RESEARCH COMMUNICATIONS RESEARCH COMMUNICATIONS Conservation implications of historic sea turtle nesting beach loss Loren McClenachan , Jeremy BC Jackson ' , and Marah JH Newman Populations of endangered Caribbean sea turtles are far more depleted than realized because current conser- vation assessments do not reflect historic nesting data. We used historical sources to analyze changes in the numbers of nesting populations and population sizes for green and hawksbill turtles on all known nesting beaches in the Caribbean over the past millennium. We present the first maps of historic nesting populations, which provide the basis for an objective measure of changes in distribution and abundance. Our results indi- cate that 20% of historic nesting sites have been lost entirely and 50% of the remaining nesting sites have been reduced to dangerously low populations. Recent conservation efforts have resulted in large population increases at several nesting sites, but loss of widespread nesting throughout the Caribbean and reductions in the Caribbean-wide population since human hunting began indicate that Caribbean turtles are far from recovered. Focusing attention on a small number of nesting populations is a risk-prone strategy; conservation programs should instead broaden their scope to protect both large and small nesting populations throughout the Caribbean. Front Ecol Environ 2006; 4(6): 290-296 By the beginning of the 20th century, green turtles (Chelonia mydas) and hawksbill turtles (Eretmochelys imbricata) had been decimated by human hunting, making both species globally endangered (Meylan and Donnelly 1999; Seminoff 2002). These species filled unique ecolog- ical roles in seagrass and coral reef ecosystems, and their removal diminished the complexity and stability of the food web, as well as the intensity of biological disturbance on seagrass beds and coral reefs (Jackson 1997; Bjorndal and Jackson 2003). Population declines and the ensuing ecological changes occurred over many centuries; without historic data, the magnitude of change has been underes- timated, a phenomenon known as the shifting baselines syndrome (Pauly 1995). Historic reductions in sea turtle populations have been recognized (Jackson 1997; Bjorndal and Jackson 2003), but pan-regional changes in numbers and distribution of nesting populations have not been systematically reviewed. To that end, we compiled a comprehensive list of historic nesting beaches for green and hawksbill turtles in the Caribbean and used the num- ber of nesting sites to refine previous estimates of historic population size and the ecological consequences of loss. Historical and archeological data provide a wealth of information that can be used to estimate early geographic ranges and population sizes of easily visible species such Center for Marine Biodiversity and Conservation, Scripps Institution of Oceanography, University of California, San Diego, Lajolla, CA, USA *(lmcclena@ucsd.edu); Geosciences Research Division, Scripps Institution of Oceanography, University of California, San Diego, La Jolla, CA, USA; Center for Tropical Paleoecology and Archeology, Smithsonian Tropical Research Institute, Apartado 2072, Balboa, Republic of Panama as sea turtles that nest on land and whose high economic value stimulated exceptional records of their exploita- tion. For hundreds of years, green turtles provided nour- ishment for European colonists and African slaves on Caribbean sugar plantations. Hawksbill turtles were prized for their shells, which were fashioned into elabo- rate hair combs and other ornaments and distributed through carefully regulated trade networks (Roberts 1827; Dampier 1968). Both species of turtle were espe- cially vulnerable to hunting during nesting (Roberts 1827; Dampier 1968; Rebel 1974; Jackson 1997). Early descriptions provide locations of nesting beaches, the magnitude of the population, sizes of adults found, and accounts of the hunt. We compiled data on nesting beach location, density of turtles, and human exploitation summarized in trade records from 163 historic sources in four historic time periods for 20 regions of the Caribbean (WebTable 1). We mapped historic nesting beaches for green and hawks- bill turtles, and used density descriptions and harvest data to categorize these sites as "major" and "minor" nesting sites (Figure 1; WebTable 2). Next, we calculated a range of Caribbean-wide population sizes for green and hawks- bill turtles by estimating the number of adult turtles sup- ported by one particularly well-documented major nest- ing site for each species and extrapolating across the region, using the total number of major and minor his- toric nesting beaches (WebTable 3). Two types of sources provided information about the size of populations: (1) observations from 20th century nesting beaches, and (2) historic harvest data. (Full materials and methods are available as Web-only material.) Finally, we refined cal- culations of historic turtle consumption in order to www.frontiersinecology.org ? The Ecological Society of America L McClenachan et al. Conservation implications of beach loss describe the ecological role of turtles in tropical marine ecosystems and the long-term effects of their removal. Historic nesting beaches and population size Historically, large nesting populations were found on beaches throughout the wider Caribbean. We found evi- dence for 59 historic green turtle nesting beaches, nine of which were considered to be major nesting populations based on density descriptions and harvest data (Figure la; WebTable 2). Green turtles in the Cayman Islands, for example, were found in "infinite numbers"; up to 50 nest- ing females could be taken in less than 3 hours (de Rochefort 1666). On the Moskito Coast of Nicaragua there were "inexhaustible supplies of the finest green tur- tle" (Roberts 1827) and settlers "turned so many that [they] were obliged to desist" (Williams 1969). For hawksbill turtles, we located 55 historic nesting beaches, seven of which supported major historic populations (Figure lb; WebTable 2). Major nesting beaches included the island of Roncador, off the Nicaraguan coast, which was "famous for the number of its turtles...the shore seemed black with turtles" (Squier 1865), and Chiriqui, Panama, which was considered to be the "most important nesting aggregation in the Caribbean" (Carr 1956). We used these geographic nesting data to calculate a range of population sizes, based on quantitative modern nesting data and historic export data. For green turtles, observations from 20th century nesting in Tortuguero, Costa Rica (Troeng and Rankin 2005) indicate that this nesting site supported an average of 130500 adults (WebTable 3). We first assumed that each of the nine his- toric major nesting sites supported populations as large as this recent Tortuguero population and that the remaining 50 minor nesting aggregations were each 10% of that value. These calculations yield an estimated historic popu- lation of 1.8 million adult green turtles (WebTable 3). Large as this number may seem compared with modern abundances of less than 300000 (Seminoff 2002), historic hunting data indicate that the 17 th century Cayman Island green turtle population alone was approximately 6.5 million adults (Jackson 1997; WebTable 3). Assuming that each of the nine major nesting beach populations was as large as the historic Cayman Island population and the 50 minor nesting aggregations were only 10% of this size yields a historic population for the Caribbean of around 91 million adult green turtles (WebTable 3). For hawksbill turtles, observations of 20th century nesting at Chiriqui Beach, Panama (Meylan and Donnelly 1999), indicate that this nesting site supported 135 000 adults in the 1950s (WebTable 3). Assuming that the seven major nesting sites each had abundances comparable to that of Chiriqui, and that each of the remaining 48 minor nesting sites had populations of only 10% of this number, the total Caribbean population was 1.6 million adult hawksbill turtles (WebTable 3), com- pared to fewer than 30000 today (Bjorndal and Jackson o U 500 nesting females (large triangles) and 100-500 (small triangles) are mapped. All modern and historic nesting data are listed in WebTable 2. 2003). However, historic export data from the 19th cen- tury Bahamian fishery (Northcroft 1900) provide an esti- mate of 936 600 adults from just this region (WebTable 3). Extrapolating across the seven major and 48 minor nesting beaches gives a historic population estimate of 11 million hawksbill turtles in the Caribbean region (WebTable3). Our green turtle population estimate based on historic nesting data is more than two times greater than the 33-39 million green turtles estimated by Jackson (1997), and our hawksbill turtle calculation is 20 times greater than the extremely conservative estimate of 540000 adult hawksbill turtles by Bjorndal and Jackson (2003). Our estimates may be high, as they assume that all major nesting sites were as large as the Bahamian and Cayman Island populations. Furthermore, some non-nesting tur- tles may have been mistaken for nesting females, poten- tially increasing the estimated number of nesting beaches. Because they are based on conservative and offi- ) The Ecological Society of America www.frontiersinecology.org Conservation implications of beach loss L McClenachan et al. cially recorded estimates of catch, however, our estimates could also be too low. For example, Cayman hunting data do not reflect exploitation by Spanish, French, Dutch, and other English settlers and pirates in the Caribbean, who did not report their catch (eg Jackson 1924). While extrapolations from anecdotal historic evidence will never be precise, using true historic data provides an accurate assessment of the order of magnitude of change that cannot be determined from traditional ecological data, particularly for severely hunted populations. Population declines and nesting beach loss Our calculations based on historic export data show that modern populations of green and hawksbill turtles are 0.33% and 0.27% of their historic numbers, respectively. These calculations assume modern populations of 300000 green and 30000 hawksbill turtles. Such stagger- ing declines in abundance have been compounded by the elimination of entire nesting populations that are extremely unlikely to become re-established (Seminoff 2002). The loss of even a single nesting site makes a per- manent, irreversible dent in the sea turtle population, but loss of nesting beaches has not been quantified, nor used as a measure of population change across the Caribbean region. Our data indicate that historic hunting com- pletely eliminated at least 17 green turtle and seven hawksbill nesting sites, including three major nesting beaches: Bermuda; Moskito Coast, Nicaragua; and Dry Tortugas, US (WebTable 2). Hunting also severely thinned turtles at the remaining sites, so that half of modern nesting populations have an uncertain future. So-called "nesting aggregations" are pitifully small, often consisting of females that nest singly. Such small sites include the once great Cayman Island green turtle rook- ery. Today, 55% of green turtle and 44% of hawksbill nesting beaches host fewer than 10 nesting females, or are described as having only rare nesting (Figure 2; WebTable 2). Historic data clearly show that each of these nesting sites supported large populations in the past. Ecosystem consequences The severe reduction of turtle numbers is of concern not only because of the turtles themselves, but also with regard to their previously important roles as ecosystem engineers in Caribbean ecosystems (Bjorndal and Jackson 2003). Green turtles feed primarily on turtle grass (Thahxssia tes- tudinum; Thayer et al. 1982) and hawksbill turtles have a unique dietary preference for marine sponges (Leon and Bjorndal 2002). Both turtle grass and sponges are impor- tant habitat-structuring species throughout the region. We estimated total food intake of historic populations of green and hawksbill turtles (WebTables 3 and 4). Ninety- one million green turtles consumed between 11 and 22 million metric tons dry mass (DM) of turtle grass, which amounts to 86% of the total area and up to 45% of the annual productivity of seagrass beds (WebTables 3 and 4). Eleven million hawksbill turtles consumed between 0.9 million and 2.0 million metric tons DM of sponges annu- ally, or 83% of the biomass and annual growth of sponges (WebTables 3 and 4). The geographic scale at which his- toric populations of turtles disturbed coral reef and seagrass communities is inconsistent with modern observations of seagrass beds that grow virtually unchecked by grazing and of coral reefs where few large predators remain. Our calcu- lations indicate that today's green and hawksbill turtle pop- ulations consume just 0.1% of the area of Caribbean sea- grass and reef sponges, respectively. In the 1830s, the great naturalist, John J Audubon described the seagrass beds of the Dry Tortugas as "cut near the roots" by vast numbers of grazing green turtles (Audubon 1926), an ecological state also described by William Dampier in the 1680s (Dampier 1968). Our cal- culations agree with these descriptions (WebTable 4). Ninety-one million grazing green turtles left behind large patches of actively growing seagrass clipped down to the blade-sheath junction (Thayer et al. 1982). The ecologi- cal extinction of green turtles transformed an ecosystem with diverse species of seagrasses dominated by large her- bivores into a detritus-based ecosystem dominated by overgrown monocultures of T testudinum, with two impor- tant conservation implications. First, the annual removal of 86% of mature seagrass blades would have greatly inhibited the spread of epiphytic organisms that characterize modern seagrass beds, and thus preempted the spread of seagrass wasting disease. The dis- ease-causing parasitic protist, Labyrinthula, attaches prefer- entially to mature seagrass blades, from which it colonizes actively growing seagrass stems (Bowles and Bell 2004). Wasting disease is now widespread and is unlikely to disap- pear unless grazing is reintroduced on an appropriately large geographic scale, even if other factors such as excess nutrients also play a role in the spread of the disease (PandolfieW. 2005). Second, in contrast to green turtles, grazing fishes and sea urchins lack the microbial symbionts that metabolize cellulose, which comprises most of the carbon in turtle grass blades. Their waste, as well as unconsumed turtle grass, is largely buried in sediments (Thayer et al. 1982), where it is no longer available to animals in the grazing food chain. The decline of green turtles has therefore resulted in a loss of productivity available to the animal food chain - including commercially exploited reef fishes - and therefore amounts to a reduction in protein-rich food available for Caribbean people. Similarly, on reefs, historic consumption of sponges by hawksbill turtles was up to 800 times higher than that of modern populations (WebTable 4); this has implications for the sponge community composition and the relative abundances of sponges and reef corals (Bjorndal and Jackson 2003). Hawksbill turtles preferentially feed on non-toxic sponges when they are available, but can sur- vive on a mix of toxic species (Leon and Bjorndal 2002). www.frontiersinecology.org ? The Ecological Society of America L McClenachan et al. Conservation implications of beach loss Thus, as turtles declined in abundance, the relative quantities of toxic sponges that each hawksbill turtle consumed should also have decreased. Historical data support this hypothesis. Observations from the 17th through the 20th century indicate that toxi- city of hawksbill turtle meat for human con- sumption has decreased over time (Table 1). This unanticipated result provides another independent measure of the extreme reduc- tion in hawksbill turtle populations, as well as indirect evidence for changes in Caribbean benthic ecosystems. The role of turtles as major agents of land- scape-scale patterns of disturbance has been questioned, based on small-scale experimen- tal studies that attempt to test the effects of turtle grazing in tiny plots (eg Bowles and Bell 2004). Such experiments cannot mimic the intensity of disturbance of tens of mil- lions of turtles across the entire Caribbean any more than clipping a few small quadrats of prairie grass could possibly recreate the effects of 30 million American bison on the Great Plains (Isenberg 2000). Our historical data indicate that centuries ago, much of the mobile animal biomass in the Caribbean was concentrated in the bodies of large animals, an ecological possibility supported by mod- ern surveys on isolated and protected reefs (Friedlander and DeMartini 2002). These data strongly suggest that the extirpation of large animals was the first step in disman- tling Caribbean marine ecosystems, and cir- cumstantial evidence - such as recent out- breaks of seagrass disease and coral overgrowth - supports the inference that breakdown in structural habitat magnified the loss of large animals. Clearly, successful conservation and management of turtles is an essential component in achieving ecosys- tem restoration. Good news for sea turtles? The protection of nesting beaches since the 1970s has resulted in extraordinary local population increases in short periods of time, particularly among green turtles (Hays 2004; Troeng and Rankin 2005). These encourag- ing results have led some to question whether green turtles are endangered within the Caribbean and greater Atlantic region (Broderick et al. 2006). Patterns in modern nesting data suggest that nesting beach con- servation efforts have indeed been highly successful in reversing downward population Trin & Tobago Haiti mainland Guyana San Andres Suriname PR Vieques PR mainland PR Culebra Costa Rica Anguila USVI St John Anguilla St Lucia PR mainland Montserrat St Lucia DR Saona DR Saona USVI St Thomas Cayman St Vin & Gren Mex Tamaulipas Nic south coast 0) 0) c o ? V V> 0) c o o ^H Green turtle ^H Hawksbill turtle j| IUCN Index Sites (green turtle) i ^?1 HH ^g^^g Mex Veracruz Nic Pearl Cays Monserrat Nic Moskito Coast Belize Belize Barbuda BVI Antigua Barbados Cuba C Largo ^MM ^^?^H e^^^ USVI St Croix Cuba Juventud PR Mona Mex Campeche Martinique Jamaica Antigua/Barbuda Trin & Tobago Mex Quin Roo BVI Pan Chiriqui Mex Quin Roo Guatemala Fr Guiana Turks & Caicos DR mainland DR mainland Aves Jamaica US Florida Mex Yucatan Mex Yucatan Mex Gulf of Mex Suriname Cuba CR Tortuguero Mex Campeche i ? 1 D? 101 102 103 104 105 Annual number of nesting females Figure 2. The number of females nesting annually on modern sites is very small and unevenly distributed, with 10% of sites hosting 90-95% of nesters. Large sites are the focus of conservation assessments such as the IUCN Red List Global Status Assessment for green turtles. The figure does not include 29 nesting sites for which reliable quantitative data do not exist; 24 of these sites are described as having rare, scattered, or infrequent nesting. All modem nesting data are listed in WebTable 2. ) The Ecological Society of America www.frontiersinecology.org Conservation implications of beach loss L McClenachan et al. Table 1. Changes in hawksbill turtle toxicity Date Location Observations of hawksbill turtle toxicity 1684 Panama "Yet these Hawks-bills, in some places are unwholesome, causing them that eat them to purge and vomit excessively..." (Dampier 1968). 1760 French Caribbean "...it is dangerous to eat of [the hawksbill's] flesh, which, though fat and delicious, is of so purgative a quality, that unless you take but a little, or are well assured that you have nothing to fear from its activity, you may expect to see yourself covered with pimples and blotches" (Jefferys 1760). 1770s Nicaragua "...this kind of Tortoise is not very agreeable to the taste, nor do we eate them" (Williams 1969). 1837 Florida "The hawksbill ... is not highly valued for food" (Williams 1837). 1884 Caribbean "The flesh of the hawksbill turtle is comparatively valueless; indeed, in the West Indies it is said that it possesses cathartic qualities in a high degree... I have seen it in Washington several times recently, both in the markets and before several restaurants in the city" (True 1884). 1900 Bahamas "All three kinds [green, hawksbill, loggerhead] are eaten. It is an unfortunate policy which takes them recklessly each season - though they are pleasant enough to eat and most nourishing - and thereby causes turtle to become scarcer each year" (Northcroft 1900). 1945 Jamaica "Formerly the chief value of the Hawksbill was for the shell, which sold at high prices. Now...the market is for meat, which finds a ready market locally" (Thompson 1945). 1974 Barbados an d Panama "Fishing is mainly for hawksbill for meat and shell... The meat and eggs of the...hawksbill turtle are taken for local consumption" (Rebel 1974). Hawksbill turtle meat was toxic until the mid to late 19th century when it began to be eaten without health consequences. These observations suggest that hawksbill turtles ate more desirable, less toxic sponge species as the turtles became less abundant and competition for food was reduced. trends on a few well-studied beaches. Our analysis of trends among IUCN-assessed nesting beaches (Seminoff 2002) suggests that the situation has improved in the past decade; nesting data collected since 1994 show a popula- tion increase when compared to data collected between 1980 and 1994 (Figure 3; P = 0.011). Despite recent conservation achievements, declaring success would be a mistake for two reasons. First, there is very little long-term data, despite the IUCN mandate to determine nesting beach trends over three generations. Instead, short-term data are extrapo- lated over longer time periods, a dan- gerous method considering that popu- lation trends are known to be quickly reversible (Hays 2004; Troeng and Rankin 2005; Broderick et a!. 2006). Time series that span more than one turtle generation exhibit significant declines; data extending over more than 40 years are highly likely to show long-term declines when compared with recent data (Figure 3; difference in results, P = 0.029). Because of this systematic difference in results, short- term data should not be used to infer long-term change. Using modem data to speculate about historic change is certain to dangerously underestimate long-term population change. Second, despite some success at a few sites, most nesting beaches across the region have suffered enormous, unmeasured losses and are much smaller than those used in conserva- tion assessments. The Caribbean component of the IUCN Red List Global Assessment (Seminoff 2002) was based on data from five nesting sites, four of which have increased over the past three decades. These increases demonstrate the efficacy of nesting beach protection, but should not be mistaken for a sign of effective conservation across the region, because these sites are anomalously large and well protected. The five IUCN sites are the largest sites in the Caribbean. Annual aggregations of nesting females range from about 300 at Aves Island in Venezuela to a few tens of thousands at Tortuguero, Costa Rica (Seminoff 2002; Figure 2; WebTable 2). Therefore, generaliza- tions based solely on these few large beaches inevitably gives a false picture of the overall status of Caribbean green turtle populations. A more accurate assessment of regional change would consider trends on all known nesting beaches, despite lack of preci- sion in numbers of nesting females on some smaller sites. As Pauly (1995) emphasized in his landmark paper on "shifting baselines", most of the big changes to large marine vertebrate populations occurred many decades to centuries ago, prior to quantitative monitoring programs. Therefore, capturing the magnitude of these changes requires the use of historic data. Our imprecise but com- prehensive historic nesting data describe changes over a www.frontiersinecology.org ? The Ecological Society of America L McClenachan et al. Conservation implications of beach loss U) C o c g Q. O Q. U) _C '?4-> (/) 200-, 150- 100- 50- -50- -100 much broader geographic and temporal scale, providing a picture of regional deple- tion that has neither stabilized nor reversed. Of the 59 green turtle nesting beach sites documented in our study, 29% have been lost and 55% of the rest are so small that they will probably disappear if not protected. A strategy that focuses attention on a few exceptional nesting beaches runs the risk of allowing the destruction of smaller, historically impor- tant nesting beaches without realizing the losses that have occurred. Determining conservation strategies for marine turtles over more than a few years inevitably involves a great deal of uncer- tainty. Important new tools, such as infor- mation gap theory, have been developed to explicitly include uncertainty when assess- ing possible management actions and deter- mining the degree of risk that can be taken to achieve desired results (Ragen et al. 2005). The currently popular focus on nest- ing data from a few major nesting beaches (eg Broderick et al. 2006) ignores the uncer- tain future of nearly 90% of the remaining nesting beaches. This risk-prone strategy does not account for factors such as extreme storms, disease, or other catastrophic events. Historically, green and hawksbill turtles were ubiquitous, abundant, and nested in high densities throughout the Caribbean. On an evolutionary timescale, widespread nesting was a risk-averse "evolutionary strategy" for the persistence of turtle species. Humans have reduced green and hawksbill turtles nesting beaches by one fifth and without proper protection, half of the remaining nesting beaches could soon be lost. Even very small nesting beaches do recover, how- ever (Hays 2004), so that protection and scientific research funding should be extended across as many beaches as possible, especially those that were once important, but are now greatly reduced. Protecting more nesting beaches is not a politically or socially simple endeavor, but it is the only way to avoid the risk of putting all the remaining turtle eggs in a very few baskets. o o O o o oto o o o oo o oo, o a. Nesting data end-date O 1995-2001 O 1980-1994 O CP o o, b 20 40 Short-term data 60 80 Long-term data 100 ?r 120 Duration of nesting data (years) Figure 3. Shifting baselines in green turtle nesting trends and recent conservation success. Data used in the IUCN Red List Global Status Assessment for green turtles (Seminoff 2002) show very large declines in numbers of nesting females over the past century, but increases on particular nesting beaches in the last decade. Each circle represents a single Index Nesting Site used in the IUCN Assessment. All time series > 40 years (right of vertical dotted line) record population declines of 65-93%. Long-term data are significantly more likely to record population declines than shorter time series (two-tailed Fisher's exact test, P = 0.029). However, short-term time series ending after 1994 (blue circles) show increases when compared to similar time series ending in 1994 or before (red circles; two-tailed Fisher's exact test, P= 0.011). We calculated percent change using past and present nesting data (Seminoff 2002) and used the midpoint when a range of years was given. Solid circles indicate Caribbean index sites. ments. This work was supported by the Center for Marine Biodiversity and Conservation's Integrative Graduate Education and Training Program, the Scripps Institution of Oceanography's William E and Mary B Ritter Chair, the History of Marine Animal Populations Program of the Census of Marine Life, and the University of California, San Diego's Institute for International, Comparative and Area Studies. Acknowledgements We thank S Sandin, E Sala, N Knowlton, H Lotze, D Pauly, K Alexander, and 5 Troeng for critical reading of the manuscript; S Sandin for suggesting analysis of recent IUCN data; K Alexander for sharing archival data; C Varella for support in Seville, Spain; and the Archivo General de Indias, the Bahamas National Archives, and the UK National Archives for providing access to docu- References Audubon JJ. 1926. Delineations of the American scenery and char- acter. New York, NY: Baker and Company. Bjorndal K and Jackson JBC. 2003. Roles of sea turtles in marine ecosystems: reconstructing the past. In: Lutz PL, Musick JA, and Wyneken J (Eds). The biology of sea turtles, volume 2. Boca Raton, FL: CRC Press. Bowles JW and SS Bell. 2004. Simulated herbivory and the dynamics of disease in Thalassia testudinum. Mar Ecol-Progr Ser 283: 127-32. Broderick AC, Frauenstein R, Glen F, et al. 2006. Are green turtles ) The Ecological Society of America www.frontiersinecology.org Conservation implications of beach loss L McClenachan et al. globally endangered? Global Ecol Biogeogr 14: 21-26. Carr A. 1956. The windward road. New York, NY: Alfred Knopf. Dampier W. 1968. A new voyage around the world. New York, NY: Dover Press. de Rochefort C. 1666. The history of the Caribby-islands. London, UK: T Dring and ] Starkey. Friedlander AM and DeMartini EE. 2002. Contrasts in density, size, and biomass of reef fishes between the northwestern and the main Hawaiian islands: the effects of fishing down apex predators. Mar Ecol-Prog Ser 230: 253-64- Hays GC. 2004. Good news for sea turtles. Trends Ecol Evol 19: 349-51. Isenberg AC. 2000. The destruction of the bison: an environmen- tal history, 1750-1920. Cambridge, UK: Cambridge University Press. Jackson JBC. 1997. Reefs since Columbus. Coral Reefs 16: S23-32. Jackson W. 1924. The voyages of Captain William Jackson (1642-1645). London, UK: Offices of the Society. Jefferys T. 1760. A natural and civil history of the French domin- ions in North and South America. London, UK: Charing- Cross. Leon YM and Bjorndal KA. 2002. Selective feeding in the hawks- bill turtle, an important predator in coral reef ecosystems. Mar EcoLProg Ser 245: 249-58. Meylan AB and Donnelly M. 1999. Status justification for listing the hawksbill turtle (Eretmochelys imbricata) as Critically Endangered on the 1996 IUCN Red List of Threatened Animals. Chelonian Conserv B: 3: 200-24- Northcroft GJH. 1900. Sketches of summerland, giving some account of Nassau and the Bahama Islands. Nassau, Bahamas: Nassau Guardian. Pandolfi JM, Jackson JBC, Baron N, et al. 2005. Are US coral reefs on the slippery slope to slime? Science 307: 1725-26. Pauly D. 1995. Anecdotes and the shifting baselines syndrome in fisheries. Trends Ecol Evol 10: 430. Regan HM, Ben-Haim Y, Langford B, et al. 2005. Robust decision making under severe uncertainty for conservation manage- ment. Ecol Appl 15: 1471-77. Rebel TP. 1974. Sea turtles and the turtle industry of the West Indies, Florida, and the Gulf of Mexico. Coral Gables, FL: University of Miami Press. Roberts OW. 1827. Narrative of voyages and excursions on the East Coast and the interior of Central America. Edinburgh, UK: Constable and Co. Seminoff JA. 2002. IUCN Red List Global Status Assessment: green turtle (Chelonia mydas). Gland, Switzerland: IUCN Marine Turtle Specialist Group. Squier EG. 1855. Waikna: adventures of the Mosquito Shore by Saml A Bard. London, UK: Sampson Low, Son and Co. Thayer GW, Engel DW, and Bjorndal KA. 1982. Evidence for short-circuiting of the detritus cycle of seagrass beds by the green turtle, Chelonia mydas L. J Ex?> Mar Biol Ecol 62: 173-83. Thompson E. 1945. The fisheries of Jamaica. Bridgetown, Barbados: Advocate Co. Troeng S and Rankin E. 2005. Long term conservation efforts con- tribute to positive green turtle Chelonia mydas nesting trend at Tortuguero, Costa Rica. Biol Conserv 121: 111-16. True FW 1884. Reptiles and batrachians. In: Goode GB (Ed). The fisheries and fisheries industries of the United States, volume 1. Washington, DC: Government Printing Office. Williams JL. 1837. The territory of Florida: or sketches of the topography, civil and natural history, of the county, the cli- mate, and the Indian tribes from the first discovery to the pre- sent time with a map, views, etc. New York, NY: AT Goodrich. Williams W. 1969. Mr Penrose: the journal of Penrose, seaman. Bloomington, IN: Indiana University Press. TAKE THIS JOURNAL TO YOUR LIBRARIAN, PLEASE Are you enjoying this issue of Frontiers? If your library had a subscription, colleagues and students could enjoy it too. Please consider recommending Frontiers in Ecology and Environment to your library. Clip or copy the form below. Thank you for your support. Library Recommendation Form To Acquisition Librarian, Serials Signature I recommend the library subscribe to: Frontiers in Ecology and the Environment (ISSN 1540-9295) To request a free sample issue of Frontiers in Ecology and the Environment, call (301) 588-4691 or email Sika Dunyoh ^^^rtjrf at sika@esa.org. Order Frontiers by contacting ESA Headquarters at (202) 833-8773, online at www.esa.org, or through your subscription agent. www.frontiersinecology.org ? The Ecological Society of America