5125:... 3.... .2. :11... :.,. .l n . . .1: .. 3.. .23.... 7.... a J... 2. .1. .z. . r 4‘ :. ... 3.. 5:. . : ?... .. .2 . 71:... ‘ 33.2.} .. . v. _:. .... i; i .. .22.? . a; a. 3,. 2.. s. 2 3.“ .... . LIBRARY ' Michigan State ’19 University This is to certify that the thesis entitled Paleoecology of Hawaiian Geese: Stable Isotope Analysis presented by Andery Ellen Calkins has been accepted towards fulfillment of the requirements for Masters degree in Environmental Geosciences MW Major professor Date L24» 1‘" 0-7639 MS U is an Affirmative Action/Equal Opportunity Institution PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE JGLZ 5% 266’s 6/01 cJClRC/DateDuopflb-DJS PALEOECOLOGY OF HAWAIIAN GEESE: STABLE ISOTOPE ANALYSIS By Andery Ellen Calkins A THESIS Submitted to Michigan State University In partial fulfillment of requirements For the degree of MASTER OF SCIENCE Department of Geological Sciences 2001 ABSTRACT PALEOECOLOGY OF HAWAIIAN GEESE: STABLE ISOTOPE ANALYSIS By Andery Ellen Calkins An evolutionary radiation of Hawaiian geese has been closely linked to the Canada Goose (Branta canedensis); the descendent spades include the extant Hawaiian Goose or Nene (B. sendvicensis), the extinct woods-walking goose (B. hylobadistes), and the extinct giant Hawaiian goose (anta sp.). Stable carbon and nitrogen isotope measurements of bone collagen were used to interpret dietary changes and habitat niche divisions between the three species. The osteological samples analyzed came from habitats that varied in precipitation, vegetation, elevation, and soil characteristics. In addition to the geese. samples of other vertebrate taxa were used to provide a broad basis for dietary comparison. The preservation of the collagen within the samples, which ranged in age from pre-contact Holocene to modern, was found to be exceptional. This is the first study in which isotopic data were used to relate speciation to changes in dietary niches of closely related bird species, and it could serve as a model for elucidating similar insular evolutionary radiations. ACKNOWLEDGEMENTS First of all, I would like to thank my committee members for helping me throughout the research and completion of this thesis. Mike Gottfried has been instrumental in finding an interesting project. and for making graduate school a more interesting experience. Peggy Ostrom, the isotope expert, has been particularly helpful with the stable isotope analysis of this research. Pamela Rasmussen, the bird expert, has provided thoughtful and intriguing comments throughout the writing of this thesis. I would also like to sincerely thank my honorary committee member, Helen James of the Smithsonian Institute. for allowing me the privilege of studying a small aspect of her ‘babies’, the Hawaiian birds. I also extend my appreciation to Loretta Knutson, the graduate secretary for the Department of Geological Sciences. She has been a great help in keeping me organized and on-time throughout my graduate career, particularly during the last couple of months. Also, I would like to thank Carla Kishinami at the Bernice P. Bishop Museum in Honolulu, Hawaii for providing this research project with modern Nene and comparison taxa. I would like to acknowledge the help of my friends, without all of them I could not have made it. They have been a support throughout this process. one that I may have fallen without. For his wonderful, insurmountable help during the last couple of months, I would like to thank with my deepest love, Adam Pituch. No one else could have calmed me down like he has. My closest friend. Amy Grohoski, I would like to thank for giving me the confidence to think I can do it, and for helping me relieve the stress when sorely needed. I would also like to iii thank two fellow graduate students who had helped me make it through my graduate career. I think it would have tumad out differently if you had not been there. Thank you very much, Tucker McNulty and Colleen Masterson. Finally, I would like to extend my appreciation to my parents. They have always believed in my abilities, even when I hadn’t. They have always been supportive and understanding, not just during my graduate career, but throughout my life. I hope that now I will reach the kind of “stable life” that they have always wished for me; I only want to make them proud. iv TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES INTRODUCTION Geology and Climatology 2 Hawaiian Archaological History 5 Insular Evolution 8 Hawaiian Terrestrial Geese 11 STABLE ISOTOPE OVERVIEW Carbon 17 Nitrogen 20 Collagen 22 METHODOLOGY RESULTS AND DISCUSSION Assessment of Isotopic Indiganeity 26 Isotopic Results 27 The Hawaiian Goose or Nana 28 Sub-fossil Geese 33 Sub-fossil and Archaological Data 35 Modern Ecosystem 37 CONCLUSIONS TABLES AND FIGURES REFERENCES vi vii 16 24 26 39 41 TABLE 1 TABLE 2 TABLE 3 TABLE 4 LIST OF TABLES List of samples analyzed for focus geese taxa. List of samples analyzed for comparison purposes. Collagenous protein yields, C to N ratios, C and N isotope values for focus geese taxa. Collagenous protein yields, C to N ratios, C and N isotope values for comparison taxa. FIGURE 1 FIGURE 2 FIGURE 3 FIGURE 4 FIGURE 5 FIGURE 6 FIGURE 7 FIGURE 8 FIGURE 9 FIGURE 10 FIGURE 11 LIST OF FIGURES Map of the Hawaiian Archipelago. Map of the Hawaiian Archipelago during presence of Maui Nui. ' Cladogram of large Canada Goose and focus species. \Mng to lag bone (radius vs. tibiotarsus) comparison for Canada Goose and focus species. Summary of methodology. Percent weight yield of bone collagen. Hawaiian Goose isotopic results. Geese averaged isotopic results. Sub-fossil geese isotopic results. Sub-fossil and archeological isotopic results. Modem ecosystem isotopic results. Introduction The Canada Goose (Branta sandvicensis), a common and widespread species native to North America, is considered to be the sister taxon to three endemic taxa of Hawaiian geese. Related to this radiation, and associated with morphological changes from the ancestral to the descendant species, I examined dietary differences and habitat partitioning between Hawaiian species of Branta. The three focus species are (1) the extant Hawaiian Goose or Nana (Branta sandvicansis), (2) the extinct woods-walking goose (B. hylobadistas), and (3) the extinct giant Hawaiian goose (Branta sp.). These geese must have been important herbivores on the islands where they occurred, yet little is known about their dietary and habitat preferences. The following questions were addressed by analyzing stable carbon and nitrogen isotopic ratios of extracted collaganous bone protein from the three focus species, and from other taxa for comparison. Samples were selected frOm different elevations, habitats, and isolated versus overlapping ranges (T ables 1 and 2). I. What were the dietary preferences of the two extinct species? ll. Did habitat partitioning occur between taxa with overlapping ranges? Ill. How do alevational and climatic differences affect resource partitioning within and between species? IV. How do environmental parameters affect the isot0pic composition of the samples? V. Do the archaological and sub-fossil data shed light on conservation issues of the extant, but very rare, Nene? A key feature of this research is that it concerns an intarspecific comparison of closely related taxa from different ecological regimes. Significant aspects of this project include: this is the first isotopic study in which the effects of smciation and colonization are interpreted in terms of their relationship to dietary niches of closely related birds; this study will help elucidate the avian palaoeoology of Hawaii and Maui before human contact, leading to a better understanding of the evolutionary history of the archipelago’s avifauna; this research provides information concaming habitat partitioning in closely related organisms with overlapping ranges, which is important for understanding insular evolution and ecology; and, finally, information on the dietary preferences and ecology of ancient Nene populations may help in the development of conservation strategies for living Nene by supplying data related to appropriate dietary and habitat essentials for long-term viability. Geology and Climatology of Hawaii The Hawaiian islands are significant for, among other reasons, being the most isolated island chain in the world (Carson, 1981). The nearest continent is North America, 3,300 kilometers (2,400 miles) away. The chain of islands has been isolated to this degree from the time it began forming during the Paleocene (Carson and Clague, 1995). This isolation has led to the very high degree of endemism seen on the islands today and in the fossil record. It has been estimated that nearly 99% of the Hawaiian biota is endemic (Olson and James, 1982b). The Hawaiian archipelago consists of a long chain of islands and seamounts which successively formed over a fixed hot spot undemaath the moving Pacific Plate (Figure 1) (Carson and Clague, 1995; Wagner and Funk, 1995). The immersed seamounts at the northwest and began forming approximately 75 to 80 mya (Carson and Clague, 1995). A band in the chain occurred 43 mya, indicating a change in direction of movement of the Pacific Plate. As the plate moved northwesterly, successively younger islands formed to the southwest (Wagner and Funk, 1995). There are currently eight topographically significant islands, the oldest being Kauai which emerged about five mya (Olson and James, 1982a), and the youngest, Hawaii, which began forming 0.5 mya. The exact spatial relationships between the islands during the geologic past are uncertain. It is difficult to determine if connections existed between islands, due to unknown island areas and changing sea levels (Carson and Clague, 1995). It is known, however, that there was a four-island complex, Maui Nui, present during the Pleistocene (Figure 2) (Carson and Clague, 1995). The connection between the islands was achieved through lowering of sea level to expose previously submerged connections between the islands of Maui, Lanai, Molokai and Kahoolawi (Olson and James, 1982a; Carson and Clagua 1995). Maui Nui began to reseparate 300,000-400,000 ya, and was completely split into four islands about 150,000 ya (Carson and Clague, 1995). It is certain that Maui and Hawaii have never been connected by a land bridge; the closest distance between them has been estimated as 13 kilometers (8 miles) (Carson and Clague, 1995; Wagner and Funk, 1995). Hawaii, the most recently formed island, is still volcanically active. This island has the greatest relief of the eight main islands, perhaps simply because it has not been as subjected to erosion and subsidence (Carson and Clague, 1995). There are three active volcanoes on this island: Mauna Loa, Kilawa, and Hualalai (Wagner and Funk, 1995). Newly forming to the southeast of Hawaii is the Lo’ihi Seamount, which is still submerged, but which will soon (in geologic time) emerge as the youngest Hawaiian island (Wagner and Funk, 1995). There have been a few slight changes in climate during the time period encompassed by the samples used in this study (up to approximately 2600 ya). Compared to the modern climatic situation, drier conditions occurred from 1250 to 700 ya, followed by wetter conditions from 700 to 50 ya (Allen, 1997). During the colonization and radiation of the terrestrial geese, the Wisconsinian Glacial Period from 10,000 to 70,000 BP produced cooler and drier conditions within the archipelago (Hotchkiss at at, 2000). At the glacial maximum approximately 18,000 ya, the sea level was 100 meters lower (Allan, 1997). The current climate in Hawaii is most strongly influenwd by trade winds and steep topography (Ralph and Van Riper, 1985). The trade winds needy constantly blow northeasterly. These winds are quite moist, and produce high amounts of rainfall on the windward side of the volcanic mountains of the islands. At elevations above 2000 m and on the leeward side of the mountains, conditions are semi-arid except for rare patches (Olson and James, 1982b; Ralph and Van Riper, 1985). Kona storms, resulting fiom rare southerly winds, occur two to seven times per year; these storms provide almost all of the annual precipitation on the leeward slopes (Ralph and Van Riper, 1985) Hawaiian Archaological History Polynesians first arrived in the uninhabited Hawaiian Archipelago as early as 1800 B.P. (Paxinos, 1998). The Polynesians discovered that the surrounding waters were not productive enough to support them on marine resources alone, so they created extensive agricultural systems with crop plants introduced from their home islands (Ralph and Van Riper, 1985). They developed tiered fish ponds, crop cultivation in masic areas, and slash-and-bum field cultivation in drier regions (Olson and James, 1982a). The lowlands of the main islands were stripped of their native forests, and some smaller islands were completely denuded. The Polynesians also intentionally introduced foreign animals to the Hawaiian archipelago, including the dog, pig, and domestic fowl. These animals have been highly destructive to island avian populations by eating or chasing ground-nastars, rooting in the soil and disturbing ground-dwellers, grazing extensively on grasses, and invading niche spaces (Ralph and Van Ripar, 1985). Not only did the Polynesians drastically changed the habitat, but they also intentionally hunted the endemic av'rfauna. A native, David Malo, explains the use of the Nene by his people, "In its moulting season, when it comes down from the mountains, is the time when the bird-catchers try to capture it in the uplands, the motive being to obtain the feathers, which are greatly valued for making Kahili (ceremonial cape). Its body is excellent eating." (Kear and Berger, 1980). The large terrestrial flightless birds, such as the geese, ibis and other waterfoM, would have been easy prey for the Polynesians. The effect the Polynesians had on the Hawaiian Islands was extraordinary. The lowland forests are believed to have contained the most diverse assemblage of plants and animals on all the islands (Olson and James, 1984). This massive degree of native habitat loss had a catastrophic effect on the avian biota. To date, 35 species of extinct birds have been recorded, and, as of 1991, perhaps 22 more sub-fossil species remain to be described. This is particularly striking considering that there are only 55 non-fossil endemic birds on the Hawaiian islands (Olson and James, 1991). The European discovery of the Hawaiian Islands dates to Captain James Cook's 1778 voyage. When the crew arrived they found the islands densely populated with natives; an estimated Polynesian population of a quarter million people inhabited the islands (Ralph and Van Riper, 1985). Early explorers described the islands as having treeless, agricultural lowlands, and the smallest islands were completely bare of trees. Shortly after Europeans arrived on the Hawaiian islands, they extended the clearing of forests to higher elevations, both for wood harvesting and for grazing introduced livestock (Ralph and Van Riper, 1985). Needy one hundred percent of the archipelago’s lowlands (<1500 m) have been altered by humans; native vegetation can no longer be found in these regions (Athens, 1997). Europeans also introduced several animals which have been detrimental to the endemic avifauna: cats, mongoose, two species of rats, cattle, goats, sheep and insects such as a carnivorous ant, parasitic flies and wasps (Ralph and Van Riper, 1985). The Europeans also brought with them domesticated fowl and other domesticated birds, which canied devastating non-native diseases, including Avian Pox Vlrus (Ralph and Van Riper, 1985). Documentation of the archipelago's endemic birds has spanned the last two centuries; during this time many rare species have likely disappeared before being recorded (Olson and James, 1991). Some birds may have been greatly reduced in numbers by the Polynesians and then finally eradicated by European occupation. For example, the Darkme Petrel was once common on all of the main islands, but is now found only at high altitudes on Maui and Hawaii. This drastic decrease is due both to hunting and predation by Polynesian- and European-introduced mammals (Olson and James, 1984). Historically, most of the endemic birds have been restricted to the higher elevation wet forests. This, however, is an artifact because fossil sites in lowland areas have representatives of the montane species, demonstrating that these birds once had a much larger range (Olson and James, 1982a). Insular Evolution Endemic species, those that are entirely restricted to one particular area, are characteristic of isolated islands. Islands offer an environment very conducive to rapid evolution and radiation. Speciation results from biological hardship, presence of an isolating mechanism, and geographic isolation (Mueller- Dombois, 1981). The extreme geographic isolation of Hawaii has provided a screen to the number and type of colonizers that reached the islands; only long- distance dispersars were capable of colonization. The native fauna lacked all non-flying mammals, land reptiles, amphibians, and freshwater animals (Mueller- Dombois, 1981 ). Once a founding pepulation arrived on an island, it then spread into the available habitat and perhaps became specialized (lie, adaptation), thus separating the one original population into subpopulations (Mueller-Dombois, 1981). Where did the island species arise? One possibility is that a waif population from a distant continent may have colonized the island by chance, and than either remained relatively unchanged or undergone evolution and sometimes diversification. If spaciation did not occur, the species would retain the approximate morphology and genetic identity of the founding population (Carson, 1981). Alternatively, radiation and spaciation may have occurred, in which a new species is produced that is younger than the island it inhabits (Carson and Clague, 1995). Dispersal theory states that the species may have evolved on an older, pre-existing island, and then moved to nearby islands. In this way, the species is older than the island it inhabits (Carson, 1981; Carson and Clague, 1995). Finally, spaciation may occur within the island itself. There is little evidence for sympatric spaciation, thus, for spaciation to occur the population must be divided (Coyne and Price, 2000). Geographic isolation may divide a population, causing genetic separation, perhaps eventually producing a new species. Some examples of within-island geographic isolation mechanisms include multiple volcanic mountains, the windward versus leeward side of the island, and separation of land areas by lava flows forming 'kipukas”, which are sections of older vegetation sunoundad by a new lava flow (Carson, 1981). A common change exhibited on islands is the loss of dispersal ability (e.g., flightlessnass). Flight is a costly investment for birds to produce and than maintain (Gill, 1995; Chatterjee, 1997). Thus, when predation pressures are removed, maintaining flight capabilities may not enhance fitness. The development of flightlessness is associated with multiple morphological changes, including development of smaller wing bones, reduced size and eventual disappearance of the stemal keel, reduction in muscles for flight, obtuse articulation angle between the scapula and coracoid, and an unossifiad area between the ilium and ischium (Gill, 1995; Chatterjee, 1997). Development of flightlessness has often been linked to paedomorphosis, or retention of juvenile characters in the adult (Livezey, 1993). The increase in size (gigantism) often associated with loss of flight, may result in non-morphological changes such as: inmased longevity and thermodynamic efficiency, and a greater capacity fto avoid starvation (Livezey, 1993). These traits would have led to greater fitness in a predator-free environment. One group of birds particularly known for their tendency to evolve flightlessness in insular environments are the rails (Rallidae) (Livezey, 1998). A number of characteristics of these birds may have facilitated flightlessness: broad habitat range, high dispersal abilities in some and sedentary, behaviorally flightless, habits in others, and an opportunistic omnivorous diet in many (Livezey, 1998; Trewick, 1997). Most of the extinct rallids were flightless, and a diversity of different species are found on different islands throughout the world (Livezey, 1998; Trewick, 1997). Most taxa of flightless rails are believed to be derviad from separate founding events by a volant species, damonstating that spaciation can occur rapidly when flight capability is lost (T rewick, 1997). The first comprehensive model to explain island ecology and evolution was presented in 1967 by R. H. MacArthur and E. O. Wilson. In their book, mg WWW they endeavored to empirically explain the unusual fauna and flora often found on islands. They developed equations based upon such variables as land area, degree of isolation, and habitat diversity. Multiple regression analyses demonstrated that island area, which is consisted to habitat diversity, accounted for mom of the variation in species numbers. Basically, the greater the diversity of habitats (and, thus, more land area), the greater the number of species an area can support (MacArthur and Vlfilson, 1967). They also determined that there is an equilibrium between immigration and extinction, which leads to a 'stable" level of diversity. They found that there seem to be limits to the number of individuals present on an 10 island, such that if one species becomes more abundant, another must likewise decline in numbers (MacArthur and Wilson, 1967). Hawaiian Terrestrial Geese The evolution of the Nana, the giant Hawaiian goose and the woods- walking goose is thought to have originated with the Canada Goose (Branfa canadensis) (Figure 3) (Paxinos, 1998). It has been suggested that a flock of large-bodied Canada Geese or a very close relation was “blown off course” while migrating, and then became established within the Hawaiian archipelago (H. James, personal communication). As these three species evolved, they developed novel methods of exploiting. the environment for resources. Instead of continuing as water-dependent geese, they became terrestrial herbivores that did not require open water for feeding or nesting. According to a molecular phylogenetic study by Paxinos (1998), the large Canada Goose is basal to the evolutionary radiation of geese that occurred in the Hawaiian archipelago. The modern Canada Goose feeds mainly upon grass seeds in the summer, and various available plant material, such as waste corn and acorns, in the winter (Hanson, 1965). Thus, the basal taxon occupied a 'browsar’ dietary niche, consuming a variety of plant material versus “gems” that consume entirely grasses. Until recently, all subspecies of Canada Geese migrated between summer and winter locales. Pair bonds formed in the 11 wintering grounds, and breeding only occurred near water environments (Hanson, 1965). The evolutionary changes evident in the focus species relative to Canada Geese represent adaptations to an insular environment At colonization by a ‘proto-typical large Canada Goose” (Paxinos, 1998), there were no terrestlial predators present in the archipelago, thus allowing the geese to reduce predator avoidance mechanisms and behaviors. The archipelago also offered the geese a year-round stable environment with abundant resources. This allowed the cessation of migratory behavior. Migration, or dispersal tendencies, would have rendered the birds unfit Thus, reduwd flight capability and even complete loss of flight arose in this insular radiation of geese (Figure 4). A convergent radiation occurred in dabbling ducks, leading to a group of very large flightless herbivores, the moa-nalos or “lost fowl” (Olson and James, 1991 ; James and Burney, 1997). In comparison to the geese, the moa-nalos had a more reduced pectoral apparatus, and showed an increase in the size of their pelvic elements and overall body (Olson and James, 1991). Moa-nalos are believed to have fed a leafy material (is. browsers), and may have been specialized for consuming ferns. The remains of moa-nalos have been found distributed on Oahu, Molokai, and Maui, and typically at low elevation scrubland and forested sites, but a few have been found in high elevation rainforest on Maui (James and Burney, 1997). Thus, the moa-nalos would have been present in the range of the Nene and Woods-walking Goose. All of these birds occupied an herbivorous niche, and, more specifically, most likely a browser dietary niche, 12 but the extent to which these flightless geese and 'ducks’ competed with one another is uncertain. Bronte sandvlcensls (None) The Nene is the only terrestrial goose to survive human arrival and colonization of the Hawaiian Archipelago. Interestingly, it is also the only endemic terrestrial goose that retained full flight capabilities; its sister taxa were all flightless. Since European colonization of the archipelago in 1778, the Nene were known only to live at high elevations on Hawaii in a habitat of scrubby trees and sparse ground cover. However, fossil discoveries have shown that before human colonization of the archipelago, the Nene was distributed throughout the main islands and was not restricted to high elevations (Olson and James, 1991). The Polynesians, therefore, evidently caused the vast reduction in Nene population numbers and range. The extent Nene is on the Federal Endangered Species list, and is the state bird of Hawaii. At the time of Cook’s landing, an estimated population of 25,000 birds inhabited the islands. In 1949, when only thirty wild and thirteen captive Nene existed, a much-needed recovery program was initiated (Black at al., 1994; Black, 1995). Up through 1993, over two thousand captive-bred Nene have been released into eight sites within Hawaii, Maui and Kauai. Their population dynamics have been closely studied in hopes of establishing a self- sustaining population. Researchers (Black, 1995; Rave, 1995) have identified two important factors needed to help ensure the Nana’s success. First, high 13 elevation habitat is not solely able to support Nene populations; food must be obtained from lower elevation sites for the geese to survive. Thus, more habitat needs to be conserved for Nene use. Second, inbreeding depression within populations must be carefully controlled by introducing geese from other lineages to keep genetic diversity at sustainable levels (Black, 1994; Rave, 1995). The extent Nene lives mostly in dry, vegetated areas on fairly recent lava flows, but also colonizes lowland masic grassland and forested areas. Nene avoid dense, masic forests. During the 18" century, they were recorded migrating to lowlands during the spring/early summer to feed on young grasses, and then returning to the uplands to feed mostly on berries. The Nene shows a dietary preference for leaves, buds, flowers, grass seed and berries (Bowler, 1994). The melting season coincides with gosling growth, resulting in the entire family being temporarily flightless, and thus being much more susceptible to predation (Berger at at, 1980). Brant: hylobadistes (Woods-mlking Goose) Remains of the woods-walking goose were discovered in montane forests of Maui from sites as young as 1,000 years old, and from Oahu (Olson and James, 1991). Its range overlapped that of the Nene in masic forests on Maui. The wing anatomy of this bird suggests that it was most likely flightless, but that it may have been able to sustain short bursts of flight (Figure 4). The morphological features of this species that suggested loss or near-loss of flight were the reduced wing size, reduction in pectoral girdle, and increased body and 14 leg size (Olson and James, 1991; Paxinos, 1998). In the present research project, woods-walking geese from Maui only were analyzed. Brant: sp. (Giant Hawaiian Goose) The first remains of the giant Hawaiian goose, which have the distinction of being the first avian fossil remains found in Hawaii, were discovemd on the western side of Hawaii in a lava tube estimated to be between 3,000 and 5,000 years old (Wetmora, 1943; Giflin, 1993; Olson and James, 1982b). The skeleton was very large and certainly flightless (for example, the wings are very reduced in comparison to body size), originally suggesting a relationship to the mea- nalos. The only remains discovered of this bird have been on Hawaii, suggesting that evolution of the flightless giant Hawaiian goose occurred only within this island. The giant Hawaiian goose overlapped in range with the Nene on this island. The original name given was Geochen muax, but osteological studies have established that it is a member of the We geese (Branta) (Olson and James, 1991). The goose’s extinction was most likely caused by the Polynesians, as supported by bone radiocarbon dates as recent as 510 SP, demonstrating the birds’ coincidence with the Polynesian colonizers (Paxinos, 1998) 15 Stable Isotopes Stable isotopes have been used for many years in ecological studies of both modern and paleontological biotas (Harrigan et al., 1989; Ambrose, 1991; Matheus, 1995; Ostrom at al., 1993). The biologically important elements (C, N, H, O, 8) have naturally varying isotopes which can be measured to aid in addressing ecological issues including nutrient cycling, dietary sources and preferences, paleoecology, trophic level, and information concaming climate and habitat (Peterson and Fry, 1987; Schwarcz, 1991; Koch at at, 1994; Koch, 1998). Isotope values are measured using a stable isotope ratio mass spachometer. They are expressed as: 0 X = ((R sample)/(R standard) - 1) X 1000 where 0Xi813C or 15M ofthe sample, and R is the ratio 13c [12C or 15N 114N, respectively. Positive values indicate that more of the trace (or heavy) isotope is present relative to the standard, whereas negative values indicate less of the trace isotope in the sample. The standard reference materials used are PDB (PeeDee Belemnite) for carbon, and atmospheric air for nitrogen (Peterson and Fry, 1 987). Isotopes of the same element behave very similarly to one another, however, they react at different rates. Molecules containing the heavier or trace 16 isotope tend to react at a slower rate due to greater stability of the molecule and higher dissociation energies (Ostrom, personal communication). The reduced rate causes the trace isotope to be discriminated against in reactions, and the lighter isotope to be concentrated in the product (Ostrom, personal communication). This isotopic discrimination, or fractionation, results in a change in the isotopic ratio between the product and the substrate of a reaction, or the r organism and its food resource. The isotopic composition of an organism's tissues strongly reflect that of their food resource. Thus, isotopic ratios are useful in ecological studies (Koch at at, 1994; Harrigan at at, 1989). The isotopic value of the animal’s tissue represents food that has been assimilated, with the exact period of time represented depending upon the turnover rate of the particular tissue analyzed (Gould at at, 1997). Carbon Stable carbon isotopes were first measured in plant tissues in the 19503, where a difference was noted between the three photosynthesis pathways (Craig, 1954; Katzenberg, 1992). The Hatch-Slack pathway (for C4 plants) is utilized predominantly by tropical, arid-adapted grasses. This pathway tends to discriminate lass against the trace isotope, leading to a 013C range of -8 to 4691» (Grecke, 1997). The majority of modem grasses are C4, but exceptions are found in habitats at high elevation and high latitude (MacFadden at at, 1994). 17 The Calvin-Benson photosynthesis pathway (for C3 plants) is utilized by temperate trees, shrubs and some grasses. This form of respiration results in a 0130 range of -23 to ~34!» (Katzenberg, 1992; Grecke, 1997). Ca grasses predominate in cool and moist environments, and do not occur below approximately 2000 m (6500 ft) elevation (MacFadden at al., 1994) The third photosynthetic pathway, which is utilized by cacti and succulents, lends intermediate 0130 values (Katzenberg, 1992). This pathway is often disregarded in ecological studies because it is used only by a select few plant groups that are not ingested by many animals (MacFadden er al., 1994; Grocke, 1997). An herbivore will have a carbon isotope value reflective of the plants it ingests (DaNiro and Epstein, 1978; Hobson and Clark, 1992; Koch at at, 1994). The 613C difference between c3 and c4 plants is substantial enough that their isotopic signatures may be recognized in consumers. Thus, carbon isotopes may be utilized to determine if an herbivore is primarily dependent upon C3 plants or C4 grasses (Magnusson at at, 1999) An herbivore that consumes a broad mix of leaves, berries, herbs, shoots and roots is termed a ‘browser", and those that consume entirely grasses are 'grazars’. This difference in herbivore dietary preference is reflected in the isotope values; browsers will have depleted 0130 values from consuming both C3 and C4 plants, and a grazer would have higher 013C from the C4 grasses isotopic signature (lacumin at al., 1998; Koch at al., 1994). 18 Fractionation of carbon isotopes occurs during biochemical cycling; the average trophic fractionation for carbon is :I: 1.0%, (DeNiro and Epstein, 1978; Koch et al., 1994; Schoeninger and DeNiro, 1984; Jahren at al., 1998; GrOcke and Bocherens, 1996; Ostrom at al., 1993), and the fractionation between diet and animal collagen is averaged at 496° (DeNiro and Epstein, 1978; Ambrose and Norr, 1998; Koch, 1998). In addition to indicating differences between Ca and C4 plants, carbon isotope values are useful in determining microhabitat conditions. They have been used to distinguish between animals living in closed-canopy forest versus open grassland habitats (GrOcke, 1997; Marra at al., 1998). This isotopic difference, or “canopy-effect,” results from depletion of 13c in closed-canopy forests due to biomass degradation (002 uptake), and a gradient in light levels causing altered plant respiration efficiencies (Koch at al., 1994; Bocherens ef al., 1996; Martinelli at al., 1998). These two factors result in increasingly depleted carbon isotope values closer to the forest floor (Martinelli at al., 1998). Elevation may also have an influence on carbon isotope ratios when using samples from a masic environment. Vitousek et el. (1990) measured the isotopic ratios of a native Hawaiian tree from a sequence of elevations (approximately 70 m up to 2500 m). The average 0130 value of all their samples was -28 to -26%o, similar to 03 plants from continental locales. An interesting correlation was found: the 0130 values increased with elevation, by almost 49!» between the lowest and highest sample sites (Vltousek at al., 1990). 19 Nitrogen Historically, 015N values have been principally used to determine trophic position (Schoeninger and DeNiro, 1984; Ostrom at al., 1993; lacumin at al., 1998). A number of studies have shown that an average 3 to 4%. increase is associated with each trophic level (DeNiro and Epstein, 1981; Peterson and Fry, 1987; Ambrose, 1991). The 3 to 4% nitrogen fractionation that occurs in animals is due to excretion of isotopically light urine, amino acid processing and synthesis which results in 15N enrichment of the product of metabolism (i.e. animal tissue) (Peterson and Fry, 1987; Ambrose, 1991). Nitrogen isotope ratios may be used to distinguish between some plant types at the base of the food web. Given that atmospheric nitrogen has a 015N value of 0%, N-fixing plants tend to have a nitrogen isotope value near zero (Peterson and Fry, 1987). For non-nitrogen-fixing plants, the 015N value reflects the isotope value of the soil, which is usually enriched in 15M (Peterson and Fry, 1987; Vltousek at al., 1989; Hegberg, 1997). The climatic variables of the habitat are recorded in the soil 615N value. High temperature, high pH, high salinity and aridity all result in higher 015N values. The type of soil may also affect the nitrogen isotope value, variables such as increased depth, higher clay content and matured soils all tend to lead to higher 015M values (Ambrose, 1991; Vitousak, 1989). Nitrogen isotopes may also be used to determine dietary reliance on terrestrial versus marine foods, because marine resources are 20 enriched in 15M in comparison to most terrestrial environments (DeNiro and Epstein, 1981; Schoeninger and DeNiro, 1984). Very hot and arid environments, however, also lead to enriched 015N values in animals due to physiological reactions during heat and water-stress (Ambrose, 1991). This factor could possibly complicate the differentiation between terrestrial and marine resources. The climate of the Hawaiian archipelago is not extremely hot or and; thus, the animals used in this study would not likely have been under heat and water- stress. The variation of nitrogen isotope values in Hawaiian soils has been investigated by Vltousek at al. (1989). They took foliage samples from volcanic sites on the island of Hawaii from elevational and primary successional gradients. Vltousek at al. (1989) found, uncommonly, that non-nitrogen-fixing plants had lower nitrogen isotope values than N-fixing plants. Nitrogen-fixing plants, as expected, had 015M values near zero, ranging from —2.0 to +06%, with a mean of -0.8%. The non-nitrogen fixers exhibited 015N values from -8.0 to -3.0% (mean -5.1%). The low values for non-N-fixing plants was attributed to: 1) soil nitrogen depleted in 15N, which was more pronounced in younger soils; 2) fractionation between soil and plant that further reduced the b15N values; and 3) negative 015N values of N in pmcipitation, often the only source of nitrogen to undeveloped soils. The latter mechanism, however, was ruled out by the Vltousek at al. (1989), because the nitrogen content within precipitation over Hawaii is vanishingly small (5 kglhalyr). The researchers found a direct relation 21 between soil age and 015N: as soil age increased, the nitrogen isotope ratio also increased (Vltousek at al., 1989). Comparing a 197-year old site to one of 67,000 years, the 015N increased by over 5% (-2.0 to +36%, respectively). This is an important consideration that will have to be taken into account when evaluating the environmental affect on the sample isotope values, especially when comparing samples originating fiom Maui versus Hawaii. It must be ensured that an isotopic difference between samples is not simply a reflection of soil age differences, but is truly reflective of the animal’s diet. Collagen The bone protein collagen is commonly used as a substrate for isotopic analysis of modern organisms (Connie and Schwarcz, 1996). Remnants of collagen (i.e. collagenous proteins) also exist within fossils, and are utilized for ancient dietary and paiaoecological analyses (Katzenberg, 1992; Ostrom at al., 1993; Koch, 1998). Collagen is the predominant bone protein, accounting for 85 to 90% of the organic matter present in bone (Katzenberg, 1992; GrOcke, 1997). The tumovar of bone tissue is very slow, such that collagen represents the animal’s average assimilated diet over a long period of time, lie. at least half a year in avian species (Hobson and Clark, 1992; Koch at al., 1994). Measurement of the enrichment in 13C of an organism relative to its diet due to fractionation during amino acid processing has given the values of. 0 to +46% (Ambrose and Norr, 1998), +2.8 to +37% (DeNiro and Epstein, 1978), and +5% 22 (Koch, 1998). in this study, the mean value of +4% will be used when determining the isotopic composition of the diet from the collagen value. It is important to understand the quality of collagen preservation in a fossil if one intends to utilize it to address paiaoecological issues. Indigenous collagen is bone protein material that has not been contaminated by outside sources, and is isotopically reflective of the diet and niche of the animal during its life. DeNiro (1985) established a criterion for assessing the indigeneity of collagen in fossil samples. He showed that collagen samples with a carbon to nitrogen ratio (CIN) between 2.9 and 3.6 are isotopically representative of protein from the living organism. If a sample is outside of this range, the carbon and nitrogen isotope values might not be representative of the dietary habits and environmental conditions of the animal during its life. Yield of organic matter can serve as another indicator of indigeneity as shown by McNulty at al. (in press) who subjected modern bone to high levels of heat for various amounts of time and humidity. The time-series of samples was analyzed for percent yield of collagen, CIN, and carbon and nitrogen isotope values. The results demonstrated that geochemical integrity of the samples was preserved even with very low percent yields of bone organic matter (collagenous protein), suggesting a 5% yield as a conservative criterion. These results are significant in application to true fossil samples because they suggest that indigenous isotopic signals can survive in fossil material, and they provide a new criterion for assessing the reliability of a sample’s representative isotopic signal. Methodology The modern samples were initially prepared by a Dennestid Beetle colony within the Michigan State University Museum to remove exterior muscle and connective tissue. Processing of all bone samples for geochemical analyses was as follows (Figure 5): the bone was first mechanically cleaned with a nylon toothbrush and e-pure water, then cut into small pieces with a key saw. The pieces were sonicated in 4°C 1N HCI for 5 minutes, rinsed with multiple washes of e-pura water, and then sonicated another 10 minutes in e-pure water. The samples were then dried overnight in an evaporatory oven at room temperature, and than powderized in a SPEX CertiPrep 6750 Freezer/Mill. For each sample, up to three grams of bone went through a pre-wash procedure of stirring for one hour in a sodium buffer solution (1N NaHzPot, pH = 6-7) at 4°C to remove humic substances. Each sample was centrifuged three times (4000 rpm) and rinsed with e-pure to remove and clean the supamatant. The samples were next demineralized in 1N HCI for approximately eighteen hours at 4°C. Separation of the collagenous pellet from the acid-soluble non- collagenous protein (NCP) was achieved by high-speed (12000-14000 rpm) cantrifugation. Each sample fraction was dialyzed at 4°C for five days in 3500 molecular weight cut-off dialysis membrane against many changes of distilled water. The samples were freeze-dried prior to geochemical analyses. Bone protein yield was calculated on the dry weight of the bone pre- and post- extraction. The collagenous and non-collagenous protein fractions were prepared for isotooic analysis using a modified Dumas combustion (Macko, 1981). Purified gases from the samples were obtained by cryogenic gas separation and subsequent isotopic measurements performed on a VG-PRISM stable isotope ratio mass spectrometer (MicroMass). Also, a Carlo-Erba elemental analyzer interfaced to the PRISM mass spectrometer was used to determine the carbon and nitrogen isotope ratios and the CIN values (Wong at al., 1992). Results 8: Discussion Assessment of Isotopic Indigeneity As mentioned eadier, indigenous collagen has a carbon to nitrogen ratio between 2.9 and 3.6. With the exception of two samples, the carbon to nitrogen ratios of the sub-fossils are within this range (T ables 3 and 4). This lends support for the use of the collagen isotope data which appears to provide an indigenous signal that can be used to accurately interpret the paleoecology. There are two outliers: the sub-fossil giant Hawaiian goose (Branta sp.) sample four (66.08.84), and the modern Short-eared Owl, or Pueo, (Asio fiemmeus) sample one (PU.01.B1), whose CIN are 3.8 and 4.7, respectively. To further evaluate the integrity of these samples, collagenous protein yield (percent by weight) was detenninad, and the 013C and 015N values of these samples were compared to other measured values for these taxa. The collagenous protein yields extracted from well-preserved samples should be between five and twenty percent. In a study of artificial diaganesis, McNulty at al. (in press) demonstrated that the isotope values of heated samples (at 100°C) with protein yields equal to or greater than 5% did not differ from those of unheated, non-diaganetically altered bone. The upper limit, 20%, is the percentage of collagen present in modern bone tissue. The majority of the sample yields were less than 20% including the modern samples, but nearly all were greater than 5% (Figure 6). On average, the sub-fossil samples produced lower collagen yields than the archaological and modern samples, which was as 26 expected considering that the bone protein had undergone a longer pedod of diaganesis. Two samples fell below 5%: giant Hawaiian geese samples 3 and 4. Sample 4 (6668. B4) also produced a carbon to nitrogen ratio outside of the accepted range. The collagenous protein yield and CIN for this sample suggest that it has undergone extensive diaganesis, and may not produce an isotopic signal representative of the indigenous value. This assumption may be further I tested by comparing the isotopic results to those of other giant Hawaiian geese. The same modern Pueo sample (PU.1) with a CIN value outside of the accepted range, also had an organic matter yield greater than 20% (24.3%). The failure of this sample to satisfy these two criteria suggests that it will yield isotopic results that are not reflective of the owl when it was alive. However, because it is a modern sample, diaganesis of the organic material is most likely not the cause for suspected non-indigeneity. Relatively little is known regarding the history of this sample; it is uncertain how long the specimen was along the roadside, or under what conditions it lived. Contamination of the bone matedal post-modem may have led to the unacceptable CIN value and protein yield. Isotopic Results Stable carbon and nitrogen isot0pic values were measured to establish the paleoecology of three closely related Branta species. A time-series (sub- fossil, archaological, and modern) of B. sandvicensis (the Hawaiian Goose, or Nene) was evaluated to determine whether the species experienced a shift in dietary niche subsequent to human presence in the archipelago. In addition, 27 isotope values of collagen from woods-walking geese (Bronte hylobadisfes) and giant Hawaiian geese (Bronte sp.) were analyzed to ascertain if a novel means of resource partitioning was developed between species to reduce direct competition with their sister taxon, the Nene. This project also endeavored to address the affect of environmental variables on the organic isotopic composition. The 013C values of the extracted collagenous protein are graphed against 015N values. The isotopic makeup of the herbawous diet may be determined by adjusting for fractionation by subtracting 4% for 013C (DeNiro and Epstein, 1973; Ambrose and Norr, 1993) and 3%. for 015N (DeNiro and Epstein, 1981, Peterson and Fry, 1987). The Hawaiian Goose or Nene (Branta sandvlcensls) Isotope data for the Hawaiian Goose is highly variable (Figure 7). The 0130 values range from -25.8 to 44.3%., and the 615M from 0.53 to 12.9%.. Such a broad span of values suggests that diet varied among individuals, and that they incorporated a wide anay of plant matedal. This is consistent with the known habits of the extant Nene (Bowler, 1994). In historical times, these geese were known to migrate seasonally beMan high and low elevations. In addition, they have a variable diet consisting predominantly of grass seed, benies, leaves, and shoots (Kear and Berger, 1980). The Nene occupies a 'browser‘ dietary niche, in contrast to grazers that predominantly consume C4 grasses. Thus, browsers would be expected to demonstrate a broad range of isotopic values. The carbon and nitrogen isotopic values suggest an herbivorous diet consisting of few, if any, C4 grasses. The isotope values show one noticeable outlier in this data set: the archaological Nene sample four (NA4). This collagenous sample yielded a very high nitrogen isotope value (12.9). This sample’s locale was near the coast and at a low elevation. Marine resources are known to cause an enrichment in 015N values, which may be reflected in this sample’s results. The values obtained for CIN and protein percent yield indicate indigeneity of the sample’s isotopic signal. This sample is the only Nene from a coastal environment; the others originate from open grassland or scrubby forested sites. This difference may be the sole reason for the enriched nitrogen isotope value. Thus, perhaps this sample should be treated as an outlier of the data set based upon the unusual 015N value and the habitat odgin. A Q-test statistical analysis for outliers was performed on this sample (Dean and Dixon, 1951; Dixon, 1951). The results indicate that the isotopic data from sample NA.4 may be rejected with 90% statistical confidence based upon the 015N value. The elevated carbon isotope value is most likely a result of a diet composed of a high percentage of C4 grasses (~90%). This sample is an outlier due to its unusual habitat, not from diaganetic alteration. One other archaological Nene had a somewhat unusual nitrogen isotope value (NA.3, 0.53%). The observation that the carbon isotope value of this sample falls within the average range of the other Nene supports the suggestion that the collagen retains an indigenous isotope signal. The carbon to nitrogen 29 ratio and collagenous protein yield data also suggest indigeneity. The sample was from a very high elevation and dry site on Hawaii. The high elevation could have caused increased carbon isotope values, and aridity tends to lead to enriched 015N values, neither of which are reflected in the result. This suggests that the isotope values are reflective of the diet of this Nene rather than a result of climatic and geographic factors. Perhaps this particular individual was unusual in that it spent most of its time in forested, closed-canopy habitats consuming non-nitrogen-fixing plants on the forest floor. Both of these conditions would result in depleted carbon and nitrogen values, respectively. The average carbon and nitrogen isotopic values for the sub-fossil Nenes are -18.73 :I: 2.29% and 7.16 :l: 1.64%, respectively. Average isotope values for the archaological samples are lower than those of the sub-fossil data set, at -206 :l: 1.75%. and 4.49 s 2.09%. for 6130 and 015N. The modern samples average -21.77 s 5.42% (0130) and 6.49 :t 0.37% (015N). The sub-fossil Nene have, on average, higher carbon and nitrogen isotopes than the archaological and modern samples (Figure 8). This result suggests that the Nene axpedenced a dietary shift overtime. The sub-fossil Nene may have had a higher C4 grass intake than the archaological geese (59.1% vs. 45.7%). This may be due to the sub-fossil geese being able to move between high elevation sites to low elevation grasslands. With the arrival of the Polynesians came extensive changes to the landscape, particulady in the low elevations. The changes may have ended the Nana's movements between elevations because the lowlands became covered with agricultural land and residential communities. The restriction in movement 30 may have caused the shift in diet reflected in the isotopic values between the sub-fossil and archaological Nene. Decrease of the archaological Nana’s range would have progressed as the Polynesians extended lowland habitat change over time and reduced population numbers through hunting and predation. The dietary change would not have been instant, hence the reduced fitness resulting from restriction of range to only high elevations would have also increased over time. Some populations of archaological Nene may have survived by finding hidden sites in which to feed and mate. The data set was subjected to statistical analysis of variance (Campbell, 1974). The results conclude that the isotope average for the sub-fossil Nene is statistically different (p < 0.05) from the archaological average based upon the 015N value, supporting the conclusion that a dietary shift occuned between the time periods. The other comparisons, archaological and sub-fossil versus modern, are not statistically significant. These comparisons may have been obscured by the variability of the Nene diet, and low sample number. More samples of the extant bird, and more information concaming captive versus wild feeding, would help in addressing the question of dietary shift over time. The modern compadson sub-set presents some interesting questions. For one, we see that the modern Nene 013C values encompass the two extremes of the sub-fossil and archaological Nene carbon isotope range. Although the sample size is small, some suggestion may be made as to what could lead to such variation. Since 1949 there has been an active conservation program to protect the dwindling population of wild Nene. The birds have been 31 bred in captivity and many have been released into the wild. The modern Nene samples analyzed may have been captive-raised birds since two of the geese ‘ (NM.1 and NM.2) were banded, and it has been indicated that birds marked with leg bands usually odginated in captive breeding programs (Stone and Prat, 1994). It is uncertain how long they were in captivity and how long they lived in the wild before death. One objective of this thesis was to assess how environmental factors may have influenced the isotopic values. To address this issue, we can compare samples that vary in elevation and/or precipitation. When this is done with the Nene-only data set, we do not see the expected pattern of enriched 013C for higher elevations or 015N for drier locales. For example, sub-fossil Nene sample five (NP.5) demonstrates an enriched 0130 value (45.1%) perhaps due to the high elevation; however, NP.2 and NA1 are needy the same in value (both are -17.0%), but they are from low elevation sites. We also see high elevation sources yielding depleted carbon isotope values (NA.3 at ~22.2%, NAB at 42.1%.). NP.5 also has a somewhat enriched 015M value (3.0%), which may be due to the drier locale. This is most likely not the cause of the enrichment, since NP.2 is from a wet site and has yielded a higher 01511 value (9.4%). Another environmental concam is soil age. In the Nene-only data set, there are samples from both Hawaii and Maui. Maui is an older island in the sequence, at approximately 1.4 mya, whereas Hawaii began forming about 0.5 mya and continues to be volcanically active (Carson and Clague, 1995). As noted earlier, Vltousek at al. (1989) found that increased soil matudty is 32 associated with incmasad nitrogen isotope values. We would expect that the samples odginating from Maui would have andched 015M values. However, this pattern is not observed; the Maui samples are very similar to the other 015N values. This lack of differentiation supports the validity of isotopic compadson between the two islands, an important consideration for this study. Sub-fossil Geese The most striking feature of the sub-fossil geese data is the isotopic separation of the giant Hawaiian goose from the other two focus geese species (Figure 9). This group has lower carbon isotope values (-24.5 to —23.2%) and somewhat depleted nitrogen isotope values (2.0 to 5.2%). Statistical analysis demonstrated that the average isotope values for the giant Hawaiian geese are significantly different (p<0.05) from the sub-fossil Nene with which it co-existed, and in 0130 (p<0.05) from the archaological Nene. The giant Hawaiian goose samples were from high elevations, two were dry sites (GG.1, GG.2) and the other two masic (GG.3, 66.4). High elevations are expected to cause an increase in carbon isotope values, which is not demonstrated in this context (Vltousek, 1990). Thus, the isot0pic separation must be due to another factor related to either the environment or diet. The depleted 0130 values are likely due to the giant Hawaiian geese focusing almost entirely on Ca plants, perhaps as much as 90% C3. This greater selectivity may have resulted from two factors: a greatly redumd or even total absence of C4 grasses in the habitat of the giant geese, or an evolutionary shift in diet to one that was less competitive with the 33 co-existing Nene. The lower 015N values of the giant geese may be due to their consumption of depleted non-nitrogen-fixing plants in a closed-canopy environment The woods-walking goose and the Nene are indistinguishable from one another based on their isotopic composition. The woods-walking goose is morphologically intermediate between the Nene and giant Hawaiian goose. The [ latter has undergwent greater evolutionary change and became more highly 5 specialized for the predator-free island environment The woods-walking goose developed greater body size and near-complete or complete loss of flight capability. The lack of isotopic difference between the Nene and the woods- walking goose may have resulted fiom the lack of substantial evolutionary differentiation in diet between the two species. If that was the case, the two species may have been competitors with one another where their habitats ovedapped in Maui. This question of competition between species is also a concam in modem avian studies, because many modern birds ovedap in range and seem to have the same dietary niche, yet they are able to successfully co- exist A detailad study of competition is difficult for extant species, but would be even harder to understand between extinct species, and in habitats and ecosystems no longer present. Another feature of the sub-fossil geese data set is that the source island of the samples (Maui versus Hawaii) is not reflected in the carbon and nitrogen isotope values, as was also the case with the Nene. In the sub-fossil geese data set, all of the giant Hawaiian geese odginated from Hawaii, the Woods-walking geese from Maui, and the Nene from both islands. An effect of soil age difference on isotope data is not supported in this instance because: 1) the 615N values of Hawaii samples are not unifome low, and 2) the samples originating from Maui do not depart isotopically from the Hawaii samples. sub-toss" and Archaological Data F The complete data set of sub-fossil and archaological samples, including comparative species, was graphed to provide a broader comparison for isotopic analysis of the focus taxa (Figures 8 and 10). As discussed above, the most E notable features of the geese data is that the Nene covers a broad range of .. isotopic values, the giant Hawaiian goose separates out isotopically, and the woods-walking goose is indistinguishable from the Nene. The Dark-rumped Petrel (DP.1, DP.2) is a pelagic species that only comes to the islands to breed. They feed upon squid, fish and crustaceans (Hawaii Audubon Society, 1993). Both the carbon and nitrogen isotope composition of the petrels were highly enriched. This is a consequence of two factors: a marine diet and camivory. Marine resources have an isotopic signature of high 6130 and c15N values compared to most terrestrial environments (DeNiro and Epstein, 1981; Schoeninger and DeNiro, 1984). Additionally, carnivores have higher 615N values, being 3%. enriched over their food source (DaNiro and Epstein, 1981; Peterson and Fry, 1987; Ambrose, 1991). The isotopic values of the two Dark-rumped Petrals seem to be entirely a result of their marine carnivorous diets. This result is expected given that petrels spend the majority of 35 their lives at sea; the terrestrial influence would be very marginal, and would have little consequence on the isotopic composition of the petrel’s tissues. The feral pig (PM) was an omnivorous species that was indiscriminate in its dietary preferences, feeding upon both animal tissue and plant material. A carnivorous diet leads to enriched nitrogen isotope values, whereas an herbivorous diet lends depleted 615N values. The metabolic mixing of the two F sources would lead to median nitrogen values, as seen in this sample. The carbon isotope value reflects a diet consisting of Ca plants and C4 grasses, or animals that etc a mixture of both. The ibis (Apten'bis sp., lB.1) was a flightless endemic to the Hawaiian archipelago that went extinct along with the woods-walking goose and giant Hawaiian goose (Olson and James, 1982b). When its ancestor arrived in the islands, specifically upon Maui Nui, it adapted to the terrestrial environment and began to change its diet away from marine and freshwater sources (James and Olson, 1991). The ibis may have been a forest browser, consuming invertebrates such as snails, land crabs and insects, and possibly berries. The diet and niche of this new form of ibis may parallel the New Zealand kiwis (Olson and James, 1991; H. James, personal communication). The isotopic composition of the ibis’ collagenous bone protein suggests, however, that it may have been an herbivore, occupying a dietary niche similar to the Nene. This does support the contention that the ibis developed a new dietary niche concommitant with the evolution of a new body form and habitat niche. The flightless Hawaiian ibis still possessed a very long, decurved bill, a bill shape 36 typically associated with probing into the soil to unearth gmbs and other soil- dwelling organisms (Gill, 1995). Perhaps the flightless Hawaiian ibis specializeded upon roots, tubers, and fungi, which would be more consistent with its herbivorous isotopic signal. Modern Ecooyotem The isotopic results for a small sub-set of the modem Hawaiian ecosystem demostrates a broad range in b13C and b15N values (Figure 11). The Nene have the lowest nitrogen isotope values, as would be expected for an herbivorous species. Modem Nene samples one and three (NM.1, NM.3) also have the lowest carbon isotope values. This is expected for a diet consisting entirely of 03 plants, which was most likely true for these two individuals. Columba livia (Rock Dove or feral Pigeon) is another herbivorous species, but it has different dietary preferences than the Nene, as reflected in the isotopic results. The Rock Dove consumes grains, small seeds, and fruit, and tends to live in close proximity to human populations (Robbins at al., 1983). Both grains and fruit have enriched 615N values compared to nitrogen-fixing plants (Schoeller et al., 1986), which is seen in the isotopic results. In addition, living near humans, the pigeons may have consumed proteinaceous foods while scavenging through our refuse, also lending and enriched 615N value. The carnivorous species in this data set, the Pueo or Short-eared Owl (Asio fiammeus), the cat (Felis cattus). and the Barn Owl (Tyto alba), all have enriched o15N values in comparison to the Nene. The Pueo, a secondary 3'7 predator consuming insects, rodents and sometimes birds (Hawaii Audubon Society, 1993), displayed an enrichment amount in comparison to the modern nene isotope average that ranges from +1.62 to 4.68%. it should be noted that Pueo sample one (PU. 1, enrichment over Nene of 1.62%) had an unusual CIN and collagenous protein yield (4.7 and 24.3%, respectively), which does not support isotopic indigeneity of the sample, thus the data of this sample should not be taken into consideration. Pueo sample three (PU.3) is more likely than sample two to have regulariy taken young Nene as their meal, since it demonstrates greater than one trophic level enrichment over the Nene (+4.88%). The cat and Barn Owl samples are similarly enriched, except for the feral cat sample two (CT.2). This sample demonstrates a +10% enrichment compared to the Nene, representing three trophic lavels, which may be an artifact 38 Conclusions This stable isotopic study of Hawaiian avifauna has given insight into the paleoecology of a specialized endemic group of terrestrial geese, and allowed the dietary preferences of the three focus geese to be interpreted. The ancient Nene experienced a slight shift in diet after human colonization occurred. This may be due to the extreme habitat destruction and change in the lowlands, thereby preventing the Nene from continuing their seasonal elevational movements. The woods-walking goose was found to have a diet indistinguishable from the sub-fossil Nene. This result suggests that the two geese may have been occupying the same dietary niche, and may have been in competition with one another where their habitats overlapped. This conclusion, however, could be strengthened with a larger sample size of woods-walking geese. The giant Hawaiian goose was found to be isotopically distinct from the other sub-fossil geese, suggesting this species had evolved new dietary preferences. This change would have allowed the giant goose and Nene to co- exist in Hawaii without competing over food resources. Speciation events following colonization of the archipelago with a proto-typical large Canada Goose, therefore, appear to have resulted in habitat partitioning between some taxa, thus reducing competition. The effect of several environmental conditions was explored in the isotopic data. Differences in elevation, soil age, and aridity tended not to shift the isotopic values. The condition of the habitats foliage (i. e. closed-canopy versus open 39 grassland) did seem to have a slight influence on the isotopic composition of the samples, although it was not easily separable. The majority of the isotopic variation has been attributed to dietary differences; specifically. the relative contribution of C3 plants and C4 grasses. The focus taxa in this study displayed a range of preferences from a predominantly C3 to a more mixed diet. A significant finding that may be of concern to the Nene Recovery Program is the evidence that a shift in the Nene’s diet occurred after human colonization of the archipelago. This supports the contention that high elevation sites alone are not able to provide the Nene with all of their required nutrients. Low elevation localities must also be available to achieve a sustainable, healthy population. 40 2222222839129. belie .cSIeNssz .egbfiszsigesbsg salsa agree-magnesia.» ..s §s§§§22222.§:§3§8§§838§5 .sasctscssassg .ngssssubsassggecsis} ggagggs .22.... .332. gasses. 28 253. 22.5... 282.2... see: .25 4.2.6.. 8 .+ 92 .. 222. 222.22.. 22.6: .8288 .880 22......2. E20. .3 255 e322 pan. .222 .3323 . as: pan. .322 .3283 .92.. :5 .522 .3253 .. asz pan. .322 .3833. .880 8.2.3385 32332.. 255 8.3.2.8 .2252. 38......» .. 2.22 :E: 22.2.. 88.8? :5: :5... 2.32.. 88.28.. . 6.. «3 .532 232.23 322 282.28 2.22 2:5 .252. 88.8? 260 55.3“. “I oo 4 5m 98 2. 22.". HI 8.;8 98... 2.. E”... 8.33 e30 .233. 2.3.... ”I 2. 4 Wm”! 98 5:60! Pal ”I 595. .naiuI a2 9532!... Asa: n2 99.622.— 43a: 12 28.625... .302 8:02 .380 52.2.63 «3.28.528 85$ |§ .96. 88.. 38. .8 courses 8358 8 cs... .2 Has. 41 .Ezzscssssggeg 2.8.8262 .egbssszscsgecosfigfis .gsfieseesmzszegcs steam .5 28.23:. Stag 2:232. 2.22.. 2.5.5... .58.... 3582 838 29.3 2. .fisstscsgsssa mgggisgggfii £3888 §8§2m¥8~§e§§ws 0.00... 83 as. 2.3. .a. :5 .322 on -+ or: E... 22.2.6... 32.29... in notch... on... 2.2 . u 25. .0 225. 4.22.... comma... a... .82. £28 3m. 9:..me .. . v a 9mm .. 8.". 4.25.... 8v... 8... . u 2.o. .O 22.0. 4.9.6: Sufism... ages. 2.862.230. walxqosi 2:83.21 00553.: $0 250 gold . 4.253.. on v tau—Ea... one gnu «San—d 4.322.. on v .38 .68.. 2.... 3.5.8, 5.65.5. .35. Educ... .3. Sauce... .322 E36... £35.... .522 E28... e30 3.3.5.5 .821. manage: 98.x 0.30. 565...: .32 530E . 33 :52 .0 .Z 33 .330... EcuoE e30 Emmy 3Q 82K on... 96. .25... .E mud .c 09:. ES. .055. .532. omvv .. on... w>§ . u 26. 205v. 4.9.6.... omwv 98 .80: was muck Eng HIM laud.— .oonoea contests .o. bonanza 32.83 .8 «a... ”N m._m<... 42 2.8. 2.... 2... as .. 3.8.8. 5.3. 8... c... «N .. 8.2.00— 0...0N. «0.0 «.0 «.05 .. «0.05.00— 5...v~... 00.5 0.0 v.0 3300 5:026: 8.0.0. in 800.0 50.00.06. 00.0 5- 00.0 0.0 0.0.. .. 00.00.95 50.05- .i ii 5.0.. .. 00.00.03 50.94.. 05.5 0.0 0.55 Aegflaifig 08080092.. 8:20 «0.00.0.5 05.50. 9.5 0.0 l. .. 00.00....2 05.0 5- 00.0 5.0 I. .. 00005.2 00.0 5- 05.5 0.0 .l .. v0.00dz 00.0 F- 00... «.0 5.3 .. 00.00....2 00.5 5.. 3.0 «.0 «Ne .. «0005.2 0N...~. 00.0 0.0 N.0.. .. 50055.2 00.3.. 05.0 «.0 0.0.. .. 00.00.oao§_ 2050.308 280.0992 5220 39.00260 .0 m..0<... cont—3&8 .9. 829 8.28. 2 new 0 new .802 z 9 0 6.0.0.5 520.... 30503.00 6 m..0<5 8.0 w- 8.0 0N N65 3.... 03.2sz 30050.5 in uztg rm.wo.m_| AB. MS. an 3. .932... £2826 5.8... 8.0. mi... 3 to. . S88 39. 8.9 2” v.9 as... 88288. 5?»... 2.23.... 5.8.3 INS... mind. 3 EN . No.89. «You. .3. 3 v.8 .28 .85 5.3258 5.8.9. 8.3- 2.: S 0.2 . 8.5.2 «I... B... 3 mt . mums... v0.55- N50 5.? QVN €50 0800.550 603$ 0:00.53. 0.6.x rm. 5.3a. 3.2- Mud. 3 0.8 . 88.50 5.2- at: «a on. $959 3.. 2... $8.26 8.9. we 5 3. . $8.5 8.9- 8.2 3 wt :8 .28. 8.8 mi“. 88.5 Jag flu mm: 3 01% 98. .EE .3 3 as": 03.8202 5332.. 9: “F $59”. =<>>>SOOI<¥ ‘ ' _>E md 2035 . O. 46 Giant Hawaiian Goose ~—— 8. hylobadistes '""“ — extinct Hawaii Nene — extinct Maui Nene Modern Nene ! Large Canada Geese i Small Canada Geese 1 Other Branta sp. FIGURE 3: Cladogram of Hawaiian geese (Paxinos, 1998). 47 . . dozen Agagna no. 2.13.35 on? 3 383800 .63ch 88¢ 25 3 coast! :33)»: 05 c. 38.32... 9E... ”v maze—n. .25. B __ an: $.31 . 8.5.5 532 r33»... .830 .330 , .Amuvunmnau 09m: .052 9333.333 mmuasozm N93 .3000 25:8 48 (Cleaned manually and by sonication) Powderized Ere-Wash (Na-buffer to remove hUrnic acids) ‘ Centrifuged Demineralization (1N HCI, 4 C, approx. 18 hrs) ‘ Centrifuged Dialysis (5 days, 4‘C, distilled H20, 3500 MWCO tubing) ‘ F reeze-dried Geochemical Analyses (Protein yield, CIN, isotopic) Figure 5: Summary of methodology. 49 .~ c5 F 83m: 5 .8925 2 893322.“ 3 >3. 3%; B 5 «so; 582“. 3053.8 6 menu: 0 n8 o 18 o J m too a :22 o mdodl 2 . Ndz o M ”gage . mazz . p I- . Po8. 3.33. 0338. ~52 K $59“. 1rd m 7 m 7 n 7 m 7 u an 5. mm- mm- R- m.< "a $56.“. m7 2. 2- E- mu. mm- R- _ . _ _ omoou. ”EEG ocuz , 8:02 252 a. , mm - a .. __ 8 a m «:02 Save: Ns 52 £95.95 -8003 4 $08 :56. ocmz O 0.. .~ Ea _ 833 s 839.. m 833633“ 3 >8. 3.3% 0338. 33w .3853 a $50.“. 1 no use 1- 2- m7 8- NN- ¢~- mm- 8- - N8- F8- 1 mg l , $2 o O... 1 v.8. Ns , Wu; 4 5.2 0 1 3.2 o No; 4 3.2 o I o 3.2 - ~52 o 53 .~ 2m _ $33 5 8265 a 2233293 8 >8. $38: 0338. 552852 Ba .3835 5P 9.8: ¢Uj 020 1 m0 8.3350 noboao 038230- $34 282 388.930 0:02 62.844 o—‘ N—v 1., m... mp :. m 7 m 7 N 7 m 7 5. fl... 3. R: 32 4 _ Q.3. 3.33. 02.88. 38.85.... as .3835 2: 9.8.“. 401 020 1 m0 8.2.58. 225.. 880 25- 22.5.6834 232 382.250 252 62.3.4 opt N—r 4p- up- mp 2. m 7 m 7 N 7 m 7 .N- «d. 3.- RN. m.oa 2 8253292.» B 23. 3.38 0332.2 Ecumxmooo Save: "2. £59“. 020 1 m U 5N- pp- mp- mp- up- mp- PN- MN- mm- Q22 l . P 22 Y. N22 I I 25 4 , NE 4 Eu 4 :so 4 , m2 4 4 ~26 . o N am .. 29. 0 NB 4 N5 ‘9 55 LITERATURE CITED Allen, J. 1997. Pro-contact landscape transformation and cultural change in windward O’ahu. In: Historical Ecology in the Pacific Islands. Eds. P. V. Kirch and T. L. Hunt, Yale University Press, p. 230-247. Ambrose, S. H. and L. Norr. 1998. Experimental evidence for the relationship of the carbon isotope ratios of whole diet and dietary protein to those of bone collagen and carbonate. Communication from the Department of Anthropology at the University of Illinois at Urbana-Champaign. Ambrose, S. H., B. M. Butler, D. B. Hanson, R. L. Hunter-Anderson and H. W. Krueger. 1997. Stable isotopic analysis of human diet in the Marianas Archipelago, Western Pacific. American Joumal of Physical Anthropology, 104: 343-361. Ambrose, S. H. 1991. Efiects of Diet, Climate and Physiology on Nitrogen Isotope Abundances in Tenestrial Foodwebs. Journal of Archaeological Science, 18: 293-317. Athens, J. S. 1997. Hawaiian native lowland vegetation in prehistory. In: Historical Ecology in the Pacific islands. Eds. P. V. Kirch and T. L Hunt, Yale University Press, p. 248-270. Black, J. M. 1995. The Nene Branta sandvicensis Recovery Initiative: Research against extinction. Ibis, 137: 153-160. Black, J. M., J. Prop, J. M. Hunter, F. Woog, A. P. Marshall, and J. M. Bowler. 1994. Foraging behavior and energetics of the Hawaiian Goose, Branta sandvicensis. VViIdfOWI, 45: 65-109. Bocherens, H., G. Pacaud P. A. Lazanev and A. Mariotti. 1996. Stable isotope abundances (‘30, 1“M in collagen and soft tissues from Pleistocene mammals from Yakutia: Implications for the palaeobiology of the Mammoth Steppe. Palaeogeography, Palaeoclimatology, Palaeoecology, 126: 31-44. Bowler. J. M. 1994. Foraging behavior and energetics of the Hawaiian Goose, Branta sandvicensis. Wildfowl, 45: 65-109. Campbell, R. C. 1974. Statistics for Biologists, Second Edition. Syndics of the Cambridge University Press, London. Carson, H. L. and D. A. Clagua. 1995. Geology and biogeography of the Hawaiian Islands. In: Hawaiian Biogeography: Evolution on a Hot Spot 56 Archipelago. Eds. W. L. Wagner and V. A. Funk. Smithsonian Institution Press, Washington. Carson, H. L. 1981. Microevolution in insular environments. In: Island Ecosystems: Biological Organization in Selected Hawaiian Communities. Eds. D. Mueller-Dombois, K W. Bridges, and H. L. Carson, The Institute of Ecology, Hutchinson Ross Publishing Company, p. 471-484. Chatterjee, S. 1997. The Rise of Birds. The Johns Hopkins University Press, Baltimore and London. Connie A. B. and H. P. Schwarcz. 1996. Effects of climate on deer bone (WM and (WC: Lack of precipitation effects on (MN for animals consuming low amounts of C4 plants. Geochimica at Cosmochimica Acta, 60: 4161- 4166. Coyne, J. A. and T. D. Price. 2000. Little evidence for sympatric speciation in island birds. Evolution, 54: 2166-2171. Craig, H. 1954. Carbon-13 in plants and the relationship between carbon-13 and carbon-14 variations in nature. Journal of Geology, 62: 115-149. Dean and W. J. Dixon. 1951. Analytical Chemistry. 636. DeNiro, M. J. 1985. Postmortem preservation and alteration of in vivo bone collagen isotope ratios in relation to palaeodietary reconstruction. Nature, 31 7: 806-809 DeNiro, M. J. and S. Epstein. 1978. Influence of diet on the distribution of carbon isotopes in animals. Geochimica at Cosmochimica Acta, 42: 495- 506. DeNiro, M. J. and S. Epstein. 1981. Influence of diet on the distribution of nitrogen isotopes in animals. Geochimica at Cosmochimica Acta, 45: 341- 351. Dixon, W. J. 1951. Ratios involving extreme values. Annals of Mathematical Statistics: 68-78. Gill, F. a. 1995. Omithology, 2"" Edition. w. H. Freeman and Company, New York. Gould, P., P. Ostrom and W. Walker. 1997. Trophic relationships of albatrosses associated with squid and large-mesh drift-net fisheries in the North Pacific Ocean. Canadian Journal of Zoology, 75: 549-562. 57 Grdcke, D. R. 1997. Stable-isotope studies on the collagenic and hydroxylapatite components of fossils: Palaeoecological implications. Lethaia, 30: 65-78. Grdcke, D. R. and H. Bocherens. 1996. Isotopic investigation of an Australian island environment. Academie des Sciences serie lIa: Paleontologie, Paleoecologie, Geochimie: 713-719. Hanson, H. C. 1965. The Giant Canada Goose. Southern Illinois University Press, Carbondale. Harrigan, P., J. C. Zieman and S. A. Macko. 1989. The base of nutritional support for the Gray Snapper (Lutianus gn'seus): an evaluation based on a combined stomach content and stable isotope analysis. Bulletin of Marine Science, 44: 65-77. Hawaii Audubon Society. 1993. Hawaii's Birds. Honolulu, Hawaii. Hobson, K A and R. G. Clark 1992. Assessing avian diets using stable isotopes I: Turnover of 13C in tissues. The Condor, 94: 181-188. Hegberg, P. 1997. Tansley review No. 95: 15N natural abundance in soil-plant systems. New Phytologist, 137: 179-203. Hotchkiss, S., P. M. Vltousek, O. A Chadwick, and J. Price. 2000. Climate cycles, geomorphological changes, and the interpretation of soil and ecosystem development. Ecosystems, 3: 522-533. lacumin, P., H. Bocherens, L. Chaix and A. Marioth. 1998. Stable carbon and nitrogen isotopes as dietary indicators of ancient Nubian populations (Northern Sudan). Journal of Archaeological Science, 25: 293-301. Jahren, A. H., L. C. Todd and R. G. Amundson. 1998. Stable isotope dietary analysis of bison bone samples from the Hudson-Meng Bonebed: effects of paleotopography. Joumal I of Archaeological Science, 25: 465-475. James, H. F. 1995. Prehistoric Extinctions and Ecological Changes on Oceanic Islands. Ecological Studies, 1.15: 88-102. James, H. F. 1990. The contribution of fossils to knowledge of Hawaiian birds. Acta )0( Congressus Intemationalis Omithologicus, 1: 420-424. James, H. J. and D. A. Burney. 1997. The diet and ecology of Hawaii’s extinct flightless waterfowl: evidence from coprolites. Biological Journal of the Linnean Society, 62: 279-297. 58 Katzenberg, M. A. 1992. Advances in stable isotope analysis of prehistoric bones. Skeletal Biology of Past Peoples. Eds. S. R. Saunders and M. A. Katzenberg. Wiley-Liss, Inc, New York. pp. 105-119. Kear, J. and A. J. Berger. 1980. The Hawaiian Goose. Buteo Books, South Dakota. Kirch, P. V. 1985. Feathered Gods and Fishhooks. University of Hawaii Press. Koch, P. L. 1998. Isotopic reconstruction of past continental environments. Annu. Rev. Earth Planet. Sci, 26: 573-613. Koch, P. L., M. L. Fogel and N. Tuross. 1994. Tracing the diets of fossil animals using stable isotopes. In: Stable isotopes in Nutrition. Eds. Lajtha, K. and R. Michener, American Chemical Society, Washington, D. C., p. 63-92. Livezey, B. C. 1998. A phylogenetic analysis of the Gruifon'nes (Aves) based on morphological characters, with an emphasis on the rails (Rallidae). Philosophical Transactions of the Royal Society of London, 353: 2077- 2151. Livezey, B. C. 1993. An ecomorphological review of the dodo (Raphus cucullatus) and solitaire (Pezophaps solitaria), flightless Columbiforrnes of the Mascarene Islands. Journal of Zoology, 230: 247-292. MacArthur, R. H. and E. O. Wilson. 1967. The Theory of Island Biogeography. Monographs in Population Biology, Princeton University Press, Princeton, New Jersey. MacFadden, B. J., Y. Wang, T. E. Carling and F. Anaya. 1994. South American fossil mammals and carbon isotopes: a 25 million-year sequence from the Bolivian Andes. Palaeogeography, Palaeoclimatology, Palaeoecology, 107: 257-268. . Macko. S. A. 1981. Stable nitrogen isotope ratios as tracers of organic geochemical processes: PhD Thesis, University of Texas at Austin, p. 181. Magnusson, W. E., M. Cannozine de Araujo, R. Cintra, A P. Lima, L. A. Martinelli, T. M. Sanaiotti, H. L. Vasconcelos, and R. L. Victoria. 1999. Contribution of 03 and C4 to higher trophic levels in an Amazonian savanna. Oecologia, 119: 91-96. Marra, P. P., K A Hobson, and R. T. Holmes. 1998. Linking winter and summer events in a migratory bird by using stable-carbon isotopes. Science, 282: 1884-1886. 59 Martinelli, L. A, S. Almeida, I. F. Brown, M. Z. Moreira, R. L. Victoria, L. S. L. Sternberg, C. A. C. Ferreira, and W. W. Thomas. 1998. Stable carbon isotope ratio of tree leaves, holes and fine litter in a tropical forest in Rondonia, Brazil. Oecologia, 114: 170-179. Matheus, P. E. 1995. Diet and Co-ecology of Pleistocene Short-faces Bears and BroWn Bears in Eastern Beringia. Quaternary Research, 44: 447-453. McNulty, T. W., A E. Calkins, P. H. Ostrom, H. Gandhi, M. Gottfried, L. D. Martin, and D. A. Gage. In press. Stable isotope interpretations of bone organic matter. An artificial diaganesis experiment and paleoecology of the Pleistocene and Holocene of Natural Trap Cave, Wyoming. Palaios. Mueller-Dombois, D. 1981. Some bioenvironmental conditions and the general design of IBP research in Hawaii. In: Island Ecosystems: Biological Organization in Selected Hawaiian Communities. Eds. D. Mueller- Dombois, K W. Bridges, and H. L. Carson, The Institute of Ecology, Hutchinson Ross Publishing Company, p. 3-34. Olson, S. L. and H. J. James. 1982a. Fossil birds from the Hawaiian Islands: evidence for wholesale extinction by man before Western contact. Science, 217: 633-635. Olson, S. L. and H. J. James. 1982b. Prodosmus of the fossil avifauna of the Hawaiian Islands. Smithsonian Contributions to Zoology, 365: 1-59. Olson, S. L. and H. J. James. 1984. The role of Polynesians in the extinction of the avifauna of the Hawaiian Islands. In: W. Eds. P. S. Martin and R. G. Klein. University of Arizona Press, Tucson. Olson, S. L. and H. F. James. 1991. Descriptions of thirty-two new species of birds from the Hawaiian Islands: Part I. Non-passerines. MW Mgnggraghs No. 45, ed. N. K Johnson, The American Omithologists’ Union, Washington D. C. Ostrom, P. H., S. A. Macko. M. H. Engel and D. A Russell. 1993. Assessment of trophic structure of Cretaceous communities based on stable nitrogen isotope analyses. Geology, 21: 491-494. Paxinos, E. E. 1998. Prehistoric anserifonn diversity in the Hawaiian islands: a molecular perspective from the analysis of subfossil DNA PhD , dissertation, Division of Biology and Medicine at Brown University. Peterson, B. J. and B. Fry. 1987. Stable isotopes in ecosystem studies. Annual Reviews in Ecological Systematics, 18: 293-320. 60 Ralph and Van Riper. 1985. Historical and current factors affecting Hawaiian native birds. In: Bird Conservation 2. Ed. S. A. Temple, The International Council for Bird Preservation, p.7-42. Rave, E. H. 1995. Genetic analysis of wild populations of Hawaiian geese using DNA fingerprinting. The Condor, 97: 82-90. Robbins, C. S., B. Bruun and H. S. Zim. 1983. A guide to field identification: Birds of North America. Golden Press, New York. Schoeller, D. A, M. Minagawa, R. Slater and T. R. Kaplan. 1986. Stable isotopes of carbon, nitrogen and hydrogen in the contemporary North American human food web. Ecology of Food and Nutrition, 18: 159-170. Schoeninger, M. J. and M. J. DeNiro. 1984. Nitrogen and carbon isotopic composition of bone collagen from marine and terrestrial animals. Geochimica at Cosmochimica Acta, 49: 625-639. Schwarcz, H. P. 1991. Some theoretical aspects of isotope paleodiet studies. Journal of Archaeological Science, 18: 261 -275. Stone, C. P. and L. W. Pratt. 1994. Hawai'i’s Plants and Animals: Biological Sketches of Hawaii Volcanoes National Park. Hawaii Natural History Association, University of Hawaii Press, Honolulu. Trewick, S. A. 1997. Flightlessness and phylogeny amongst endemic rails (Aves: Rallidae) of the New Zealand region. Philosophical Transactions of the Royal Society of London, 352: 429-446. Vltousek, P. M., c. a. Field and P. A. Matson. 1990. Variation in foliar 5‘30 in Hawaiian Metrbsidsros polymorpha: a case of internal resistance? Oecologia, 84: 362-370. Vltousek, P. M., a. Shearer and o. H. Kohl. 1989. Foliar "N natural abundance in Hawaiian rainforest: patterns and possible mechanisms. Oecologia, 78: 383-388. Wagner, W. L. and V. A. Funk. 1995. Hawaiian Biogeography.: Evolution on a Hot Spot Archipelago. Smithsonian Institution. Wetmore, A 1943. An extinct goose from the island of Hawaii. The Condor, 45: 146-148. Wong, W. W., L. L. Clark, G. A Johnson, M. Llaurador, and P. D. Klein. 1992. Comparison of two elemental analyzer gas isotope ratio mass 61 spectrometer systems in the simultaneous measurement of 130/12C ratios and carbon content in organic sample. Analytical Chemistry, 64: 354-358. 62 HIHHHHHIH 293 02318 2540