peninsula dur ing the early Holocene. Anas cf. discors Linnaeus 1766, Blue-winged Teal Material.Proximal end and shaft of left coracoids.

PAGE 229

213 Remarks.This locally fairly common nearctic winter migrant is rarely seen on salt water, but mostly on lakes, ponds, and freshw ater marshes. Up to 3200 individuals have been observed on an artificial re servoir, Santa Elena Peninsula, where this species is most common between October a nd April. Osteologically, A. discors is indistinguishable from A. cyanoptera , a resident and migrant now virtuall y extirpated from Ecuador. Recent records of A. cyanoptera from southwestern Ecuador may have been vagrant individuals from Peru. It is not clear why this species is absent from the Ecuadorian coast, since the migratory subspecies septentrionalium is still abundant on its breeding grounds in North America (Ridgely and Greenfield 2001a). Anas cf. bahamensis Linnaeus 1758, White-cheeked Pintail Material.Proximal end of left scapula and proximal end of left tibiotarsus. Remarks.This duck is common in wester n Ecuador, with its highest population on the Santa Elena Peninsula where up to 2000-4000 individuals have been observed. Presently populations are increasing due to artificial lagoons. Alt hough it is likely that most of the intermediate-size Anas bones come from this species, much of the bone material is too fragmentary fo r species-level iden tification. The possibility of the past presence of other migratory species, such as A. clypeata , cannot be excluded (Chapter 2). One coracoid has cut marks, evidence that this species was used for food. Anas georgica Gmelin 1789, Yellow-billed Pintail Material.Distal fragments of two right a nd two left coracoids, distal end of right humerus, and distal end of right ulna. Remarks.In the northern part of this pint ailÂ’s range, it is presently restricted to high Andean lakes. There are no records from the east Pacific coast north of southern

PAGE 230

214 Chile (Madge and Burn 1988). Apart from cool air temperatures, this duck may require cold bodies of water to feed in. All the cor acoids are more massive compared to modern skeletons, and there is a possibility that the archaeological specimens belong to the extinct duck Anas amotape Campbell 1979 (Chapter 2). Order ACCIPITRIFORMES Family ACCIPITRIDAE Buteogallus anthracinus (Deppe 1830), Common Black-Hawk Material.Distal segment of shaft from ri ght ulna and phalanx 2 of a left hallux. Remarks.A rare to uncommon hawk asso ciated with mangroves, populations of Buteogallus anthracinus appear to be declining from disappearance of mangroves; it may have been more common in the past. Buteo cf. nitidus (Latham 1790), Gray Hawk Material.Phalanx 4 of right fourth digit. Remarks.This phalanx belongs to a medium -sized hawk comparable in size to Buteo nitidus (Gray Hawk) or B. platypterus (Broad-winged Hawk). On the Santa Elena Peninsula, B. nitidus is more likely due to its pref erence for deciduous woodland. On the other hand, B. platypterus is a nearctic migrant that may easily stray from its mostly Andean wintering range that includes much of the lower foothills. Overall proportions make B. nitidus the more likely identification (Chapter 2). Family Falconidae Herpetotheres cachinnans (Linnaeus 1758), Laughing Falcon Material.Proximal end of left carpometacarpus.

PAGE 231

215 Remarks.This falcon is uncommon to fairly common in a variety of wooded and forested habitats, being more common in w oodlands and rarely away from continuous forest. Most often this species is seen in isol ated trees or open vege tation where it is more easily detected. Typically inha biting humid rather than in dry forest (Brown and Amadon 1968), the Laughing Falcon still occurs on Santa Elena Peninsula (Marchant 1958). Order GRUIFORMES Family RALLIDAE Rallus cf. longirostris Boddaert 1783, Clapper Rail Material.Phalanges 1 (damaged proxima l end) and 2 (proximal end) of second digit, phalanges 1 and 2 of third digit, and phalanges 1, 3 (proximal end), and 4 (proximal end and shaft) of fourth digit of a left foot. Remarks.The phalanges can be identified as Rallus . The only member of this genus in coastal Ecuador is R. longirostris . In addition, all other members of this genus (with the exception of R. elegans , which does not occur sout h of central Mexico) are considerably smaller. A preference for sa ltwater marshes and mangroves make it likely that the specimens belong to R. longirostris. There are no recent reco rds, but this rail may still occur in a few mangrove s in western Ecuador. Porzana carolina (Linnaeus 1758), Sora Rail Material.Shaft of right co racoid, nearly complete and complete left coracoid, distal ends of right humerus and two left humeri, proximal end of right ulna and nearly complete left ulna, distal end of left carpomet acarpus, distal ends of five tibiotarsi, shaft of right tibiotarsus, and proximal ends of two right tarsometatarsi.

PAGE 232

216 Remarks.A rare to locally uncommon near ctic migrant that prefers freshwater marshes, damp grassy areas, and rice fields. This species has not been recorded in the drier parts of western Ecuador, in cluding Santa Elena Peninsula. Neocrex cf. erythrops (Sclater 1867), Paint-billed Crake Material.Distal end of right ulna. Remarks.Two species of Neocrex currently are found in western Ecuador. Neither has been recorded closer than 170 km from Site 80. Both species are rare and local. Neocrex erythrops occurs in moist tall grass, along ponds, and presently rice fields in the dry lowlands of southwestern Ecuador. Neocrex columbianus prefers the more humid environments of northwestern Ecuador, main ly Esmeraldas province south to southern Pichincha province, ca. 260 km north of Site 80 (Figures 3-1, 3-2, and 3-3). This species also occurs in the foothills of the Ande s. Based on current habitat preferences, N. erythrops is more likely to have inhabited the ar id Santa Elena Penins ula in the past. cf. Pardirallus sp. (Boddaert 1783), Spotted Rail Material.Proximal end of right humerus. Remarks.These rare rails can be found in dry southwestern Ecuador as far north as southern Los Ríos province (Figure 3-3). The archaeological specimen of site 80 is larger than the humeri of modern skeletons of Pardirallus maculatus , however, and published weights on this species (Dunn 1993) suggest th at the species shoul d be smaller than Gallinula chloropus . The specimens are nearly identical in size. Identification of this specimen is therefore tentative.

PAGE 233

217 Order COLUMBIFORMES Family COLUMBIDAE Columbina cruziana , Ground-Doves Material.Cranial fragment of sternum and distal ends of right and two left coracoids. Remarks.Three species of ground-doves occur in western Ecuador of which C. buckleyi (Ecuadorian Ground-Dove) and C. cruziana (Croaking Ground-Dove) are both possible based on size. Generally, C. cruziana has a preference for dry scrub and C. buckleyi tends to prefer more humid secondary vegetation. Zenaida auriculata (Des Murs 1847), Eared Dove Material.Cranial fragments of five st erna, proximal ends of three scapulae, complete left coracoid, proximal ends of four right and one left coracoids, distal ends of four right and five left coracoids, proximal e nds of one right and four left humeri, distal ends of two right and two left humeri, proxima l ends of three right and five left ulnae, distal ends of three right and th ree left ulnae, shafts of one ri ght and two left ulnae, distal end of right radius, proximal and distal ends of carpometacarpi, proximal ends of three right and one left femora, dist al ends of two right and one left femora, and proximal and distal ends of tibiotarsi. Remarks.Common in agricultural areas and urban settlements, especially in the Andes, Zenaida auriculata also is common on the Santa Elena Peninsula, where its numbers fluctuate widely within and across years (Marchant 1958).

PAGE 234

218 Zenaida meloda (Tschudi 1843), West Peruvian Dove Material.Proximal end of right coracoid, and proximal ends of two right and distal ends of two left humeri. Remarks.This uncommon and local dove inha bits riparian areas, desert scrub, and mangrove edges in the arid lowlands of Ecua dor and Peru. Marchant (1958) noted that it is most common in wooded valleys with some tree cover. The West Peruvian Dove does not flock extensively in the post-breeding seas on, and thus may be more difficult to hunt than the Eared Dove. Order PSITTACIFORMES Family PSITTACIDAE Aratinga erythrogenys (Lesson 1844), Red-masked Parakeet Material.Proximal end of right ulna. Remarks.This parakeet is rare to lo cally fairly common throughout the dry and semi humid lowlands and foothills preferri ng the canopy and borders of deciduous forest and woodland. It also frequents agricultural lands and occasionally is seen in desert scrub. On Santa Elena Peninsula its presence can be noted after the breeding season when small flocks are common in large wooded valleys (Marchant 1958). It may have been much more common in the past when non-br eeding populations may have numbered in the thousands, before extensive trapping fo r the pet trade. In August 2004, I observed a flock of about 25 individuals in Salinas, ca. 15 km west of Site 80 (Figure 3-1). Amazona farinosa (Boddaert 1783), Mealy Parrot Material.Fragmented shaft of right coracoid including a portion of facies articularis sternalis and cotyla scapularis.

PAGE 235

219 Remarks.The coracoid comes from a large Amazona that in size agrees only with that of A. farinosa . The only two species of Amazona in western Ecuador are A. farinosa and the smaller A. autumnalis . Currently, neither occurs in western Guayas province (Figure 3-3). Amazona autumnalis has a slightly wider dist ribution, especially with regard to the western margin of its distributi on which extends into some of the drier parts of western Ecuador. On the other hand, A. farinosa tends to be more common than A. autumnalis . Amazona farinosa has been recorded in the Reserva Ecológica Manglares Churute within ca 140 km of Site 80 whereas A. autumnalis may be found as close as ca. 30 km (Ridgely et al. 2003; Figure 3-2). Order STRIGIFORMES Family STRIGIDAE Asio cf. clamator (Vieillot 1808), Striped Owl Material.Phalanx 2 of s econd digit of right foot. Remarks.The phalanx clearly belongs to an Asio sensu lato, but is stronger and larger than that in A. flammeus (Short-eared Owl). Despite being about the same overall size as A. flammeus , the Striped Owl has powerful tal ons and long claws enabling it to take larger prey than Asio flammeus , which would explain the difference in foot size (König et al. 1999). This rare and local owl prefers semi open habitats ranging from pastures, grasslands, and marshes with scattere d trees or brushy vegetation to fairly open woodlands, occasionally inhabiting areas around houses. It is currently not found in the driest parts of Santa Elena Peninsula. The cl osest recent record comes from the foothills of the Cordillera Chongón-Colonche about 80 km to the north, a lthough favorable habitat may be found within 40 km of Santa Elena (Figure 3-2).

PAGE 236

220 Sturnella bellicosa Filippi 1847, Peruvian Meadowlark Material.Complete left humerus and pr oximal ends of right and left humerus. Remarks.This is a common and conspicuous bird in pastures, agricultural areas, and arid scrublands. In August 2004, I observed this species near humid pastures that were artificially watered, but it was absent in the large area of dry scrub surveyed by car (ca. 6 km2 of potential habitat). It is fairly common and tolerant of human habitat alteration, although it may be restricted to areas with some nearby water. Discussion The avian assemblage from Site 80 include s species that are still common locally (within a radius of 20 km), but also species that occur outside a radius of 100 km. The overall composition indicates that Vegas pe ople hunted birds in mangroves as well as freshwater marshes, ponds, lakes and terrestria l habitats. In contrast to Early Holocene sites from the more arid Peruvian coast, marine birds are absent. Mangrove-associated species are presently absent from the peninsula due to the lack of suitable habitat. Throughout the Holocene, mangroves have declined from a combination of tectonic uplift and anth ropogenic habitat destruction (Ferdon 1981, Stothert et al. 2003). Tectonic uplift has reduced much of th e shallow, near-shore waters, and much of the vegetation on the Santa Elen a Peninsula has been cut for fire wood as the human population has increased over the la st century (Marchant 1958, Stothert et al. 2003). Freshwater ponds and marshes ar e required to varying degrees by Tigrisoma lineatum , Dendrocygna bicolor , Anas discors / cyanoptera , Porzana carolina , Neocrex sp., and Pardirallus sp. Apart from mangroves, some freshwater wetlands may have been present during the Early Holocene. These wetla nds would have required either a higher amount of rainfall or a higher water table. Fe rdon (1981) speculated that a shallow water

PAGE 237

221 table could have formed near mangroves as a freshwater lens sitting on saltwater that percolates inland at sea level. On the flat te rrain of the peninsula, shallow depressions in the ground could have been filled by this wate r, providing a freshwater environment at least for several months after the rainy season. Table 3-2. Minimum number of individuals (M NI) and number of individual specimens (NISP) for two broad habitat classes fo r three depth categories, at Site 80, Santa Elena Peninsula, Ecuador. Wetlands refers to any freshwater habitat, including marshes, swamps, and ponds. Depth below datum (cm) Habitat 50 – 90 90 – 100 100 – 160 MNI Mangroves/Wetlands 5 (20.8%) 10 (50.0%) 13 (48.2%) Terrestrial 19 (79.2%) 10 (50.0%) 14 (51.8%) NISP Mangroves/Wetlands 5 (11.1%) 18 (36.0%) 21 (48.8%) Terrestrial 40 (88.9%) 32 (64%) 22 (51.2%) A discontinuous sediment layer containing the mangrove clam Anadara maculosa , which had been recognized previously as a marker for the transition between the two Vegas periods (Stothert 1985), was indistinguis hable from the early period. I hypothesize that this layer (90-100 cm) represents an ev ent of marine transgre ssion that may have caused a temporary localized in crease of mangroves (Stother t et al. 2003). During this time, Vegas people temporarily took advantage of the increased availability of clams and other mangrove resources, but then shifted to the exploitation of other marine and terrestrial habitats as transgression was followed by uplift that reduced the suitable area for mangrove formation. This study supports the notion of a changing Early Holocene

PAGE 238

222 environment due to tectonic uplift and sea-le vel rise. Although there is no clear evidence of climatic change, the presence of two tropical bird species, Tinamus cf. major (Great Tinamou) and Amazona cf. farinosa (Mealy Parrot), suggests that conditions could potentially have been more humid than in the present as neither species inhabits the arid environment of the Santa Elena Peninsula. These bones also could have arrived through trade, but at this time there are no data to support either hypothesis. Being located at the northern extreme of the area under the influence of the cold Peru Current, Site 80 has outstanding potenti al for comparison with other Early Holocene archaeological sites from Peru to the south (more arid) and more northward (more humid) in Ecuador. A caveat to this comparison among ar chaeological sites is that differences in taphonomy, depositional environment, bone pr eservation, and the methods used to recover bone material may bias the compositi on of the assemblage (Stahl 2003). In terms of methodology, the use of large mesh sizes fo r screening sediments also distorts the abundance and richness of sma ll-bodied taxa (Quitmyer 2004). As at the Late Pleistocene sites of La Carolina and La Talara, Peru (Campbell 1976, 1979), marine birds are absent from th e Site 80 avifauna. On the other hand, marine birds dominate the assemblages of the Peruvian sites at La Paloma (E. J. Reitz unpublished report, Matthiesen 1988) and Qu ebrada Tacahuay (deFrance et al. 2001, deFrance 2005; Figure 3-4). At Quebrada Tacah uay marine birds are the most important faunal component (deFrance 2005). Middle Ho locene sites along the Peruvian coast, including Huaca Prieta, Los Gavilanes, and the Moche Valley also show a predominance of marine birds (E. J. Re itz unpublished report, Pozorski 1983, Matthiesen 1988; Figure 3-4). Such an emphasis on marine faunas ma y have resulted from both high resource

PAGE 239

223 abundance and relative ease of procurement. Specific differences among sites may be due to cultural biases as well as resour ce specialization (deFrance et al. 2001). With the exception of snook ( Centropomus sp.), Vegas people almost exclusively fished species that can be found in estuarie s, near shore environments, or in mangroves (Stothert et al. 2003). On the Santa Elena Peninsula, terrestrial ecosystems and mangroves probably provided sufficient and easily obtainable sources of protein to keep Vegas people from engaging in much off-shore fishing. The dominance of Zenaida auriculata is a good example. These doves easily become habituated to humans and occur in large flocks, especi ally after the bree ding season. Low biomass (compared to deer or other large mammals) is offset by th e little amount of effort required to hunt a large number of these birds. Ease of hunting made the Passenger Pigeon highly attractive to hunters in North America, for exampl e (Schorger 1955), and several small-bodied species of small birds and mammals are more important in the diets of some indigenous people in Ecuador than large game (Hames and Vickers 1982). Hunting efficiency by Vegas people may further have been enha nced by a close association, possibly domestication, of a wild canid, Dusicyon sechurea (Wing 1988). Thus, there were no energetic incentives to adopt a maritime lifestyle. Domesticated plants may have played an im portant role in the diets of some Vegas inhabitants (Stothert 1988, Pi perno and Stothert 2003), al though analyses of human skeletal material suggest that agriculture wa s not intensive (Ubelake r 1988). In any case, Las Vegas people had a wide variety of food sources available to them, making fishing a relatively less essentia l way of acquiring food than furt her south in the more arid Peruvian coast.

PAGE 240

224 Figure 3-4. The locations of several Peruvian ar chaeological sites mentioned in the text in relation to Site 80. This study of bird bones from Site 80 underscores the need for species-level identifications in zooarchaeological studies. Three results would not have been possible without the taxonomic detail provided here. 1) Many mangrove-associated bird species had wider distributions in the past. As mangroves diminished during the Holocene (Ferdon 1981, Heusser and Shackleton 1994), these species have not been able to utilize the surrounding terrestrial habitat. If the remaining mangroves in southwestern Ecuador continue to be converted to shrimp farms, several species of birds may become extirpated from Ecuador. 2) The presence of the humid forest species, Tinamus cf. major and Amazona cf. farinosa , is the first potential evidence that the Early Holocene environment was possibly more humid than presently. Al ternatively, these species hint at the

PAGE 241

225 possibility of trade with dist ant human groups in the more humid lowlands to the north. 3) Beyond the exploitation of clams and fis h, Vegas people hunted birds and possibly mammals in mangroves. During the second occupational phase, the significance of mangroves and freshwater wetlands decreases. This change in birds is much more dramatic than in clams as reported by Stothert et al. (2003), and app ears to mark a change in the types of foods consumed during the tw o periods. Future research on similar Vegas sites that are farther from the coast shoul d provide additional in sights into how local conditions may have shaped resource use, in particular with regard to marine vs. terrestrial resources.

PAGE 242

226 CHAPTER 4 REMARKABLY HIGH NUMBERS OF SONGBIRDS AND GROUND DOVES AT THE EL AZÚCAR ARCHAEOLOGICAL SITE, COASTAL ECUADOR: AN UNUSUAL HUNTING ADAPTATION TO AN ARID ENVIRONMENT? Introduction Agriculture has played an important role in western Ecuador for more than 6000 years (Piperno and Pearsall 1998). Many prehistoric societies reli ed to a very large degree on corn, supplementing their diet with game and fish. Intensive exploitation of marine resources only began once people settled and began to grow crops. Farming also converted native dry forests to agricultural lands generating a mosa ic of fragments and anthropogenic habitats surrounding hu man settlements (Stahl 1991, 2000). By 200 AD the Guangala people inhabited mu ch of the Santa Elena Peninsula. On the coast nutrition was supplemented by marine resources whereas in the interior hunting played a major role (Reitz and Masucci 2004). In the arid environment of the peninsula, agriculture was made possible by building walk -in wells that extended water availability during the long dry season. The relative importance of marine resources vs. game is still a matter of much debate. Reitz and Masucci (2004 ) conclude that in th e El Azúcar valley, about 25 km from the coast, marine resour ces played an important role, although hunting provided most of the animal protein. Agriculture on the Santa Elena penins ula poses significant challenges. The environment is very seasonal and susceptible to frequent pr olonged droughts associated with inter-annual variation in the El Ni ño Southern Oscillation (ENSO). Long-term droughts have resulted in the abandonment of inland sites in favor of coastal areas

PAGE 243

227 (Paulsen 1976). Given the higher reliability of co astal areas, it appears difficult to explain the settlement of interior s ites and to place their significan ce within the regional social context. Demand for foods with higher fat c ontent and diet diversif ication are thought to have contributed to the specialization on hunti ng large mammals, such as deer, providing the necessary incentive to move inland (Reitz and Masucci 2004). On the other hand, before the intensive uti lization of marine re sources associated with the onset of agriculture, settlements on the peninsula may have been distributed more evenly on the Santa Elena Peninsula. So me pre-ceramic Vegas sites were located at a considerable distance from the ocean (P iperno and Stothert 2003, K. E. Stothert personal comment). From the st andpoint of cultural evolution coastal people may have diverged from a largely terrestrially based society. In this cas e terrestrial resource utilization can be considered the ancestral character state; inland communities simply continued to base their socio-economic relationships on agriculture and hunting. Whatever the motivation to occupy inland si tes, the continued existence of small agricultural settlements for hundr eds of years in light of resource fluctuation requires further explanation. Did inland peoples subsis t completely by agriculture and hunting or were other economic activities also important ? The abundance of shell beads in the El Azúcar sediments suggests that the production of crafts may have played an important role in the acquisition of food via trade (Masuc ci 1995). The dearth of deer elements from parts of the body that yield the highest amount s of meat also suggests that large game may not have been consumed locally, but may have been an important trade item as well (Reitz and Masucci 2004). Thus, economic dive rsification may have allowed inland sites to be occupied for long periods of time. A dditional evidence for trade and exchange of

PAGE 244

228 information comes from the suggestion th at nonlocal materials and motifs were incorporated into El Azúcar cera mics (Masucci and Macfarlane 1997). As agriculture played an important role , habitat conversion should have left a signature in the relative a bundance of different types of animals in the archaeological deposits (Stahl 2000). More specif ically, species that are associ ated primarily with forest habitat should decrease in numbers over time whereas those species that benefit from eating crops such as corn should increase. Intensification of agriculture may have multiple causes (Piperno and Pearsall 1998). 1) Increase of yields due to favorable climatic conditions. 2) Increased accessibili ty of water during the dry season. 3) An increase in population. 4) Changes in the subsistence patterns of local communities. Conversely, a decline in agricultural lands ma y be due to unfavorable conditions, water shortage, a decrease in p opulation size, and a reliance on other economic activities. Hunting probably focused on large game such as deer and other large bodied species of mammals and birds as suggested by zooarchaeological analysis (Reitz and Masucci 2004) and recent ethnographic stud ies (Hames and Vickers 1982, Redford and Robinson 1987). Most birds probably were ta ken opportunistically (Steadman et al. 2003). Within the prehistoric landscapes of the arid coastal areas , people could choose from a variety of birds for food, feathers, or captivity. Presently, coastal and upland habitats of Santa Elena Peninsula harbor tinamous, herons, gulls, ducks, pigeons, conures, motmots, and songbirds. In the more humid foothills of the Cordillera ChongónColonche, large-bodied guans a nd parrots can also be found. The accumulation of bird bones in archaeo logical deposits depends on many factors such as human selectivity, consumption patt erns, and the conditions of burial and bone

PAGE 245

229 preservation (Stahl 2003). Birds used as f ood include most large-bodied species, such as tinamous, guans, ducks, seabirds, hawks, and he rons. In tropical America smaller species with a large pectoral muscle mass relative to body size, such as quail, pigeons and doves, were (and are) also frequently consum ed (Hames and Vickers 1982, Redford and Robison, Kent et al. 1998, Steadman et al . 2003, Chapter 3, Chapter 5). Species that provide colorful plumage in clude parrots, trogons, hummingbi rds, and songbirds. Some of these may also have been kept as pets . In many archaeological sites, large bodied species predominate (Byrd 1976, Stahl 1991, 2000). Whether this trend is due to preferential exploitation of large species or simply an artifact of sampling methodology or post-depositional processes that affect bone preservation has to be evaluated on a caseby-case basis (Stahl 2003). Paleoenvironmental interpretations based on the zooarchaeological record are often limited because of the dearth of species-level identifications. Generic identifications are often useful, although species in some genera have widely different habitat preferences. The objective of this study is to use species level identificat ions (when possible) of birds to characterize the environment in the El Azú car valley and to aid in evaluating the role of the site within a regional social context. Birds are particularly well suited because of the wealth of information on habitat selection and distributional patterns as compared to other taxonomic groups (Stotz et al. 1996, Ridgely and Greenfield 2001a,b). Having a better understanding of the prehis toric surroundings may then he lp to interpret patterns in the utilization of game animals, non-local marine resources (for the production of crafts), and agricultural crops.

PAGE 246

230 Methods Study Site The El Azúcar valley is located on the Santa Elena Peninsula, about 25 km from the Pacific Ocean and is part of the Río Ch anduy drainage basin (Figure 4-1). The area is highly seasonal in terms of rainfall and temp eratures because of th e peninsula’s location at the northern limit of the cold Peru Current. Annual cycl es of cloudy, dry periods (when upwelling along the coast is st rong) alternating with s unny, hot periods with heavy downpours (when the Equatorial countercurrent br ings humid air to th e southern coast of Ecuador; Chapter 1) dominate the climate. ENSO is a predominant climatic feature that brings years of very high rain fall but also ones of droughts. In prehistoric times the vegetation was dominated by lowland dry scrub, savanna, and lowland deciduous forest (Hidalgo Nistri 1998, Piperno and Pearsall 1998; vegetation classification after Sierra 1999). Because of deforestation during the twentieth century, much of the area is now dominated by dry sc rub and near desert conditions (Marchant 1958). El Azúcar is located on the southern limits of the Cordillera Chongón-Colonche which marks the boundary between humid vegeta tion types found in the coastal mountain ranges and the Río Guayas basin and the xeri c vegetation of the peninsula (Piperno and Pearsall 1998, Veintimilla 2000). Local people co mment that as recently as 30 years ago humid forests could be found just north of the valley (M. A. Masucci personal comment). Excavation Details on the excavation of sites 30 and 47 in the El Azúcar valley are described in Masucci (1992). Archaeological field work was conducted between 1986 and 1988. Site 30 consists of a ca. 1.5 ha area in which a tren ch of four 1x1 m units was excavated next to a previously excavated 1x1 m unit. Excavatio n at arbitrary 10 cm levels (due to the

PAGE 247

231 lack of a discernable stratigraphy) to a dept h of 50-60 cm yielded a total sediment volume of 2.8 m3. Site 47 covers ca. 2.5 ha within whic h an area equivalent to 33 1x1 m units were excavated (see Reitz and Masu cci 2004, Figs. 5.3 and 5.4) yielding 45 m3 of sediments. Excavation in some units was done in arbitrary 5 or 10 centimeter levels (depending on unit) unless a natu ral stratigraphy was detected. Figure 4-1. Map showing major towns ( ) as well as the El Azúcar ( ), La Ponga ( ), Loma Alta ( ), Valdivia Village ( ), and Real Alto ( ) archaeological sites on the Santa Elena Peninsula, Ecuador. The shaded area represents the Cordillera Chongón-Colonche above the 200 m contour line.

PAGE 248

232 In this study I analyzed the bones found in si te 30 and in trenches X and H of site 47. The two trenches consist of a total of 11 1x1 m units which reached depths of 130180 cm below surface, equivalent to 15.2 m3 of sediment. All sediment was sieved through ¼ and 1/8 inch stacked screens (Masucci 1992). Reitz and Masucci (2004) divi ded the different units in to “lots” depending on the age of the deposits. Age for trench X was determined by radiocarbon dating and ceramic cross-dating. Lots in trench H were identified by ceramic similarities between the two trenches. Age of Lot B/C (X trench) was estimated to span a period between 60±150 BC and AD 280±85, deposits of Lot A (X and H trenches) accumulated between AD 280±85 and AD 370±80, and Lot B (H trench) was esti mated to have lasted from AD 370±80 to AD 600/650. Radiocarbon dates are corrected and calibrated. The three lots correspond to the Early, Middle, and Late Guangala occupati ons at this site (Re itz and Masucci 2004). Since sedimentation rates were fast and the site consists of a primary midden, I follow Reitz and Masucci (2004) in assuming that ve ry little vertical mi xing has occurred and that bones in the lots and levels are still in chronol ogical order (Masucci 1992). Analysis of Bird Bones Bird bones were identified using comparativ e material from the Florida Museum of Natural History (FLMNH) applying a qualitati ve approach based on shared and unique traits that characteri ze each taxon (Chapter 2). Once identified, specimens were assigned by their provenience to one of the three lots. Since pigeons and doves (columbids) ar e often associated with agricultural areas throughout Ecuador (personal observation), I compared the numbers of columbids to those of songbirds (passerines) for each lot us ing a Wilk’s G-test. I performed the test on both minimum number of individuals (MNI) an d number of individual specimens (NISP)

PAGE 249

233 as these two values represent a low and a high estimate for the true abundance of individuals in the archaeologi cal deposits (i.e., they repr esent the range of abundances that includes the true abundance). I focus on the numbers of columbids and passerines assuming that the presence of pigeons and doves is indicative of agri cultural activity, and that passerines tend to decr ease in abundance in arid ag ricultural habitats (M. P. Tellkamp, personal observation). As a caveat, some passerines may also benefit directly or indirectly from agricultural activity by f eeding on crops or by finding better habitat in open country. Pigeon and doves, however, are s eed predators that respond much more readily to a change in the availability of grains and seeds, which in Guangala Period would have been most abundant in corn fields. For comparative reasons I also reanalyzed the bird bones from the archaeological sites of Valdivia Village (2370±60–2030±60 14C yr BP; Stothert 1997, Ubelaker 1997), Loma Alta (ca. 3000–2300 14C yr BP; Stahl 1991, Zeidler 2003), and La Ponga (ca. 3200–2800 14C yr BP; Stahl 1991, Zeidler 2003) on the northern limit of the peninsula in the coastal foothills of the Co rdillera Chongón-Colonche. I hen ceforth refer to these sites collectively as the Valdivia Valley. I also re vised the bird bones from Real Alto, a site located near the mouth of the Río Chanduy (6195±215–3845±240 14C yr BP; Zeidler 2003; Figure 4-1). Despite the inherent difficulties in comparing the samples from different sites due to different excavation t echniques as well as site specific taphonomic conditions (Stahl 2003, Quitmyer 2004), these da ta allow making some inferences about prehistoric bird exploitation and gene ral paleoenvironmental conditions. The zooarchaeological record of the Valdivia Vall ey sites has been interpreted so far as

PAGE 250

234 Table 4-1. Minimum number of individuals (M NI) and number of individual specimens (NISP) for all species identified from the El Azúcar archaeological site, Santa Elena Peninsula, Ecuador, divided by “l ot.” Lots are groupings of sediment layers according to age of the deposits (Reitz and Masucci 2004). Age of Lot B/C was estimated to span a pe riod between 60±150 BC and AD 280±85, deposits of Lot A covers the time between AD 280±85 and AD 370±80, and Lot B was estimated to have la sted from AD 370±80 to AD 600/650. Radiocarbon dates are corrected and ca librated. The three lots correspond to the Early, Middle, and Late Guangala o ccupations at this site. See text for more details. MNI NISP MNI NISP Species B/C A B 30 B/C A B B/C Total Total TINAMIDAE Crypturellus transfasciatus 0 6 4 1 0 6 8 1 11 15 ARDEIDAE Ardea alba 1 0 0 0 1 0 0 0 1 1 Ixobrychus cf. exilis 0 1 0 0 0 1 0 0 1 1 ANATIDAE Anas bahamensis 0 1 0 0 0 1 0 0 1 1 ACCIPITRIDAE Parabuteo unicinctus 0 1 1 0 0 1 7 0 2 8 Buteo nitidus 0 0 1 0 0 0 2 0 1 2 FALCONIDAE Falco sparverius 0 1 0 0 0 1 0 0 1 1 COLUMBIDAE Zenaida auriculata 9 32 3 14 11 70 6 27 58 114 Zenaida meloda 1 10 2 3 7 20 2 7 16 36 Columbina minuta 1 8 0 2 1 9 0 2 11 12 Columbina buckleyi 1 1 3 3 1 1 3 4 8 9

PAGE 251

235 Table 4-1. Continued. MNI NISP MNI NISP Species B/C A B 30 B/C A B B/C Total Total COLUMBIDAE Columbina cruziana 5 17 5 11 5 33 7 16 38 61 Columbina sp. 1 9 0 2 1 11 0 2 12 14 Claravis pretiosa 0 1 0 0 0 1 0 0 1 1 Leptotila cf. verreauxi 0 1 0 0 0 1 0 0 1 1 Leptotila sp. 0 4 0 3 0 6 0 4 7 10 Geotrygon montana 0 1 0 0 0 1 0 0 1 1 UnID sp. 0 5 0 1 0 6 0 2 6 8 PSITTACIDAE Ara cf. ambigua 1 0 0 0 8 0 0 0 1 8 Forpus cf. coelestis 1 1 0 3 2 2 0 4 5 8 CUCULIDAE Crotophaga sulcirostris 0 1 0 0 0 1 0 0 1 1 THAMNOPHILIDAE Taraba major 1 0 0 0 1 0 0 0 1 1 FORMICARIIDAE Grallaria sp.* 0 1 0 0 0 2 0 0 1 2 TURDIDAE Turdus cf. maculirostris 0 1 0 1 0 1 0 1 2 2 MIMIDAE Mimus cf. longicaudatus 1 2 0 1 1 2 0 1 4 4

PAGE 252

236 Table 4-1. Continued. MNI NISP MNI NISP Species B/C A B 30 B/C A B B/C Total Total HIRUNDINIDAE Riparia riparia 0 1 0 0 0 2 0 0 1 2 CARDINALIDAE Spiza americana 0 1 0 0 0 1 0 0 1 1 EMBERIZIDAE Volatinia jacarina 0 0 1 0 0 0 1 0 1 1 Sporophila cf. peruviana 0 1 0 0 0 1 0 0 1 1 Sicalis cf. tackzanowskii 1 2 0 1 1 2 0 1 4 4 ICTERIDAE Dives cf. warszewiczi 0 1 0 0 0 2 0 0 1 2 Quiscalus mexicanus 1 1 1 0 1 1 1 0 3 3 Icterus cf. graceannae 0 0 0 1 0 0 0 1 1 1 Sturnella bellicosa 2 8 1 1 4 10 1 1 12 16 NON-IDENTIFIED PASSERIFORMES size of Tiaris 0 1 0 0 0 1 0 0 1 1 size of Sporophila 0 5 0 0 0 5 0 0 5 5 size of Sicalis 6 9 2 1 13 10 2 1 18 26 size of Tangara 2 0 0 2 4 0 0 4 4 8 size of Thraupis 0 2 0 0 0 2 0 0 2 2 size of Campylorhynchus 0 1 0 0 0 1 0 0 1 1 size of Saltator 4 17 1 6 7 28 1 6 28 42 size of Mimus 2 14 1 3 2 22 1 5 20 30

PAGE 253

237 Table 4-1. Continued. MNI NISP MNI NISP Species B/C A B 30 B/C A B B/C Total Total NON-IDENTIFIED PASSERIFORMES size of Furnarius 1 4 2 0 1 5 2 0 7 8 size of Sturnella 10 29 0 12 14 108 0 21 51 143 size of Icterus 4 9 0 2 4 13 0 2 15 19 size of large Saltator 1 1 0 0 2 1 0 0 2 3 size of Turdus 1 9 0 5 1 10 0 5 15 16 size of Dives 0 5 3 1 0 5 5 2 9 12 size of Cyanocorax 0 1 0 0 0 1 0 0 1 1 Passeriformes sp. 0 4 0 4 0 4 0 4 8 8 TOTAL 58 232 32 85 93 414 50 125 407 682 * Grallaria guatemalensis or G. watkinsi . suggesting that people at inland settlement s were themselves fishers who provisioned their communities with fish and mollusks (Byrd 1976, Stahl 1991). Thus, the general social context may have been similar to the one observed at El Azúc ar. Other sites on the Santa Elena Peninsula may have followed a sim ilar pattern, but little or no bird material was recovered from those sites (Byrd 1976). Results The bird assemblages from sites 30 and 47 at El Azúcar are highly unusual. Of the 686 individual specimens, 373 belong to the Passeriformes, surpassing even the archaeologically ubiquitous Columbidae which are second in abundance with 273 specimens (Table 4-1). The 31 identified spec ies belong to 16 families and nine orders.

PAGE 254

238 Despite the abundance of identified specime ns, 11 species are represented by only one bone. In contrast, the 63 bird bones from the Valdivia Valley site s include only eight individual specimens of Passeriformes and 11 specimens of Columbidae (Table 4-2). The 21 species belong to 16 families and 12 orders. The relative abundances of the Valdivia Valley assemblage are more evenly distributed than those from El Az úcar. At Real Alto only 12 specimens of ducks, represen ting at least two species, were found. As elsewhere, tinamous are an important component of the El Azúcar assemblage, but to a much lesser degree than at the highland site of La Chimba (Chapter 5; Table 4-1). Marine birds are absent from El Azúcar but present at the co astal sites of the Valdivia Valley. Shorebirds and wading birds are rare in the El Azúcar assemblage. Raptors and parrots are slightly more common. As is th e case for Site OGSE-80 near the town of Santa Elena (Chapter 3) a nd La Chimba (Chapter 5), Zenaida auriculata is by far the most common species of columbid. Several families are represented by a singl e individual or specimen. These are the Anatidae, Falconidae, Rallidae, Cuculid ae, Thamnophilidae, Hirundinidae, and Cardinalidae. The presence of a single duck stands in contrast to the bird assemblage of Real Alto which consisted of only 12 sp ecimens of dabbling ducks (Table 4-3). Identification of the 12 species of Passerifo rmes was possible due to the good condition of many bones, including ma ny identifiable fragments from the mandible and the premaxillary. Due to the difficulty of identifying songbirds I grouped many bones into size classes. I named the classes according to likely candidate species. These classes

PAGE 255

239 suggest that Sturnella bellicosa and Cacicus spp. were common species or culturally important at El Azúcar. A B 0 20 40 60 80 B/CAB Lot% MNI 0 20 40 60 80 B/CAB Lot% NISP Figure 4-2. A. Percentage minimum number of individuals (MNI) and B. percentage number of individual specimens (NISP) of passerines (songbirds; ), columbids (ground doves; ), and other birds ( ) for each lot at site 47, El Azúcar. As expected, columbids increase wher eas passerines decrease in abundance over time at Site 47 (Figure 4-2). The relative ch anges are, however, non-significant for MNI (Wilk’s G2 = 2.727; df = 2; p = 0.256) and only ma rginally significant for NISP (Wilk’s G2 = 5.795; df = 2; p = 0.055). Low sample size of birds in Lot B may be responsible for low statistical power. For Site 30, sample sizes also were too low for statistical analysis, but the abundance of doves tends to in crease over time as well (Figure 4-3). Systematic List In this section I summarize the bone a ssemblages and place each species in a biogeographic context. Sequence of nomencla ture and the current distributional data, unless otherwise noted, are from Ridgely a nd Greenfield (2001a,b). In accordance with recent molecular evidence (Che sser 2004), I treat the genus Tityra as part of the

PAGE 256

240 Cotingidae. The comparative osteology and just ifications for the identifications are found in Chapter 2. A B 0 20 40 60 80 54321 Level% MNI 0 20 40 60 80 54321 Level% NISP Figure 4-3. A. Percentage minimum number of individuals (MNI) and B. percentage number of individual specimens (NISP) of passerines (songbirds; ), columbids (ground doves; ), and other birds ( ) for each excavation level at site 30, El Azúcar. Order TINAMIFORMES Family TINAMIDAE Crypturellus cf. transfasciatus (Sclater and Salvin 1878), Pale-browed Tinamou Material.El Azúcar : Cranial end of sternum, proximal end of left scapula, complete left coracoid, distal e nds of right and left coracoid, complete left humerus, distal end of right humerus, proximal end of left ulna, carpometacarpus lacking metacarpal III, proximal end of right carpometacarpus, complete left phalanx 1 of digit II, distal end of left carpometacarpus, shaft of metacarpal II of left carpometacarpus, proximal end of left femur, distal end of left tibiotarsus. The 15 specimens represent a minimum of 11 individuals. Loma Alta : Complete left carpometacarpus, right carpometacarpus lacking metacarpal III, and proximal end of left femur, representing three individuals.

PAGE 257

241 Remarks.This common tinamou of d eciduous woodland and regenerating low secondary vegetation can be found in southwes tern Ecuador in two distinct populations. Loma Alta and El Azúcar are located at th e western and southwestern margins of the northern population, respectively. During my vis it to Loma Alta and La Ponga in August 2004, several individuals were calling incessant ly from low secondary vegetation on the bottom of small valleys. At El Azúcar no singing was detected. There may be two reasons for the apparent absence of C. transfasciatus from the surroundings of the Azúcar valley. 1) The area is heavily deforested and may not provide adequate conditions for this species anymore. 2) Although lacking climatol ogical data, the area around El Azúcar was decidedly drier than the Valdivia Valley. Whereas the vegetation around Loma Alta and La Ponga was lush and green, forest patches ne ar El Azúcar were very dry with most of the deciduous trees in a leafle ss state (Figure 4-4). I suggest that El Azúcar possibly was at least temporarily more humid during the Guangala occupation of the site. Alternatively, deforestation may not have been as severe. Th e local population of tinamous could also have a different breed ing phenology compared to those of the Valdivia Valley population. Order PROCELLARIIFORMES Family PROCELLARIIDAE Puffinus griseus (J. F. Gmelin 1789), Sooty Shearwater Material.La Ponga : Distal end of right humerus. Valdivia Village : Abraded complete right coracoid, shaft of left coraco id, and proximal end of left phalanx 1 of digit II, representing at le ast two individuals.

PAGE 258

242 Table 4-2. Minimum number of individuals (M NI) and number of individual specimens (NISP) for all species identified from the Valdivia Valley archaeological sites, just north of Santa El ena Peninsula, Ecuador. Valdivia La Ponga Loma Alta Species MNI NISP MNI NISP MNI NISP TINAMIDAE 0 0 0 0 0 0 Crypturellus transfasciatus 0 0 0 0 3 3 PROCELLARIIDAE 0 0 0 0 0 0 Puffinus griseus 2 3 1 1 0 0 PHALCROCORACIDAE 0 0 0 0 0 0 Phalacrocorax brasilianus 0 0 0 0 1 1 Phalacrocorax bougainvillii 1 3 0 0 0 0 ARDEIDAE 0 0 0 0 0 0 Ardea alba 0 0 0 0 1 1 Egretta cf. thula 0 0 0 0 1 1 ANATIDAE 0 0 0 0 0 0 cf. Netta erythrophthalma 0 0 0 0 1 1 Anatinae sp. 0 0 0 0 1 1 ACCIPITRIDAE 0 0 0 0 0 0 Ictinia plumbea 0 0 1 1 0 0 Buteo swainsoni 0 0 0 0 2 3 Buteogallus meridionalis 0 0 1 1 0 0 Parabuteo unicinctus 1 1 0 0 1 1 FALCONIDAE 0 0 0 0 0 0 Falconidae sp. 0 0 1 1 0 0

PAGE 259

243 Table 4-2. Continued. Valdivia La Ponga Loma Alta Species MNI NISP MNI NISP MNI NISP CRACIDAE 0 0 0 0 0 0 Ortalis erythroptera 0 0 0 0 1 1 Penelope cf. purpurascens 0 0 1 1 1 1 Cracidae sp. 0 0 0 0 1 1 SCOLOPACIDAE 0 0 0 0 0 0 Tringa flavipes 1 2 0 0 0 0 LARIDAE 0 0 0 0 0 0 Larus sp.‡ 1 4 0 0 1 1 Sterna sandvicensis 1 1 0 0 0 0 COLUMBIDAE 0 0 0 0 0 0 Columba cayennensis 0 0 1 1 0 0 Columbina buckleyi 1 1 0 0 0 0 Zenaida auriculata 0 0 1 2 2 4 Zenaida meloda 0 0 1 1 1 1 Leptotila sp. 0 0 0 0 1 1 PSITTACIDAE 0 0 0 0 0 0 Aratinga erythrogenys 0 0 1 3 0 0 Amazona autumnalis 0 0 0 0 1 1 Amazona farinosa 0 0 0 0 1 2 TYTONIDAE 0 0 0 0 0 0 Tyto alba 0 0 0 0 2 2

PAGE 260

244 Table 4-2. Continued. Valdivia La Ponga Loma Alta Species MNI NISP MNI NISP MNI NISP FURNARIIDAE 0 0 0 0 0 0 Furnariidae sp. 0 0 1 1 0 0 COTINGIDAE 0 0 0 0 0 0 Tityra semifasciata 0 0 1 7 0 0 CORVIDAE 0 0 0 0 0 0 Cyanocorax cf. mystacalis 0 0 0 0 1 1 Total 8 15 11 20 24 28 ‡ Larus atricilla or L. pipixcan . Remarks.This shearwater is a fairly common offshore species and can be seen regularly from land. Large die-offs ha ve been reported during times of “ peste ” (an unspecified disease) when bodies have b een reported washing up on shore (Marchant 1958, p. 359). If the peste is a phenomenon dating to prehis toric times, dead individuals may also have been picked up on the beaches and taken to the settlement for closer inspection. Alternatively, Valdivia fishers may have caught this species in nets as it dove in pursuit of prey. Given its reluctance to fly over land, the presence of this shearwater in La Ponga is probably due to human tr ansport rather than local hunting. Order PELECANIFORMES Family PHALACROCORACIDAE Phalacrocorax brasilianus (J. F. Gmelin 1789), Neotropical Cormorant Material.Loma Alta : Proximal end of left carpometacarpus.

PAGE 261

245 Remarks.In western Ecuador this cormor ant is a rare to uncommon visitor of streams, lakes, lagoons, and coastal areas. Large flocks can be seen, es pecially in the Río Guayas drainage basin. I have seen several groups of up to 30 individuals at the Represa El Azúcar. No breeding colony has been found to date in Ecuador. Phalacrocorax bougainvillii (Lesson 1837), Guanay Shag Material.Valdivia Village : Caudal fragment of rostrum premaxillaris, complete right humerus, and proximal end of right radi us, representing at least one individual. Remarks.Modern fishing practices and several recent strong El Niños probably have led to a sharp decline in the population, which may have been more common in the past. Presently, only few vagran ts reach the coasts of Ecua dor. Marchant (1958) observed varying numbers of this species every year he spent on Santa Elena Peninsula during the 1950s. From 8000 to 5800 14C yr BP, when upwelling off the Ecuadorian coast was strong and El Niños infrequent (Chapter 1), it is conceivable that breeding grounds, now reaching as far north as Isla Macabi, Peru (Johnsgard 1993), ca. 660 km south of Santa Elena Peninsula, could have expanded north, bringing the source populat ion closer to the settlement of Valdivia. Order CICONIIFORMES Family ARDEIDAE Ixobrychus exilis (J. F. Gmelin 1789), Least Bittern Material.El Azúcar : Fragment of distal end of left tarsometatarsus. Remarks.This bittern has been recorded only rarely in historic times in western Ecuador, making its presence in the El Azúcar bird assemblage a surprise. Perhaps this species has been more common in the past.

PAGE 262

246 The subspecific identity and status of the individuals seen at a few marshes in El Oro and Manabi provinces have not been determined. Ridgely and Greenfield (2001a) speculate that they may be either the Nearctic migrant I. e. exilis or the coastal, Peruvian I. e. peruvianus . The large size of trochlea metatarsi III suggests that this specimen can be attributed to I. e. peruvianus , the largest subspecies (Hancock and Kushan 1984). Ardea alba Linnaeus 1758, Great Egret Material.El Azúcar : Proximal end of left carpometacarpus. Loma Alta : Distal end of carpometacarpus. Remarks.This uncommon egret can be found in a variety of wetland habitats throughout western Ecuador, bei ng most common in the southw est. Its ability to wander to distant areas outside its normal range suggest s that Great Egrets were probably adept at exploiting the many walk-in well s scattered around Santa Elena Peninsula. I thus suggest that the bird was hunted loca lly. Around Loma Alta the pres ence of rivers and a nearby estuary provided more suitable habitat, sugge sting that the egret from that site was probably a local bird. Order ANSERIFORMES Family ANATIDAE Anas cf. discors Linnaeus 1766, Blue-winged Teal Material.Real Alto : Distal end with shaft of left coracoid. Remarks.This locally fairly common nearctic winter migrant occurs on lakes, ponds, and freshwater marshes. Up to 3200 i ndividuals have been observed on the Santa Elena Peninsula, where this species is most common between October and April. Osteologically, A. discors is indistinguishable from A. cyanoptera , a resident and migrant

PAGE 263

247 now virtually extirpated from Ecuador. The iden tification is thus tentative. The coracoid belonged to a juvenile bird as judged by its pitted appearance, sugge sting that breeding occurred in the past. Marchant (1958) suspected that A. discors may still have bred on Santa Elena Peninsula in the twentieth century. Anas cf. bahamensis Linnaeus 1758, White-cheeked Pintail Material.El Azúcar : Proximal end of left scapula. Real Alto : Cranial fragment of sternum, proximal end of right scapula, comple te right coracoid, pr oximal ends of three right coracoids, distal end of left coracoid, proximal end of left humerus, proximal end of left ulna, proximal end of a right carpomet acarpus. The 10 specimens represent at least seven individuals. Remarks.This is presently the most co mmon duck in western Ecuador. Its highest population occurs on the Santa Elena Peninsula where up to 2000-4000 individuals have been observed on artificial lagoons. Anatinae sp. Material.Loma Alta : Proximal end of right coracoid. Remarks.The proximal end of the coracoid cannot be identified with confidence. It is similar in size to that of a 739 g specimen (UF 42037) of Anas acuta (Northern Pintail). Cairina moschata (Muscovy Duck) is unlikely based on size (Stahl 2005), but unfortunately no specimens of small wild females were available. Currently, no Ecuadorian species of the genus Anas is as large as A. acuta . Duck bones found in the excavations of Real Alto, about 23 km south of El Azúcar (Figur e 4-1, Table 4-3), are also very large and possibly belong to A. acuta or A. clypeata . Neither species is known south of northern Colombia during the borea l winter (Madge and Burn 1988). On the

PAGE 264

248 other hand, as with the migratory A. cyanoptera , the migratory populations of A. acuta and/or A. clypeata may have become extirpated. Table 4-3. Minimum number of individuals (M NI) and number of individual specimens (NISP) for all species at the Real Alto archaeological site, Santa Elena Peninsula, Ecuador. Species MNI NISP ANATIDAE Anas cf. bahamensis 7 10 Anas cf. discors 1 1 Netta erythrophthalma (Wied 1832), Southern Pochard Material.Loma Alta : Distal end of left coracoid. Remarks.This is a rare and local duck found on lakes and freshwater marshes, with no recent records from western Ecuador. Th is species must have been more common in the past as a group of “hundreds” of i ndividuals was observed once at a dam on the Santa Elena Peninsula (Marchant 1958, p. 364). Order ACCIPITRIFORMES Family ACCIPITRIDAE Ictinia plumbea (J. F. Gmelin 1788), Plumbeous Kite Material.La Ponga : Distal end of right humerus. Remarks.An uncommon to locally fairly common kite that has been recorded in western Ecuador south to southeastern Guay as province. A few records are from the Cordillera Chongón-Colonche. Buteogallus meridionalis (Latham 1790), Savanna Hawk Material.La Ponga : Proximal end of left scapula.

PAGE 265

249 Remarks.This hawk is uncommon to locally fairly common in areas with scattered trees and in edges of deci duous and humid forest. La P onga was probably covered by semi-humid forest before the settlement by pe ople in the Holocene, as this site in the foothills of the Cordillera Chongón-Colonche receives abundant mist. If hunted locally, the presence of B. meridionalis suggests considerable conversi on of forest to agricultural lands during the Late Formative. Parabuteo unicinctus (Temminck 1824), Harris’s Hawk Material.El Azúcar : Distal end of right humerus, pr oximal and distal ends of left humeri, damaged left proximal end and dama ged right distal end of tarsometatarsi, complete left first phalanx of the hallux, and two complete left claws. The eight specimens represent at leas t 2 individuals. Loma Alta : Proximal end of left ulna. Remarks.An inhabitant of the arid lowl ands and inter-Andean valleys, this species is uncommon to locally fairly common. Duri ng my visit of El Azúcar I observed two soaring individuals. Buteo nitidus (Latham 1790), Gray Hawk Material.El Azúcar : Shaft of left coracoid and di stal end of right tibiotarsus, representing at le ast one individual. Remarks.Rare to locally uncommon, th is hawk can be found throughout xeric woodlands and scrub as well as forest cleari ngs and borders of more humid areas of the northern Guayas drainage basin and southwestern Andean foothills. Buteo swainsoni Bonaparte 1838, Swainson’s Hawk Material.Loma Alta : Proximal end of left humerus, distal end of right coracoid, and distal end of left tibiotarus, representing at least two individuals.

PAGE 266

250 Remarks.This is a rare Nearctic migr ant with very few records in Ecuador, although several sightings near the town of Sa nta Elena have been reported, more than from any other site in Ecuador. Family FALCONIDAE Falco sparverius Linnaeus 1758, American Kestrel Material.El Azúcar : Distal end of left carpometacarpus. Remarks.A common inhabitant of the dry in ter-Andean valleys, the kestrel is rare in coastal Ecuador. Several recent sightings s uggest that this speci es is only recently colonizing western Guayas provi nce. The bone at El Azúcar suggests that this species may have had a wider distribution in the pa st. There is no obvious reason why this species should have disappear ed following the Guangala pe riod or why it is recolonizing now. Order GALLIFORMES Family CRACIDAE Penelope cf. purpurascens Wagler 1830, Crested Guan Material.La Ponga : Distal end of right humerus. Loma Alta : Proximal end of right tarsometatarsus. Remarks.Rare to locally common, this larg e guan still persists at several sites in the Cordillera Chongón-Colonche, including Loma Alta (Becker and López 1997). Its persistence is noteworthy since this gamebird has been hunted there for several thousands of years. Ortalis cf. erythroptera (Sclater and Salvin 1870), Rufous-headed Chachalaca Material.Loma Alta : Proximal end of left femur.

PAGE 267

251 Remarks.Uncommon to locally fairly co mmon this chachalaca prefers humid or semihumid forest. It tolerates some deforest ation and can be found in forest fragments. Interestingly, locals in western Ecuador cont end that this species is difficult to hunt. Order CHARADRIIFORMES Family SCOLOPACIDAE Arenaria interpres (Linnaeus 1758), Ruddy Turnstone Material.Valdivia Village : Proximal and distal ends of left humeri, representing one individual. Remarks.This migratory shorebird is a common visitor throughout western Ecuador, where it is most common on roc ky beaches of Santa Elena Peninsula. Nonbreeding birds may stay th roughout the northern summer. Family LARIDAE Larus pipixcan , Wagler 1831, FranklinÂ’s Gull Material.Loma Alta : Complete left coracoid. Valdivia Village : Complete right coracoid, distal end of right humerus, proxima l end of right ulna, and proximal end of a right carpometacarpus, representing one individual. Remarks.Great intraspecific variability makes these two species very difficult to distinguish. Both are fairly common migrants along the coast of Ecuador and both may fly inland. Large concentrations of both sp ecies can be observed on the Santa Elena Peninsula, especially in coastal settings. Sterna hirundinacea Lesson 1831, South American Tern Material.Valdivia Village : Distal end of right tarsometatarsus.

PAGE 268

252 Remarks.This tern is a casual migrant to the coast of Ecuador where it rarely can be seen on sandy beaches and lagoons. This species breeds only as far north as 15 S, but can be regularly seen along the co ast of central Peru (Harrison 1983). Order COLUMBIFORMES Famly COLUMBIDAE Columba cf. cayennensis Bonnaterre 1792, Pale-vented Pigeon Material.La Ponga : Proximal end with shaft of right femur. Remarks.Uncommon to locally common in humid forest and semi humid woodlands, forest borders, and mangroves. Despite its moderately high modern abundance, only one bone from La Ponga can be attributed to this species. Being a canopy species and given the relative eas e of hunting the more terrestrial Zenaida spp., this pigeon probably was pursued only very opportunistically. Zenaida auriculata (Des Murs 1847), Eared Dove Material.El Azúcar : Cranial end of three sterna, pr oximal ends of four right and five left scapulae, two complete right cora coids, proximal end of right and three left coracoids, distal ends of four right and five left coracoid s, shafts of right and left coracoids, whole right and left humeri, proxima l ends of five right and four left humeri, distal ends of four right and three left humeri, proximal ends of four right and two left ulnae, distal ends of three ri ght and four left ulnae, proxi mal ends of two right and one left radii, proximal ends of right and two left carpometacarpi, distal ends of two left and three right carpometacarpi, five right and tw o left carpometacarpi lacking metacarpal III, two complete phalanges 1 of digit II, complete left femur, proximal ends of two right and three left femora, distal ends of three right and two left femora, proximal ends of two

PAGE 269

253 right and three left tibiot arsi, distal ends of two right and f our left tibiotarsi, shaft of left tibiotarsus, complete left tarsometatarsus, proximal ends of three right and two left tarsometatarsi, distal ends of two right and two left tarsom etatarsi, and shaft of right tarsometatarsus. The 114 specimens repr esent at least 58 individuals. La Ponga : Distal end of left coracoid and complete left fe mur, representing one individual. Loma Alta : Nearly complete right coracoid, distal end of left coracoid, complete right ulna, and distal end of left ulna. The four specimens represent at least two individuals. Remarks.A common dove in agricultural areas and urban settlements on the Santa Elena Peninsula, its numbers fluctuate widely within and between years (Marchant 1958). In the non-breeding season it tends to form large flocks. Zenaida meloda (Tschudi 1843), West Peruvian Dove Material.El Azúcar : Cranial fragment of sternum, proximal ends of two right and five left scapulae, two complete left coracoids, proximal ends of two left coracoids, distal ends of right and three left co racoids, proximal ends of right and left humeri, distal ends of two right and one le ft humerus, complete right ulna, proximal e nds of right and left ulna, distal end of left ulna, proximal ends of two right and left radius, distal end of distal right radius, two left carpometacarpi without metacarpal III, proximal and distal ends of left femora, distal end of le ft tiobiotarsus, whole right tarsometatarsus, proximal and distal ends of right tarsometatarsi. The 36 speci mens represent at least 16 individuals. La Ponga : Complete right coracoid. Loma Alta : Proximal end of right femur. Remarks.This is an uncommon and local dove inhabiting ripari an areas, desert scrub, and near mangroves in the more arid lo wlands. It shows a marked preference for wooded valleys with some tree cover (Mar chant 1958). The West Peruvian Dove does

PAGE 270

254 not flock as extensively in the post-breedi ng season as its more common congener and may thus be more difficult to hunt. I have observed this species on artificially watered pastures. Walk-in wells and ir rigated fields probably attrac ted these doves to prehistoric agricultural areas. Columbina minuta (Linnaeus 1766), Plain-breasted Ground Dove Material.El Azúcar : Two proximal ends of right s capulae, proximal end of right coracoid, proximal end of right humerus, shaft of right humerus, distal ends of two left ulnae, right and left carpometacarpi lack ing metacarpus III, proximal end of right carpometacarpus, proximal end of left femur, and proximal end of left tarsometatarsus. The 12 specimens represent at least 11 individuals. Remarks.Rare to uncommon and local, this tiny dove appears to have colonized the Ecuadorian arid lowlands only recently. As with Falco sparverius , this species probably contracted in range for unknown reasons in pr ehistoric times. Columbina buckleyi (Sclater and Salvin 1877), Ecuadorian Ground Dove Material.El Azúcar : Two whole left coracoids, dist al end of right coracoid, whole right and left humeri, proximal end of left ul na, distal end of right femur, and proximal ends of right and left tars ometatarsus. The nine specimens represent at least eight individuals. Valdivia Village : Nearly complete right ulna. Remarks.This is a fairly common ground dove associated with clearings, gardens, agricultural areas, and young secondary forest in more humid areas of the western Ecuadorian lowlands. El Azú car seems too dry today to support this species. I hypothesize that C. buckleyi can expand its range during a succession of wet years and thus occur temporarily at El Azúcar.

PAGE 271

255 Columbina cruziana Prévost 1842, Croaking Ground Dove Material.El Azúcar : Proximal ends of two right a nd four left scapulae, complete right and complete left cor acoids, damaged complete righ t coracoid, proximal end of right coracoid, four distal ends of right coracoids, six dist al ends of left coracoids, complete left humerus, proximal end of right humerus, distal ends of two right and two left humeri, whole right ulna, pr oximal ends of two right and th ree left ulnae, distal ends of two right and four left ulnae, three right and three left carpometacarpi lacking metacarpal III, proximal end of left carpom etacarpus, complete left femur, proximal end of left and distal end of right femur, proximal end of right ti biotarsus, distal ends of two right and four left tibi otarsi, the proximal ends of thr ee right tarsometat arsi, and distal ends of right and left tarsomet atarsus. The 61 specimens repres ent at least 38 individuals. Remarks.Possibly the most common columb id in the arid lowlands of western Ecuador, this dove is known to breed dur ing all months on Santa Elena Peninsula (Marchant 1958). This species is replaced in more humid areas by C. buckleyi with which it overlaps extensively in some areas. Claravis pretiosa (Ferrari-Pérez 1886), Blue Ground Dove Material.El Azúcar : Complete right tarsometatarsus. Remarks.This dove is uncommon to locally fairly common this dove and rather inconspicuous and nomadic inhabitant of shrubby and tall secondary vegetation, and humid forest borders. Its presen ce at El Azúcar may be e xplained by its nomadic habits which would have allowed this species to have responded fairly rapidly to increased humidity on the Santa Elena Peninsul a during a succession of wet years.

PAGE 272

256 Leptotila cf. verreauxi (Bonaparte1855), White-tipped Dove Material.El Azúcar : Proximal end of left femur. Remarks.A common to fairly common dove in semi-humid and deciduous forest it occurs on or near the ground. It is especially common in ag ricultural areas with shrubby areas and woodlots. Although this species vent ures out into the open, it is most often heard from dense vegetation making it le ss attractive for hunting than the common Zenaida doves. Several other bones may belong to th is species, but given the presence of three species of Leptotila in western Ecuador and the lack of comparative material, identification to species is difficu lt. One of the other two congeners, L. pallida , usually replaces L. verreauxi in more humid areas, whereas L. ochraceiventris is rare and even less likely to occur outside of forest. Geotrygon montana (Linnaeus 1758), Ruddy Quail-Dove Material.El Azúcar : Distal end of right tibiotarsus. Remarks.Uncommon to fairly common in the interior of humid forest, this dove may have reached El Azúcar during extremely wet years. Order PSITTACIFORMES Family PSITTACIDAE Ara cf. ambiguus (Bechstein 1811), Great Green Macaw Material.El Azúcar : Distal end of left tarsom etatarsus, four whole pedal phalanges, and two left claws, repr esenting at least one individual. Remarks.This currently rare macaw is the only member of its genus ranging into the arid and semi-humid portions of western Ecuador. A small population still persists in the Cordillera Chongón-Colonche. During the Gu angala period when large areas of the

PAGE 273

257 Santa Elena Peninsula was likely to be cove red by deciduous forest, the population was probably as close as the southwestern lim its of the cordillera. Hunting this canopydwelling bird, however, was probably difficult. Forpus coelestis (Lesson 1847), Pacific Parrotlet Material.El Azúcar : Complete right and left coracoid s, distal end of left coracoid, complete right humerus, proximal end of left humerus, proximal ends of left and right ulna, proximal end of left carpometacarpus, and complete right tarsometatarsus. The eight specimens represent at least five individuals. Remarks.This is a common to very comm on species in xeric habitats as well as deciduous forest and urban areas. Presently it occurs at great number near El Azúcar with flocks of up to 30 individuals being a comm on sight throughout the day (M. P. Tellkamp personal observation). Amazona autumnalis (Linnaeus 1758), Red-lored Parrot Material.Loma Alta : Distal fragment of left ulna. Remarks.This is a rare to locally fa irly common parrot found in the canopy of humid and deciduous forest. It can still be found in the Cordillera Chongón-Colonche whereas in other areas its numbers are decreasing. Amazona farinosa (Boddaert 1783), Mealy Parrot Material.Loma Alta: Proximal end of ri ght ulna and distal end of left ulna, representing one individual. Remarks.An uncommon to locally comm on parrot of the canopy and borders of humid forest. Currently it is found no closer than 130 km northwest of Loma Alta. This species was also found at site OGSE-80 in Ve gas culture context (C hapter 3). Although

PAGE 274

258 birds may have reached Loma Alta and OGSE80 through trade, it seems plausible that its range formerly included much of the Cordillera Chongón-Colonche, before deforestation and hunting le d to local extirpation. Order CUCULIFORMES Family CUCULIDAE Crotophaga sulcirostris Swainson 1827, Groove-billed Ani Material.El Azúcar : Proximal end of right scapula. Remarks.Presently this ani is common to fairly common in agricultural areas and the arid region of western Ecuador. It is a resident of Santa Elena Peninsula where breeding may not occur during very dry years. Order STRIGIFORMES Family TYTONIDAE Tyto alba (Scopoli 1769), Barn Owl Material.Loma Alta : Proximal and distal end of left ulnae, representing two individuals. Remarks.This cosmopolitan owl is an uncommon to locally fairly common species in agricultural and na tural semi-open areas. The bones from Loma Alta probably are from a local population of Tyto alba . Order PASSERIFORMES Family THAMNOPHILIDAE Taraba major (Vieillot 1816), Great Antshrike Material.El Azúcar : Proximal end with shaft of left femur.

PAGE 275

259 Remarks.An uncommon to fairly common bird of dense secondary vegetation and forest border, this antshrik e does not occur in the most arid parts of Santa Elena Peninsula. El Azúcar presently seems too dry to support a local populatio n of this species, which throughout prehistory probably expande d to the area during extended periods of very wet years. Family FORMICARIIDAE Grallaria sp., Antpitta Material.Distal ends of two right tibiotarsi. Remarks.The range of Grallaria guatemalensis (Scaled Antpitta) includes El Azúcar, whereas G. watkinsi (Watkin’s Antpitta) occurs with in 15 km of the site in the Cordillera Chongón-Colonche (Ridgely et al. 2003). Both species ar e associated with forest habitat, G. guatemalensis occurs in tall dry and humid forest and G. watkinsi is more common in secondary vegetation of humid forest. Family COTINGIDAE Tityra semifasciata (Spix 1825), Masked Tityra Material.La Ponga : Proximal and distal ends of ri ght humeri, distal end of left humerus, complete left ulna, nearly complete left ulna, proximal end of left ulna, and distal end of right carpometacarpus , representing one individual. Remarks.A common and conspicuous canopy bird of humid forests and clearings with scattered trees, it can still be found at Loma Alta (Becker and Lopez 1997). Family CORVIDAE Cyanocorax cf. mystacalis (I. Geoffroy Saint Hilair e 1835), White-tailed Jay Material.Loma Alta : Distal end with shaft of left femur.

PAGE 276

260 Remarks.This jay is an uncommon to fair ly common jay inhabitant of the arid and semi arid regions of western Ecuador, I obs erved a group of seven individuals about 10 km south of El Azúcar. Ridgl ey et al. (2003) place the nor thern limit near the Valdivia Valley, but at Loma Alta this species has not been observed recently, possibly related to the recent declines throughout its range. Family TURDIDAE Turdus cf. maculirostris Berlepsch and Tackzanow ski 1883, Ecuadorian Thrush Material.El Azúcar : Distal end of right coracoid and proximal end of left femur. The 2 specimens represent at least two individuals. Remarks.This fairly large thrush is uncommon to locally fairly common in a variety of humid and semi humid habitats, bu t is more tolerant of dry conditions than other large lowland thrushes. Pr esently it has not been record ed at El Azúcar, although it may occur there during a succession of wet years. Family MIMIDAE Mimus cf. longicaudatus Tschudi 1844, Long-tailed Mockingbird Material.El Azúcar : Proximal end of right humerus, complete ulna, proximal end of left ulna, and distal en d of right tarsometatarsus, representing four individuals. Remarks.A common bird in arid and ev en barren areas, it is easily observed throughout the Santa Elena Peninsula and is very abundant today at El Azúcar. Family HIRUNDINIDAE Riparia riparia (Linnaeus 1758), Collared Sand Martin Material.El Azúcar : Complete right and left humer i, representing one individual.

PAGE 277

261 Remarks.This is an uncommon migrant swallow in western Ecuador. Not having a local breeding population with vulnerable fl edglings and being difficult to catch, the presence of this bird in the El Azúcar assemblage is surprising. Family CARDINALIDAE Spiza americana (J. F. Gmelin 1789), Dickcissel Material.El Azúcar : Incomplete ramus and portion of mandibular symphysis. Remarks.The mandibular fragment is the first evidence for this species from Ecuador. Family EMBERIZIDAE Volatinia jacarina (Linnaeus 1766), Blue-black Grassquit Material.El Azúcar : Rostrum maxillaris with os nasale. Remarks.A common to very common bird in agricultural areas, grassy habitats, and rural settlements, it proba bly was a common inhabitant of the prehistoric El Azúcar settlement. Sporophila cf. peruviana (Lesson 1842), Parrot-billed Seedeater Material.El Azúcar : Rostrum maxillaris with os nasale. Remarks.This is an uncommon to fairly common seedeater with similar microhabitat preferences as Volatinia jacarina , but it is more restricted to arid areas. During the non-breeding season this species forms flocks. cf. Sicalis tackzanowskii Sharpe 1888, Sul phur-throated Finch Material.El Azúcar : Four rostra maxillaris, tw o with include os nasale, representing four individuals.

PAGE 278

262 Remarks.Presently a rare species in we stern Ecuador, it may have been more common in the past as Marchant (1958) observed “vast” flocks in the 1950s. The population has since declined. Since breeding has not been recorded on the Santa Elena Peninsula, the erratic appearance of this species may hint at post-breeding migratory movements. Family ICTERIDAE cf. Dives warszewiczi (Cabanis 1861), Scrub Blackbird Material.El Azúcar : Distal end with shaft of right humerus and distal end of right tarsometatarsus, representing one individual. Remarks.This fairly common to comm on bird of open habitats, including agricultural areas under semi humid conditions, avoids most of the arid Santa Elena Peninsula. Again, its presence in the El Azúcar assemblage may be due to unusually wet conditions lasting for several years. Quiscalus mexicanus (J. F. Gmelin 1788), Great-tailed Grackle Material.El Azúcar : Cranial portion and symphysis of mandible, proximal end of right humerus, and distal end of left car pometacarpus, representing three individuals. Remarks.This is a fairly common species in and around mangroves, along beaches, shorelines, and coastal rivers. El Azúcar specimens are probably derived from traded live or dead birds. In the past the Chanduy area may have been the most important trading center for Guangala people at El Azú car (Reitz and Masucci 2004). I suggest this is the source area for the grackle bones from El Azúcar. The estuary of the Río Zapotal near Chanduy (Figure 4-1) presently has no ma ngrove stands, but a few saplings can be

PAGE 279

263 seen on the mudflats (Figure 45) suggesting that the condi tions are favorable. Pearsall (1988, Mapa 5) shows most of the Zapotal estuary as being covered by mangroves. Icterus cf. graceannae Cassin 1867, White-edged Oriole Material.El Azúcar : Caudal portion of left mandibular ramus. Remarks.An uncommon to fairly common bi rd in the xeric environments of the Santa Elena Peninsula, this species proba bly was common around El Azúcar during the Guangala Phase occupation. Sturnella bellicosa Filippi 1847, Peruvian Meadowlark Material.El Azúcar : Rostral end of rostrum maxillaris, partial rostrum maxillaris with os nasale, two mandibular symphyses, cau dal ends of right and two left mandibular rami, whole left coracoid, whole left humeru s, whole but partially broken left humerus, proximal ends of three humeri, complete le ft ulna, proximal end of right ulna, and complete right femur. The 16 specimens represent at l east 12 individuals. Remarks.This is a common and conspi cuous bird around pastures, agricultural areas, and in arid scrublands. In August 2004, I observed this species near humid pastures that were artificially watere d, but it was absent in the larg e area of dry scrub surveyed by car (ca. 6 km2 of potential habitat). Although it is fairly common and tolerant of dry conditions, it may prefer humid areas such as irrigated agricultural fi elds and the vicinity of walk-in wells. Discussion The unusual avian assemblage in sites 30 and 47 of El Azúcar reflects the settlement’s specialization in the social we b of the Guangala people on the Santa Elena Peninsula. Being removed 25 km from the sea, El Azúcar was a site of craft production and possibly hunting for trad e (Reitz and Masucci 2004). These activities may have

PAGE 280

264 affected the people’s selectivity for particular bird species such as songbirds and pigeons for local consumption. Most of the species from the El Azúcar sites are open-country birds that are presently associated with human settlements in coastal Ecuador. Apart from the tinamou and the macaw, none of the birds can be considered true forest birds. This contrasts sharply with the small assemblage obtained from the Valdivia Valley which does contain some humid forest birds, including guans and two species of Amazona parrots. A. B. B Figure 4-4. A. The valley of El Azúcar taken from the road about 2 km from the modern town. B. A side valley of the Río Valdivia near La Ponga. Photographs were taken 31 July 2004 and 30 July 2004, respectively.

PAGE 281

265 Another notable aspect of the El Azúcar bi rds is the lack of marine species. Apart from Quiscalus mexicanus , Ardea alba , and Anas bahamensis , none of the birds has a close association with any coastal habitat. The latter two species can also take advantage of inland walk-in wells and ponds. The Valdivia Village inland sites, La Ponga and Loma Alta, both contain marine birds, including gulls and Puffinus griseus , which rarely come near shore. Guangala Fishers, Farmers, Hunters, and Bird Trappers? One of the main questions regarding the Gu angala people of El Azúcar is whether they fished themselves or traded marine resources from skilled coastal fishers (Reitz and Masucci 2004). Based on a zooarchaeological an alysis of Loma Alta and La Ponga, Stahl (1991) concluded that inland farmers provided their settlements with fish themselves. These two sites are about 10 km inland. The di versity and quantity of fish and mussels at El Azúcar make it doubtful that the inhab itants were both fishers and farmers since catching and transporting fish are very tim e consuming activities that reduce any time spent on agriculture, crafts and hunting deer (Reitz and Masucci 2004). Marine resources were probably obtained via trad e. The lack of large birds from El Azúcar despite good preservation of even small bones can be e xplained by trade as we ll, which would have resulted in deposition of the bones far away from El Azúcar. Hunting techniques for pigeons, doves, and s ongbirds need also be considered. The predominance of small-bodied species that liv e close to the ground, especially in shrubby vegetation, and/or form flocks ( Columbina spp., Zenaida spp., Forpus coelestis , many finches, and Sturnella bellicosa ) suggest to an ornithologist familiar with mist-netting birds that some forms of nets or bird lime were used. Other possible techniques to obtain large numbers of small songbirds, consid ered less likely, include rock throwing,

PAGE 282

266 slingshots, and blow guns. Although questio ned by some archaeologists (Byrd 1976), Guangala people probably used fishing nets as suggested by the presence of schooling fish at the Valdivia Village site (Stother t 1997, Reitz and Masucci 2004). These nets also could have been used for birds. The Paleoenvironment of El Azúcar The near lack of humid forest birds at El Azúcar suggests that the environment around the settlement consisted of dry scr ub and deciduous forests. Occasionally, more humid conditions due to ENSO may have allowed for the temporary expansion of some open-habitat species from humid areas into the El Azúcar valley. Most of the assemblage resembles the modern avifauna around El Azúcar. Many of the common species in the archaeological record are common today as we ll. During my brief stay in the area during August 2004, I observed nine of the 31 species identified from bones. Adding the species observed by Marchant (1958) and mentioned in Ridgley and Greenfield (2001a,b) as inhabiting the most arid portions of the Sant a Elena Peninsula, the list ascends to 23 species. The remaining eight species can be divided into two groups. The first includes wetland species that may have used walk-i n wells or were traded. The second group consists of the open-habitat birds from the more humid por tions of the peninsula that extended their ranges toward El Azúcar dur ing very wet years. The presence of these species suggests that, during the Guangala occupation, El Niño years must have brought great amounts of rainfall to this area. Between 60 BC a nd 600 AD, El Niños were especially frequent around 100 AD and be tween 400 and 600 AD (Moy et al. 2002). Although during the Real Alto occupation (ca. 6200-3800 14C yr BC; Zeidler 2003) climate was similar to that of El Azú car (Pearsall 1988, Moy et al. 2002, Reitz and Masucci 2004), macrobotanical data from this site show that some humid forest trees

PAGE 283

267 existed on the Santa Elena Peninsula in the past (Pearsall 1988). These species now occur ca. 60 km away from Real Alto in the Cordillera Chongón-Colonche, suggesting that conditions were at l east temporarily (and possibly local ly around rivers) humid enough to sustain humid forest trees and thus birds from humid vegetation types. Figure 4-5. The mudflats of the Río Chanduy estuary with several Rhizophora sp. saplings. The near absence of large-bodied species at El Azúcar may be due to three factors that have a direct bearing on paleoenvironmenta l interpretations. 1) The area was arid and supported only an avifauna that could tolerate very dry conditions. If other habitats had been nearby, species characteristic of to those habitats should have been recorded in the archaeological record. This notion is suppor ted by bone assemblages from Chiapas, Mexico (Steadman et al. 2003), the Valdiv ia Valley sites (Byrd 1976, Stahl 1991, this study), the Jama Valley (Stahl 2000), and La ke Titicaca, Bolivia (Kent et al. 1999), in which species from all locally available habitats are represen ted. It is possible, however, that 2) many birds were trad ed and thus not obtained locally or that 3) many birds were

PAGE 284

268 hunted only opportunistically and consumed at distant butcheri ng sites during hunting trips (Kent 1993). It is not possible to distinguish among these hypotheses at this time. Small species of vertebrates, such as songbirds, are rarely taken by traditional neotropical hunters in any great numbers (Hames a nd Vickers 1982, Redford and Robinson 1987). Many of the passerine bone s found at El Azúcar were charred suggesting that these birds were used for food. Given that hunting of large game and trading of large quant ities of heavy merchandise was probably a male activity (Hames and Vickers 1982, Redford and Robinson 1987, Re itz and Masucci 2004), I consider it likely that these birds were mainly hunted in agricultural areas by women and children who were tending the fields. Unfortunatel y, ethnozoological studies usually focus on hunting of large game by men and ignore the possibility that children and women may often kill and consume small mammals and birds near settlements. From my own experience in Ecuador with colonos inhabiting the foothills of the Andes, children often times kill as many as 20 songbirds per day, ma ny of which are roasted over fire and eaten. Children often measure themselves with other children accordi ng to their abilities to kill small, colorful birds. The increase in columbids relative to songbi rds suggests (Figures 4-2 and 4-3) that agriculture did become more extensive ove r time, but the initial large number of columbids indicates that forest conversi on around the settlement was probably swift during the first years of occupation. Rapid inte nsification of agricultu re as a response to favorable climatic conditions has been observe d as well for the Lake Titicaca area in the Bolivian altiplano (Binford et al. 1997).

PAGE 285

269 The bird assemblage of the two archae ological sites of El Azúcar reveals distributional aspects of seve ral presently uncommon species. Ixobrychus exilis , Falco sparverius , and Columbina minuta are considered recent invaders or vagrants in western Ecuador (Ridgely and Greenfield 2001a). This study suggests that these birds may have been more common in the past, but then disappeared for unknown reasons before the recent range expansions. The Valdivia Valley assemblage also includes at least one species whose range may have contracted from recent habitat loss. Although some individuals may have reached the Santa Elena Peninsula by trade, it is equally like that Amazona farinosa had a greater range in western Ec uador that included at least the western portion of the Cordillera Chongón-Colonche. Conclusion Species-level identifications of bird s have made possible novel hypotheses regarding the exploitation of birds by preh istoric people. At the same time this study supports many conclusions made previously re garding the Guangala settlement of El Azúcar (Masucci 1995, Reitz and Masucci 2004). Among the new ideas are the notion that most birds were probably hunted and consumed locally by women and children using innovative approaches such as fishing nets or bird lime. The large number of small birds suggests that Guangala people adapted in a p eculiar way to resource limitations during dry years by exploiting large numbe rs of low-protein songbirds. Previous conclusions supported by this st udy are the socio-economic position of the settlement within a regional trading networ k. Although circumstantial, the evidence is at least consistent. Future research will have to focus on some of the alternative hypotheses. This study also weakly underscores the conclusion that agriculture became increasingly important with time during the occupation of El Azúcar (Reitz and Masucci 2004). The

PAGE 286

270 Late Phase coincides with increased frequencie s of El Niño events. Increased rainfall and higher annual yields may have driven this change. An intensification of agriculture requires a large amount of effort, taking away time from other activ ities such as craft production. High output of cr ops would also eventually have eliminated the need to hunt small birds, resulting in the observed declin es in bone and bead densitity (Reitz and Masucci 2004). Whether the unusual characteristic s of the El Azúcar archaeological sites simply represent the idiosyncrasies of a single settlement or were pa rt of a more general pattern awaits further studies. Apart from good preservation, sampling with fine meshed screens is one of the main reasons why the bone material of small birds is so rich compared to other sites (Byrd 1976, Stahl 1991, 2000). Had a larger mesh size been used at El Azúcar, the recovered avifauna would have containe d only a few individuals and species. Archaeological excavations using fine meshed screens at other sites will reveal how unusual the high abundance of passerines and ground doves at El Azúcar really is.

PAGE 287

271 CHAPTER 5 A PREHISTORIC BIRD ASSEMBLAGE FROM TH E ANDEAN TREE LINE AT LA CHIMBA, NORTHERN ECUADOR Introduction Although early ecologists attempted to explain the distri bution of páramo grasslands in terms of climatological f actors such as low precipitation and soil characteristics, most geographers and botan ists now view páramos and the forest – páramo ecotone as a product of both biol ogical and anthropogeni c factors (Gade 1999, Sarmiento and Frolich 2002). Some palynolog ical work above 3700 m in the southern Ecuadorian Andes shows that an extensive páramo vegetation dominated by ferns and terrestrial bromeliads was present during th e Late Pleistocene and declined at the beginning of the Holocene (Hansen et al. 2003) . Most of the dec line was due to the disappearance of species that require high levels of humid ity. Grasses increased during the Holocene as a response to a drier climat e and a higher incidence of fires (Hansen et al. 2003, Fig. 6 and 9). Fires are attributed to climate rather than the presence of humans although the highlands may have sustained people during th e early Holocene (Ardila y Politis 1989). Evidence from palynological sites in northern Ecuador at ca. 2800 m do not show any indication of páramo, but suggest that ma ny elements of montane forests such as Podocarpus where present in an area almost co mpletely devoid of mature forest vegetation today (Colinvaux et al.1988). A rece nt pollen study from a small marsh south of Quito shows that the arboreal compone nt diminished only within the last 250 14C yr

PAGE 288

272 BP with a simultaneous increase in charcoal. Before that there are some indications that the woody vegetation had moved upslope due to an increase in precip itation (Wille et al. 2002). A general increase of Alnus pollen throughout the nort hern and southern Andes after 2000 14C yr BP may be due to increased deforestation (often accompanied by regeneration of unused areas) and a generally warmer climate. Toward the last 500 years of the Holocene Alnus declined throughout its range, probably from clearing large areas of secondary (and primary) forest (Weng et al. 2004). The picture that emerges from these pollen studies is that by the Mid to Late Holocene hum ans subsisted in a landscape matrix of natural habitats (both forest and páramo) mixed with agri cultural lands, not in a basically modern landscape matrix with very little natural vegetation as Stahl and Athens (2001) suggest. Humans undoubtedly have had an environmental impact since the Pleistocene/Holocene boundary (Gnecco 2003), l eaving us with no clear benchmark as to the exact floristic and faunal attributes of what we now call páramo. In addition, throughout the Pleistocene, high Andean ve getation coexisted with a mammalian fauna that was characterized by large-bodied species that became extinct just prior to the Holocene (Hoffstetter 1986). Thus, Ecuadorian páramos may be a fairly recent vegetation association that did not exist in the absence of humans. The magnitude of the changes due to the peopling of the Andes is not known as palynological studies are not al ways conclusive regarding th e presence of human signals in the pollen record, unless some cultivated sp ecies such as maize can be found. There is a great need to study the asso ciation between patterns in hum an resource exploitation and changes in the local environment. Archaeo logical excavations often result in the

PAGE 289

273 collection of a large number of specimens of vertebrates that can add to the body of evidence regarding the past climate and ve getational change. Unfortunately, the focus tends to be on taxa with large energetic cont ributions to the diet of early people that may have imposed limitations on human carrying capacities (Gnecco 2003). The identifications are often left at the familial or generic level. Although many species of the same genus or family share similar habita t requirements, many do not. Some of the crucial information for paleoecological recons tructions is lost. Becau se of a wealth of information on their habitat preferen ce, birds are potentially excellent paleoenvironmental indicators. In most instances, however, birds are the least studied component of the archaeological faunas. The bones recovered from La Chimba, nor thern Ecuador, are exceptional in terms of their quality and quantity. Of the 40,000 well preserved vertebrate specimens, nearly 300 belong to birds. The object of this study is to identify the birds from this site, and then to use these identifications to make inferences about prehistoric, anthropogenic vegetational change of páramo grasslands. This is one of the first studies of its kind in the Neotropics and the first for the north ern Andes (see Steadman et al. 2003). Methods Study Site La Chimba is located in northeastern Pichincha province of Ecuador at 3180 m elevation at the bottom of a small, fertile valley on the western slopes of the eastern Cordillera (Figure 5-1). The valle y widens toward the west a nd ends in a large plateau on which the towns of Ayora, Olmedo, Pesillo, a nd La Chimba are located. The western end is delimited by Volcán Cusín. Toward the south a narrow valley gently descends and broadens toward the town of Cayambe. Toward the northwest a narrow, hilly valley cut

PAGE 290

274 by deep gorges runs between the eastern Cord illera, Volcán Imbabura, and Volcán Cusín, eventually dropping into the valley of Ibarra. Otavalo can be reached over a low pass between Volcán Imbabura and Volcán Cusí n. To the east the La Chimba valley is delimited by the ascending slopes of the eastern Cordillera and a pass at 4000 m elevation. The landscape is dominated by th e glacier clad Volcán Cayambe about 10 kilometers south-southeast of the archaeological site. The valleys are covered today by agricultural lands, occasional Eucalyptus plantations, and interspersed small towns. To the northwest and north, respectively, the Laguna de San Pablo and the Laguna Yaguarc ocha are the area’s two major wetlands. Along many stretches of the eas tern Cordillera, humid high Andean shrublands form a band separating the dry valley and the páramo grasslands. In the La Chimba valley a direct transition between agricultural lands and páramo can be observed. In some side valleys this vegetation may be as high as 5 m due to a more humid micro-climate, as I observed during a visit in August 2004. Patches of mature Andean forest can be found when ascending the Cordillera to the east surr ounded by páramo grassl ands. Larger tracts of forest cover the eastern slopes as they de scend toward the Amazon basin. In this area the Laguna de San Marcos provides habita t for several aquatic bird species. La Chimba is presently the earliest know n Ecuadorian ceramic site north of Quito. The ceramic styles can be divided into three periods: Early La Chim ba is characterized by “punctate bowl rims (punctate de sign under exterior lips)” at 2640-2390 14C yr BP; Middle La Chimba ceramics are distinguished by “fine line diagonal or crosshatch incision on bowl exterior (often zoned, sometimes with white lip” at 2390-1994 14C yr BP; and Late La Chimba is exemplified by “ bowls and jars with exterior red paint or

PAGE 291

275 lipped bowls with excision at 1994-1700 14C yr BP (Stahl and Athens 2001, Table 1). The site was occupied continuously during this time. A total of five test units, TP-A, TP-1, TP -2, TP-3, and TP-4, were excavated by S. J. Athens and A. J. Osborn in 1972 and 1974. During a third field season in 1989, S. J. Athens excavated three additional units, TP -5, TP-6, and TP-7. Except for the poorly stratified TP-6 all units yielded a high concentration of bones (Athens 1990). The sediments of TP-5 were screened using a 1/ 8 inch mesh, bagged, and later water screened and sorted. Screening for TP-7 was done with a 1/4 inch mesh. Starting at level seven, 36 liter subsamples where wet screened usi ng a 1/8 inch mesh. No information on the excavation procedures of the other pits is currently available. Bone Identification Bird bones were identified using comparativ e material from the Florida Museum of Natural History (FLMNH) using a qualitativ e approach based on the identification of shared and unique traits that characterize each taxon. Morphometric measurements were taken to help in the identification of several large passerine bones. These data were then analyzed using a principle components analys is (PCA) on XLSTAT software (Chapter 2). Once identified, specimens were assigned by their provenience to one of La ChimbaÂ’s three ceramic periods. The minimu m number of individua l (MNI) as well as the number of individual specimens (NISP) wa s recorded for each period. For TP-5 and TP-7, these measures where also calculated for each excavation level and then converted to MNI and NISP per 100 liters of sediments to control for sampling bias. Sediment volumes were taken from Athens (1990).

PAGE 292

276 Figure 5-1. A map of the study ar ea with contour lines at 400 meter intervals; the dark square in the inset shows the locatio n of the study area within Ecuador. All the birds were classified as either bei ng associated with humid forest or forest edge, dry/open habitats, páramo, or wetlands (b oth riverine as well as lacustrine). This classification is a heuristic a pproximation of exis ting habitats given the constraints of sample size. A 2 likelihood ratio test was used to test for an association between presence in a particular type of habitat and period. For each period the percentage of each category was calculated using the MNI and NISP.

PAGE 293

277 Results At least 39 species belonging to 18 families and 11 orders were identified from the archaeological excavations at La Chimba (Tab le 5-1). The most common species was the tinamou Nothoprocta cf. curvirostris which was found from nearly every excavation unit that contained bone material. The tinamou N. pentlandii , the smallest highland tinamou in Ecuador, was the second most common species . The next most abundant species was the thrush Turdus fuscater , present in each period although not in each unit. Owls (Tytonidae and Strigidae) also were recovered in each period. On the other hand, hawks (Accipitridae) and parrots (Psittacidae) o ccurred only in Early and Middle La Chimba deposits. At least six species pr obably are not a local component of the avifauna, i.e., they come from outside a 30 kilometer radius from the site: the White-tailed Kite ( Elanus leucurus ), Purple Gallinule ( Porphyrula martinica ), Scaled Pigeon ( Columba speciosa ), Chestnut-fronted Macaw ( Ara severa ), Black-headed Parrot ( Pionites melanocephala ), and Mealy Parrot ( Amazona farinosa ). The period with the highest number of MN I, NISP, and species is Middle La Chimba (Table 5-1). This is due, in part, to a larger sediment volume sampled for this period (Figure 5-2). The only specimens of guans (Cracidae) and toucanet ( Aulacorhyncus prasinus ) are from this period. Middle La Chimba also contains several individuals of the oropendola Psarocolius angustifrons . For both units TP-5 and TP-7, the concentration of bird bones is also highest in these levels. That the concentration of bones for Middle La Chimba for TP-7 is low comp ared to that of TP-5 (Table 5-2) is probably an artifact of screening mostly with a 1/4 inch mesh in the former.

PAGE 294

278 Table 5-1. Minimum number of individuals (M NI) and number of individual specimens (NISP) for each species at the La Ch imba archaeological site, by cultural periods. MNI NISP FAMILY/Species Early Middle Late Early Middle Late TINAMIDAE Nothoprocta cf. curvirostris 14 17 5 26 33 11 Nothoprocta cf. pentlandii 8 10 7 13 22 15 ANATIDAE Anas andium 0 1 0 0 1 0 Merganetta armata 0 2 1 0 2 1 ACCIPITRIDAE Elanus leucurus 1 0 0 1 0 0 Geranoaetus melanoleucus 1 1 0 4 2 0 Buteo cf. platypterus 1 0 0 1 0 0 FALCONIDAE Phalcoboenus carunculatus 1 0 0 1 0 0 Falco sparverius 0 0 1 0 0 1 CRACIDAE Penelope montagnii 0 1 0 0 1 0 cf. Aburria aburri 0 2 0 0 2 0 cf. Chamaepetes goudotii 0 1 0 0 1 0 RALLIDAE Rallus aequatorialis 0 1 0 0 1 0 Porphyrula martinica 0 1 0 0 1 0 Fulica ardesiaca 1 0 0 1 0 0

PAGE 295

279 Table 5-1. Continued. MNI NISP FAMILY/Species Early Middle Late Early Middle Late SCOLOPACIDAE Tringa melanoleuca 0 0 0 0 0 0 Gallinago sp. 0 0 1 0 0 1 THINOCORIIDAE Attagis gayi 0 1 0 0 1 0 COLUMBIDAE Columba speciosa 0 0 1 0 0 1 Columba cf. Plumbea 0 1 0 0 1 0 Zenaida auriculata 2 4 4 2 8 8 PSITTACIDAE Ara severa 0 1 0 0 1 0 Aratinga cf. Wagleri 0 0 0 0 0 0 cf. Leptosittaca branickii 1 0 0 7 0 0 Pionites melanocephala 1 0 0 6 0 0 Amazona farinosa 0 1 0 0 2 0 Amazona mercenaria 0 1 0 0 1 0 TYTONIDAE Tyto alba 2 2 0 2 3 0 STRIGIDAE Bubo virginianus 2 3 0 2 4 0 Asio cf. stygius 1 3 1 3 3 8 Unidentified sp. 0 1 0 0 3 0

PAGE 296

280 Table 5-1. Continued. MNI NISP FAMILY/Species Early Middle Late Early Middle Late RAMPHASTIDAE Aulacorhynchus prasinus 0 1 0 0 1 0 CORVIDAE 0 0 0 0 0 0 Cyanolyca sp. 1 1 0 1 3 0 TURDIDAE Turdus fuscater 2 6 2 2 7 3 Turdus serranus 0 0 1 0 0 1 TROGLODYTIDAE cf. Henicorhina leucophrys 0 0 1 0 0 1 ICTERIDAE cf. Zarhynchus wagleri 0 1 0 0 1 0 Psarocolius cf. angustifrons 1 7 0 1 16 0 THRAUPIDAE Sericossypha albocristata 0 1 0 0 1 0 Buthraupis montana 0 1 1 0 1 2 TOTAL 28 69 23 56 133 53 The sampling bias generated by the diffe rent amounts of sediments screened (Figure 5-2) is reflected in the peak abundance of NISP and MNI for the Middle La Chimba period (Figure 5-3). An association between cultural peri od and composition of the avifauna was apparent, however. The a ssociation between pe riod and habitat is insignificant for MNI (WilkÂ’s G2 = 4.468, df = 4; p = 0.346), but significant for NISP (WilkÂ’s G2 = 10.426, df = 4; p = 0.034). Although MNI and NISP track each other

PAGE 297

281 closely (see also Grayson 1984) , non-significance for MNI is probably the result of low sample size and hence low statistical power. Sample sizes for NISP are much larger. The significant changes in the avifauna are summar ized as follows: 1) more páramo species are present than expected by chance during the Early period; more forest birds are present during the Middle period; and 3) the Late period is ch aracterized by a higher than expected abundance of open habitat species. Table 5-2. Densities of MNI, NISP, a nd species per 100 liters of sediment. Site/Period MNI/100L NISP/100L Species/100L TP-5 Middle 0.86 1.27 0.26 Early 0.55 0.69 0.41 TP-7 Late 0.24 0.51 0.15 Middle 0.39 0.73 0.21 Early 0.18 0.28 0.11 Systematic List This section summarizes the material rec overed from each species and places each species in a biogeographic context. Seque nce of nomenclature and the current distributional data, unless ot herwise noted, are from Ridge ly and Greenfield (2001a,b). The comparative osteology and justifications for the identifications are found in Chapter 2.

PAGE 298

282 Order TINAMIFORMES Family TINAMIDAE Nothoprocta cf. curvirostris Sclater and Salvin 1873, Curve-billed Tinamou Material.Complete premaxillary, occipital area of skull, six cranial fragment of sterna, complete right and left scapulae, pr oximal right and proximal left, three complete right and three complete left coracoids, proxima l end of left coracoi d, two distal ends of right coracoids and distal end of left coraco id, shaft of right cor acoid, complete right humerus, proximal end of right humerus, near ly complete left humerus, proximal end of two left humeri, distal ends of two left hum eri, complete right ulna, two complete left ulnae, nearly complete left ulna, proximal end of left ulna, distal e nds of two left ulnae, proximal ends of two radii, distal end of right radius, complete right and left carpometacarpi, proximal ends of two carpometacarpi, proximal end of left carpometacarpus, proximal ends of two right an d three left femora, distal ends of two right femora, complete left tibiotarsus, proximal ends of three right tib iotarsi, distal ends of three right and two left tibiotarsi, complete right and three left tarsometatarsi, proximal ends of right and two left tarsometatarsi, and distal end of right tarsometatarsus. The 70 specimens represent at least 36 individuals. Remarks.This rare to locally uncommon tin amou inhabits the tree-line in Ecuador and is usually found in the páramo/forest ecotone throughout the northern and central Andes of Ecuador. The northernmost record comes from Reserva Guanderas in Carchi province where it is rare (C resswell et al. 1999). Before 1994, it had not been observed north of Volcan Pichincha (Robbins et al. 1994 ). Given its apparent rarity nowadays, the large quantity of bones from La Chimba s eems puzzling. Living near the tree line, La

PAGE 299

283 Chimba people encountered this species mo re frequenty than m odern ornithologists visiting this habitat sporadic ally. In addition, since this tinamou requires some woodland, it may be fair to speculate about the presence of large tracts of high Andean forests and an expanded páramo/forest ecotone. A. TP-50 5 150200250300350400 Sediment VolumeMinimum Number of Individuals Middle Early B. TP-7-0.5 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 100200300400500600700800900 Sediment VolumeMinimum Number of Individuals Late Middle Early Figure 5-2. Relationship between sediment volume and NISP (number of individuals specimens) for A) TP-5 and B) TP-7.

PAGE 300

284 Nothoprocta cf. pentlandii (G. R.Gray 1867), Andean Tinamou Material.Crianial fragments of three st erna, proximal ends of two scapulae, three complete right and seven left coracoids, near ly complete left coracoid, proximal end of left coracoid, distal end of left coracoid, n early complete right hume rus, distal ends of two right humeri, two complete right and two left ulnae, distal end of left ulna, proximal ends of two right radii, complete left carpometacarpi, nearly complete right carpometacarpus, proximal end of right carpomet acarpus, proximal ends of four right and two left femora, distal end of right femur, proximal ends of tw o left tibiotarsi, distal ends of three left tibiotarsi, complete right and left tarsometatarsi, and thr ee distal ends of left tarsometatarsi. The 50 specime ns represent 25 individuals. Remarks.This uncommon and local tinamou prefers dry habitats and agricultural areas between 1000-2300 m in southern Ecuador . Currently, the closest population to La Chimba is located more than 350 km to the so uth. This species is absent from central and northern Ecuador despite the availability of suitable, albeit fragmentary habitat. Order ANSERIFORMES Family ANATIDAE Anas andium (Sclater and Salvin 1873), Andean Teal Material.Proximal end and shaft of right humerus. Remarks.Presently this duck is one of the most common wetland species in the Andes. It requires lacustrine environments a nd therefore may not have been caught in the vicinity of La Chimba. The nearest lakes ar e Laguna San Marcos (ca. 10 km away) and Lago Puruanta (ca. 11 km away) on the easte rn slopes of the cord illera, and Lago San

PAGE 301

285 Pablo in the Otavalo valley (ca. 20 km aw ay). Lagunas San Marcos and Puruanta are accessible through a strenuous walk within a day’s walking distance. Merganetta armata Gould 1842, Torrent Duck Material.Shaft of right fe mur, complete right tarsometatarsus, and complete left tarsometatarsus. The three specimens re present a minimum of three individuals. Remarks.The Torrent Duck is a locally co mmon species that occurs at mid to high elevations. It can be found in rapidly flowing streams in páramo as well as montane forest. Río La Chimba provides enough water fo r this species to have occurred in the La Chimba valley in the past. Order ACCIPITRIFORMES Family ACCIPITRIDAE Elanus leucurus Vieillot 1818, White-tailed Kite Material.Right carpometacar pus lacking metacarpal III. Remarks.This is a presently rare to unc ommon species in Ecuador that inhabits open and semi-open agricultural lands in the western and easter n lowlands. Wandering individuals have been seen as high as 3300 m in the northern Andes, near the Colombian border. Presently, his species appears to incr ease its range in res ponse to deforestation and most records are recent. Buteo cf. platypterus (Vieillot 1823), Broad-winged Hawk Material.Right complete ulna. Remarks.In the central Andean valley this uncommon boreal winter resident has recently been observed up to 2,800 meters although it is usually found at lower

PAGE 302

286 elevations. This hawk prefers the forest border. Until modern specimens of B. albigula become available, this identif ication will remain tentative. Geranoaetus melanoleucus (Vieillot 1819), Black-chested Buzzard Eagle Material.Distal end of right tibiotarsus, distal ends of right and left tarsometatarsus, and three complete phalanges of the left foot. Remarks.This large raptor is fairly co mmon at high elevations in páramo or upper Andean forest. It prefers areas with rocky cliffs, such as those found east of the La Chimba valley. Family FALCONIDAE Phalcoboenus carunculatus Des Murs 1853, Carunculated Caracara Material.Complete right humerus. Remarks.Locally a fairly common the páramo highlands, the Carunculated Caracara has a preference for areas with steep cliffs. The highlands around Volcan Cayambe provide ample habitat for this species in the present. Falco sparverius Linnaeus 1758, American Kestrel Material.Left scapula with most of the blade missing. Remarks.This is a common falcon in the mostly dry highlands although deforestation has led to a recent expansion to the humid slopes of the Andes. At the time that the La Chimba site was occupied, this falcon was probably restricted to the drier parts of the arid upper Rio Guayllaba mba watershed and cultivated areas.

PAGE 303

287 Order GALLIFORMES Family CRACIDAE Penelope montagnii (Bonaparte 1856), Andean Guan Material.Shaft of left tarsometatarsus. Remarks.The Andean Guan is the most common cracid in high elevation Andean forest. Although it tolerates deforestation a nd selective cutting, it is never found away from forest. This species usually does not overlap in range with Chamaepetes goudotii , which tends to occur at lower elev ations (Remsen and Cardiff 1990). cf. Chamaepetes goudotii (Lesson 1828), Sickle-winged Guan Material.Nearly complete right femur. Remarks.This is a common guan at lower el evations that tolerates a fair amount of deforestation. The presence at La Chimba is surprising given that this guan usually does not overlap in range with P. montagnii . The nearest current populations can be found almost equidistant from La Chimba on the western slopes of the western Andes and the eastern slopes of the eastern Andes. Whether the bones come from individuals caught in the potentially, formerly more humid Cayamb e valley or from a more distant location cannot be evaluated at this time. No compar ative material was available to confirm the identification; however, the spec imen is very different from Ortalis and Penelope and less so from Aburria , with which it has been previously allied based on morphology (Chapter 2). The only other ge nus possible in Ecuador is Chamaepetes . cf. Aburria aburri (Lesson1828), Wattled Guan Material.Proximal left radius and dist al left tibiotarsus. The two specimens repesent two individuals.

PAGE 304

288 Remarks.This is a secretive and uncomm on large guan that inhabits the forested slopes of the Andes from 1200 to 2100 mete rs. The two bones in the La Chimba assemblage are surprising given its current distribution. Wattled Guans may be very tame where hunting pressures are low (M. P. Tellkam p personal observation). As was the case for C. goudotii , the origin of the birds is unclear, although a warm and formerly forested Cayambe valley may be a potential source. Order GRUIFORMES Family RALLIDAE Rallus aequatorialis Sharpe 1894, Ecuadorian Rail Material.Complete right coracoid. Remarks.This rail is locally fairly common around high elev ation lakes. The requirement for a lacustrine habitat suggests a source area as close as Laguna San Marcos (a locality with no known records, however) or more distant sites such as Lago San Pablo. Other, smaller lakes may also occur in the páramo highlands east of the La Chimba valley. Porphyrula martinica (Linnaeus 1766), Purple Gallinule Material.Distal end of right ulna. Remarks.This species is rare in marshe s and lake shores of the Ecuadorian Andes where its status is not fully understoo d. Most records come from the early 20th century from wetlands that are drained today. Alt hough it may reach elevations of ca. 2800 m, no adequate habitat exists near La Chimba. Th e most likely source was Lago San Pablo in the highlands, the now drained marshes of Quito, or lowlands to the west or east. This

PAGE 305

289 species may be indicative of long-distance trad e in animal remains, or of local wetlands that no longer exist. Fulica ardesiaca Tschudi 1843, Andean Coot Material.Proximal end of right tibiotarsus. Remarks.In Ecuador this coot is comm on in highland lakes with abundant reed along the shores. As for R. aequatorialis the source areas may be Laguna San Marcos (where I observed this species on 03 August 2004) or Lago San Pablo. Order CHARADRIIFORMES Family SCOLOPACIDAE Tringa melanoleuca (Gmelin 1789), Greater Yellowlegs Material.Distal end of right humerus. Remarks.This fairly common Nearctic migr ant occurs in wet habitats such as the shores of lakes, marshes, fields, and rivers. Although it is most comm on on the coast, this sandpiper can regularly been seen at 3300 m elevation at Parque Nacional Cotopaxi. Gallinago sp. (large), Snipe Material.Distal end of left humerus. Remarks.Three species of large snipes occur in the highlands of Ecuador. They are G. nobilis , G. jamesoni , and G. imperialis . Without comparative materials I cannot determine which species the archaeol ogical specimen belongs to. Whereas G. nobilis favors bogs and marshes in páramo grasslands, G. jamesoni prefers shrubby páramo, pastures, and the edge of Polylepis woodlots. Gallinago imperialis is more of a forest bird that occurs in fore st and timberline habitats.

PAGE 306

290 Family THINOCORIDAE Attagis gayi I. Geoffroy Saint-Hilaire and Lesson 1831, Rufous-bellied Seedsnipe Material.Cranial segment of ster num and complete right humerus. Remarks.This is an uncommon species from the nearly bare slopes of several volcanos including Volcán Cayambe. Order COLUMBIFORMES Family COLUMBIDAE Columba speciosa Gmelin 1789, Scaled Pigeon Material.Cranial fragment of ster num and proximal end of right humerus representing two individuals. Remarks.This rare to uncommon pigeon o ccurs near the Andes in eastern Ecuador up to 1100-1200 m elevation. A lthough there are records at 1700 m in southeastern Ecuador, this species is mostly found in the lowlands below 400 m. The Scaled Pigeon is a canopy species of humid forest, frequenting ed ge and secondary forest habitats. In the west, this pigeon occurs as far south as northern Peru during the rainy season. Columba cf. plumbea (Vieillot 1818), Plumbeous Pigeon Material.Proximal end of right humerus. Remarks.This common pigeon usually is encountered in the canopy of mature or secondary montane forest. On the eastern slopes of the Andes some individuals may range as high as 3000 m. Zenaida auriculata (Des Murs 1847), Eared Dove Material.Cranial fragments of two sterna , proximal end of right scapula, complete right and left coracoids, complete humerus, proximal ends of thr ee right and two left

PAGE 307

291 humeri, complete right and left ulnae, near ly complete left ulna, left carpometacarpus lacking metacarpal III, two incomplete synsacra , distal end and shaft of left tibiotarsus, and complete right tarsometatarsus. The 18 bones represent at least 10 individuals. Remarks.A common species of arid habita ts and agricultural lands throughout the inter-Andean valley and the dry southwestern coast of Ecuador, this dove may be found as high as 3,500 meters. Presently in the La Ch imba valley, where the climate is generally dry, this species is at least seasonally absent or rare (M. P. Tellkamp personal observation). A. B. 0 10 20 30 40 50 EarlyMiddleLate La Chimba Period% MNI 0 10 20 30 40 50 60 EarlyMiddleLate La Chimba Period% NISP Figure 5-3. Change of percentages of A) MN I (minimum number of individuals) and B) NISP (number of individuals specimens) of bird bones for the three periods of La Chimba classified by habitat association: páramo vegetation ( ), dry/open habitat ( ), humid forest/forest edge ( ), and wetlands ( ). Order PSITTACIFORMES Family PSITTACIDAE Ara cf. chloroptera G. R. Gray 1859, Red-and-green Macaw Material.Complete left ta rsometatarsus (Villalba 1988).

PAGE 308

292 Remarks.In Ecuador this uncommon macaw only occurs in the humid eastern lowlands, below 400 m. It is a forest bird th at frequently uses na tural and anthropogenic forest edge. Ara severa (Linnaeus 1766), Cestnut-fronted Macaw Material.Proximal end of right tarsometatarsus. Remarks.This small macaw is uncommon to fairly common in the western and eastern lowlands. Although usually prefer fo rest below 600 m, ther e are severa l records up to 1800 m in the northeastern Andes. This macaw favors forest edge as well as clearings with scattered trees. Leptosittaca branickii Berlepsch and Stolzmann 1894, Golden-plumed Parakeet Material.Rostral end (i ncluding the frontals, nasals , and lachrymals) of the cranium. Remarks.This is a rare and nomadic parakeet whose numbers have declined recently due to deforestation. Its elevati onal range encompasses the altitudinal belt between 2400 -3400 m of the outer slopes of th e cordilleras as well as the inter-Andean valley. There are a few records from norther n Ecuador, both north of La Chimba. This parakeet was probably caught near the prehistoric human settlement. Pionites melanocephala (Linnaeus 1758), Black-headed Parrot Materials.Mandibular symphysis, complete ri ght and left coracoids, complete left humerus, complete right ulna, complete left carpometacarpus. Remarks.This common species of the eastern lowlands below 400 m prefers both intact forest and forest borders.

PAGE 309

293 Amazona farinosa (Boddaert 1783), Mealy Amazon Material.Abraded left humeru s with most of the shaft and distal end of left femur. Remarks.This large parrot usually occu rs in the lowlands of northwestern and eastern Ecuador. Some records indicate that this species occasi onally moves to 1200 m along the Andes. The Mealy Amazon prefers edges of humid forest and is tolerant of low levels of deforestation. Amazona cf. mercenaria (Tschudi 1844), Scaly-naped Amazon Material.Whole le ft carpometacarpus. Remarks.This parrot is an uncommon sp ecies of lower montane to high Andean forests on both slopes of the Andes. A lthough usually found up to 2600 m, there are several records from areas above 3000 m. Order STRIGIFORMES Family TYTONIDAE Tyto alba (Scopoli 1769), Barn Owl Material.Cranium, shaft of right humerus, distal end of right ulna, distal end of right radius, and proximal end of right carpometacarpus. The fi ve specimens represent at least four individuals. Remarks.As its common English name suggests, this is a locally common species does well in agricultural areas and other open habitats. Altho ugh in Ecuador this species occurs mainly below 2,000 meters, in Colomb ia it is found up to 3,000 meters (Hilty and Brown 1986), and Stotz et al. ( 1996) put its upper limit at 4,000 meters. It is likely that the source area for the bones was in the va lleys below La Chimba or the La Chimba

PAGE 310

294 valley itself, rather than the eastern slopes of the Andes where this species is currently absent. Family STRIGIDAE Bubo virginianus (J. F. Gmelin 1788), Great Horned Owl Material.Occipital area of cranium, co mplete right radius, two pedal phalanges, and two ungual phalanges. The six specimens represent at least five individuals. Remarks.A locally uncommon species, th is owl inhabits semi-open woodlands and páramo as well as somewhat disturbed ag ricultural areas. Its elevational range (32004500 meters) places it well with in the range of prehistoric hunters from the La Chimba settlement. Asio cf. stygius (Wagler 1832), Short-eared Owl Material.Proximal end of right ulna, comp lete left ulna, complete left radius, nearly complete carpometacarpus (lacking processus extensorius), complete left femur, complete left tarsometatarsus, two proximal pedal phalanges, intermediate pedal phalanx, and four ungual phalanges. The 13 specimens represent at least four individuals. Remarks.In Ecuador this is an unco mmon owl of open and forested habitats. Although the La Chimba bones show a very close fit to A. flammeus , they are sufficiently different for me to ypothesize that they belong to a different species. Asio stygius is the most likely species given distributi on and osteology (Chapter 2). Both A. stygius and flammeus may reach above 3100 m. A . flammeus is an open habitat/páramo species, less dependent on forest habitat than A. stygius .

PAGE 311

295 Order PICIFORMES Family RAMPHASTIDAE Aulacorhynchus prasinus (Gould 1834), Emerald Toucanet Material.Complete left carpometacarpus. Remarks.A locally fairly common toucan et that occurs to up to 3,250 meters below Oyacachi about 30 kilometers south, south east of La Chimba. Observations at that altitude are rare, however, and most records come from forest between 1500-2600 m elevation. Order PASSERIFORMES Family CORVIDAE Cyanolyca sp., Jay Material.-Complete left femur, the proximal ends of right and left tibiotarsi, and the distal end of left tibiotarsus. The four sp ecimens represent at least two individuals. Remarks.Two similar species of jays can be found at high elevations in Ecuador. Cyanolyca turcosa is common between 2000-3000 m, ra nging as high as 3500 m. This species has also been recorded in th e inter-Andean valley. On the other hand, C. armillata is rare and very local on the eastern slopes of the Andes (Center for Research on the Cultural and Biological Diversity of Andean Rainforests 2000). Although both species are plausible for occu rring at or near La Chim ba, based on current knowledge about the distribution of the two species, C. turcosa is much more likely to have been hunted there in prehistoric times.

PAGE 312

296 Family TURDIDAE Turdus fuscater d’Orbigny and Lafresnaye 1837, Great Thrush Material.Two complete right coracoids, complete right humerus, proximal and distal ends of right humeri, shaft of le ft humerus (including portions of the double pneumatic fossa), complete right ulna, comp lete carpometacarpus, di stal end of right femur, proximal end of left femur, distal e nd of left tibiotarsus, distal end of right tarsometatarsus, complete left tarsometat arsus, and shaft of left tarsometatarsus (including the foramen vasculare distale). The 15 specimens represent at least 10 individuals. Remarks.This is a common thrush in forest border, secondary vegetation, agricultural lands, and the forest/páramo ecot one. Presently it occu rs at extremely high densities in the La Chimba valley (M. P. Tellkamp personal observation). Getting accustomed rapidly to human presence, the Great Thrush was probably hunted in the vicinity of the prehistoric settlement. Turdus serranus Tschudi 1844, Glossy-black Thrush Material.Complete right ulna. Remarks.This common thrush is more restricted to forest than its congener T. fuscater . Although it is a bird of the forest canopy and midstory, it can be observed near the ground in edge and pastures, and is es pecially common in variably fragmented landscapes (M. P. Tellkamp personal observation). Most common between 1500 and 2800 meters, this thrush occurs as high as 3100-3350 m on the eastern slopes of the eastern Andes.

PAGE 313

297 Family TROGLODYTIDAE cf. Henicorhina leucophrys (Tschudii 1844), Grey-breasted Wood Wren Material.Distal end and shaft of right tibiotarsus. Remarks.This common understory bird occurs between 1500 and 3000 meters. Not having been recorded on the inter-Andean slopes of the Ecuadorian Andes, the source area of this wren is probably th e eastern slopes of the eastern Andes. Henicorhina leucophrys may have had an inter-Andean population, however, when forests were more widely spread (Hidalgo Nistri 1998, Chapter 1). Family THRAUPIDAE Sericossypha albocristata (Lafresnaye 1843), White-capped Tanager Material.Proximal end of right ulna. Remarks.Being locally uncommon, this tanager occurs be tween 1750-3000 m in forest and forest edge of the eastern slopes of the Andes. Most r ecent Ecuadorian records are from the south, but this species also o ccurs in northern Ecuador. The source area for this bird probably was the forested landscap e to the east of La Chimba although a local prehistoric population cannot be ruled out. Buthraupis montana (dÂ’Orbigny and Lafresnaye 1837), Hooded Mountain-Tanager Material.Complete rostrum with att ached nasals, complete right ulna, and complete right carpometacarpus. Remarks.A common tanager found in forest with less preference for edge, and adjacent clearings between 1900-3200 m. It prob ably occurred in the vicinity of the La Chimba settlement as its wide use of habita ts makes a past inter-Andean population very likely (Robbins et al. 1994).

PAGE 314

298 Family ICTERIDAE cf. Zarhynchus wagleri (G. R Gray 1845), Chestnut-headed Oropendola Material.Distal end of right femur. Remarks.Usually occurring below 700 m in northwestern Ecuador, this rare to uncommon oropendola was probably transported by people from the western lowlands of Esmeraldas or Manabí provinces, as th ere are no known populati ons in the eastern lowlands. The identification is somewhat tentative due to the lack of additional comparative material. Psarocolius cf. angustifrons (Spix 1824), Russet-backed Oropendola Material.Cranial fragment of sternum, complete right and le ft coracoids, two proximal ends of right humeri, two distal e nds of right humeri, complete left humerus, two proximal ends of left humeri, complete left ulna, complete left carpometacarpus, proximal end of right carpometacarpus, distal end of right femur, distal end of right tibiotarsus, proximal end of left tarsometatarsus, and distal end of left tarsometatarsus. The 19 specimens represent at least 8 individuals. Remarks.This species of oropendola doe s presently not occur above 2100 m today on either western and eastern sl opes of the Andes. Whereas it is a possibility that these birds were obtained on the eastern slopes (see account of Ara sp.), it is also possible that P. angustifrons may have occurred in humid forests just below Cayambe in the interAndean valley (see account of Aulacorhynchus prasinus ). Discussion It seems clear that open grasslands wh ere present near La Chimba 2600-1700 yr BP. The high abundance of Nothoprocta cf. curvirostris would even suggest a presence of páramo vegetation with some shrub cover, agricultural areas, and/or dry temperate

PAGE 315

299 vegetation. In northern Ecuador this species has recently been recorded at Guandera Biological Reserve (0°36’N, 77°41’; elevati on 3000-4100 m) in agricu ltural lands with abundant hedges and shrubs, but not in the bunchgrass dominated páramo (Cresswell et al. 1999). At Oyacachi, south of La Chimba, this tinamou is absent possibly due to the higher humidity on the eastern than on the western slopes of the eastern cordillera (Robbins et al. 1994, Center for Research on the Cultural and Biological Diversity of Andean Rainforests 2000). I suspect that th e rarity of the Curve-billed Tinamou today may reflect the absence of mixed páramo ve getation consisting of bunch grasses and high elevation shrubs (as presently found in the s outhern province of Loja). Over time the relative abundance of bones from this species decreases and that of N. pentlandii increases, suggesting degradation of pára mo vegetation and possibly drier conditions toward the Late period. Interestingly, N. pentlandii does not occur in northern Ecuador anymore, possibly the result of nearly comp lete conversion of natural habitats to agricultural lands and/or dry scrub throughout the north ern inter-Andean valleys. The composition of the avifauna changes ove r time. Forest and páramo associated birds decrease whereas species that pref er open areas increase (Figure 5-3), although there is only partial statistical support. The bi rd bones from La Chimba not only represent a local tree line and páramo fauna, but also include an increasing component of dry temperate and humid subtropical forests a nd some wetland species. This unusual mix may result from several cultural and natural factors. Most species in the sample are typical of high Andean forest and páramo (Robbins et al. 1994, Creswell et al. 1999, Center for Research on the Cultural and Biological Diversity of Andean Rainforests 2000). Th e source areas for the páramo birds are

PAGE 316

300 probably the grasslands above the La Chimba valley. Many of these birds are also found in agricultural areas and forest borders and may therefore be local in origin. Representatives of this category are Nothoprocta curvirostris , Geranoaetus melanoleucus , Phalcoboenus carunculatus , and Attagis gayi . The forest avifauna is more difficult to a ssign to a particular source area. The interAndean slopes of the eastern cordillera ma y have had sufficient humid vegetation cover for montane birds to be present locally. That the areas surrounding La Chimba were mostly forested and close to the tree line is suggested by the analysis of the writings of early chroniclers (Hidalgo Nistri 1998, Chapter 1). In addition, pockets of humid forest in some small, inaccessible latera l valleys feeding into the La Chimba valley, such as in Quebrada Turucucho (Figure 5-4), suggest that climatic conditions may have been favorable for forests in the past as well. Al though the presence of fo rest may be due to favorable microclimatic conditions, deforest ation of these valle ys was probably less severe due to the steepness of the slopes and the inaccessibility of the area; only fairly recently has a road into the valley been co mpleted. In addition, a band of secondary scrub can be seen along most of the eastern cordille ra just below the páramo grasslands. At a distance these look more like mesic than xeric fo rmations. Closer analysis is required, but together these lines of thought suggest that se veral forest species in the assemblage were probably local in origin, including Penelope montagnii , Columba cf. plumbea , cf. Leptosittaca branickii , Amazona cf. mercenaria , Bubo virginianus , Cyanolyca sp., Turdus fuscater , and Buthraupis montana . With the exception of C. plumbea and A. mercenaria , all of these species have been found at Cerro Mongus (0°27’N, 77°52’; elevation 3400-3650 m) to the no rth of La Chimba on the western slopes of the eastern

PAGE 317

301 cordillera (Robbins et al. 1994). The species of Cyanolyca at Cerro Mongus was C. turcosa . Many of the birds from this category c ould have been collected on the eastern slopes of the eastern Andes. The presence of a macaw ( Ara severa), two parrots, a kite, and a pigeon most commonly encountered in the Andean foothills or lowlands suggests either contact with eastern population, the pres ence of a trade route to the lowlands, or possibly multiple-day hunting expeditions into this area. Trading probably included Cotocollao (now northern suburbs of Quito, Figure 5-1), where Ara cf. chloropterus was found in archaeological contexts. This macaw only occurs in the eastern lowlands of Ecuador. This study provides the first eviden ce for extensive tradi ng of wild birds in prehistoric Ecuador. Figure 5-4. A side valley of the La Chimba valley, showing the dark green remnants of humid forest as indicated by the arrows. The steep terrain and freque nt eruptions of Volcán Ca yambe, which during the La Chimba occupation were directed toward th e east-northeast of the volcano (Samaniego et al. 1998), make permanent habitation in th e subtropical vegetation belt unlikely. The recovery of a Saimiri squirrel monkey and some Cosa nga-style pottery, on the other

PAGE 318

302 hand, suggest some exchange with the eastern lowlands (Stahl and Athens 2001). Other species from the eastern slopes of the Andes are Aulacorhynchus prasinus , Sericossypha albocristata , and Psarocolius cf. angustifrons . As suggested by writings of early chroniclers, parts of the Cayambe valley contained patches of humid forest ( monte ) and most of the western slopes of the eastern Andes in northern Ecuador was forested (Hidal go Nistri 1998). In addition, the flanks of the Mojanda massive may have been clad in dense forest creating an east-west corridor through the inter-Andean valley. Thus, presen tly dry areas may have provided ample habitat for tropical montane birds between the two cordilleras. The deep valley cut by the Rio Guayllabamba may have provided a corridor for dispersal of we stern birds into the inter-Andean valley and onto the valley of Ca yambe (see Chapter 1). If this hypothesis is correct, the biogeographical implications are si milar to those for southwestern Andes of Ecuador where western races and species r each the upper western slopes of the eastern Andes, at least up to 2,700 meters (Best and Kessler 1995). Any bird s caught to the west, below La Chimba should be of the western (s ub-) species. This scen ario is a distinct possibility for Psarocolius cf. angustifrons . If this line of reasoning were true, then A. haematopygus should have been encountered below La Chimba. Since the immediate suroundings of La Chimba are a bove the upper altitudinal limit of A. prasinus , however, the carpometacarpus of this species was probably obtained on the east slope. Several bird species in the assemblage are associated with the dry vegetation characteristic of much of the modern la ndscapes of the Ecuadorian Andes. These are Nothoprocta pentlandii , Falco sparverius , Tyto alba , and Zenaida auriculata . All three species can also be found currently in recently deforested humid montane and

PAGE 319

303 premontane forest areas. Although the exact orig in of these specimens cannot be stated with certainty, their re gular occurrence in the bone assemblage s uggests at least a nearby dry temperate vegetation formation and/or considerable prehis toric deforestation. Most of the wetland-associated species we re probably not local with the exception of Merganetta armata . The Torrent Duck can be found in fast flowing streams such as Río La Chimba. The other aquatic species ( Anas andium , Rallus aequatorialis , Porphyrula martinica , and Fulica ardesiaca ) could have been collected at Lagunas San Marcos and Puruanta in the high Andes or La guna San Pablo at intermediate elevations. Given the current distribution of R. aequatoriali s and P. martinica , Laguna San Pablo is more likely for these rallids ( P. martinica may also have been collected in the western lowlands). The presence of several species fr om the lowlands or foothills ( Elanus leucurus , Columba speciosa , Ara severa , Amazona farinosa , Zarhynchus wagleri , Psarocolius angustifrons ) as well as distant lakes ( Porphyrula martinica ) shows that bird trade was common during prehistoric Ecuador. Bird trade was often reported by Spanish chroniclers about 500 years a go as well (Newson 1995). Extens ive trade of birds, many of which may have been alive at the time of transport, needs to be considered a potential factor determining bird species di stributions throughout the Holocene. The large geographic region that served as source area for the avifauna (this study) as well as mammals (Stahl and Athens 2001) in the bone assemblage is as surprising as the location of the La Chimba settlement. Being located too high for productive corn cultivation, having harsh climat e compared to the valleys be low, and lacking two of the most common highland domestic ates (llama and cuy), this site seems highly unusual.

PAGE 320

304 Several hypotheses may explain the founding of a settlement on the fringes of the most productive areas in northern Ecuador. 1) Human population growth pushed some populations into marginal habitat (Athens 1990); 2) relatively dry conditions between 4,000 – 2,500 14C yr BP forced people to colonize more humid areas in the mountains (Colinvaux et al. 1988, Mayewski 2004); 3) the settlement was a food procurement site in the form of meat for the inter-Andean valle ys (as judged by the vast amount of bone at this site [Stahl and Athens 2001]); or 4) La Chimba was a stopover site for people visiting satellite areas in the east that may have been used to cultivate subtropical plants (see Gade 1999). At this point it is difficult to determine the most likely explanation. However, it needs to be pointed out that c onceptualization of th ese competing hypotheses would not have been possible without species le vel identifications of faunal materials. In addition, the identifications presented in this study allow for the conclusion that humans probably did transport bird remains or live bi rds from medium to l ong distances. Most of the source areas appear to be at lower elevatio ns in the east. Even though marine shells in the archaeological record are strong evidence for trade with the western lowlands, no bird species in the assemblage is une quivocally from that area. More archaeological work in the study area is required to shed light on the significance of La Chimba in the contex t of the Formative period of Ecuador. Palynological studies on Ecuador ian lakes are also needed for more precise paleoclimatic reconstructions. Finally, natural sites of bone deposition should also be searched for in the northern Andes of Ecuador to find a dditional data concerning the late Holocene biogeography of birds in this region.

PAGE 321

305 CHAPTER 6 PREHISTORIC EXPLOITATION AND BIOG EOGRAPHY OF BIRDS IN ANDEAN AND COASTAL ECUADOR Introduction Because so few paleoecological studies exist for the western lowlands and the northern Andes of Ecuador (Chapter 1), th e large number of archaeological sites spanning all cultural periods throughout th e Holocene (ca. 10,000 14C BP to present) contributes disproportionately to our knowledge about past environments and climates as well as the patterns in prehistoric use of na tural resources. Systematic archaeobotanical and zooarchaeological studies are only a recen t addition to many archaeological research programs, producing novel result s on the origin and spread of domesticated plants and animals (Piperno and Pearsall 1998, Piperno and Stothert 2003, Pearsall 2004, Stahl 2004). Despite the potential for paleoenvironmen tal reconstruction, only a small number of zooarchaeological studies have had the taxonomic rigor to make sound interpretations (Stahl 2004). Species-level iden tifications of bones are much more informative than those to family or genus. For instance, whereas the Family Psittacidae (macaws, parakeets, parrots) is nearly cosmopolitan, its members differ widely in terms of habitat selection, distribution, and ecology (Stotz et al. 1996). Paleontological and zooarchaeological inve stigations from the western lowlands and northern Ecuadorian Andes have focuse d on Late Pleistocene birds and megamammals from Santa Elena Peninsula (Edmund 1965, Campbell 1982) as well as mammals from various sites in the A ndes (Hoffstetter 1986, Coltorti et al. 1998).

PAGE 322

306 Zooarchaeological studies for several coas tal and Andean sites have focused on the nutritional contribution of mollusks, fish, and deer, or the origin and spread of domesticated animals (Stahl 2004). Few of th ese include information on birds, which in most cases are identified only to family level (Stahl 2003). The analyses for six archaeofaunas presented in this dissertation (Chapters 3, 4, and 5) ar e a big step toward filling this taxonomic void. They are the fi rst studies of this kind for Ecuador, contributing to our present knowledge a bout prehistoric bird exploitation and biogeography for the Neotropics (Steadman et al. 2003). I will often refer to the six sites analyzed collectively as the Ecuadorian sites. Five of the sites are from the lowlands and one from the highlands (Table 6-1, Figure 6-1). Table 6-1. Summary statistics for the six archaeo logical sites studied in Chapters 3 to 5, providing age range, cultur al period, elevation, pote ntial current vegetation (after Sierra et al. 1999b ), and key references. Site Age Range (14C yr BP) Cultural Period Elevation (m) Potential Current Vegetation References OGSE-80 10,000–6600 Vegas 30 Lowland dry brush Stothert 1988 El Azucar 2030–1670 Guangala 40 Lowland dry brush/Savanna Reitz and Masucci 2004 Loma Alta ca. 3000–2300 Chorrera 60 Lowland deciduous forest Stahl 1991 La Ponga ca. 3200–2800 Machalilla 60 Lowland deciduous forest Stahl 1991 Valdivia Village 2370–2030 Guangala 5 Lowland dry brush/Savanna Stothert 1997, Ubelaker 1997 La Chimba 2640–1700 La Chimba 3200 Evergreen high montane forest of the western Andes Stahl and Athens 2001 Although much remains to be learned about their habitat preferences, birds are relatively well studied compar ed to other Neotropical ta xonomic groups. They thus lend themselves to paleoenvironmental analyses. Several caveats need to be addressed to

PAGE 323

307 avoid misinterpretations of zooarchaeol ogical data. As with any bone deposit, the taphonomy needs to be evaluated to detect biases in the composition of avian assemblages. In zooarchaeological studies the taphonomy can be particularly complex. Humans as predators select their prey base d on not only biological necessity, but also cultural imperatives. Trade of live animals and animal parts is important in many cultures (Stahl and Norton 1987, Reitz a nd Masucci 2004). In addition, the bone assemblage in a midden can include species that were either hunt ed or that were attracted to the refuse around human settlements (Stahl 2000, 2003). Met hodological limitations in the recovery of the specimens, such as screening of sedi ments with different mesh sizes, introduce an additional layer of bias (Quitmyer 2004). For instance, coarse-meshed screens are inadequate to retrieve the bones of small birds. Despite these limitations, a critical evaluation of each species will allow a prelim inary assessment of prehistoric patterns of bird exploitation and distributions. For now, any uncertainties and alternative hypotheses need to be highlighted to point to new directions for future research. The purpose of this review is to place th e recent zooarchaeological studies into the broader context of resource utilization pattern s and avian bioge ography during the Holocene in Ecuador. I will also provide a fi rst outline of a historical perspective for present conservation efforts in the western lowlands and the northern Andes of Ecuador. Throughout the review I will compare and contra st the results from the Ecuadorian sites with those obtained for Paso de la Amada, Chiapas, Mexico (Steadman et al. 2003), and Chiripa, Lago Titicaca, Bolivia (Kent et al. 199 9), the only other Neotropical, continental archaeological sites where dive rse sets of bird bones have been studied in detail.

PAGE 324

308 Figure 6-1. Map showing the locatio n of all archaeological sites ( ) and major cities and towns ( ) mentioned in the text. Inset show s enlarged Santa Elena Peninsula. See Chapters 3, 4, and 5 for site specific maps .

PAGE 325

309 Prehistoric Exploitation of Birds in Ecuador At all Ecuadorian archaeological sites analyz ed so far (Table 6-1), birds are a minor component of the bone assemblages (Byr d 1976, Chase 1988, Stothert 1997, Stahl and Athens 2001, Stothert et al. 2003, Reitz and Masucci 2004, Stahl 2004). Sites OGSE-80 (Site 80), OGSE-MA-172 (Valdivia Village), La Ponga, Loma Alta, and El Azúcar are located within 25 km of the P acific Ocean but belong to different cultural periods (Table 6-1, Figures 1-1 and 6-1), and are dominate d by mangrove-associated mollusks, fish, and mammals. The archaeofauna of the high Andean site of La Chimba consists mostly of mammals, with rabbit and deer being by far the most important components. The low number of bird bones may be due to their fragile nature (generally hollow and thin-walled), which makes them susceptibl e to fracture under the pressure exerted by the surrounding sediments. I suspect, however, that the relative scarcity of bird material is not just due to taphonomic factors. At La Chimba, hollow rabbit bones are extremely abundant and bird bones also are well preser ved. Under these conditions many small bird bones probably were preserved as well, but mo st of these were almost certainly lost through the 1/4 inch mesh of sediment sc reens. This supposition is supported by the material recovered from El Azúcar with 1/8 inch screens. Abrasion of bones and breakage was more severe at this dry, coas tal site than at La Chimba, but the vast majority of the nearly 700 avian specimens belonged to small doves and songbirds. In general, the bones of large birds preserve be tter than those of sma ll birds, yet at two Ecuadorian sites they make up only a small fract ion of bird faunas (Chapters 3 and 4). In contrast, large bird bones are very common at sites with intense prehistoric exploitation of marine birds in coastal Peru (Chapter 3). Obviously, non-taphonomic factors must play an important role at these sites.

PAGE 326

310 Studies in resource utilization by modern but traditional Amerindian groups from the Neotropical lowlands have produced s uggestive results that can help in the interpretation of zooarchaeological data. Comb ined hunting data from 15 tribes show that mammals are numerically the most important taxonomic group (Redford and Robinson 1987, Table 6-2). In five tribes, however, bi rds were more important. The most hunted families of birds, in decreasing order of importance, were guans (Cracidae), toucans (Ramphastidae), parrots (Psittacidae), tinam ous (Tinamidae), a group consisting of unidentified birds, trumpeters (Psophiidae ), and pigeons (Columbidae; Redford and Robinson 1987). The order of importanceamong these families was highly variable and depended biological factors such as prey av ailability and prey density, and cultural factors such as hunting technique, taboos , hunting regulation, and degree of human habitat alteration near settlements. I consid er their framework, based on a comparison of 15 different human groups and data -rich studies, especially useful for the interpretation of the Ecuadorian zooachaeologica l assemblages of birds. Prey availability Most of the species found at the Ecuador ian archaeological sites can be found in their immediate surroundings today (Ridgely and Greenfi eld 2001a,b; M. P. Tellkamp personal observation). The number of species obtained through trad e or during extended hunting trips (> 20 km; see below), referre d to by Redford and Robinson (1987, p. 664) as “trekking”, is comparatively low. On the other hand, some species that undoubtedly were available are not represented in the archaeofauna. Among these are marine birds near Site 80 (Ridgely and Greenfield 2001a,b; M. P. Tellkamp personal observations; Chapter 3) and condors ( Vultur gryphus ) in the páramos east of La Chimba. (I saw a juvenile condor ca. 10 km east of La Chimba on 04 August 2004; Chapter 5). Both

PAGE 327

311 marine birds and condors are found elsewhere in Peruvian and Ecuadorian archaeological contexts (Chapter 3 and 4; P. W. Stahl personal comment). Prey density Except for Loma Alta, the bird assemblages are dominated by one or a few species. At sites OGSE-80 and El Azúcar, Zenaida auriculata is a dominant component, with nearly 47% and 17% of the sp ecimens, respectively. Smaller Columbina doves are also common at both sites. Passerines are the most abundant group (54% of all specimens) at El Azucar. At La Chimba, the tinamou Nothoprocta cf. curvirostris makes up 49% of all specimens. The most common species in the archaeofauna are thus small to mediumsized (30–300 g), open-country birds with th e ability to thrive in anthropogenic landscapes (Stotz et al. 1996, Ridgely and Greenfield 2001a; Table 6-1). Opportunistic hunting of these species near prehistoric sett lements probably reflects a trade-off between availability/ease of hunting a nd nutritional value. Large game birds, such as guans, may be preferred because they provide a large amount of protein (Hames and Vickers 1984, Redford and Robinson 1987), but their associ ation with the fore st canopy makes them relatively difficult to hunt compared to pigeons that forage in corn fields near people’s homes. Relatively low caloric content is offs et by the low amount of energy required to hunt a large number of i ndividuals (Chapter 3). Hunting Technique Despite their importance for human prey choice, little is known about hunting techniques employed by Holocene people in Ec uador. Projectiles, snares, and traps may have been used to hunt medium-sized a nd large mammals (Byrd 1976). Conceivably, snares could also have been efficient for the capture of small mammals and terrestrial birds. The Matses in northeaste rn Peru are skilled at killi ng prey with hand-held weapons

PAGE 328

312 such as clubs and digging sticks (Redford and Robinson 1987), a technique available to prehistoric people as well. The predominance of terrestrial and understory birds as well as species readily found in low brushy vegetati on suggests that hunting techniques were not very effective at procuring birds from th e forest canopy (Table 6-1). In Chapter 3, I suggest that Vegas people may have increas ed hunting efficiency by using a (semi-) domesticated fox (Wing 1988) and in Chapter 4, I speculate that woven nets were used to catch gregarious birds foraging low in agri cultural fields. The a bundance of tinamous at La Chimba, in turn, might be explained as a by-product of catching rabbits with snares (Chapter 5). Hopefully, future archaeological work will recover some artifactual evidence of hunting techniques, although it is likely th at most weapons were made of wood and other organic materials that seldom preserve in archaeological sites except in extremely arid areas such as coastal Peru. Taboos Food taboos are common in Amerindian so cieties (Redford and Robinson 1987). It is nearly impossible to say anything about th e importance of taboos in prehistoric times. The absence of condors at La Chimba is ha rd to explain, given suitable nearby habitat and the abundance of this species at so me other Ecuadorian archaeological sites (Chaullyabamba, P. W. Stah l personal comment). Perhap s taboos inhibited people from hunting this huge car rion-eating bird. Hunting Regulation Around permanent settlements, preferred game species usually decrease in abundance through time. Three common mechan isms to deal with dwindling game numbers are 1) “hunting zone rotation”, 2) satellite or “o utlier” settlements, and 3) extended hunting trips, or trekking (Redford and Robinson 1987, pp. 663-664). An

PAGE 329

313 important distinction between the three mechanisms concer ns the deposition of bones at archaeological sites. Whereas in the hunting zo ne rotation approach most game is brought back to the main settlement, the use of satel lite camps and trekking lead to at least some deposition of bone far away from the main mi dden. The species least likely to be brought back to the main settlement are those used as provision during the hunting trips. If birds were hunted opportunistically and used as pr ovision, many individuals can be expected to have been deposited elsewhere. The specimens from the midden thus may not accurately reflect the whole importance of birds for prehis toric people. It is im portant to investigate potential satellite settlements or temporary ca mps. More than 30 small sites that could be of either type have been found for the Ea rly Holocene Vegas cultu re, but no bird bones have been found or identified (Stothert et al . 2003). A detailed faunis tic analysis at the species level for these and similar sites would enhance our ability to interpret patterns in prehistoric bird exploitation. Human Habitat Alteration On the American continents, humans have been modifying the environment ever since their arrival during the Late Pleistoc ene (Martin and Steadman 1999). With the onset of agriculture, conversi on of the original land cover to agri cultural lands has accelerated, creating in many instances co mplex landscape matrices with forest fragments, agricultural fields, and sec ondary vegetation (Stahl 1991, 2000, Veintimilla 2000). Many species of birds and mammals are attracted to agricultu ral lands, secondary growth, and forest edges (S totz et al. 1996, Ridgely and Greenfield 2001a). Some Amazonian tribes even establish fields mainly to attract game species (Redford and Robinson 1987). The increased availability a nd density near human habitations makes these species especially attr active prey (see above). Some of the species attracted to

PAGE 330

314 agricultural and modified lands are gr ound-foraging doves, tinamous, and several frugivores, including parrots, toucans, a nd tanagers. Being highly desirable for Amazonian (Redford and Robinson 1987, Vick ers 1987) and other traditional peoples (Steadman 1997), frugivores are well represente d at La Chimba and the Valdivia Valley sites. Site 80 and El Azúcar contain few species from this guild, which in the arid Santa Elena Peninsula is represente d by a parrotlet, parakeet, and several small songirds. The dearth of frugivores at El Azúcar is proba bly related to difficulty of identifying small tanagers, although some bones have been tent atively assigned to the thraupid genera Tangara and Thraupis . In conclusion, the low abundance of bird s at the Ecuadorian sites suggests that birds were not the primary object of hunting or collecting trips. Most birds in the middens probably came from the immediate vicinity of each site. The dearth of forest-associated birds can be explained, at least in part, by c onsumption of birds away from settlements as provisions during hunting trips. The few forest birds recovere d are colorful species that might have been hunted for their plumage, rare guans that may have persisted close to the settlements, and individuals that we re traded from distant areas. Comparison with other Neotropical Sites The archaeofaunas from two other Neotropical sites, Paso de la Amada, Chiapas, Mexico, and Chiripa, Lake Titicaca, Bolivia, su ggest similar trends with regard to the presence of species readily available and abundant in the surroundings, and a few species from more distant source areas (Kent et al. 1999, Stead man et al. 2003). The avian archaeofaunas from several Peruvian sites are also similar in that locally common species that are relatively easily hunted dominate the bird assemblage, but are quite different in

PAGE 331

315 that birds, in particular s eabirds, are a major component of the archaeofauna (Pozorski 1983, Matthiesen 1988, deFrance et al . 2001, deFrance 2005; Chapter 3). The Lake Titicaca Site has fewer species a nd a lower diversity th an the sites at La Chimba, El Azúcar, and Site 80, but future an alyses will increase the number of species (Kent et al. 1999, D. W. Steadman persona l comment). Sixteen of 27 species are associated with either aquatic or wetland habitats. Aquatic species make up 82% of all bone specimens and are also numerically the most important group of birds. Most of these dive for food and thus can be caught in fishing nets as a bycatch. Bird hunting in this case might be the unintended by-product of the fishing activity or another case of specialized extractive activit y (deFrance 2005). Despite high present densities, highly edible tinamous are only a small component of the archaeofauna. With the great availability of camelids and fish apart from agricultural products, there probably was not much incentive to hunt these birds. The diversity of species at El Paso de la Amada is extremely high compared to that at the Ecuadorian sites (Fi gure 6-2; Steadman et al. 2003). I calculated Fisher’s alpha diversity index using the EstimateS software (Colwell 2005) because of its discriminant ability and low sensitivity to unequal samp le sizes (Magurran 1988). Not only does the site have very high species richness, but it al so lacks a truly dominating species. Most of the species are mediumto large-sized birds. Given that some of the Ecuadorian sites have very low sample sizes, diversity measur es are not directly comparable. I therefore constructed species accumulation curves us ing Ecosim (Gotelli and Entsminger 2004) based on the number of individual specimens (NISP) to improve the comparison among

PAGE 332

316 sites. Due to a lack of data for Chiripa, I was not able to include this site in the comparison. A steeply increasing curve repr esents a rich and even assemblage. 0 10 20 30 40La Ch i mba El Azúcar O G S E -80 C hi a pasSiteFisher's Alpha Figure 6-2. Fisher’s alpha bi odiversity index for the Ecua dorian sites of OGSE-80, El Azucar, and La Chimba, as well as the Mexican site of Paso de la Amada. Error bars give 95% confidence interval. A shallow curve portrays a community that is dominated by a few species. Among the Ecuadorian sites, only Loma Alta has a steep curve comparable to Paso de la Amada (Figure 6-3). This site is similar in the n ear absence of small species. With only four specimens, Zenaida auriculata (Eared Dove) is only a minor component of the bird fauna. At Paso de la Amada Zenaida macroura (Mourning Dove), which is similar in size to Z. auriculata , is represented by only four specimens. The smaller ( Columbina -sized) Scardafella inca (Inca Dove) is the most abundant spec ies, but hardly more so than the large raptor Poyborus plancus (Crested Caracara). This near lack of common small species that could easily be hunt ed near the settlements (Cha pters 3 and 4) has a twofold explanation. First, the use of 1/4 inch meshed screens to sieve the sediments may have biased the fauna toward large species. As the excavation at El Azúcar has shown, small

PAGE 333

317 mesh size can yield large numbers of small songbirds (Passeriformes) and doves ( Columbina ; Chapter 4). Secondly, hunters at Paso de la Amada might have been more selective when hunting birds than their Ec uadorian counterparts. A high abundance of non-avian food sources may have made hunti ng small birds unprofitable in energetic terms. In addition, taboos c ould have played a role. A. 0 10 20 30 40 50 60 70 050100150200250300 Number of SpecimensNumber of Species OGSE-80 El Azúcar La Chimba Chiapas B. 0 5 10 15 20 25 040 Number of SpecimensNumber of Species OGSE-80 El Azúcar La Chimba Valdivia Village La Ponga Loma Alta Chiapas Figure 6-3. Species accumulation curves base d on the number of identified specimens (NISP) for A) OGSE-80, El Azúcar, La Chimba, and Paso de la Amada; and B) for OGSE-80, El Azúcar, La Chimba, and Paso de la Amada truncated at 40 specimens, as well as Valdivia Village, La Ponga, and Loma Alta.

PAGE 334

318 Table 6-2. Species identified at archaeological sites OGSE-80 ( O80 ), El Azucar ( EA ), Valdivia Valley ( VV ), and La Chimba ( LC ), Ecuador. For each species habitat preference ( Habitat ), foraging stratum ( FoSt ), and location of the site with regard to the current species range are provided ( Range ). Data are from Stotz et al. (1996) and Ridgely a nd Greenfield (2001a,b). For habitat preference and foraging stratum, abbr eviations follow Stotz et al. (1996). Habitat preference is shown as A1 (Freshwater marshes); A2 (Saltwater/brackish marshes); A3 (Coastal sand beaches/mudflats); A4 (Coastal rocky beaches); A5 (Riverine sand beaches); A6 (Freshwater lakes and ponds); A8 (Rivers); A9 (Streams); A11 (Coastal waters); A12 (Pelagic waters); F1 (Tropical lowland evergreen forest); F3 (River-edge forest); F4 (Montane evergreen forest); F5 (Elfin forest); F7 (Tropical deciduous forest); F8 (Gallery forest); F14 (Mangrove forest); F15 (Secondary forest); N1 (Arid lowland scrub); N2 (Arid montane scrub); N3 (Semihumid/humid montane scrub); N6 (Low, seasonally wet grassland); N10 (Paramo grassland); N11 (Riparian thickets); N13 (Pasture/agricultural lands); N14 (Second-growth scrub). An E modifies any one of the above to indicate preference for edge. Foraging Stratum is indicated by T (Terrestrial); U (Understory); M (Midstory); C (canopy); A (Aerial); W (Water). The location of the site and a 20 km radius in relation to the current range is indicated by a + if in current range; a [+] if in current range, but near boundary, or species rare or sporadically present; and – if not within current range. Species O80 EA VV LC Habitat FoSt Range TINAMIDAE Tinamus cf. major x F1 T – Crypturellus transfasciatus x x F7, F14 T + Nothoprocta cf. pentlandii x N2, N13 T – Nothoprocta cf. curvirostris x N10, N2, N13 T + PROCELLARIIDAE Puffinus griseus x A12 W +m PHALACROCORACIDAE Phalacrocorax brasilianus x A11, A6, A8 W + Phalacrocorax bougainvilli x A11 W [+]v ARDEIDAE Ixobrychus cf. exilis x A1, A2, F14 U [+]v

PAGE 335

319 Table 6-1. Continued. Species O80 EA VV LC Habitat FoSt Range ARDEIDAE Tigrisoma fasciatum x A8, A9 T-W – Tigrisoma lineatum x A1 T-W – Ardea alba x x A1, A2, A3, A4, A6 T-W + ANATIDAE Dendrocygna autumnalis / bicolor x A1 / A1, A6 T-W + Merganetta armata x A8, A9 W + Anas flavirostris x A1, A6, A8 W + Anas bahamensis x x A1, A2, A6, A11 W + Anas georgica x A1, A6 W – Anas cf. discors / cyanoptera x A1, A6 W [+]m/–m Netta erythrophthalma x A6 W [+] ACCIPITRIDAE Elanus leucurus x N13, N14, N6, N1, N2 C-A + Ictinia plumbea x F1, F7, F8 C-A +m Buteogallus meridionalis x N6, N14 T-C + Buteogallus anthracinus x F14, F7, F8, F1, C – Parabuteo unicinctus x x N1, N2, F7, N6 C + Geranoaetus melanoleucus x N1, N2, N7 T + Buteo nitidus x F8, F7, F1E, F3 C + Buteo cf. platypterus / albigula x F1, F4, F7 / F4, N2, N1 C +m/+ Buteo swainsoni x N1 T [+]m

PAGE 336

320 Table 6-1. Continued. Species O80 EA VV LC Habitat FoSt Range FALCONIDAE Phalcoboenus carunculatus x N10, N14 T + Falco sparverius x x N1, F7, F8, N13, N10 T-C + Herpetotheres cachinnans x F1E, F8, F3, F2, F7, N1 C + CRACIDAE Ortalis erythroptera x F7, F14 T-C + Penelope montagnii x F4, F15 M-C + Penelope cf. purpurascens x F1, F7, F8, F4 U-C + cf. Aburria aburri x F4, F1 C + cf. Chamaepetes goudotii x F4, F4E M-C – RALLIDAE Pardirallus cf. maculatus x A1 T – Rallus cf. longirostris x A2, F14 T – Rallus aequatorialis x A1 T + Neocrex cf. erythrops x N6, N13, A6 T [+] Porzana carolina x A1 T [+]m Porphyrula martinica x A1 T – Fulica ardesiaca x A1, A6 T-W [+] SCOLOPACIDAE Tringa melanoleuca x A1, A2, A3, A5 T +m Arenaria interpres x A3, A4, A2 T +m Gallinago sp. x N10, A10, F5 T +

PAGE 337

321 Table 6-1. Continued. Species O80 EA VV LC Habitat FoSt Range THINOCORIDAE Attagis gayi x N10, rocky N10 type T + LARIDAE Larus cf. atricilla / pipixcan x A11, A3, A4 T-W +m Sterna cf. hirundinacea x A11 W + COLUMBIDAE Columba speciosa x F1, F4, F8 C – Columba cayennensis x F8, F3, F1E, F15 C + Zenaida auriculata x x x x N14, N1, N2 T-M + Zenaida meloda x x x F7, F8, N11, N14, N1 T-M + Columbina minuta x N1, N6, N2, N14 T [+] Columbina buckleyi x x F7, N1 T + Columbina cruziana x N14, N1, N11, N2 T + Claravis pretiosa x F1E, F15, F3, F8 T-M + Leptotila cf. verreauxi x x F7, F8, F15, F3, F1E T/U + Geotrygon montana x F1, F4, F7 T – PSITTACIDAE Ara cf. ambigua x F1 C – Ara severa x F3, F8, F1E,F2 C – Aratinga erythrogenys x x F7, F8 C + cf. Leptosittaca branickii x F4, F5 C + Forpus cf. coelestis x F7, F8, F15, N1 C + Pionites melanocephala x F1, F2 C –

PAGE 338

322 Table 6-1. Continued. Species O80 EA VV LC Habitat FoSt Range PSITTACIDAE Amazona autumnalis x F1E, F7, F8, F15 C + Amazona cf. mercenaria x F4, F1 C + Amazona farinosa x x x F1 C – TYTONIDAE Tyto alba x x N14, N2, N1 C + STRIGIDAE Bubo virginianus x F4, F1, F7, F8, N1, N2 C + Asio cf. stygius x F4, F1 C + Asio cf. clamator x N6, N14 C / T – CUCULIDAE Crotophaga sulcirostris x N14, N11 T-M + RAMPHASTIDAE Aulacorhynchus prasinus x F4, F1 M-C + THAMNOPHILIDAE Taraba major x F1E, F15, F8, N11, N14 U + FORMICARIDAE Grallaria cf. guatemalensis / watkinsi x F1, F4, F7 / F7, F15 T +/– COTINGIDAE Tityra semifasciata x F1, F4, F15 C + CORVIDAE Cyanolyca sp. x F4, F15 M-C +

PAGE 339

323 Table 6-2. Continued. Species O80 EA VV LC Habitat FoSt Range CORVIDAE Cyanocorax cf. mystacalis x F7, F8 C [+] TURDIDAE Turdus fuscater x F4E, F15, F4, N3, N14 T-C + Turdus cf. serranus x F4, F4E, F15 U-C –? Turdus cf. maculirostris x F7, F7E, N13, F4E T-M + MIMIDAE Mimus cf. longicaudatus x N1, N2, F8, F7E C + HIRUDINIDAE Riparia riparia x N14, N13, A8, A9 A +m TROGLODYTIDAE Henicorhina cf. leucophrys x F4, F15 U – CARDINALIDAE Spiza americana x N13, N1, F7E T –m EMBERIZIDAE Volatinia jacarina x N14, N6, N1, N13, N11 T-U + Sporophila cf. peruviana x N1, N11 U + Sicalis cf. tackzanowskii x N1 T-C [+] THRAUPIDAE Buthraupis montana x F4, F4E M-C + Sericossypha albicristata x F4 C + ICTERIDAE Psarocolius angustifrons x F3, F4E, F1E, F15 C –

PAGE 340

324 Table 6-2. Continued. Species O80 EA VV LC Habitat FoSt Range ICTERIDAE Dives cf. warszewiczi x F8, F15 T-M + Quiscalus mexicanus x x F14, N14, N13, F15 T-C – Icterus cf. graceannae x F7, F8, N1 C + Sturnella bellicosa x x N13, N2, N1, A2 T/U + m nearctic of austral migrant v present sporadically as vagrant Several Peruvian sites are very different from the Ecuadorian sites in that mostly marine birds are represented in their bone assemblages. At the Early Holocene sites of Ostra Base Camp (Reitz and Sandweiss 2001) and Quebrada Tacahuay (deFrance et al. 2001), almost all bird specimens represent co mmon species of coastal waters. The Early to Middle Holocene site of La Paloma cont ains a few bones of herons (Ardeidae) in addition to those of marine birds (originally studied by E. J. Reitz and later identified to species by Matthiesen 1988). The Middle to La te Holocene sites of Moche Valley, Huaca Prieta, and Los Gavilanes (E. J. Reitz unpublished report, Pozorski 1983, Matthiesen 1988) are also dominated by marine birds. Huaca Prieta also contains a wetland component of herons, storks, and ducks. In this arid region of Peru, where terrestrial resources are scant, people turned mainly to marine resources. Late Pleistocene humans along the Peruvian coast probably fished a nd hunted marine mammals as well as large numbers of marine birds (deFrance 2005). As many of the marine birds dive for food, they were probably caught in nets. As oppos ed to Chiripa, birds were not taken incidentally along the Peruvian coas t, but were pursued purposefully.

PAGE 341

325 Historical Biogeography of Birds in We stern Ecuador and the Northern Andes Coastal Sites The composition of the bird fauna at th e Ecuadorian archaeological sites suggests that most of the terrestrial birds were hunted in the immediat e surroundings of human habitations. Of the 38 forest sp ecies (40.0% of all species), on ly four are restricted to a single forest habitat (Tables 6-2 and 6-3), al though each is to some degree tolerant to human disturbance. The remaining 34 forest spec ies inhabit at least two types of forested habitats (beginning with F in Table 6-2), which in some instances include secondary forest or forest edge. Most of these speci es are commonly found today in landscapes with a large proportion of agricultural lands (R idgely and Greenfield 2001a; M. P. Tellkamp personal observation). Only 23.2% of all bird s are associated with wetlands/mangroves. Even if the species found only at Chimba, wh ere wetlands are more distant and smaller in size, are excluded, the proportion of wetland/ma ngrove species barely increases to 23.4% which is still well below the the 48% observe d at Paso de la Amada and 51.6% found at Chiripa. Only 3.3% of all birds (4.7% when excluding the species limited to La Chimba) are marine birds. This is in stark contrast to the Peruvian sites (deFrance et al. 2001; see above). Numerically the most important group of birds (40.0%) are those species that utilize forest habitats (coded by F and FF in Table 6-2). When excluding La Chimba, this number drops to 34.4%. Species using at least one type of open habitat (coded by an N in Table 6-2) make up 40.6% of the bird a ssemblage on the coast (36.8% for all sites combined) underlining the notion of opportuni stic hunting around human settlements. A lower percentage of these birds at La Chimba could be explained by the presence of large forested areas around the settlement.

PAGE 342

326 Most species of birds from the archaeol ogical sites were probably available to prehistoric people within a 20 km radius of their settlements. Currently most of the species found at the archaeological sites are st ill present within a radius of 20 km. Nearly 22% of the species, however, probably came from a more distant source area, or had wider ranges in the past. These species pres ently occur more than 20 if not 10 km from the archaeological sites, in some instance mo re than 100 km (Chapter 3). Some of these species might have been traded, but in some instances natural or anthropogenic vegetation change probably led to the local extirpation of populations. Species likely to be traded whole or in parts are those with colorful plumage (e.g., parrots) or other exceptional morphological features , such as large talons or large feathers (in raptors; Table 6-3) . Other species might have been tr aded alive as pets. Whereas dead birds are easily transported over long distan ces, live birds probably were traded mostly within the lowlands and within the highl ands, respectively. I re gard long-distance transport of live birds between the warm lowl ands and the cold highlands as less likely. The difficult terrain and the drastic climatic changes would have made it difficult to keep the birds alive. A third distinct possibility is the transport of live birds across the Andes if exposure to the cold highland climate is limite d to a few days. Zoogeographers need to be aware of the possibility that some bird species, especially parrots and macaws (Psittacidae) as well as some pigeons (Col umbidae) and songbirds (Passeriformes; see Chapter 5), may have been traded alive over long distances and that, if kept as pets, some of these could have escaped and established new local populations. Pets are a common feature of Native American communities, often with a focus on species that are colorful, power ful, or good singers. For instance, the Siona and Secoya of

PAGE 343

327 the Ecuadorian Amazon keep toucans and pa rrots (Vickers 1989) . In 1994, I saw an impressive Harpy Eagle ( Harpia harpia ) and a Black Vulture ( Coragyps atratus ) at a Huaorani community along the Río Yasuní in eastern Ecuador. Given the availability of game near sett lements, I do not believe that birds were transported long distance for food in prehistoric Ecuador. Th us, the large tinamou at Site 80 may have been hunted as close as the Cordillera Chongón-Colonche. The cordillera currently represents the westernmost exte nsion of the Chocó biogeographic region in Ecuador, with humid forests nestled on mountain tops func tioning as habitat islands surrounded by deciduous and semi-deciduous forest s (Sierra et al. 1999b). It is possible that the ranges of the large tinamou and the Mealy Parrot ( Amazona farinosa ) once extended into this cordillera. Tinamus major has been reported from Manglares Churute, ca. 80 km to the east (Ridgely and Greenfield 20 01a), suggesting that this species used to have larger range. The combined effects of low population size in forest fragments, hunting, and anthropogenic habitat change might have caused local extirpation of these two species. Table 6-3. Number of species (percentage) in five broad habitat categories for all Ecuadorian sites combined, combined coastal sites, and La Chimba. Abbreviations are: M – Marine species; W – wetland and mangrove species; F – Restricted to one forest type; FF – found in several forest types, but not open habitats; and N – found in, but not restricted to open habitats. Since the categories are not mutually exclusive, the percentages should not sum to one. M W F FF N All Ecuadorian sites 3(3.2) 22(23.2) 4(4.2) 34(35.8) 35(36.8) All coastal sites 3(4.7) 15(23.4) 3(4.7) 19(29.7) 26(40.6)) La Chimba 0(0) 6(17.1) 1(2.9) 15(42.9) 12(34.2)

PAGE 344

328 The gradual disappearance of mangroves during the Holocene from natural and anthropogenic factors (Ferdon 1981, Heusser an d Shackleton 1994, Stothert et al. 2003) is responsible for the local exti rpation of a variety of bird sp ecies that prefer this habitat (Chapter 3). Freshwater wetla nds probably developed as rainwater formed lenses above the saltwater percolating into the soil behind mangroves (Ferdon 1981). Several species of rails as well as Buteogallus anthracinus subtilis (Mangrove Hawk) and Quiscalus mexicanus (Great-tailed Grackle) disappeared from Santa Elena Peninsula together along with mangroves. Another 11% of the species from the ar chaeological sites presently have low population sizes, are near their range limits, or occur as rare vagrants ([+] under the range column in Table 6-2). Many of these ha d much larger populations 50 years ago (Marchant 1958) and two, Ixobrychus exilis and Columbina minuta , have been increasing over the last decades of the twentieth century, leading to the suggestion that their presence in Ecuador is a recent phenome non (Ridgely and Greenfield 2001a). The zooarchaeological data show that that is not the case and that dist ributional patterns are much more dynamic than previously assumed. Sh ifts in distributions are also suggested by a small humid habitat component in the otherwise arid surroundi ngs at El Azúcar (Chapter 4). Periods of fre quent El Niños brought higher amounts of rainfall to the southwestern Ecuadorian coast and probably allowed for the temporary expansion of species associated with more humid condi tions. A similar expansion of dry habitat species probably occurred duri ng years of prolonged droughts. Northern Andes La Chimba is of special interest as it ma y shed some light on prehistoric conditions in the now arid, completely deforested in ter-Andean valley as well as the páramo

PAGE 345

329 grasslands. The presence of severa l bones of Olive-backed Oropendola ( Psarocolius angustifrons ) opens the possibility of considerably more humid conditions in the interAndean valley. The nearest populations of this racous, colorful species are presently more than 25 km away on the outer slopes of the we stern and eastern Cordillera of the Andes. Three hypotheses might explain the occurren ce of this species at La Chimba. 1) Prehistoric people made hunting trips to the eastern slopes of the Andes. Two bones of Ara cf. militaris (uncommon between 700 and 1500 m elev ation) suggest that hunting or trading trips to this area more than 25 km aw ay were made. 2) Birds were traded from or hunted on the western slopes of the Andes, about 25 km west of La Chimba. 3) The interAndean valley was partially covered by humi d forest that sustained many species now restricted to the outer slopes of the Ande s. At the time the Spanish arrived, some remnants of these forests remained near Cayambe (Hidalgo Nistri 1998, Chapter 3). Although above the current elevat ional range of the oropendola, the inter-Andean valley west of Cayambe lies partially in a rainsh adow that creates a warm micro-climate. Ancient DNA analyses could be used to test these hypotheses. Whereas the former presence of humid fore st in the inter-Andean valley is still debatable, this study clearly suggests that xeri c ecosystems have existed in the valley for at least 2000 years, possibly throughout the Holocene. During the La Chimba occupation, dry-habitat species even app ear to increase, which suggest s a regional increase of arid environments. Another important implication of the La Chimba site is the suggestion of more structurally complex páramo vegetation than is found today in most of the Ecuadorian Andes. I claimed above that prehistoric peopl e in Ecuador hunted bi rds opportunistically,

PAGE 346

330 especially common birds in the immediat e surroundings. The present scarcity of Nothoprocta curvirostris seems to contradict the notion of high availability near human habitations, though. I suspect that a combina tion of high frequency of burns in the Ecuadorian paramo and massive deforestation has significantly diminished habitat quality for the tinamou along the tree line during the last century. Burning is a commonly used practice to improve the quality of the forage (Sarmiento and Frolich 2000) and leads to a low and sparse grass cover with little or no brushy vegetation (personal observation) required by the tinamou (Ridgely and Greenfiel d 2001a). Wille et al. (2002) have shown that burning increased in frequency after the Spanish arrived in the Andes and that deforestation occurred mostly during the last fi ve centuries. Presently very little suitable habitat for the tinamou remains. The fact that the few recent records of this species come from areas with a structur ally diverse paramo, includi ng abundant brush and nearby forest, supports the notion that a recent loss of habitat has led to th e population declines. Migratory Species Are migratory birds disproportionally pr one to being hunted? Migration imposes great energetic demands on birds. During a dverse climatic conditions, migrants can become extremely weakened. It is not unus ual to find migratory birds on the ground unable to fly, or evading pe ople sluggishly. During this se nsitive period migratory birds are very susceptible to being hunted. In th is study, however, the proportion of migratory species in the archaeological sites is nearly the same as that of the entire territory of Ecuador (10.5% vs. 10.0%). Lessons for Bird Conservation With more than 1600 species, the bird fauna of Ecuador is among the richest in the world (Ridgely and Greenfield 2001a,b). Much of this diversity is now threatened by

PAGE 347

331 habitat destruction, especially by conversion of forests to agricultural lands (Sierra et al. 1999a). Because alpha diversity is highest in the Amazonian lowlands, most of the protected areas in Ecuador have been estab lished in this sparsely populated area. The majority of the species considered to be th reatened by IUCN standards, however, come from the western lowlands and the Andes (IUCN 2004), which contain a large number of endemic species (Cracraft 1985a, Stotz et al. 1996), and where severe deforestation over the last 500 years (Dodson and Gentry 1991, Hi dalgo Nistri 1998, Sierra et al. 1999a) now threatens many endemic species. Extinction ri sk of birds in the Neotropics has been correlated with a series of ecological species traits, such as body size, feeding guild, geographic position with relation to range lim its, range size, habita t specialization, and population size to name a few (Lande 1988, Ka ttan et al. 1994, Fjeldså 2001). Although these associations are of value in setting priorities for conservation, we still lack a historical perspective on the l ong-term effects of past natu ral and anthropogenic habitat change on the distribution of birds. For inst ance, recent studies have suggested that many parts of Amazonia and other areas of “undist urbed” rainforest worldwide have supported significant human populations (Heckenberger et al. 2003, Willis et al. 2004). These areas still harbor diverse bird communities. Obvious ly, many species have an ability to rebound after severe habitat degrada tion and hunting pressure. Resi lience of birds to habitat change is a key aspect for habitat restoration projects as well as the management of conservation corridors. These considerations are important given the recent work on predicting extinctions from simple models based on species area rela tionships suggesting that forest birds will not be able to survive habitat loss and fragmentation (Harri s and Pimm 2004). The

PAGE 348

332 resilience shown by the avifauna s surrounding the archaeological sites provides us with a more optimistic outlook in which extinctions of Neotropical birds may be averted by a more dynamic approach to the management of diverse landscape mosaics by mimicking Holocene patterns of deforest ation and habitat recovery. Habitat fragments that may temporarily maintain diverse bird faunas (Leck 1979, Kattan et al. 1994, Renjifo 1999) and that were probably common in the past (S tahl 2000), may have played a central role in habitat recovery. In additi on, estimates of forest cover in prehistoric times assume the absence of humans (Brooks et al. 1999, Manne et al. 1999, Harris and Pimm 2004), but the peopling of the Americas has had a cons iderable impact on the vegetation beginning in the Late Pleistocene (Chapter 1). Approxima tions for the size of prehistoric forests are therefore probably overestimates. Also problematic for the design of conserva tion strategies is th e lack of a true benchmark for measuring the success of speci fic conservation measures. What should an ideal preserve look like to protect representative habita t for perpetuity? What does representative mean in the light of a con tinuously changing natural world? What is the meaning of “pristine” or “nat ural” given that humans had co lonized most of the planet by 10,000 years ago? Paleoecological studies based on polle n, phytoliths, paleontology, phylogeography, and zooarchaeology will all contribute to an swering these questions. The studies of prehistoric bird faunas from Ecuador so far a llow me to draw this rough outline of several broad patterns. 1) Even though humans modified the natural environment throughout the Holocene, many currently rare species had much larger ranges for mo st of the last 10,000

PAGE 349

333 years. Massive deforestation, excessive hunti ng, and burning of non-forest habitats, such as páramo, are probably a fairly recent threat to bird dive rsity in Ecuador. Many species of birds, such as the Crested Guan ( Penelope purpurascens ), have been hunted for thousands of years and yet still occur in ma ny severely degraded areas. The ability to persist, however, might be compromised in the future by the magnitude of the change faced by many species. 2) The ability of many species to live in partially fragmented, constantly changing landscapes, as evidenced by the zooarchaeologi cal (Chapters 3, 4, and 5) and ecological data (Tellkamp 1999), should be seen as an opportunity for conservation. Conservation efforts should not only be directed at pristine areas, but also at la ndscape mosaics with a great potential for restoration. This is especially true for centers of endemism, many of which are heavily deforested (Fjeldså 2001). 3) A benchmark for the Holocene needs to take into account anthropogenically generated habitat diversity. This diversity should not be seen as a large-scale modification of landscapes as is the case pr esently, but as the intr oduction of localized patchworks of agricultural la nds and secondary vegetation type s within a mostly forested landscape. The fact that many species had larg er ranges for thousands of years means that suitable habitat must have been more wide ly available. Because of the bias of archaeofaunas toward common species of di sturbed habitats, deposits of bones not derived by people need to be studied to find stronger evidence of recent range contractions (Steadman et al. 2003). Conclusion Prehistoric people obtained most of the birds locally around th eir settlements or during hunting trips. In some instances a loca l source is unrealistic, and bones deposited

PAGE 350

334 at the archaeological sites pr obably arrived there through tr ade (Chapters 3 and 5). The predominance of Zenaida auriculata at Site 80 and El Azúcar, and of Nothoprocta cf. curvirostris at La Chimba, suggests that people ob tained most of the birds for local consumption in the immediate vicinity of th e settlement in proportion to their availability. The species selected by prehistoric people in Ecuador appear to represent the outcome of a compromise between bird size, availabil ity (density), and ease of capture. Recent changes in the distributions of species are due habitat dest ruction as well as natural processes (especially when considering mangroves). Impli cations for conservation are both encouraging and worrisome. On one hand, restoration projects may be successful as species are able to persist in altered hab itats and rebound after conditions become more favorable. Current destruction of habitats, how ever, may be more severe than during any other period of the Holocene and spell wide-r anging extinctions if action is not taken soon. This is particularly true for the western lowlands a nd northern Andes of Ecuador.

PAGE 351

335 APPENDIX LIST OF MUSEUM SPECIMENS USED. Appendix. Specimens used for comparative osteology (Chapter 2) in alphabetical order. Additional specimens were view ed when larger series were necessary. Species Sex/Age Specimen Number Accipiter cooperi female UF 40309 Accipiter superciliosus superciliosus female UF 23871 Actitis macularia female UF 11714 Aimophila ruficauda ruficauda male UF 38259 Aimophila ruficeps duponti female UF 28834 Amazona autumnalis autumnalis male UF 25771 Amazona farinosa inornata female UF 25818 Ammodramus savannarum pratensis male UF 28869 Anas acuta female UF 23483 Anas acuta male UF 42037 Anas bahamensis male UF 23487 Anas bahamensis male UF 42469 Anas bahamensis bahamensis male UF 41188 Anas bahamensis rubrirostris male UF 23488 Anas clypeata male UF 23509 Anas clypeata female UF 23511 Anas flavirostris female UF 39299 Anas flavirostris oxyptera male UF 33735

PAGE 352

336 Species Sex/Age Specimen Number Anas georgica female UF 39353 Anas georgica male UF 39339 Andigena laminirostris male UF 26904 Ara macao imm. UF 25869 Ara severa male UF 22234 Aramides cajanea cajanea unknown UF 38835 Aratinga erythrogenys female UF 25878 Aratinga leucophthalma female UF 25904 Aratinga mitrata female UF 25897 Arenaria interpres unknown UF 21713 Arremon aurantiirostris saturatus male UF 38278 Arremonops conirostris richmondi male UF 29317 Asio flammeus domingensis female UF 41995 Asio flammeus flammeus female UF 41962 Atlapetes albinucha coloratus unknown UF 29320 Atlapetes brunneinucha brunneinucha unknown UF 29329 Atlapetes brunneinucha suttoni male UF 29326 Atlapetes rufinucha baroni male UF 43411 Aulacorhynchus derbianus derbianus male UF 39871 Aulacorhynchus haematopygus male UF 26898 Aulacorhynchus prasinus male UF 21482 Aulacorhynchus prasinus male UF 21487 Aythya affinis female UF 41954

PAGE 353

337 Species Sex/Age Specimen Number Bartramia longicauda female UF 35211 Bartramia longicauda unknown UF 33818 Basileuterus tristriatus tristriatus male UF 43430 Brotogeris chrysopterus male UF 25949 Bubo virginianus female UF 40842 Buteo brachyurus female UF 40124 Buteo magnirostris male UF 38955 Buteo magnirostris unknown UF 28954 Buteo magnirostris griseocauda female UF 38897 Buteo nitidus female UF 33746 Buteo nitidus plagiatus Male UF 23882 Buteo platypterus female UF 40752 Buteo platypterus male UF 42046 Buteo swainsoni female UF 41965 Buteogallus aequinoctialis male UF 33556 Buteogallus anthracinus anthracinus female UF 33744 Buteogallus urubitinga ridgwayi female UF 33554 Buteroides virescens male UF 40641 Buteroides virescens male UF 40762 Buthraupis montana male UF 29661 Buthraupis montana cyanonota male UF 43519 Cacicus haemorrhous haemorrhous female UF 33073 Cacicus holosericeus male UF 33277

PAGE 354

338 Species Sex/Age Specimen Number Cacicus melanicterus male UF 33076 Cacicus uropygialis microrhynchus male UF 33062 Calidris canutus female UF 24733 Campylorhynchus zonatus vulcanius male UF 38308 Catotrophorus semipalmatus male UF 24632 Charadrius vociferous female UF 40323 Chondrohierax uncinatus uncinatus female UF 33737 Ciccaba virgata squamulata female UF 26465 Cinnycerthia olivascens olivascens female UF 43407 Circus cyaneus female UF 40350 Circus macrourus male UF 38886 Cistothorus platensis stellaris male UF 33478 Claravis pretiosa female UF 11550 Claravis pretiosa female UF 38787 Claravis pretiosa male UF 11559 Coccyzus americanus americanus female UF 18072 Coccyzus erythropthalmus unknown UF 26281 Columba cayennensis cayennensis male UF 38773 Columba fasciata fasciata male UF 32308 Columba fasciata monilis female UF 32305 Columba fasciata monilis female UF 32306 Columba fasciata monilis female UF 32307 Columba nigrirostris male UF 38774

PAGE 355

339 Species Sex/Age Specimen Number Columba plumbea delicata male UF 39869 Columba speciosa male UF 13618 Columba speciosa male UF 25412 Columba speciosa male UF 38772 Columbina buckleyi female UF 41837 Columbina minuta elaeodes male UF 32371 Columbina minuta minuta female UF 38786 Columbina picui female UF 38967 Columbina picui male UF 38968 Crotophaga ani male UF 38970 Crotophaga sulcirostris male UF 14543 Crotophaga sulcirostris sulcirostris male UF 33855 Crypturellus cinnamomeus praepes male UF 36835 Crypturellus soui panamensis female UF 38948 Cyanocorax affinis affinis male UF 32105 Cyanocorax sanblasianus male UF 38204 Cyanocorax yncas yncas female UF 39867 Cyanocorax yncas yncas male UF 32137 Cyanolyca viridicyana jolyaea female UF 43404 Cyanolyca viridicyana jolyaea male UF 43499 Daption capense female UF 38161 Daptrius ater female UF 23973 Dendrocincla tyrannina tyrannina female UF 43463

PAGE 356

340 Species Sex/Age Specimen Number Dendrocygna autumnalis discolor male UF 23206 Dendrocygna autumnalis discolor unknown UF 23208 Dendrocygna bicolor bicolor male UF 23198 Deroptyus accipitrinus female UF 25965 Diglossa carbonaria brunneiventris male UF 29770 Dives dives female UF 33071 Dolichonyx oryzivorus male UF 31151 Elanoides forficatus female UF 44182 Elanus leucurus majusculus unknown UF 23819 Emberizoides herbicola sphenurus male UF 38272 Eubucco bourcierii male UF 33403 Eudromia elegans male UF 22258 Falco columbarius female UF 42034 Falco columbarius female UF 42034 Falco peregrinus female UF 24044 Falco peregrinus anatus female UF 24044 Falco rufigularis rufigularis male UF 38880 Falco rufigularis rufigularis male UF 38880 Falco sparverius male UF 40316 Falco sparverius male UF 40316 Formicarius analis umbrosus male UF 27303 Fulica americana americana female UF 16941 Fulica americana americana male UF 16940

PAGE 357

341 Species Sex/Age Specimen Number Fulica ardesiaca female UF 39760 Fulica ardesiaca male UF 39208 Fulica ardesiaca male UF 39222 Furnarius leucopus male UF 38987 Gallinago gallinago delicata male UF 41111 Gallinago paraguaiae paraguaiae male UF 24685 Gallinula chloropus female UF 40968 Gallinula chloropus male UF 40343 Gampsonyx swainsonii leonae female UF 38890 Gampsonyx swainsonii leonae female UF 38890 Geotrygon frenata frenata female UF 43349 Geotrygon montana female UF 33846 Geranoaetus melanoleucus immature UF 33557 Geranoaetus melanoleucus unknown UF 38062 Geranospiza caerulescens balzarensis male UF 23847 Geranospiza caerulescens livens male? UF 23846 Glaucidium gnoma californicum unknown UF 33851 Grallaria przewalskii male UF 43384 Grallaria przewalskii male UF 43473 Harpagus bidentatus fasciatus female UF 33751 Hemispingus atropileus auricularis female UF 43415 Henicorhina leucoptera male UF 43405 Herpetotheres cachinnans cachinnans male UF 38874

PAGE 358

342 Species Sex/Age Specimen Number Heteroscelus incanus male UF 24644 Heteroscelus incanus unknown UF 24642 Heterospizias meridionalis meridionalis male UF 38903 Icterus chrysaster chrysaster male UF 38242 Icterus icterus ridgwayi male UF 30895 Icterus mesomelas male UF 38244 Ictinia plumbea female UF 23826 Jacana jacana hypomelaena male UF 33820 Larosterna inca male UF 35305 Larus atricilla female UF 40876 Larus delawarensis male UF 40351 Larus fuscus unknown UF 33830 Larus pipixcan female UF 24819 Larus serranus male UF 39357 Laterallus jamaicensis male UF 40370 Laterallus leucopyrrhus unknown UF 24349 Leistes militaris militaris male UF 33128 Leptotila verreauxi bangsi male UF 41093 Leptotila verreauxi verreauxi female UF 32311 Limnodromus griseus male UF 12891 Limosa fedoa female UF 24580 Lipaugus vociferans male UF 27754 Megarhynchus pintagua mexicanus male UF 40791

PAGE 359

343 Species Sex/Age Specimen Number Melopyrrha nigra unknown UF 29242 Merganetta armata leucogenis male UF 43435 Metriopelia melanoptera unknown UF 39212 Micrastur semitorquatus male UF 38875 Mimus gundlachii gundlachii female UF 40106 Mimus saturninus male UF 39010 Molothrus bonariensiss minimus male UF 34002 Muscisaxicola albifrons female UF 38631 Myiodynastes maculatus nobilis male UF 38619 Netta erythrophthalma brunnea female UF 23525 Nothocercus nigrocapillus nigrocapillus female UF 43347 Nothocercus nigrocapillus nigrocapillus female UF 43432 Nothocercus nigrocapillus nigrocapillus female UF 43433 Nothoprocta cinerascens male UF 38951 Nothoprocta cinerascens male UF 38952 Nothura maculosa salvadorii female UF 22260 Oryzoborus angolensis female UF 36159 Oryzoborus angolensis male UF 36157 Oryzoborus funereus male UF 29234 Oxyura jamaicensis female UF 43781 Parabuteo unicinctus harrisi male UF 33745 Pardirallus maculata insolinus female UF 24348 Pardirallus maculata insolinus unknown UF 24340

PAGE 360

344 Species Sex/Age Specimen Number Passerina cyanea female UF 17473 Passerina cyanoides cyanoides male UF 29451 Phalaropus tricolor female UF 21179 Phalaropus tricolor male UF 21177 Phalcoboenus megalopterus male UF 38873 Phalcoboenus megalopterus male UF 38873 Pheucticus chrysopeplus male UF 34034 Phlegopsis nigromacul ata nigromaculata female UF 39865 Phrygilus plebejus plebejus female UF 35648 Piaya cayana thermophila male UF 41094 Pionites melanocephala male UF 26068 Pionopsitta haematotis male UF 38767 Pionus fuscus male UF 26165 Piranga olivacea male UF 19659 Pluvialis dominica dominica female UF 24531 Pluvialis squatarola female UF 24518 Polyborus plancus plancus female UF 23980 Polyborus plancus plancus female UF 23980 Poospiza melanoleuca female UF 41540 Porphyrula martinica female UF 40324 Porphyrula martinica male UF 42418 Porzana carolina female UF 12832 Porzana carolina male UF 40370

PAGE 361

345 Species Sex/Age Specimen Number Procellaria aequinoctialis unknown UF 36673 Psarocolius angustifrons female UF 33080 Psarocolius decumanus female UF 33078 Psarocolius decumanus male UF 30855 Psarocolius decumanus male UF 30856 Psarocolius decumanus male UF 30857 Psarocolius decumanus male? UF 33077 Psarocolius decumanus decumanus male UF 30858 Psarocolius viridis viridis female UF 33079 Psarocolius wagleri male UF 30861 Pterodroma inexpectata unknown UF 22486 Pteroglossus torquatus torquatus male UF 13023 Puffinus bulleri female UF 22568 Puffinus griseus unknown UF 39651 Pulsatrix perspicillata perspicillata female UF 38705 Querula purpurata male UF 27744 Quiscalus mexicanus mayor female UF 31043 Quiscalus nicaraguensis female UF 33132 Rallus limicola female UF 24324 Rallus limicola male UF 24325 Rallus limicola male UF 39723 Rallus longirostris wayni male UF 11696 Rallus longirostris wayni male UF 24304

PAGE 362

346 Species Sex/Age Specimen Number Ramphocelus flammigerus icteronotus male UF 29630 Recurvirostra americana male UF 24483 Rhinoptynx clamator female UF 26479 Riparia riparia female UF 43748 Rissa brevirostris unknown UF 34353 Rostrhamus sociabilis sociabilis unknown UF 33741 Rupicola peruvianus sanguinolentus male UF 27753 Sakesphorus canadensis male UF 11561 Saltator albicollis albicollis unknown UF 29446 Saltator aurantiirostris albociliaris female UF 35402 Saltator maximus magnoides female UF 40808 Scolopax minor male UF 41931 Selenidera culik male UF 33398 Sericossypha albocristata male UF 43413 Sicalis flaveola female UF 29164 Sicalis flaveola unknown UF 38271 Sicalis olivascen chloris male UF 35654 Sicalis uropygialis female UF 39226 Spiza americana male UF 33139 Sporophila telasco male UF 38274 Sterna maxima unknown UF 25221 Sterna nilotica female UF 25096 Sterna sandvicensis male UF 39705

PAGE 363

347 Species Sex/Age Specimen Number Sterna sandvicensis male UF 40643 Sterna sandvicensis acuflavidus male UF 41793 Synallaxis brachyura male UF 38988 Syndactyla subalaris lineata unknown UF 27387 Tabara major male UF 38991 Tangara vitriolina male UF 29725 Tangara xanthocephala venusta male UF 43422 Tapera naevia excellans male UF 26249 Thinocorus rumicivorus female UF 39352 Thraupis episcopus cana male UF 29646 Thraupis episcopus cana unknown UF 29641 Thripadectes rufobrunneus unknown UF 27388 Thryothorus leucotis albipectus male UF 28139 Thryothorus modestus modestus male UF 40794 Thryothorus nigricapillus castaneus male UF28146 Thryothorus sinaloa sinaloa male UF 38302 Tiaris olivacea olivacea male UF 29249 Tityra semifasciata costaricensis male UF 27632 Tityra semifasciata personata female UF 38624 Tringa flavipes male UF 33760 Tringa melanoleuca male UF 21955 Turdus chiguanco female UF 39215 Turdus grayi female UF 33980

PAGE 364

348 Species Sex/Age Specimen Number Turdus serranus serranus female UF 43409 Tyrannus melancholicus occidentalis female UF 27663 Tyto alba female UF 38974 Vanellus resplendens male UF 39341 Veniliornis fumigatus sanguinolentus female UF 38667 Volatinia jacarina splendens male UF 29166 Xema sabini male UF 33833 Xiphorhynchus promeropirhynchus compressirostris male UF 43464 Xiphorhynchus triangularis intermedius female UF 43373 Zarhynchus wagleri ridgwayi male UF 33081 Zarhynchus wagleri ridgwayi male UF 30860 Zenaida asiatica asiatica male UF 40008 Zenaida asiatica asiatica male UF 40641 Zenaida auriculata female UF 39347 Zenaida auriculata male UF 38961 Zonotrichia capensis septentrional female UF 38261

PAGE 365

349 LIST OF REFERENCES ANDRUS, C. F. T., D. E. CROWE, D. H. SANDWEISS, E. J. REITZ, AND C. S. ROMANEK. 2002. Otolith 18O record of Mid-Holocene sea surface temperatures in Peru. Science 295: 1508-1511. ANDRUS, C. F. T., D. E. CROWE, D. H. SANDWEISS, E. J. REITZ, AND C. S. ROMANEK. 2003. Response to comment on “Otolith 18O record of Mid-Holocene sea surface temperatures in Peru.” Science 299: 203b. ARDILA, G., AND G. POLITIS. 1989. Nuevos datos para un viejo problema. Boletín del Museo de Oro (Bogotá, Colombia) 23: 3-45. ATHENS, J. S. 1990. Prehistoric agricultural expansi on and population growth in northern highland Ecuador: interim report for 1989 fi eldwork. International Archaeological Research Institute, Honolulu, Hawaii. ATHENS, J. S., AND J. V. WARD. 1999. The late Quaternary of the western Amazon; climate, vegetation and humans. Antiquity 73: 280-287. BAUMEL, J. J., A. S. KING, J. E. BREAZILE, H. E. EVANS, AND J. C. VANDEN BERGE [Eds.]. 1993. Handbook of avian anatomy: nomina anatomica avium, second edition. Publications of the Nutta ll Ornithological Club 23. BECKER, C. D., AND B. L"PEZ. 1997. Conservation value of Garua forest in the dry season: a bird survey in Reserva Ecológi ca Loma Alta, Ecuador. Cotinga 8: 66-74. BEHLING, H., H. HOOGHIEMSTRA, AND A. J. NEGRET. 1998. Holocene history of the Chocó rain forest from Laguna Piusbi, s outhern Pacific lowlands of Colombia. Quaternary Research 50: 300-308. BERGER, A. J. 1968. Appendicular myology of Kirtland’s Warbler. Auk 85: 594-616. BERTELLI, S., AND L. M. CHIAPPE. 2005. Earliest tinamous (Ave s: Palaeognathae) from the Miocene of Argentina and their phylogenetic position. Contributions in Science, Natural History Museum of Los Angeles County 502: 1-20. BERTELLI, S. AND A. L. PORZECANSKI. 2004. Tinamou (Tinamidae) systematics: a preliminary combined analysis of morphology and molecules. Ornitología Neotropical 15 (Supplement): 1-7.

PAGE 366

350 BEST, B. J., AND M. KESSLER. 1995. Biodiversity and conserva tion in Tumbesian Ecuador and Peru. BirdLife International, Cambridge, United Kingdom. BINFORD, M. W., A. L. KOLATA, M. BRENNER, J. W. JANUSEK, M. T. SEDDON, M. ABBOT, AND J. H. CURTIS. 1997. Climate variation and the ri se and fall of an Andean civilization. Quaternary Research 47: 235-248. BOCK, W. J. 1962. The pneumatic fossa of the hume rus in the passeres. Auk 79: 425-443. BRODKORB, P. 1955. The avifauna of the Bone Valley formation. Florida Geological Survey, Report of Investigations 14: 1-57. BROOKS, T. M., S. L. PIMM, V. KAPOS, AND CORINNA RAVILIOUS. 1999. Threat from deforestation to montane and lowland bi rds and mammals in insular South-east Asia. Journal of Animal Ecology 68: 1061-1078. BROWN, L., AND D. AMADON. 1968. Eagles, hawks and falcons of the world, volume 2. McGraw-Hill Book Company, New York, New York. BUSH, M. B., P. A. COLINVAUX, M. C. WEIMANN, D. R. PIPERNO, AND K.-B. LIU. 1990. Late Pleistocene temperature depression and vegetation change in Ecuadorian Amazonia. Quaternary Research 34: 330-345. BUSH M.B., M.C. MILLER, P.E. DE OLIVEIRA, AND P.A. COLINVAUX. 2002. Orbital forcing signal in sediments of two Amaz onian lakes. Journal of Paleolimnology 27 : 341–352. BUSH, M. B., D. R. PIPERNO, AND P. A. COLINVAUX. 1989. A 6,000 year history of Amazonian maize cultivation. Nature 340: 303-305. BUSH, M. B., M. R. SILMAN, AND D. H. URREGO. 2004. 48,000 Years of climate and forest change in a biodiversity hot spot. Science 303: 827-829. BYRD, K. M. 1976. Changing animal utilization patte rns and their implications: southwest Ecuador (6500 B.C.-A.D. 1400). Ph.D. Di ssertation, University of Florida, Gainesville, Florida. CAMPBELL, K. E. 1976. The late Pleistocene avifauna of La Carolina, southwestern Ecuador, p. 155-168. In S. L. Olson [Ed.], Collected papers in avian paleontology honoring the 90th birthday of Alexander Wetmore. Smithsonian Contributions to Paleobiology 27. CAMPBELL, K. E. 1979. The non-passerine Pleistocene av ifauna of the Talara tar seeps, northwestern Peru. Royal Ontario Museum, Life Sciences Contribution 118: 1-203. CAMPBELL, K. E. 1982. Late Pleistocene events along the coastal plain of northwestern South America, p. 423-440. In G. T. Prance [Ed.], Biologi cal diversification in the tropics. Columbia University Press, New York, New York.

PAGE 367

351 CANTALAMESSA, G., AND C. DI CELMA. 2004. Origin and chronology of Pleistocene marine terraces of Isla de la Plata and of flat, gently dipping surfaces of the southern coast of Cabo San Lorenzo (Manab, Ecuador). Journal of South American Earth Sciences 16: 633. CENTER FOR RESEARCH ON THE CULTURAL AND BIOLOGICAL DIVERSITY OF ANDEAN RAINFORESTS. 2000. Oyacachi – la gente y la biodiversidad. Centro para la Investigacin de la Center for Research on the Cultural and Biological Diversity of Andean Rainforests (DIVA), Copenhage n, Denmark, and Ediciones Abya-Yala, Quito, Ecuador. CHAPMAN, F. M. 1926. The distribution of bird-life in Ecuador a contribution to a study of the origin of Andean bi rd-life. Bulletin of the American Museum of Natural History 55: 1-784. CHASE, T. 1988. Restos faun sticos, pages 171-178. In K. E. Stothert, La prehistoria temprana de la Pennsula de Santa El ena, Ecuador: cultura Las Vegas (with contributions by T. Chase, D. Piperno, E. Wing, and D. Ubelaker). Miscelnea Antropolgica Ecuadoriana, Serie Monogrfi ca (Museums of the Central Bank of Ecuador, Guayaquil, Ecuador). CHESSER, R. T. 2004. Molecular systematics of New World suboscine birds. Molecular Phylogenetics and Evolution 32: 11. CHURCHER, C. S. 1959. Fossil Canis from the tar pits of La Brea, Peru. Science 130: 564565. CHURCHER, C. S. 1966. The insect fauna from the Ta lara tar-seeps, Peru. Canadian Journal of Zoology 44: 985-993. CLEMENT, A. C., M. A. CANE, AND R. SEAGER. 2001. An orbitally dr iven tropical source for abrupt climate change. Journal of Climate 14: 2369-2375. CLEMENT, P. 2000. Thrushes. Princeton University Press, Princeton, New Jersey. COLINVAUX, P. A., M. B. BUSH, AND M. STEINITZ-KANNAN. 1997. Glacial and Postglacial Pollen Records from the Ecuadorian Ande s and Amazon. Quaternary Research 48 : 69. COLINVAUX, P. A., K. OLSON, AND K.-B. LIU. 1988. Late-glacial and Holocene pollen diagrams from two endorheic lakes of the inter-Andean plateau or Ecuador. Review of Palaeobotany and Palynology 55: 83-99. COLINVAUX, P. A., AND E. K. SCHOFIELD. 1976. Historical ecology in the Galpagos Islands: I. a Holocene pollen r ecord from El Junco Lake, Isla San Cristbal. Journal of Ecology 64: 989-1012.

PAGE 368

352 COLTORTI, M., G. FICCARELLI, H. JAHREN, M. MORENO-ESPINOSA, L. ROOK, AND D. TORRE. 1998. The last occurrence of Pleisto cene megafauna in the Ecuadorian Andes. Journal of South American Earth Sciences 11: 581-586. COLWELL, R. K. 2005. EstimatS: Statistical estimati on of species richness and shared species from samples. Version 7.5. Universi ty of Connecticut, Storrs, Connecticut. 15 April 2005. CORREAL, G. 1999. Evidencias culturales pleistocn icas y del Temprano Holoceno en la Cordillera Oriental de Colombia : periodizacin tentativa, p. 63-73. In P. Ledergerber-Crespo [Ed.], Formativo S udamericano, una revaluacin. Ediciones Abya-Yala, Quito, Ecuador. CRACRAFT, J. 1968. The lacrimal-ectethmoid bone co mplex in birds: a single character matrix. American Midland Naturalist 80: 316-359. CRACRAFT, J. 1985a. Historical biogeography and pa tterns of differen tiation within the South American avifauna: ar eas of endemism, p. 49-84. In P. A. Buckley, M. S. Foster, E. S. Morton, R. S. Ridgely, a nd F. G. Buckley [Eds.]. Neotropical ornithology. Ornithological Monographs 36. CRACRAFT, J. 1985b. Monophyly and phylogenetic relations hips of the Pelecaniformes: a numerical cladistic an alysis. Auk 102: 834-853. CRESSWELL, W., M. HUGHES,R. MELLANBY, S. BRIGHT, P. CATRY, J. CHAVES, J. FREILE, A. GABELA, H. MARTINEAU, R. MACLEOD, F. MCPHIE, N. ANDERSON, S. HOLT, S. BARABAS, C. CHAPEL, AND T. SANCHEZ. 1999. Densities and habitat preferences of Andean cloud-forest birds in pristine and degraded habitats in north-eastern Ecuador. Bird Conservation International 9: 129-145. DEFRANCE, S. D. 2005. Late Pleistocene marine birds from southern Peru: distinguishing human capture from El Nioinduced windfall. Journal of Archaeological Science 32: 1131-1146. DEFRANCE, S. D., D. K. KEEFER, J. B. RICHARDSON, AND A. UMIRE ALVAREZ. 2001. Late paleo-indian coastal foragers: speciali zed extractive behavior at Quebrada Tacahuay, Peru. Latin American Antiquity 12: 413-426. DEMENOCAL, P. B. 2001. Cultural responses to climat e change during the Late Holocene. Science 292: 667-673. DODSON, C. H., AND A. H. GENTRY. 1991. Biological extincti on in western Ecuador. Annals of the Missouri Bo tanical Garden 78: 273-295. DUNNING, J. B. [Ed.]. 1993. CRC handbook of avian body masses. CRC Press, Boca Raton, Florida.

PAGE 369

353 EDMUND, A. G. 1965. A late Pleistocene fauna from the Santa Elena Peninsula, Ecuador. Royal Ontario Museum, Life Sc iences Contribution 63: 1-21. EISENBERG, J. F., AND REDFORD. 1999. Mammals of the Neotropics, volume 3. University of Chicago Press, Chicago, Illinois. EMMONS, L. H., AND F. FEER. 1997. Neotropical rainforest mammals, 2nd Edition. University of Chicago Press, Illinois. FEDUCCIA, A. 1973. Evolutionary trends in th e Neotropical ovenbirds and woodhewers. Ornithological Monographs 13: 1-69. FERDON, E. N. 1981. Holocene mangrove formations on the Santa Elena Peninsula, Ecuador: pluvial indicators or ecological response to physiographic changes. American Antiquity 46: 619-626. FICCARELLI, G., M. COLTORTI, M. MORENO-ESPINOSA, P.L. PIERUCCINI, L. ROOK, AND D. TORRE. 2003. A model for the Holocene extinction of the mammal megafauna in Ecuador. Journal of South Amer ican Earth Sciences 15: 835-845. FJELDS, J. 2001. Cartografiar la avif auna andina: una base ci entfica para establecer prioridades de conservacin, pages 125-152. In M. Kappelle and A. D. Brown [Eds.], Bosques nublados del neotrpic o. Instituto Nacional de Biodiversidad, Santo Domingo de Heredia, Costa Rica. GADE, D. W. 1999. Nature and culture in the Ande s. University of Wisconsin Press, Madison, Wisconsin. GARCA-MORENO, J., P. ARCTANDER, AND J. FJELDS. 1998. Pre-Pleistocene differentiation among chat-tyrants. Condor 100: 629-640. GEORGE, J. C., AND A. J. BERGER. 1966. Avian myology. Academic Press, New York, New York. GNECCO, C. 2003. Against ecological reductionism: La te Pleistocene hunter-gatherers in the tropical forests of northern South Am erica. Quaternary International 109: 13. GOTELLI, N.J., AND G.L. ENTSMINGER. 2004. EcoSim: Null models software for ecology. Version 7.7. Acquired Intelligence In c. & Kesey-Bear, Jericho, Vermont. 5 December 2004. GRAU, E. T., S. L. PEREIRA, L. F. SILVEIRA, E. HFLING, AND A. WAJNTAL. 2005. Molecular phylogenetics and biogeography of Neotropical piping guans (Aves: Galliformes): Pipile Bonaparte, 1856 is synonym of Aburria Reichenbach, 1853. Molecular Phylogenetics and Evolution 35: 637.

PAGE 370

354 GRAYSON, D. K. 1984. Quantitative zooarchaeology: topics in the analysis of archaeologica faunas. Academic Press, New York, New York. HAMES, R. B., AND W. T. VICKERS. 1982. Optimal diet breadth theory as a model to explain variability in Amazonian hunting. American Ethnologist 9: 358-378. HAMON, J. H. 1964. Osteology and paleontology of the passerine birds of the Reddik, Florida, Pleistocene. Geol ogical Bulletin 44: 1-210. HANCOCK, J., AND J. KUSHLAN. 1984. The herons handbook. Harper and Row Publishers, New York, New York. HANSEN, B. C. S., D. T. RODBELL, G. O. SELTZER, B. LE”N, K. R. YOUNG, AND M. ABBOTT. 2003. Late-glacial and Holocene vegetati onal history from two sites in the western Cordillera of southwestern Ec uador. Palaeogeography, Palaeoclimatology, Palaeoecology 194: 79-108. HARRIS, G. M., AND S. L. PIMM. 2004. Bird species’ tolerance of secondary forest habitats and its efffects on extinction. Conservation Biology 18: 1607-1616. HARRISON, P. 1983. Seabirds, an identificati on guide. Houghton Mifflin Company, Boston, Massachusets. HAUG, G. H., K. A. HUGHEN, D. M. SIGMAN, L. C. PETERSON, U. RHL. 2001. Southward migration of the Intertro pical Convergence Zone thr ough the Holocene. Science 293: 1304-1308. HECKENBERGER, M. J., A. KUIKURO, U. T. KUIKURO, J. C. RUSSELL, M. SCHMIDT, C. FAUSTO, B. FRANCHETTO. 2003. Amazonia 1492: Pristin e Forest or Cultural Parkland? Science 301: 1710-1714. HEINE, K. 2000. Tropical South America during th e Last Glacial Maximum: evidence from glacial, periglacial a nd fluvial records. Quaterna ry International 72: 7-21. HEUSSER, L. E., AND N. J. SHACKLETON. 1994. Tropical climatic va riation on the pacific slopes of the Ecuadorian Andes based on a 25,000-year pollen record from deepsea sediment core Tri 163-31B. Quaternary Research 42: 222-225. HIDALGO NISTRI, F. 1998. Los antiguos paisajes forestales del Ecuador: una reconstruccin de sus primitivos ecosist emas, 2nd edition. Ediciones Abya-Yala, Quito, Ecuador. HILTY, S. L., AND W. L. BROWN. 1986. Birds of Colombia. Princeton University Press, Princeton, New Jersey. HOFFSTETTER, R. 1986. High Andean mammalian fauna s during the Plio-Pleistocene, p. 218-245. In F. Vuilleumier and M. Monasterio [Eds.], High altitude tropical biogeography. Oxford University Press, New York.

PAGE 371

355 HFLING, E., AND H. M. F. ALVARENGA. 2001. Osteology of the shoulder girdle in the Piciformes, Passeriformes and related groups of birds. Zoologischer Anzeiger 240 : 196. HOOGHIEMSTRA, H., AND T. VAN DER HAMMEN. 1998. Neogene and Quaternary development of the neotropica l forest: the forest refugi a hypothesis, and a literature overview. Earth-Science Reviews 44: 147-183. HOWARD, H. 1929. The avifauna of Emeryville Sh ellmound. University of California Publications in Zoology. 32: 301-391. INTERNATIONAL UNION FOR THE CONSERVATION OF NATURE AND NATURAL RESOURCES (IUCN). 2004. 2004 IUCN Redlist of thre atened species. The IUCN Species Survival Commission, Gland, Switzerland. 14 December 2004. ISLER, M. L., AND P. R. ISLER. 1987. The tanagers: natural history, distribution, and identification. Smithsonian Inst itution Press, Washington, DC. JACKSON, C. S., AND A. J. BROCCOLI. 2003. Orbital forcing of Arctic climate: mechanisms of climate response and implications for continental glaciatio n. Climate Dynamics 21: 539-557. JOHNSGARD, P. A. 1993. Cormorants, darters, and pelicans of the world. Smithsonian Institution Press, Washington, DC. KATTAN, G. H., H. ALVAREZ-LOPEZ, AND M. GIRALDO. 1994. Forest fragmentation and bird extinctions: San Antonio eighty year s later. Conservation Biology 8: 138-146. KENT, A. M., T. WEBBER, AND D. W. STEADMAN. 1999. Distribution, relative abundance, and prehistory of birds on the Taraco Peninsula, Bo livian Altiplano. Ornitologa Neotropical 10: 151-178. KENT, S. 1993. Variability in faunal assemblages: the influence of hunting skill, sharing, dogs, and mode of cooking on faunal remain s at a sedentary Kalahari community. Journal of Anthropologic al Archaeology 12: 323-385. KNIG, C., F. WEICK, AND J.-H. BECKING. 1999. Owls: a guide to the owls of the world. Yale University Press, New Haven, Connecticut. KOUTAVAS, A., J. LYNCH-STIEGLITZ, T. M. MARCHITTO, AND J. P. SACHS. 2002. El Ninolike pattern in Ice Age tropical Pacific s ea surface temperature. Science 297: 226230. LANDE, R. 1988. Genetics and demography in biological conserva tion. Science 241: 1455-1460.

PAGE 372

356 LECK, C. F. 1979. Avian extinctions in an isolated tropical wet-forest preserve, Ecuador. Auk 96: 343-352. LEMON, R. R. H., AND C. S. CHURCHER. 1961. Pleistocene geology and paleontology of the Talara region, Peru. American Journal of Science 259: 410-429. LIU, K.-B., AND P. A. COLINVAUX. 1985. Forest changes in the Amazon basin during the last glacial maximum. Nature 318 : 556. LIVEZEY, B. C. 1998. A phylogenetic analysis of th e Gruiformes (Aves) based on morphological characters, with an emphasi s on the rails (Ralli dae). Philosopihical Transactions of the Royal Societ y of London, Series B. 353: 2077-2151. LYNCH, T. F., AND S. POLLOCK. 1981. La arqueologa de la cueva negra de Chobshi. Miscelnea Antropolgica Ecuatoriana 1: 92-119. MACFADDEN, B. J., AND B. J. SHOCKEY. 1997. Ancient feeding ecology and niche differentiation of Pleistocene mammalia n herbivores from Tarija, Bolivia: morphological and isotopic evid ence. Paleobiology 23: 77-100. MADGE, S., AND H. BURN. 1988. Waterfowl: an identificatio n guide to the ducks, geese and swans of the world. Houghton Mifflin Company, Boston, Massachusetts. MAGURRAN, A. E. 1988. Ecological diversity and its meas urement. Princeton University Press, Princeton, New Jersey. MANEGOLD, A., G. MAYR, AND C. MOURER-CHAUVIR. 2004. Miocene songbirds and the composition of the European Passeriform avifauna. 121: 1055-1060. MANNE, L. L., T. M. BROOKS, AND S. L. PIMM. 1999. Relative riskof extinction of passerine birds on continents and islands. Nature 399: 258-261. MARCHANT, S. 1958. The birds of Santa Elena Peninsul a, S. W. Ecuador. Ibis 100: 349387. MARCHETTI, K., T. PRICE, AND A. RICHMAN. 1995. Correlates of wing morphology with foraging behaviour and migra tion distance in the genus Phylloscopus . Journal of Avian Biology 26: 177. MARTIN, P. S., AND D. W. STEADMAN. 1999. Prehistoric extinctions on islands and continents, p. 17-55. In R. D. E. MacPhee [Ed.], Extin ctions in near time. Kluwer Academic/Plenum Publishers, New York. MASUCCI, M. A. 1992. Ceramic change in the Guangala Phase, southwest Ecuador: a typology and chronology. Ph.D. Dissertati on, University Microfilms, Ann Arbor, MI.

PAGE 373

357 MASUCCI, M. A. 1995. Marine shell bead production and the role of domestic craft activities in the economy of the Gua ngala phase, southwest Ecuador. Latin American Antiquity 6: 70-84. MASUCCI, M. A., AND A. MACFARLANE. 1997. An application of geological survey and ceramic petrology to provenance studies of Guangala Phase ceramics of ancient Ecuador. Geoarchaeology 12: 765. MATTHIESEN, D. G. 1988. Preceramic animal use at Huaca Prieta, pages 18-30. In E. S. Wing and J. C. Wheeler [Eds.], Economic prehistory of the Central Andes. BAR International Series 427. MAYEWSKI, P. A., E. E. ROHLING, J. C. STAGER, W. KARLN, K. A. MAASCH, L. D. MEEKER, E. A. MEYERSON, F. GASSE, S. VAN KREVELD, K. HOLMGREN, J. LEETHORP, G. ROSQVIST, F. RACK, M. STAUBWASSER, R. R. SCHNEIDER, E. J. STEIG. 2004. Holocene climate variability. Quaternary Research 62: 243-255. MOLINA-CRUZ, A., AND J. C. HERGUERA. 2002. Paleoceanographic evolution of backwater in the Nazca region, southeastern Pacific, during la te Pleistocene. Revista Mexicana de Cienci as Geolgicas 19: 252-259. MOY, C. M., G. O. SELTZER, D. T. RODBELL, AND D. M. ANDERSON. 2002. Variability of El Nio/Southern Oscillation activity at millennial timescales during the Holocene epoch. Nature 420: 162-165. NEWSON, L. A. 1995. Life and death in early colonial Ecuador. University of Oklahoma Press, Norman, Oklahoma. PADUANO, G. M., M. B. BUSH, P. A. BAKER, S. C. FRITZ, AND G. O. SELTZER. 2003. A vegetation and fire history of Lake T iticaca since the Last Glacial Maximum. Palaeogeography, Palaeoclimat ology, Palaeoecology 194: 259-279. PAULSEN, A. C. 1976. Environment and empire: clima tic factors in prehistoric Andean culture change. Worl d Archaeology 8: 121-132. PEARSALL, D. M. 1988. La produccin de alimentos en Real Alto: la aplicacin de las tcnicas etnobotnicas al problema de la subsitencia en el perodo formativo ecuatoriano. Biblioteca Ecuatoriana de Arqueologa, Volumen 2. Corporacin Editora Nacional, Quito, Ecuador. PEARSALL, D. M. 2004. Plants and people in ancient Ecuador: the ethnobotany of the Rio Jama valley. Case Studies in Archaeol ogy Series, Wadsworth/Thompson Learning, Belmont, California. PERREIRA, S. L., A. J. BAKER, AND A. WAJNTAL. 2002. Combined nuclear and mitochondrial DNA sequences resolve generi c relationships within the Cracidae (Galliformes, Aves). Systematic Biology 51: 946-958.

PAGE 374

358 PIPERNO, D. R., T. C. ANDRES, AND K. E. STOTHERT. 2000a. Phytoliths in Cucurbita and other Neotropical Cucurbitaceae and their occurrence in early archaeological sites from the lowland American tropics. J ournal of Archaeological Science 27: 193208. PIPERNO, D. R., AND D. M. PEARSALL. 1998. The origins of agriculture in the lowland Neotropics. Academic Press, San Diego, California. PIPERNO, D. R., A. J. RANERE, I. HOLST, AND P. HANSELL. 2000b. Starch grains reveal early root crop horticulture in the Pana manian tropical forest. Nature 407: 894-897. PIPERNO, D. R., AND K. E. STOTHERT. 2003. Phytolith evidence for early Holocene Cucurbita domestication in southwes t Ecuador. Science 299: 1054-1057. POZORSKI, S. 1983. Changing subsistance priorities and early settleme nt patterns on the north coast of Peru. Jour nal of Ethnobiology 3: 15-38. PRADO, J. L., M. T. ALBERDI, B. AZANZA, B. SNCHEZ, AND D. FRASSINETTI. 2005. The Pleistocene Gomphotheriidae (Proboscidea ) from South America. Quaternary International 126-128: 21-30. QUITMYER, I. R. 2004. What kind of data are in the back dirt? An experiment on the influence of screen size on optimal data recovery. Archaeofauna 13: 109-129. RAYMOND, J. S. 2003. Social formations in the west ern lowlands of Ecuador during the Early Formative, pages 33-67. In J. S. Raymond and R. L. Burger [Eds.], Archaeology of formative Ecuador. Du mbarton Oaks Research Library and Collection, Washington, DC. REDFORD, K. H., AND J. G. ROBINSON. 1987. The game of choice: patterns of indian and colonist hunting in the Neotropics. American Anthropologist 89: 650-667. REITZ, E. J., AND M. A. MASUCCI. 2004. Guangala fishers and farmers: a case study of animal use at El Azcar, southwestern Ecuador. University of Pittsburgh Press, Pittsburgh, Pennsylvania, and Edic iones Libri Mundi, Quito, Ecuador. REITZ, E. J., AND D. H. SANDWEISS. 2001. Environmental change at Ostra Base Camp, a Peruvian pre-ceramic site. Journa l of Archaeological Science 28: 1085. REMSEN, J. V., AND S. W. CARDIFF. 1990. Patterns of elevational and latitudinal distribution, including a “nic he shift,” in some guans (Cracidae) of the Andes. Condor 92: 970-981. RENJIFO, L. M. 1999. Composition Changes in a Subandean avifauna after long-term forest fragmentation. Cons ervation Biology 13: 1124-1139.

PAGE 375

359 RIDGELY, R. S., T. F. ALLNUTT, T. BROOKS, D. K. MCNICOL, D. W. MEHLMAN, B. E. YOUNG, AND J. R. ZOOK. 2003. Digital Distribution Maps of the Birds of the Western Hemisphere, version 1.0. Na tureServe, Arlington, Virginia. RIDGELY, R. S., AND P. J. GREENFIELD. 2001a. The birds of Ecuador, volume 1: status, distribution and taxonomy. Cornell Univ ersity Press, Ithaca, New York. RIDGELY, R. S., AND P. J. GREENFIELD. 2001b. The birds of Ecuador, volume 2: field guide. Cornell University Press, Ithaca, New York. RIESING, M. J., L. KRUCKENHAUSER, A. GAMAUF, AND E. HARING. 2003. Molecular phylogeny of the genus Buteo (Aves: Accipitridae) based on mitochondrial marker sequences. Molecular Phylogene tics and Evolution 27: 328. ROBBINS, M. B., N. KRABBE, G. H. ROSENBERG, AND F. SORNOZA. 1994. The tree line avifauna at Cerro Mongus, Prov. Carchi, nor theastern Ecuador. Proceedings of the Academy of Natural Sciences of Philadelphia 145: 209-216. RODBELL, D. T., G. O. SELTZER, D. M. ANDERSON, M. B. ABBOT, D. B. ENFIELD, AND J. H. NEWMAN. 1999. An ~15,000-year record of El Nio-driven alluviation in southwestern Ecuador. Science 283: 516-520. SAMANIEGO, P., M. MONZIER, C. ROBIN, AND M. L. HALL. 1998. Late Holocene eruptive activity at Nevado Cayambe Volcano, Ec uador. Bulletin of Volcanology 59: 451459. SNCHEZ, B., J. L. PRADO, AND M. T. ALBERDI. 2004. Feeding ecology, dispersal, and extinction of South American Pleist ocene gomphotheres (Gomphotheriidae, Proboscidea). Paleobiology 30: 146-161. SARMIENTO, F. O., AND L. M. FROLICH. 2002. Andean cloud forest tree lines: naturalness, agriculture and the human dimension. M ountain Research and Development 22: 278. SCHIMMELMANN, A., C. B. LANGE, AND B. J. MEGGERS. 2003. Palaeoclimatic and archaeological evidence for a ~200-yr recurrence of floods and droughts linking California, Mesoamerica and South Ameri ca over the past 2000 years. Holocene 13: 763-778. SCHORGER, A. W. 1955. The Passenger Pigeon: its natural history and extinction. University of Wisconsin Press, Madison, Wisconsin. SIBLEY, C. G., AND J. E. AHLQUIST. 1990. Phylogeny and classification of birds: a study in molecular evolution. Yale Univer sity Press, New Haven, Connecticut. SIEGEL-CAUSEY, D. 1988. Phylogeny of the Phalacrocoracidae. Condor 90: 885-905.

PAGE 376

360 SIERRA, R. [Ed.]. Propuesta preliminar de un sist ema de clasificacion de vegetacion para el Ecuador continental. Proyecto IN EFAN/GEF-BIRF and EcoCiencia, Quito, Ecuador. SIERRA, R., F. CAMPOS, AND J. CHAMBERLIN. 1999a. reas prioritarias para la conservacin de la biodiver sidad en el Ecuador continen tal. Un estudio basado en la biodiversidad de ecosistemas y su or nitofauna. Ministerio de Medio Ambiente, Proyecto INEFAN/GEF-BIRF, EcoCiencia , and Wildlife Conservation Society, Quito, Ecuador. SIERRA, R., C. CER”N, W. PALACIOS, AND R. VALENCIA. 1999b. Mapa de vegetacin del Ecuador continental. Escala 1:1. 000. Proyecto INEFAN/GEF-BIRF, Wildlife Conservation Society and EcoCiencia, Quito, Ecuador. STAHL, P. W. 1991. Arid landscapes and environm ental transformations in ancient southwestern Ecuador. Wo rld Archaeology 22: 346-359. STAHL, P. W. 1996. Holocene biodiversity: an ar chaeological perspective from the Americas. Annual Review of Anthropology 25: 105-126. STAHL, P. W. 2000. Archaeofaunal accumulation, frag mented forests, and anthropogenic landscape mosaics in the tropical lowlands of prehistoric Ecuador. Latin American Antiquity 11: 241-257. STAHL, P. W. 2003. The zooarchaeological record from formative Ecuador, pages 175212. In J. S. Raymond and R. L. Burger [Eds.], Archaeology of formative Ecuador. Dumbarton Oaks Research Librar y and Collection, Washington, DC. STAHL, P. W. 2004. Neotropical zooarchaeology in Ecuador, p. 203-220. In G. L. Mengoni [Ed.], Zooarchaeology of South Am erica. BAR International Series 1298. STAHL, P. W. 2005. An exploratory osteolog ical study of the muscovy duck ( Cairina moschata ) (Aves: Anatidae) with implications for neotropical archaeology. Journal of Archaeological Science 32: 915-929. STAHL, P. W., AND J. S. ATHENS. 2001. A high elevational zooa rchaeological assemblage of the northern Andes of Ecuador. Jo urnal of Field Archaeology 26: 161-176. STAHL, P. W., AND P. NORTON. 1987. Precolumbian animal domesticates from Salango, Ecuador. American Antiquity 52: 382-391. STEADMAN, D. W. 1980. A review of the osteology a nd paleontology of turkeys (Aves: Meleagridinae). Contributions in Science, Natural History Museum of Los Angeles County 330: 131-207. STEADMAN, D. W. 1982. The origins of Darwin’s finc hes (Fringillidae, Passeriformes). Transactions of the San Diego So ciety of Natural History 19: 279-296.

PAGE 377

361 STEADMAN, D. W. 1997. Extinctions of Polynesian bird s: reciprocal impacts of birds and people. In P. V. Kirch and T. L. Hunt [Eds.] , Historical ecology in the Pacific islands: prehistoric environmental and landscape change. Yale University Press, New Haven, Connecticut. STEADMAN, D. W., AND G. S. MORGAN. 1985. A new species of bullfinch (Aves: Emberizinae) from a Late Quaternary cav e deposit on Cayman Brac, West Indies. Proceedings of the Biological Society of Washington 98: 544-553. STEADMAN, D. W., M. P. TELLKAMP, AND T. A. WAKE. 2003. Prehistoric expoitation of birds on the Pacific coast of Ch iapas, Mexico. Condor 105: 572-579. STOTHERT, K. E. 1983. Review of the early preceramic complexes of Santa Elena Paninsula, Ecuador. American Antiquity 48: 122-127. STOTHERT, K. E. 1985. The preceramic Las Vegas culture of coastal Ecuador. American Antiquity 50: 613-637. STOTHERT, K. E. 1988. La prehistoria temprana de la Pennsula de Santa Elena, Ecuador: cultura Las Vegas (with contributions by T. Chase, D. Piperno, E. Wing, and D. Ubelaker). Miscelnea Antropolgica Ecua doriana, Serie Monogrfica (Museums of the Central Bank of Ecuador, Guayaquil, Ecuador) 10: 1-271. STOTHERT, K. E. 1997. Un sitio de Guangala Temprano en el suroeste del Ecuador, second edition. Banco Central del Ecuador, Guayaquil, Ecuador. STOTHERT, K. E., D. R. PIPERNO, AND T. C. ANDRES. 2003. Terminal Pleistocene/Early Holocene human adaptation in coastal Ecuador: the Las Vegas evidence. Quaternary Interna tional 109-110: 23-43. STOTZ, D. F., J. W. FITZPATRICK, T. A. PARKER, AND D. K. MOSKOVITS. 1996. Neotropical birds: ecology and conservation. University of Chicago Press, Chicago, Illinois. TELLKAMP, M. P. 1999. The understory bird community of a fragmented premontane forest in the western Andes of Ecuador . MS Thesis, University of Florida, Gainesville, Florida. THOMAS, M. F. 2000. Late Quaternary environmental changes and the alluvial record in humid tropical environments. Quat ernary International 72: 23-36. TOMIATI C., AND L. ABBAZZI. 2002. Deer fauna from Pleistoc ene and Holocene localities of Ecuador (South America). Geobios 35: 631. UBELAKER, D. H. 1988. Restos de esqueletos humanos del sitio OGSE-80, pages 105132. In K. E. Stothert, La prehistoria tempra na de la Pennsula de Santa Elena, Ecuador: cultura Las Vegas (w ith contributions by T. Chase, D. Piperno, E. Wing, and D. Ubelaker). Miscelnea Antro polgica Ecuadoriana, Serie Monogrfica (Museums of the Central Bank of Ecuador, Guayaquil, Ecuador).

PAGE 378

362 UBELAKER, D. H. 1997. Restos humanos esquelticos de OGSE-MA-172, un sitio Guangala Temprano en la costa del Ec uador, second edition. Banco Central del Ecuador, Guayaquil, Ecuador. VAURIE, C. 1968. Taxonomy of the Cracidae (Aves). Bulletin of the American Museum of Natural History 138: 131-260. VEINTIMILLA, C. I. 2000. Reconstruccin paleo-ambiental y e volucin agrcola en el valle del Ro Jama, Provincia de Manab , Ecuador. Miscelnea Antropolgica Ecuatoriana, Revista del Museo Antropol gico del Banco Central del EcuadorGuayaquil 9: 135-151. VICKERS, W. T. 1989. Los Sionas y Secoyas, su ad aptacin al ambiente amaznico. Coleccin 500 Aos. Ediciones Abya-Yala and Movimientos Laicos para Amrica Latina, Quito, Ecuador. VILLALBA, M. 1988. Cotocollao: una aldea formativa del valle de Quito. Miscelnea Antropolgica Ecuatoriana, Serie Monogrf ica 2. Museo del Banco Central del Ecuador, Quito, Ecuador. VON DEN DRIESCH, A. 1976. A guide to the measurement of animal bones from archaeological sites. Peabody Museum of Archaeology and Ethnology Bulletin 1: 1-137. WENG, C., M. B. BUSH, AND J. S. ATHENS. 2002. Holocene climate change and hydrarch succession in lowland Amazonian Ecua dor. Review of Palaeobotany and Palynology 120: 73-90. WENG, C., M. B. BUSH, AND A. J. CHEPSTOW-LUSTY. 2004. Holocene changes of Andean alder ( Alnus acuminata ) in highland Ecuador and Pe ru. Journal of Quaternary Science 19: 685. WILLE, M., H. HOOGHIEMSTRA, R. HOFSTEDE, J. FEHSE, AND J. SEVINK. 2002. Upper forest line reconstruction in a deforested area in northern Ecuador based on pollen and vegetation analysis. Jour nal of Tropical Ecology 18: 409. WILLIAMS, M. I. 1999. Systematics of Neotropical pa rrots (Psittacidae) with emphasis on Deroptyus and Triclaria . MSc Thesis, University of Fl orida, Gainesville, Florida. WILLIS, K. J., L. GILLSON, AND T. M. BRNCIC. 2004. How “virgin”is virgin rainforest?. Science 304: 402-403. WING, E. S. 1988. Dusicyon sechurae , en contextos arquelgicos tempranos, pages 179185. In K. E. Stothert, La prehistoria tempra na de la Pennsula de Santa Elena, Ecuador: cultura Las Vegas (w ith contributions by T. Chase, D. Piperno, E. Wing, and D. Ubelaker). Miscelnea Antro polgica Ecuadoriana, Serie Monogrfica (Museums of the Central Bank of Ecuador, Guayaquil, Ecuador).

PAGE 379

363 WOOLFENDEN, G. E. 1961. Postcranial osteology of the wa terfowl. Bulletin of the Florida State Museum 6: 1-129. ZEIDLER, J. A. 2003. Appendix A: Formative Period ch ronology for the coast and western lowlands of Ecuador, pages 487-527. In J. S. Raymond and R. L. Burger [Eds.], Archaeology of formative Ecuador. Du mbarton Oaks Research Library and Collection, Washington, DC. ZEIDLER, J. A., AND J. S. ISAACSON. 2003. Settlement process and historical contingency in the western Ecuadorian Formative, p. 69-123. In J. S. Raymond and R. L. Burger [Eds.], Archaeology of formative Ecuador. Dumbarton Oaks Research Library and Collection, Washington, DC. ZUSI, R. L., AND G. D. BENTZ. 1978. The appendicular myology of the Labrador Duck ( Camptorhynchus labradorius ). Condor 80: 407-418.

PAGE 380

364 BIOGRAPHICAL SKETCH Ecuadorian by birth, I was raised in Ec uador, Germany, and El Salvador. My undergraduate studies took me to the University of Victoria in British Columbia where I was planning to become a botanist. By the time I was ready for graduate education, I had spent several years in Ecua dor working on several conservation projects and as a naturalist guide. These activities opened my ey es to the beauty of birds. I began my graduate career at the Department of Wild life Ecology and Conservation, University of Florida. My MS thesis focused on bird communities in fragmented landscapes of the Ecuadorian Andes, a subject sti ll of great interest to me. A fascination with biological structure and historical bioge ography led me to this study in zooarchaeology on bird bones. Exploring nature and cu ltures is something I am looking forward to do in future, together with my wife and children.