People of the sea, or of the soil? How the balance of marine and terrestrial resource availability informs maximum population on four Polynesian islands

People of the sea, or of the soil? How the balance of marine and terrestrial resource availability informs maximum population on four Polynesian islands | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article People of the sea, or of the soil? How the balance of marine and terrestrial resource availability informs maximum population on four Polynesian islands Cedric O. Puleston, Jennifer G. Kahn, Oliver A. Chadwick, Nick Belluzzo, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4797211/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Most studies of the food resource potential of early Polynesian populations focus exclusively on agricultural potential, and specifically starchy staples, despite the importance of the marine world to the Polynesians. In an attempt to more accurately estimate total precontact food resource availability, we characterized the terrestrial and near-shore marine environments of four Polynesian islands: Moʻorea, Maupiti, Mangareva and Taravai. We estimate the agricultural potential of each island after a consideration of the ecological factors related to productivity. We also estimate the productivity of the near-shore marine environment as a function of surface area. Using a range of fish yield scenarios from Pacific subsistence systems we scale relative measures of faunal food biomass derived from local archaeological excavations to generate absolute biomass estimates. We convert our estimates of agricultural and fished/foraged food potential into a maximum population size based on calorie availability. Moʻorea’s fertile valleys and wetlands would potentially generate sufficient food energy, mostly from starchy staples, to support a considerable population, many times larger than the other islands. In the most likely fish yield scenario Moʻorea gets only 0.4% of its calories from marine sources, while the others range from 7-18%. These relative inputs reflect the vast superiority in productivity (in terms of calories per km 2 ) of agricultural vs near-shore marine zones. We also find that population size on islands with smaller fringing reefs, as illustrated by Moʻorea, may have been limited by a lack of fish protein. Moʻorea’s maximum population is approximately halved when diet breadth is considered. Agroecology Ecological Modeling Population Biology Marine and Freshwater Ecology Terrestrial Ecology Anthropology Archaeology Geographic Information Systems French Polynesia marine productivity subsistence agriculture pre-contact population INTRODUCTION Elements of a common culture, ancestral language and a dependence on a core set of common plants and animals unite many Pacific populations across stretches of open ocean that span more than 70 degrees of latitude, from the Hawaiian Archipelago to the southern tip of Aotearoa New Zealand. However, despite these powerful commonalities, early European visitors were often struck by how much the islands differed from one another, in the population sizes or densities of occupation, the availability of food and goods for trade, and the apparent quality of life. This variability has been noted by anthropologists and archaeologists. Certainly, some of this variation is a consequence of historical accident, including the effect of choices individuals and groups made in the thousand or more years since the more remote islands were first settled. But an additional source of variability lies in the nature of the islands themselves: in the wide variation in ecological characteristics, including climate, geologic substrate, the likelihood of catastrophic weather events, proximity to other islands or landmasses, the presence of fertile valley bottoms and the extent of a surrounding reef, each of which influenced how the human population lived. Thus, island biogeography and other ecological factors have played a key role in debates concerning precontact sociopolitical complexity and human ecodynamics in the Pacific Islands. Scholarly efforts to identify the features of Polynesian islands that are most predictive of population size and density have, for good reason, focused on island area and resource availability. Larger islands are generally host to larger human populations, but the relationship does not appear to be linear. In a careful study of several Polynesian outlier atolls Tim Bayliss-Smith (1974) proposed that human populations were limited by their ability to produce the staple crops with a high return in starch calories to labor inputs. In a response, Stephen Beckerman (1977) proposed the “protein hypothesis,” in which he argued that access to coconuts and near-shore protein sources were limiting, supporting his argument with data showing a strong correlation between island circumference and estimates of precontact population size. Rosalind Hunter-Anderson and Yigal Zan persuasively argued that Beckerman’s analysis was flawed and claimed instead, that “the upper limit of population size appears to be conditioned by the presence of intensifiable resources. Under certain geographical conditions, these resources might involve protein, and under others, starch. In small tropical high islands, the intensifiable resources tend to be starch” (Hunter-Anderson and Zan 1985). Their argument was in line with Bayliss-Smith’s findings in the small atolls he studied. But Bayliss-Smith described a method of estimating populations on the islands using just starch calories and a correction factor for uncalculated starch inputs, then modified by a correction factor to account for discrepancies between estimates of theoretical human carrying capacity and observed population numbers. Hunter-Anderson and Zan only went so far as to describe their model conceptually. Our study addresses the question of limits on population size and ecological variables, notably the balance of resource availability between the land and the water. This study is unique in its effort to quantify and interpret the combined population consequences of two features of islands: the extent and typology of land suitable for traditional agriculture, and the extent of near-shore shallow water protected by reef, which provides access to marine food resources. As there exists no way to test the importance of these factors in a controlled manner, our study makes use of Polynesia as a “natural experiment” (Kirch 1980), meaning that we observe a sample of islands and attempt to understand how much our two ecological factors influence estimates of maximum population size or density of occupation. By choosing islands that are similar with respect to a number of the other confounding ecological characteristics, we minimize the risk of missing an important factor. The specific question we ask here is: To what extent does the availability of terrestrial resources vs. near-shore marine resources drive estimates of island population density on Polynesian islands? To answer the question, we selected four islands in Eastern Polynesia in relatively close proximity, thus having similar climate, and with a range of agricultural resources and varying degrees of access to near-shore marine resources. In addition, each of the islands has archaeologically excavated samples of residential or food-preparation sites, providing island-specific information on the relative distribution of terrestrial and marine animal food sources. We estimated agricultural production by assigning island area to either non-agricultural use, or to the inferred agricultural use based on slope and substrate, as well as on ethnohistoric accounts of land use. We then used multiple published sources to estimate the nutritional value and expected yields of each agricultural zone. We estimated near-shore marine resources by quantifying the surface area of shallow reef-enclosed water. Then, given the uncertainty of fish yields in early Polynesia, we used three scenarios of reef productivity, spanning the range observed in modern times using various fishing tools and methodologies. The reef fish yield scenarios were used to convert the relative faunal contribution data from the excavations into absolute biomass numbers. METHODS Land Classification We imported topographical data for the islands of Moʻorea, Maupiti (Society Islands, French Polynesia), and Mangareva and Taravai (Gambier Islands, French Polynesia) and soil maps of Moʻorea and Mangareva into a GIS format. We extended the Moʻorea soil classes to Maupiti and the Mangareva soil classes to Taravai following Kirch et al. (2022), who verified the existing soil maps and the extrapolations with a lab analysis of soil samples and confirmed the expected nutrient availability as indicated by base saturation. We then characterized the land area by suitability for agriculture, and if suitable, the most likely use according to a rubric using slope, aspect, and substrate as variables. These productivity zones were then corrected, where necessary, by Kahn and Kirch, who have familiarity with landscape use and agricultural site distribution specific to particular ecotones on each island. Small associated coral islands ( motu ) were excluded from calculations except where evidence existed for raised-bed agriculture. (See the note in “Sources of potential error” in the Results and Discussion section, below.) On Mangareva lithosols and lateritic soils (26.2% and 33.1% of area, respectively, per Tercinier (1974)) were deemed unproductive. On Moʻorea, ferralitic soils (5.6% of area, per Buffard-Morel and Thuilier (1994)) were also eliminated from our productivity estimates. Additionally, on Moʻorea and Maupiti, perennial streams, which are not found on the other two islands, were buffered by 15 meters and removed from the productivity estimates (with one exception for wet taro, described below). On all four islands, the remaining land with a slope in excess of 50% was deemed only suitable for upland semi-cultivation of low-yield foods, while zones with slope >30% and ≤50% were classed for shifting cultivation. While some of the upland terrain was certainly too steep even for famine foods, we argue that the broad classification is a useful approximation, even if it contributes to what is probably an overestimate of this relatively minor contribution to terrestrial food resources (see below). Land with slope >5% and ≤30% was designated for arboriculture and shifting cultivation, except on Mangareva, which assigned eutrophic soils with slope 0% to ≤30% to this category. Land with slope ≤5% and calcareous soils on Mangareva were assigned to breadfruit and coconut cultivation. Exceptions occur on Mangareva and Moʻorea, where hydromorphic soils (0.9% of Mangareva’s area) and ≤5% slope land with evidence of raised beds (0.5% of Moʻorea’s area) were assigned to wet taro production. On Maupiti, a small portion of the ≤5% slope area abutting Haranai Stream (0.4% of main island) was classed as suitable for wet taro and the island’s motu included 0.2 km 2 for raised beds of wet taro. See Tables 1 and 2 for a summary. Estimating Terrestrial Agricultural Productivity Based on our analysis of ethnohistoric and ethnographic sources, we defined a total of six different land use classes. Each was characterized by the dominant precontact agricultural products which could be cultivated there. In the absence of complete information on soil characteristics and the climatic data required to parameterize crop models, and given the paucity of such models for many of the traditional Polynesian food crops, we relied on best estimates of typical crop yields from the literature in similar contexts where traditional subsistence agriculture was observed. The land use classes are: Wet taro: Hamilton and Kahn (2007) drew on Kirch (1994) and Spriggs (1981, 1984) to estimate that wet taro cultivation on Moʻorea yielded 25 U.S. tons of wet corms per hectare per year. To estimate caloric content of taro, we averaged six sources to estimate a yield of 56,950 kcal/ha/day, of which 89.8% come from carbohydrates and 4.8% from protein (Supplemental materials, Table S1). These energy estimates account for a 20% fallow span in wetland taro cultivation. Arboriculture: We drew from Hamilton and Kahn (2007, Table 8.5, p. 149), who for Moʻorea adapted Kirch’s (1994) description of Futuna (a Polynesian Outlier) arboriculture. The system relies on breadfruit, coconut, banana, Alocasia aroids, Colocasia taro, Dioscorea yams and sweet potato. We added coconut to the food crops, as it was not included in the previous starch-based analyses. For Moʻorea and Maupiti we assumed a breadfruit to coconut tree ratio of 40:60 and for Mangareva and Taravai we used a ratio of 75:25. Breadfruit occurs at an average of 45 trees/ha on the former two islands and 50 trees/ha on the latter, and the difference in the breadfruit:coconut ratio is achieved mostly through a reduction in the prevalence of coconut. Planting density of yams and sweet potato was assumed to be equal. Using the nutritional data in Supplemental materials Table S1 we estimate this zone yields 37,712 kcal/ha/day on Moʻorea and Maupiti, and 37,421 kcal/ha/day on Mangareva and Taravai (Table 3). Carbohydrates contributed 32,974 and 34,692 kcal/ha/day, respectively, and protein contributed 1,737 and 1,856 kcal/ha/day, respectively. Shifting cultivation: Kirch (1994, Table 12, p. 183) describes a shifting agricultural system on Futuna, which we modified to include sweet potato in place of 50% of the yam yield (yams not being as important in Eastern Polynesia). The dryland aroids were assumed to be exclusively dryland taro. We assume 3 years of cultivation and 10 years of fallow, following Hamilton and Kahn (2007, p. 146). This zone yields an average of 12,031 kcal/ha/day after accounting for fallow, of which 10,809 derive from carbohydrates and 584 derive from protein (Table 4). Arboriculture and shifting cultivation: We assume the area in this zone is divided evenly between the two agricultural types. After accounting for fallow, as above, this zone contributes on average 24,872 kcal/ha/day on Moʻorea and Maupiti, of which 21,891 derive from carbohydrates and 1,160 from protein. On Managareva and Taravai the yield is 24,726 kcal/da/day, with 22,751 and 1,220 coming from carbohydrates and protein, respectively. Breadfruit and coconut: While breadfruit was an important source of starch in the Polynesian diet, coconut was used primarily for oil and milk in cooking, providing little in the way of starch. Duke (2001) estimates that coconut trees provide 200 kg of copra (dried meat from which oil is extracted) annually per ha in Polynesia. At 6% water content, this would provide 3,440 total kcal/ha/day in monoculture, of which 82.6% come from fat (Haytowitz et al. 2005). If only the fat is consumed it yields 2,842 kcal/ha/day. We assume this zone is dedicated to 40% breadfruit and 60% coconut by number of trees, following Kirch (1994, Table 11, p. 182), except on Mangareva and Taravai, where breadfruit makes up 75% of the trees. On Moʻorea and Maupiti breadfruit interspersed with coconut provides an estimated 19,471 kcal/ha/day, with 15,470 and 1,067 coming from carbohydrates and protein, respectively. On Mangareva and Taravai the yield is 19,180 kcal/ha/day, with 17,189 and 1,186 from carbohydrates and protein, respectively. Upland semi-cultivation: This zone is made up of wild and semi-cultivated foods in the regions less accessible and otherwise unsuitable for intensive agriculture. Hamilton and Kahn (2007, p. 149) follow Massal and Barrau (1956, p. 17) to estimate 1.25 U.S. tons of starchy foods per ha per year. After accounting for a fallow of three years for each year of harvesting, the average yield is 979 kcal/ha/day, with 830 and 50 coming from carbohydrates and protein, respectively. To estimate the contribution of agricultural production to the diet of the islanders we scale the agricultural zone productivity (Table 5) by the zone-specific area on each island (Table 2). The potential agricultural energetic output by island is found in Table 6. Marine and Reef Classification Our study is the first to include productivity estimates of the near-shore marine ecology while modeling island productivity in relation to precontact human populations. To estimate near-shore contributions of fish and shellfish to the Polynesian diet on these islands, we examined maps and satellite images to divide the reef-associated area into “productive” and “non-productive” zones. Submerged area was designated as productive if it was shallower than 30m of depth, including the outer edge of the reef. Areas with a substrate identified as mud were excluded. The 30m limit is common in studies of reef productivity (see Craig et al. 2008, p. 238) and corresponds roughly to the depth at which light penetration declines sharply and primary marine productivity approaches zero. Where depth charts were not available we used visible-light satellite images imported into ArcMap to estimate the 30m contour. We found that where we had both visible-light imagery and bathymetric data (Institut Géographique National 2015), for Mangareva and Taravai only, the 30m contour corresponded well to the line at which the visible lagoon substrate graded into the color of deep water. For consistency we used the same visual spectrum method of lagoon and reef classification for all four islands. We manually corrected any errors and removed modern structures which extended into the marine zones, such as airport runways or docks. The productive reef zone area for each island is shown in Table 7. As Mangareva and Taravai share a fringing reef and lagoon, we allocated this productive marine area proportional to their land area for our calculations. Estimating Marine and Reef Productivity Estimates of nearshore yields of fish and shellfish derive mainly from two sources of information. The first includes studies of observed yields in populations practicing subsistence fishing and foraging using a range of methods and equipment, from those similar to what precontact Polynesians used to those that more resemble modern small-scale commercial fishing. The second set of sources seeks to estimate sustainable or maximum harvests using whatever methods might be available, including modern highly extractive commercial means. Both sources are informative, the first providing evidence of what people actually do, and the second summarizing data from multiple locations and fishing methods to estimate potential yield. We divide marine inputs into five categories: nearshore finfish, shellfish (mollusks and crustaceans), pelagic (open-water) fish, turtles, and marine mammals. Dalzell and Adams (1997) drew on data from 43 locations in the South Pacific to determine that contemporary observed yields range from 0.3 to 64 metric tons live weight of reef finfish (excludes shellfish) per km 2 per year, with a mean of 7.7 mt. They estimate the maximum sustainable yield (MSY) varies by location between 6-20 mt/km 2 /year, with a typical value of about 16. Newton et al. (2007) expanded on Dalzell and Adams (1997) and included mollusks and crustaceans in a study of 79 data sets from coral islands around the world compiled by the U.N. FAO. The yields ranged from 0.2 to 40 mt/km 2 /year, with a median of approximately 3. For their analyses the authors assumed a “more realistic” MSY of 5 mt/km 2 /year, but bracketed that value with scenarios of MSY at 1 and 10 mt/km 2 /year. Leenhardt et al. (2012) reviewed six studies of Moʻorea finfish catches and argued that observational studies of fishers and of roadside sales significantly underestimate the true rate of removal because these fish rarely get tallied. They estimated the yield at the time to be about 25 mt/km 2 /year, exceeding the MSY of 23 mt/km 2 /year for Moʻorea calculated by Galzin (1987). Dalzell and Adams (1997) predict higher maximum sustainable yields than Newton et al. (2007), despite the fact that they exclude mollusks and crustaceans from their analysis. However, their data come exclusively from the South Pacific. We use Dalzell and Adams’ (1997) estimate of 16 mt/km 2 /year for reef-associated fish in the South Pacific to represent the region’s maximum sustainable yield, an ecological measure of fish carrying capacity. We also examine the values of 5 and 1 mt/km 2 /year as moderate and low estimates of what subsistence fishing economies in remote parts of Polynesia may yield. Using modern yield data to estimate the potential contribution of mollusks and crustaceans is more problematic than is the case with finfish. The mix of mollusks and crustacean species targeted for precontact subsistence might be different than those that appear in yield records recorded in the FAO database. In a study of the outer islands of American Samoa, Craig et al. (2008) used conversions for estimating edible mass from live weight (including shells), determining that invertebrates made up approximately 17% of the annual harvest of reef-associated live weight (after excluding shell weights). The species captured included octopus (49.7% of invertebrate mass), polychaete worms (26.6%), spiny lobster (13.0%, including shell), turban snail (6.5%, without shell), giant clam (3.6%, without shell) and sea urchin (0.7%, without shell). In a comparison of fishing practices and yields between 1982 and 2002 in a remote subsistence fishery in Fiji, Kuster et al. (2005) found that in both years finfish made up 76% of the live weight caught and invertebrates were 24%. While we have estimates of marine contributions to diet from archaeological excavations, the data from our four islands excludes shellfish and mollusks (other than crustaceans from Moʻorea and Maupiti) due to excavation methodologies. Therefore, we average these two results and assume that shellfish and mollusks make up 20.5% of the combined live weight of the nearshore catch, equivalent to 25.8% of the finfish biomass (20.5%/79.5%=25.8%). Finally, open-water, or pelagic, fish and sharks were a culturally important food in many locations across Polynesia. Their contribution to the diet, however, varied across Polynesia (Allen 2017). Fraser (2001, p. 127) argued that this variation “reflect[s] cultural-historical, rather than natural, processes.” Fraser found in a study of data from 21 excavated Polynesian sites that the bones of offshore species were rare, with three notable exceptions, all from Eastern Polynesia (see also Allen 2017). Following this, we used biomass estimates from excavated faunal remains from three archaeological sites on Maupiti and two on Moʻorea. We limited our analysis to data from layers dated from AD 1400-1800, and eliminated data from one location on each island (MAU-11 on Maupiti and ScMo-350 Block 3 on Moʻorea) that appeared to be high-status or ceremonial. We focus on the later phase of precontact occupation, when some of the foods favored by the earliest arrivals were no longer abundant and human populations were larger. We used these biomass estimates to reconstruct the relative contributions of faunal inputs to the traditional diet, first averaging across time within each location and then across locations to get an island-wide estimate. For Mangareva and Taravai we used biomass estimates from excavations on Agakauitai (Kirch et al. 2015), a small island which shares the lagoon with Mangareva and Taravai (Kirch et al. 2010). As before, we limited our analysis to layers dated AD 1400-1800, which consisted of two on Agakauitai and one from Taravai. Biomass distributions were averaged across the two layers at Agakauitai and the result was averaged with the Taravai data. We used this distribution to characterize both Mangareva and Taravai’s relative faunal inputs. The data show evidence of modest pelagic input on Moʻorea and Maupiti, despite numerous bones of nearshore fish. Requiem sharks were recovered in almost all precontact cultural deposits, accounting for the largest fraction of estimated biomass from pelagic species. On Maupiti requiem sharks made up 17.0% of the fish biomass, and 5.6% on Moʻorea. Pelagic mackerel and tuna, which like the requiem shark can sometimes be caught close to shore, contributed 1.3% of estimated fish biomass at Maupiti and 0.7% at Moʻorea. Reef-associated fish made up the largest part of the marine-derived diet on both islands (Table 8). At Mangareva and smaller islands in the lagoon pelagic fish are sparse in the archaeological record (Weisler and Green 2013; Rurua 2015). Family Scombridae, which includes pelagic tuna and mackerel, represented < 1% of the number of identified specimens (NISP) at Mangareva, and was the 13 th -most common family observed (Weisler and Green 2013). Biomass estimates from Kirch and associates’ excavations on Agakauitai and Taravai do not include differentiation below the level of class. For these islands we adopt an estimated pelagic fish fraction of 5% of total fish biomass. Other Animal Inputs The distribution of estimated faunal biomass in the excavation data provides insight into the relative importance of the various food sources. Here we use the same subset of data from Kahn (Kahn, In press) for Maupiti and Moʻorea, and from the excavations on Agakauitai (Kirch et al. 2015) and Taravai (Kirch et al. 2010). In all cases the biomass estimates do not include shellfish or mollusks, which we instead estimate at 25.8% of the fish biomass, as described above. The two Maupiti commoner residential sites (MAU-1 and MAU-2) and intermediate-status (MAU-5) site differ somewhat from each other, and across time to an extent, but in aggregate biomass is dominated by finfish (46.5%) and mammals (45.3%), with turtles contributing 6.6% and birds 1.7%. Shellfish and mollusks were not included in the biomass analysis. The relative proportions of rat, dog and pig biomass are fairly stable, even across the status gradient, meaning that although the higher-status sites had greater biomass of identified mammals, the ratio of pig to dog biomass, for example, was more or less constant. Rat biomass was 5.8% of mammal biomass, dog was 8.1% and pig predominated with 86.1% of the estimated mammal biomass. Marine mammals contributed 0.1%. At the non-elite residential sites on Moʻorea there is a general trend over time toward less mammal and more fish biomass in pre-contact cultural deposits; this reflects tapu restrictions in the later period which excluded commoners largely from eating pig and dog (Kahn 2024; Oliver 1974, 276). However, averaging across the period 1400-1800 among commoner sites we find the biomass contribution of mammals was 37.8% and of finfish was 57.0%, with turtle contributing 4.5% and birds 0.8%. Of the estimated mammal biomass that could be identified from the Moʻorea sites, 5.3% came from rats, 28.0% from dogs, 54.8% from pigs and 11.9% from marine mammals. Averaged across the two layers on Agakauitai and one layer on Taravai the biomass estimates from the Mangareva island group are 79.5% finfish, 0.8% turtle, 8.4% bird and 11.3% mammal, identified remains of which were made up mostly of pig (9.0% of faunal biomass) and rat (2.3%), while dog represented 0.1%. Turtle and marine mammal remains are absent from the samples. Estimates of Faunal Dietary Inputs The estimated relative faunal contributions to late-precontact diet on the four islands are presented in Table 8, after including the shellfish contribution as a fixed 25.8% of the total fish biomass (see above). Being a function of reef-associated surface area, estimates of nearshore marine productivity are the only biomass inputs available that can be scaled across the four islands. We adopt values of 1, 5 and 16 mt/km 2 /year of nearshore finfish as our expected subsistence yields, representing the range of likely values. Converting Faunal Biomass to Food In estimating the fraction of live weight that made its way into the local diet, what Lyman (1979) called “consumable meat,” we assume that the bulk of muscle (flesh) is available, while viscera, hide and bones are not consumed. Edible fat is assumed to be limited to that found in the meat, with the addition of blubber. We estimate the edible fraction of reef fish to be 50% of live weight, an approximation of fish yields of similar species (Buchanan 2023). Yield fractions of moderate to small sized fish vary widely, from as low as 30% to as much as 90% of live weight, but in general, smaller fish and fish with large, bony heads yield less. Pelagic fish, characterized by tuna, are assumed to yield 70% of live weight as edible meat (Smith 2011), and marine invertebrates are assumed to be entirely edible, as live weight excludes shells and exoskeletons. The edible portion (meat) of pigs is estimated to be 70% of live weight (White 1953). There are few measurements on the meat yields of dogs and rats, but Smith (2011) estimates these at 60% and 70% of live weight, respectively. These dogs were usually specially bred for eating and fattened on human foods, including breadfruit and taro (Titcomb and Pukui 1969; Bay-Petersen 1983). Lockyer (1991) includes an exhaustive study of 206 dissected whales from 9 species and all parts of the globe. The study found that 40% of whale live weight came from muscle and 26% came from blubber and we adopt these fractions for all marine mammals. The live weight of turtles includes the mass of the carapace and the edible fraction is estimated to be 50%, following a calculation in Frazier (1980, Table 1). Birds, like fish, vary widely in their particulars, but we follow Smith (2011) in the estimate of 70% edible fraction for small to medium birds, including terrestrial and wetland/marine species, in precontact New Zealand. The two nutritional components of greatest importance are the energy these foods provided and their protein content. Meat carbohydrates exist in small quantities, but are often recorded as zero in nutritional tables, and we will assume their contribution is negligible. The data we used to make nutritional component calculations, along with their sources, are included in the supplemental materials and are summarized in Table 9. The values reflect the nutrient content of multiple samples of meat (generally muscle tissue), which include incorporated fat. The blubber of cetaceans is an exception, and the values for birds includes the skin. To calculate the expected faunal energy available to the population we scale the faunal biomass distributions (Table 8) to the reef finfish amount, then multiply by the island-specific productive reef zone area (Table 7) to give metric tons of faunal biomass per metric ton of reef fish. We then multiply through by the edible fraction and caloric value of each food type (Table 9) to give food energy per metric ton of reef fish. Finally, we multiply these values by the scenario-appropriate value for reef fish catch per km 2 /year and then convert this to a daily value (kcal/day). The faunal energy yields under the three reef productivity scenarios are shown in Tables 10-12. RESULTS AND DISCUSSION Combining Marine and Terrestrial Productivity To better understand the estimated marine and terrestrial food yields we converted them into the maximum human population that might have been supported on those amounts. We adopted the simplifying assumptions that each person consumed 2700 kcal/day, representing the caloric requirement of a well-fed, active individual, and that 30% of the yield of plants and animals was unrealized, being either lost or diverted from the human diet. The 30% deduction represents a compromise between Bayliss-Smith (1974), who estimated populations at 70-80% of carrying capacity, and Spriggs and Kirch (1992), who estimated realized food production at approximately 50% of potential production, both in Pacific subsistence contexts. The results by reef fishing scenario, including a breakdown of food contributions by plant vs. animal sources, are summarized in Table 13. The same results are broken down by marine vs. terrestrial food sources in Table 14. The scenarios in the tables range from a low reef productivity level of 1 mt/km 2 /year of reef fish caught, through 5 and 16 mt/km 2 /year. These scenarios are labeled A-C, respectively. Recall that the archaeologically excavated faunal remains provide relative abundance data, which we convert to absolute values by pegging them to reef fish catch rate by scenario. As the estimate of fish catch increases, so does the mass of all other faunal food sources. Putting the pieces together allows us to compare human population estimates across the varying islands. We have broken down the initial estimates of maximum population by the contributions of animal vs. plant resources (Table 13), and also by marine vs. terrestrial resources (Table 14). We find Moʻorea’s population maximum is many times larger than that of any of the other islands under all scenarios. But Moʻorea’s population ceiling does not increase much with estimates of fish abundance. The estimate of the island’s maximum population increases only from 43,575 to 44,368 even as reef fish inputs increase 16-fold. Population estimates among the other islands increase by almost 50% (Maupiti; from 1,489 to 2,210), 74% (Mangareva; from 2,954 to 5,139), or 94% (Taravai; from 701 to 1,364) across the scenarios. In the breakdown of faunal vs. agricultural food resources (Table 13) Moʻorea’s mix remains heavily agricultural in all cases, while the other three islands stand to gain much more — in terms of maximum population size — from better fishing success. Zooming out to examine the importance of ecological variables and food sources on population density, here we provide two measures. The first is whole-island population density (Table 15), which is population estimate by scenario divided by total land area, excluding motus. The second is density as a function of agricultural land area (Table 16), in which we divide by all land suitable for agriculture, excluding semicultivation zones, following Table 2. If questioning whether agricultural land is more or less important to food availability than marine resources, the answer is clear in Table 17. This table captures the tremendous value of terrestrial resources on all four islands under the most likely fish yield scenario, 5 mt/km 2 /y. Agricultural land is 150-200 times as productive as the near-shore marine area in terms of caloric availability. In all cases, fish and reef resources are important, but their potential contribution to the caloric budget is dwarfed by the terrestrial component. To this point we have calculated potential population purely on the basis of total calories, but the tremendous disparity in terrestrial and marine resources on Mo‘orea (Table 14) raises the question of dietary balance. Human diets have a great deal of plasticity, but are ultimately constrained in their composition at the extremes. Fat and carbohydrates can be exchanged for one another by the body, within limits, but protein is different. Humans lack the ability to produce a number of essential amino acids, and dietary sources of protein are the only way to acquire them. Nutritionists suggest that the acceptable macronutrient distribution range (AMDR) is 45-65% of total calories from carbohydrates, 20-35% from fat (for ages 4+), and 10-35% from protein (for ages 18+) (Institute of Medicine 2005), and most human populations settle on diets that fall within these ranges. A number of Pacific subsistence/diet studies have observed starch percentages in excess of 65%, but even those include a protein fraction of 9.4% or greater (e.g., Bayliss-Smith 1974; Ross 1976:579; Lindeberg and Vessby 1995:48). Recommended dietary allowances (RDAs) are a second way to assess macronutrient needs. These age- and sex-specific macronutrient mass values are assumed to capture the minimum requirements of 97 to 98% of the population (Institute of Medicine 2005). We used the UN Model Life Tables – West to generate the age and sex structure of a hypothetical stable early Polynesian population and determined that the average RDA for protein in this population was almost exactly 45 g per person per day. This corresponds to 8% of a recommended 2,250 kcal daily diet. We have assumed that the subsistence Polynesian populations required 2,700 kcal per day (see above), and if protein makes up at least 8% of the energy in the diet, that comes to a minimum of 54g protein per day per person. We will consider both approaches to minimum protein requirements: 10% of the total kcal, and 54 g/person/day, corresponding to 8% of the total kcal. First, we calculate the number of people each island’s protein supply could support at the 10% minimum, independent of other macronutrients or calories, given the fish yields in Scenario B (Table 18). For example, Mangareva is expected to be able to supply enough protein to feed 6,632 people, but the maximum population based on total calories is only 3,537, so we assume protein is not limiting on Mangareva and the maximum population remains at the lower number. This is the case on all the islands, with the notable exception of Mo‘orea, whose marine resources are far outstripped by its terrestrial ones, leading to the high likelihood of protein limitation. We estimate that the maximum population, fed with food procured on or nearby the island, would be 22,906, or 52.3% of the calorie-only estimate. In the second approach we relax the protein requirement to a minimum of 8% of the food energy (equivalent to 54 g/d/person), resulting in a maximum population on Mo‘orea of 28,632 under the assumptions of fishing Scenario B. This is 65.4% of the naïve population estimate. While there is an argument to be made for either the 8% or 10% minimum protein requirement, seeing that the dietary studies support a protein minimum closer to 10% of total kcals, we regard 22,906 as more likely. Sources of Potential Error In comparison to estimated population densities on other islands in Polynesia before European contact, our values in Tables 15 and 16 are quite high. Puleston and Ladefoged (2022) concluded that estimates of precontact Polynesian population densities with respect to agricultural area ranged from about 50 to 250 p/km 2 . The estimates in Table 16 are several times greater than even that upper bound. There are several sources of variation, or error, that should be considered relating to our case studies. These include: (1) overestimates of agricultural productivity in one or more important agricultural types, (2) the mis-designation of less-productive land into a higher-productivity class, (3) underestimation of food diversion and losses due to ceremonial use, waste and spoilage, (4) the assumption that labor supply never limits productivity, and (5) the assumption that the populations were always working to maximize total production. First, it is possible our methodology overestimates productive land area, but we have been careful to avoid doing so. We defined our land classifications precisely, but the suitability of land at the local scale is heterogeneous so perhaps only some fraction of the actual area was used in the manner we designated. However, our estimates of productivity by agricultural activity come from similar non-industrial systems in similar contexts, so the effect of heterogeneity is already incorporated into those calculations. One exception might be a modest overestimate of semicultivation zone area (see methods), but the productivity in these areas is so low that we do not expect it to make a significant difference to our results. Finally, density of population as a function of agricultural land should be less vulnerable to this form of error because overestimation of agricultural land should increase the denominator (agricultural land area) faster than the numerator (population size) in the calculation, particularly as our estimates of population include non-agricultural (marine) resources. It should be noted that our decision to include only known motu agricultural areas introduces a modest potential bias toward an underestimation of agricultural area by ignoring motu locations which might have hosted agriculture in the past, but are not currently known. We argue, however, that the likelihood of unrecognized raised-bed areas is small, and that the contribution of any such unincluded zones is unlikely to alter our results. In reference to the second possible source of error, it is possible that we overestimated potential production by classifying agricultural area as more productive than it might actually have been. This effect would be most pronounced in the zones we designated as appropriate for wet taro and mixed arboriculture and shifting. In terms of the third potential source of error, food crops are subject to stunting and also loss before and after harvest from diverse sources including damage from domestic and wild animals and insect pests, rainfall variability and drought, and rot after harvest. In addition, across Polynesia food offerings for ceremonial purposes diverted calories to the gods that might have otherwise been available to people. Furthermore, the cost of feeding agricultural calories to domestic animals convert one type of food into a more socially valuable one, but this is inefficient in terms of food energy. Obligatory tribute in the form of food resources was also common in Polynesia before European contact; the end result is a diversion of food from the people producing it into the hands and mouths of others. All of these play a role in decreasing food available to the agriculturalist population from a theoretical maximum. In most cases the yield figures we use account for losses in the fields, but the losses and diversions after that point need to be estimated separately. We have assumed 30% of potential production never makes it to the mouths of the human population, but some have argued that this is an underestimate. This correction factor also includes the effect of the fifth source of error, below. It should also be noted that a variable food supply, which includes the production of rain-fed systems, can have complex effects on population dynamics (Lee et al. 2009). Rare droughts and crop failures can have immediate catastrophic effects that may take generations to recover from, leading to a significant depression of mean population size vs. the theoretical maximum under the assumption of consistent yields. Societal features, including the requirement of tribute or labor diversion can exacerbate this effect (Winterhalder and Puleston 2018; Puleston and Winterhalder 2019). The fourth source of error centers around a potential shortage of laborers limiting the productivity of an agricultural system. Work by demographer Shripad Tuljapurkar and others, dubbed food-limited demography, (Lee and Tuljapurkar 2008; Puleston and Tuljapurkar 2008; Puleston and Winterhalder 2019) considered the role of labor in a rain-fed preindustrial agricultural system, focusing on the dynamic characteristics of the human populations that rely on that food. Such modeling work has found that in a natural-fertility population working to maximize productivity, labor is limiting while the population is growing and is more likely to be found in excess in a population living near its equilibrium or carrying capacity (Puleston, et al. 2014; Winterhalder and Puleston 2018). It is of course possible that ongoing diversions of labor into societal features like monumental architecture, the maintenance of infrastructure, or the presence of a highly specialized labor force, or a standing army, would make field labor scarcer, but in an established population the effect is small. The fifth assumption may be the most important one. Anthropologist Tim Bayliss-Smith (1980) has considered the tradeoff between population size and the intensity of labor effort. He and others found that in agricultural systems like the ones that dominated these islands there is a fairly common expectation as to how much labor a person contributes to agricultural work in a week. People would work less in many cases if they could, and in some cases work considerably more than the expectation of 10-20 hours a week (Bayliss-Smith 1978) in subsistence agriculture. However, in many places we do not really know what the local expectations of hours of field labor were. We also expect that the population, as individuals and as a group, might shift their energy investments to the activities that provided the greatest return, which might be measured in calories or in the more subjective currency of food desirability. The conclusion across a variety of scenarios is that while the number of laborers might be not be a limiting factor, the hours they work might be. The result is a level of total production, and a population size, diminished from its theoretical maximum by something on the order of 30%, as mentioned above. The point of Bayliss-Smith’s and food-limited demography’s findings is that there are tradeoffs between quality and quantity of life, and if the agricultural population has the power to make its own choices it may choose a smaller population with a better quality of life, in terms of work-hours, lifespan, infant mortality and food availability. CONCLUSIONS Our estimates of the maximum sustainable populations for the four islands are summarized in Table 19. Moʻorea’s fertile valleys and wetlands would potentially generate sufficient food energy to support a maximum population > 40,000 people, regardless of fish yield scenario. The other three islands are much more dependent on estimates of fish yields, and the population numbers are much smaller, ranging from less than 1,000 to more than 3,500 in the most likely fish yield scenario. Clearly, ecology does matter: Moʻorea gets only 0.4% of its caloric potential from marine sources, while Maupiti gets 7.3%, Mangareva gets 15.0% and Taravai gets 18.4%. These relative inputs reflect the vast difference in productivity (in terms of calories per km 2 ) in agricultural vs near-shore marine zones. Agricultural land is some 150 times as productive as the marine area at all four islands. We conclude that while terrestrial resources were much more plentiful, and thus played a more significant role, marine inputs could be important, particularly on smaller land-poor islands, and should not be ignored. Further, while terrestrial production had greater potential, these crops were both more variable and more vulnerable than the more passively maintained pool of marine resources, pointing to a promising area of inquiry for future study. Available studies suggest that traditional subsistence-level exploitation of near-shore marine systems led to sustainable use, despite the significant initial impact of human arrival and some evidence of long-term decline in the size of fish and invertebrates (e.g., Allen 2017). In contrast, local extirpations of high-value pigs and dogs on small or isolated islands, thought to be a result of intentional choices to reduce trophic competition with human populations, hints at the vulnerability of the terrestrial agricultural systems on smaller islands (Kirch 2000). Consider also the natural variability of rain-fed agriculture, which can be highly unpredictable (Lee et al. 2006). That many Pacific human populations were able to simultaneously rely on these two independent food systems surely contributed to their own resilience. It is important to note that in cases of extreme imbalance in the terrestrial vs. marine potential, such as is observed on the large, but geologically young island of Mo‘orea, the abundance of valued carbohydrate resources cannot be fully realized by the local population without sufficient protein availability. Our estimate of the island’s maximum population drops by almost half when we consider diet composition. Such constraints might well be relevant to other regions, like the Marquesas Islands and Rapa Nui in Eastern Polynesia, which have small fringing reefs due to young geological age, or to other ecological factors (e.g., the Humboldt Current in the Marquesas Islands). Even the starch-loving Polynesians required some quantity of fish, or meat, to make a proper meal (Kirch and O’Day 2003). Put another way, one cannot live by breadfruit alone. Declarations Acknowledgements Susan Andrews is thanked for calculating the initial biomass estimates from Kahn and Kirch’s faunal materials. We thank Neil Davies for helping to acquire GIS data from the Service de l’urbanisme of French Polynesia, whom we also thank. We thank Diane Ragone for generously sharing advice and information regarding breadfruit in Polynesia. Competing Interests The authors declare that there are no competing interests. Ethical Approval There are no applicable ethics declarations. Funding Funding for this work was provided by a National Science Foundation Grant (CNH # 1642894), awarded to Kahn and Kirch. References Ab-Lah, Roslizawati, Joshua Smith, Dale Savins, Ashley Dowell, Daniel Bucher, and Kirsten Benkendorff. 2017. “Investigation of Nutritional Properties of Three Species of Marine Turban Snails for Human Consumption.” Food Science & Nutrition 5 (1): 14–30. https://doi.org/10.1002/fsn3.360. Allen, Melinda S. 2017. “Spatial Variability and Human Eco-Dynamics in Central East Polynesian Fisheries.” https://academic.oup.com/edited-volume/34661/chapter/295345817. Ann, Yong-Geun. 1999. “Dog Meat Foods in Korea.” Korean Journal of Food and Nutrition 12 (4): 397–408. Arai, Yoro, and Shigeru Sakai. 1952. “Whale Meat in Nutrition.” The Scientific Reports of the Whales Research Institute 7:51–67. Barrau, Jacques. 1958. Subsistence Agriculture in Melanesia . Bulletin 219. Honolulu: Bernice P. Bishop Museum. https://cir.nii.ac.jp/crid/1570009750817146624. Bayliss-Smith, T. P. 1980. “Population Pressure, Resources and Welfare: Towards a More Realistic Measure of Carrying Capacity.” In Population-Environment Relations in Tropical Islands: The Case of Eastern Fiji , edited by H. C. Brookfield, 61–93. Paris: UNESCO. Bayliss-Smith, Tim. 1974. “Constraints on Population Growth: The Case of the Polynesian Outlier Atolls in the Precontact Period.” Human Ecology 2 (4): 259–95. https://doi.org/10.1007/BF01531318. Bayliss-Smith, Tim P. 1978. “Batiki in the 1970s: Satellite of Suva.” In The Small Islands and Reefs , edited by H. C. Brookfield, 67–128. Canberra: Australian National University, Development Studies Centre for UNESCO. Bay-Petersen, Jan. 1983. “Competition for Resources: The Role of Pig and Dog in the Polynesian Agricultural Economy.” Journal de La Société Des Océanistes 39 (77): 121–29. Beckerman, Stephen. 1977. “Protein and Population in Tropical Polynesia.” The Journal of the Polynesian Society 86 (1): 73–79. Bradbury, J. Howard, and Warren D. Holloway. 1988. Chemistry of Tropical Root Crops: Significance for Nutrition and Agriculture in the Pacific . ACIAR Monographs 6. Canberra: Australian Centre for International Agricultural Research. https://ageconsearch.umn.edu/record/118050/files/6.pdf. Buchanan, David. 2023. “Fish Fillet Butchering Yields.” Fish Fillet Butchering Yields. 2023. https://www.chefs-resources.com/seafood/seafood-yields/. Burtenshaw, Mike, and Graham Harris. 2007. “Experimental Archaeology Gardens Assessing the Productivity of Ancient Māori Cultivars of Sweet Potato, Ipomoea Batatas [L.] Lam. in New Zealand.” Economic Botany 61 (3): 235–45. https://doi.org/10.1663/0013-0001(2007)61[235:EAGATP]2.0.CO;2. Craig, P., A. Green, and F. Tuilagi. 2008. “Subsistence Harvest of Coral Reef Resources in the Outer Islands of American Samoa: Modern, Historic and Prehistoric Catches.” Fisheries Research 89 (3): 230–40. Dalzell, P., and T. J. H. Adams. 1997. “Sustainability and Management of Reef Fisheries in the Pacific Islands.” In Proc. 8th Int. Coral Reef Symp , 2:2027–32. https://www.spc.int/digitallibrary/doc/fame/reports/adams_96_panama.pdf. Dignan, Cecily, Barbara Burlingame, Shailesh Kumar, and William Aalbersberg. 2004. The Pacific Islands Food Composition Tables. Ed. 2. Food and Agriculture Organization of the United Nations (FAO). https://www.cabdirect.org/cabdirect/abstract/20043191618. Duke, James A. 2001. “ Cocos Nucifera L. ” In Handbook of Nuts , 1st ed., 100–109. Boca Raton, FL: CRC Press. FAO. 2016. “INFOODS Global Food Composition Database for Fish and Shellfish Version 1.0-uFiSh1. 0.” Rome, Italy. https://www.fao.org/3/a-i6655e.pdf. Fraser, K.L. 2001. “Variation in Tuna Fish Catches in Pacific Prehistory.” International Journal of Osteoarchaeology 11 (1–2): 127–35. https://doi.org/10.1002/oa.551. Frazier, J. 1980. “Exploitation of Marine Turtles in the Indian Ocean.” Human Ecology 8 (4): 329–70. https://doi.org/10.1007/BF01560999. Galzin, Rene. 1987. “Potential Fisheries Yield of a Moorea Fringing Reef (French Polynesia) by the Analysis of Three Dominant Fishes.” Atoll Research Bulletin . https://repository.si.edu/bitstream/handle/10088/5853/00305.pdf. Hamilton, Brenda K., and Jennifer G. Kahn. 2007. “Pre-Contact Population in the ’Opunohu Valley, Mo’orea: An Integrated Archaeological and Ethnohistorical Approach.” In The Growth and Collapse of Pacific Island Societies: Archaeological and Demographic Perspectives , edited by Patrick V. Kirch and J.L. Rallu. University of Hawaii Press. Haytowitz, David, Jaspreet Ahuja, Xianli Wu, Meena Somanchi, Melissa Nickle, Quyen Nguyen, Janet Roseland, et al. 2005. “USDA National Nutrient Database for Standard Reference, Legacy Release.” Nutrient Data Laboratory, Beltsville Human Nutrition Research Center, ARS, USDA. https://doi.org/10.15482/USDA.ADC/1529216. Hipsley, Eben H., and Nancy E. Kirk. 1965. Studies of Dietary Intake and the Expenditure of Energy by New Guineans. Technical Paper 147. Noumea, New Caledonia: South Pacific Commission. https://www.cabidigitallibrary.org/doi/full/10.5555/19662900948. Huang, A. S., C. A. Titchenal, and B. A. Meilleur. 2000. “Nutrient Composition of Taro Corms and Breadfruit.” Journal of Food Composition and Analysis 13 (5): 859–64. Hunter-Anderson, Rosalind L., and Yigal Zan. 1985. “‘Where’s the Fish?’ Or, What Limited the Size of Aboriginal Populations in Small Tropical High Islands?” The Journal of the Polynesian Society 94 (1): 61–65. Institut Géographique National. 2015. “French Polynesia Topographic Maps.” Paris: Polynésie Française, Service de l’urbanisme, Section topographie. Institute of Medicine. 2005. Dietary Reference Intakes for Energy, Carbohydrate, Fiber, Fat, Fatty Acids, Cholesterol, Protein, and Amino Acids . Washington, D.C.: National Academies Press. https://doi.org/10.17226/10490. Jones, Andrew Maxwell Phineas, Diane Ragone, NuG Tavana, D. W. Bernotas, and Susan J. Murch. 2011. “Beyond the Bounty: Breadfruit ( Artocarpus Altilis ) for Food Security and Novel Foods in the 21st Century.” https://scholarspace.manoa.hawaii.edu/handle/10125/21022. Jones, Kevin L., and R. Garry Law. 1987. “Prehistoric Population Estimates for the Tolaga Bay Vicinity, East Coast, North Island, New Zealand.” New Zealand Journal of Archaeology 9:81–114. Kahn, Jennifer G. 2024. “Pig and Dog Use in the Pre‐contact Society Island Chiefdoms: Integrated Ethnohistoric, Archaeological and Use‐web Analyses.” Archaeology in Oceania , April, arco.5314. https://doi.org/10.1002/arco.5314. ———. In press. Fenua and Fare, Marae and Mana: The Archaeology of Ancient Tahiti and the Society Islands . Honolulu, HI: University of Hawaiʻi Press. Kirch, Patrick, and Sharyn Jones O’Day. 2003. “New Archaeological Insights into Food and Status: A Case Study from Pre-Contact Hawaii.” World Archaeology 34 (3): 484–97. https://doi.org/10.1080/0043824021000026468. Kirch, Patrick V. 1980. “Polynesian Prehistory: Cultural Adaptation in Island Ecosystems: Oceanic Islands Serve as Archaeological Laboratories for Studying the Complex Dialectic between Human Populations and Their Environments.” American Scientist 68 (1): 39–48. Kirch, Patrick V., Eric Conte, Warren Sharp, and Cordelia Nickelsen. 2010. “The Onemea Site (Taravai Island, Mangareva) and the Human Colonization of Southeastern Polynesia.” Archaeology in Oceania 45 (2): 66–79. https://doi.org/10.1002/j.1834-4453.2010.tb00081.x. Kirch, Patrick V., Jennifer G. Kahn, and Oliver A. Chadwick. 2022. “Soils, Agriculture, and Land Use in Island Socio‐ecosystems: Three Case Studies from Southeastern Polynesia.” Geoarchaeology 38 (1): 20–34. https://doi.org/10.1002/gea.21934. Kirch, Patrick V., Guillaume Molle, Cordelia Nickelsen, Peter Mills, Emilie Dotte‐Sarout, Jillian Swift, Allison Wolfe, and Mark Horrocks. 2015. “Human ecodynamics in the Mangareva Islands: a stratified sequence from Nenega-Iti Rock Shelter (site AGA-3, Agakauitai Island).” Archaeology in Oceania 50 (1): 23–42. https://doi.org/10.1002/arco.5050. Kirch, Patrick Vinton. 1994. The Wet and the Dry: Irrigation and Agricultural Intensification in Polynesia . University of Chicago Press. ———. 2000. “Pigs, Humans, and Trophic Competition on Small Oceanic Islands.” In Australian Archaeologist: Collected Papers in Honour of Jim Allen , edited by A. Anderson and T. Murray, 427–39. Canberra: Australian National University, Centre for Archaeological Research and Department of Archaeology and Natural History. Kuster, C., V. C. Vuki, and L. P. Zann. 2005. “Long-Term Trends in Subsistence Fishing Patterns and Coral Reef Fisheries Yield from a Remote Fijian Island.” Fisheries Research 76 (2): 221–28. Lee, Charlotte T., Cedric O. Puleston, and Shripad Tuljapurkar. 2009. “Population and Prehistory III: Food-Dependent Demography in Variable Environments.” Theoretical Population Biology 76 (3): 179–88. Lee, Charlotte T., and Shripad Tuljapurkar. 2008. “Population and Prehistory I: Food-Dependent Population Growth in Constant Environments.” Theoretical Population Biology 73 (4): 473–82. Lee, Charlotte T., Shripad Tuljapurkar, and Peter M. Vitousek. 2006. “Risky Business: Temporal and Spatial Variation in Preindustrial Dryland Agriculture.” Human Ecology 34 (6): 739–63. https://doi.org/10.1007/s10745-006-9037-x. Leenhardt, Pierre, R. Madi Moussa, and René Galzin. 2012. “Reef and Lagoon Fisheries Yields in Moorea: A Summary of Data Collected.” http://www.spc.int/DigitalLibrary/Doc/FAME/InfoBull/FishNews/137/FishNews137_27_Leenhardt.pdf. Lindeberg, Staffan, and Bengt Vessby. 1995. “Fatty Acid Composition of Cholesterol Esters and Serum Tocopherols in Melanesians Apparently Free from Cardiovascular Disease-the Kitava Study.” Nutr Metab Cardiovasc Dis , 45. Lockyer, Christina. 1991. “Body Composition of the Sperm Whale, Physeter Catodon , with Special Reference to the Possible Functions of Fat Depots.” Rit Fiskideilda 12 (January). Lyman, R. Lee. 1979. “Available Meat from Faunal Remains: A Consideration of Techniques.” American Antiquity 44 (3): 536–46. Massal, Emile, and Jacques Barrau. 1955. “Sweet Potato.” In Pacific Subsistence Crops , 10–14. Bulletin 5. South Pacific Commission. ———. 1956. Food Plants of the South Sea Islands . Vol. 94. Noumea, New Caledonia: South Pacific Commission. https://library.wur.nl/WebQuery/titel/446787. McKee, Hugh Shaw. 1957. Some Food Problems in the Pacific Islands . Technical Paper 106. Noumea, New Caledonia: South Pacific Commission. https://library.wur.nl/WebQuery/titel/401355. Miller, Carey D., and Florence Pen. 1959. “Composition and Nutritive Value of Palolo ( Palola Siciliensis Grube ).” Pacific Science 13:191–94. Murai, Mary, Carey D. Miller, and Florence Pen. 1958. Some Tropical South Pacific Island Foods; Description, History, Use, Composition, and Nutritive Value . Honolulu: Univ. of Hawaii Press. Newton, Katie, Isabelle M. Côté, Graham M. Pilling, Simon Jennings, and Nicholas K. Dulvy. 2007. “Current and Future Sustainability of Island Coral Reef Fisheries.” Current Biology 17 (7): 655–58. O’Hara, Todd M., Paul Hoekstra, Cyd Hanns, Derek Muir, Dana Wetzel, and John Reynolds. 2004. “A Preliminary Assessment of the Nutritive Value of Select Tissues from the Bowhead Whale Based on Suggested Nutrient Daily Intakes.” International Whaling Commission Scientific Committee Report SC/56 E 2:1–19. Oliver, Douglas L. 1974. Ancient Tahitian Society . University of Hawaii Press. https://books.google.com/books?hl=en&lr=&id=kVrGDwAAQBAJ&oi=fnd&pg=PR2&dq=oliver+ancient+tahitian+society&ots=yL2sgRocjC&sig=lFqDzxvwN4-DkEZdvunxsnDbGTo. Peters, F. E. 1957. Chemical Composition of South Pacific Foods: An Annotated Bibliography . Technical Paper 100. Noumea, New Caledonia: South Pacific Commission. Puleston, Cedric O., and Thegn N. Ladefoged. 2022. “Ecology Limits Population, But Interaction with Culture Defines It: Carrying Capacity on Rapa Nui.” In The Prehistory of Rapa Nui (Easter Island) , edited by Valentí Rull and Christopher Stevenson, 22:521–51. Developments in Paleoenvironmental Research. Cham: Springer International Publishing. https://doi.org/10.1007/978-3-030-91127-0_20. Puleston, Cedric O., and Shripad Tuljapurkar. 2008. “Population and Prehistory II: Space-Limited Human Populations in Constant Environments.” Theoretical Population Biology 74 (2): 147–60. Puleston, Cedric, Shripad Tuljapurkar, and Bruce Winterhalder. 2014. “The Invisible Cliff: Abrupt Imposition of Malthusian Equilibrium in a Natural-Fertility, Agrarian Society.” PLoS One 9 (1): e87541. Puleston, Cedric, and Bruce Winterhalder. 2019. “Demography, Environment, and Human Behavior.” In Handbook of Evolutionary Research in Archaeology , edited by Anna Marie Prentiss, 311–35. New York: Springer International Publishing. https://doi.org/10.1007/978-3-030-11117-5_16. Ragone, Diane, and Catherine G. Cavaletto. 2006. “Sensory Evaluation of Fruit Quality and Nutritional Composition of 20 Breadfruit ( Artocarpus, Moraceae ) Cultivars.” Economic Botany 60 (4): 335–46. https://doi.org/10.1663/0013-0001(2006)60[335:SEOFQA]2.0.CO;2. Ross, Harold M. 1976. “Bush Fallow Farming, Diet and Nutrition: A Melanesian Example of Successful Adaptation.” In The Measures of Man: Methodologies in Biological Anthropology , edited by E. Giles and J.S. Friedlander, 550–615. Cambridge, Massachusetts: Peabody Museum Press. Rurua, Vahine. 2015. “Biodiversité et Exploitation Des Ressources Marines En Polynésie Française Sur La Longue Durée : Étude Comparative Des Archipels Des Marquises et Des Gambier.” These en préparation, Polynésie française. https://www.theses.fr/s140157. Smith, Ian. 2011. “Meat Weight, Nutritional and Energy Yield Values for New Zealand Archaeofauna.” Department of Anthropology & Archaeology; University of Otago. https://ourarchive.otago.ac.nz/bitstream/handle/10523/5942/OALR_No8.pdf?sequence=1. Spriggs, Matthew. 1984. “Taro Irrigation Techniques in the Pacific.” In Edible Aroids , edited by S. Chandra, 123–35. Oxford: Clarendon Press. Spriggs, Matthew James Thomas. 1981. “Vegetable Kingdoms: Taro Irrigation and Pacific Prehistory.” Canberra: The Australian National University. https://www.researchgate.net/profile/Matthew-Spriggs-2/publication/35933028_Vegetable_kingdoms_taro_irrigation_and_Pacific_prehistory/links/55ff430108aeafc8ac8b9a21/Vegetable-kingdoms-taro-irrigation-and-Pacific-prehistory.pdf. Spriggs, Matthew, and Patrick V. Kirch. 1992. “‘Auwai, Kanawai, and Waiwai: Irrigation in Kawailoa-Uka.” In Anahulu: The Anthropology of History in the Kingdom of Hawaii , edited by M. Sahlins and P.V. Kirch, 2:118–64. Chicago: University of Chicago Press. Tercinier, G. 1974. “Les sols de l’île de Mangareva (Gambier): étude pédologique témoin d’une île haute de la Polynésie Française.” In Cahiers du Pacifique , 18:341–57. Paris: Fondation Singer-Polignac. Thomas, Frank R. 2003. “Shellfish Gathering in Kiribati,Micronesia: Nutritional,Microbiological, and Toxicological Aspects.” Ecology of Food and Nutrition 42 (2): 91–127. https://doi.org/10.1080/03670240390202246. Titcomb, Margaret, and Mary Kawena Pukui. 1969. Dog and Man in the Ancient Pacific, with Special Attention to Hawaii . Bishop Museum Special Publication 59. Honolulu, HI: Bernice P. Bishop Museum. https://cir.nii.ac.jp/crid/1130000796938777600. Weisler, Marshall I., and Roger C. Green. 2013. “Mangareva Fishing Strategies in Regional Context: An Analysis of Fish Bones from Five Sites Excavated in 1959.” Journal of Pacific Archaeology 4 (1): 73–89. White, Theodore E. 1953. “A Method of Calculating the Dietary Percentage of Various Food Animals Utilized by Aboriginal Peoples.” American Antiquity 18 (4): 396–98. Wills, Ron B. H., Jessie S. K. Lim, Heather Greenfield, and Tim Bayliss‐Smith. 1983. “Nutrient Composition of Taro ( Colocasia Esculenta ) Cultivars from the Papua New Guinea Highlands.” Journal of the Science of Food and Agriculture 34 (10): 1137–42. https://doi.org/10.1002/jsfa.2740341015. Winterhalder, Bruce P., and Cedric O. Puleston. 2018. “The Exchequer’s Guide to Population Ecology and Resource Exploitation in the Agrarian State.” Cliodynamics, 9 (2) . https://www.academia.edu/download/58096716/Winterhalder_and_Puleston_2018.pdf. Tables Tables 1 to 19 are available in the Supplementary Files section Additional Declarations The authors declare no competing interests. Supplementary Files SUPPLEMENTALMATERIALS.docx Tables.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-4797211","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":331375738,"identity":"bfe9eb4d-623c-4388-b91d-7dfa73b59070","order_by":0,"name":"Cedric O. Puleston","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA5klEQVRIie3OMQuCQBTA8SeBLgetLwz6CoKQSYJfxUO4SVqCaApDuJai1Y9RBM2F0OQHuEaXppY2hYYkmipO2hruzx0PjveDA1Cp/rJW/Bzt+uLzAFgNRKtJANCJfybW4bXfSByk86IqvZEtwkJUfOCDkexRRtyUJjYJ2LgvmO2uONKYnCZSYgnKTQgyuheRjiTHADDqN5FFVdZkl0Z6556jD71rI+FAarLBSDfJFLUYiZy4yyIxCWM0zS/2sDtFygkbD2TEMcLjrfQ8ul6Exflqzfy2kW2F9GMfL7ps/TtRqVQq1XsPuP1GPJ+PH58AAAAASUVORK5CYII=","orcid":"","institution":"University of California, Davis","correspondingAuthor":true,"prefix":"","firstName":"Cedric","middleName":"O.","lastName":"Puleston","suffix":""},{"id":331375739,"identity":"05594c54-d1ed-446b-adeb-3ea96a9387f8","order_by":1,"name":"Jennifer G. Kahn","email":"","orcid":"","institution":"The College of William \u0026 Mary","correspondingAuthor":false,"prefix":"","firstName":"Jennifer","middleName":"G.","lastName":"Kahn","suffix":""},{"id":331375740,"identity":"fe5c836c-fa92-4fa4-bb47-746c6a38db7a","order_by":2,"name":"Oliver A. Chadwick","email":"","orcid":"","institution":"University of California, Santa Barbara","correspondingAuthor":false,"prefix":"","firstName":"Oliver","middleName":"A.","lastName":"Chadwick","suffix":""},{"id":331375741,"identity":"02afa0f5-dc22-4732-a93b-28405db3833f","order_by":3,"name":"Nick Belluzzo","email":"","orcid":"","institution":"The College of William \u0026 Mary","correspondingAuthor":false,"prefix":"","firstName":"Nick","middleName":"","lastName":"Belluzzo","suffix":""},{"id":331375742,"identity":"126d7c73-ff9c-4e6c-b635-07c8e8f83f3b","order_by":4,"name":"Patrick V. Kirch","email":"","orcid":"","institution":"University of Hawaiʻi at Mānoa","correspondingAuthor":false,"prefix":"","firstName":"Patrick","middleName":"V.","lastName":"Kirch","suffix":""}],"badges":[],"createdAt":"2024-07-24 17:50:31","currentVersionCode":1,"declarations":{"humanSubjects":false,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-4797211/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4797211/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":61300366,"identity":"9a5b3729-43a5-4644-b0ca-234af71e5dcd","added_by":"auto","created_at":"2024-07-29 08:51:19","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":463425,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4797211/v1/b85c0056-4f5a-4f99-b393-28d46f19f660.pdf"},{"id":61299470,"identity":"96a64dcc-d998-455c-9392-0f8ee900fc1d","added_by":"auto","created_at":"2024-07-29 08:43:19","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":244426,"visible":true,"origin":"","legend":"","description":"","filename":"SUPPLEMENTALMATERIALS.docx","url":"https://assets-eu.researchsquare.com/files/rs-4797211/v1/08a0f8ba67883c1f382d8d8e.docx"},{"id":61299471,"identity":"b22c985c-9bf1-4b7d-9492-629b6e5a743c","added_by":"auto","created_at":"2024-07-29 08:43:19","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":822356,"visible":true,"origin":"","legend":"","description":"","filename":"Tables.docx","url":"https://assets-eu.researchsquare.com/files/rs-4797211/v1/989fc16f90cf378e48f1d653.docx"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003ePeople of the sea, or of the soil? How the balance of marine and terrestrial resource availability informs maximum population on four Polynesian islands\u003c/p\u003e","fulltext":[{"header":"INTRODUCTION","content":"\u003cp\u003eElements of a common culture, ancestral language and a dependence on a core set of common plants and animals unite many Pacific populations across stretches of open ocean that span more than 70 degrees of latitude, from the Hawaiian Archipelago to the southern tip of Aotearoa New Zealand. However, despite these powerful commonalities, early European visitors were often struck by how much the islands differed from one another, in the population sizes or densities of occupation, the availability of food and goods for trade, and the apparent quality of life. This variability has been noted by anthropologists and archaeologists. Certainly, some of this variation is a consequence of historical accident, including the effect of choices individuals and groups made in the thousand or more years since the more remote islands were first settled. But an additional source of variability lies in the nature of the islands themselves: in the wide variation in ecological characteristics, including climate, geologic substrate, the likelihood of catastrophic weather events, proximity to other islands or landmasses, the presence of fertile valley bottoms and the extent of a surrounding reef, each of which influenced how the human population lived. Thus, island biogeography and other ecological factors have played a key role in debates concerning precontact sociopolitical complexity and human ecodynamics in the Pacific Islands.\u003c/p\u003e\n\u003cp\u003eScholarly efforts to identify the features of Polynesian islands that are most predictive of population size and density have, for good reason, focused on island area and resource availability. Larger islands are generally host to larger human populations, but the relationship does not appear to be linear. In a careful study of several Polynesian outlier atolls Tim Bayliss-Smith (1974) proposed that human populations were limited by their ability to produce the staple crops with a high return in starch calories to labor inputs. In a response, Stephen Beckerman (1977) proposed the “protein hypothesis,” in which he argued that access to coconuts and near-shore protein sources were limiting, supporting his argument with data showing a strong correlation between island circumference and estimates of precontact population size. Rosalind Hunter-Anderson and Yigal Zan persuasively argued that Beckerman’s analysis was flawed and claimed instead, that “the upper limit of population size appears to be conditioned by the presence of intensifiable resources. Under certain geographical conditions, these resources might involve protein, and under others, starch. In small tropical high islands, the intensifiable resources tend to be starch” (Hunter-Anderson and Zan 1985). Their argument was in line with Bayliss-Smith’s findings in the small atolls he studied. But Bayliss-Smith described a method of estimating populations on the islands using just starch calories and a correction factor for uncalculated starch inputs, then modified by a correction factor to account for discrepancies between estimates of theoretical human carrying capacity and observed population numbers. Hunter-Anderson and Zan only went so far as to describe their model conceptually.\u003c/p\u003e\n\u003cp\u003eOur study addresses the question of limits on population size and ecological variables, notably the balance of resource availability between the land and the water. This study is unique in its effort to quantify and interpret the combined population consequences of two features of islands: the extent and typology of land suitable for traditional agriculture, and the extent of near-shore shallow water protected by reef, which provides access to marine food resources. As there exists no way to test the importance of these factors in a controlled manner, our study makes use of Polynesia as a “natural experiment” (Kirch 1980), meaning that we observe a sample of islands and attempt to understand how much our two ecological factors influence estimates of maximum population size or density of occupation. By choosing islands that are similar with respect to a number of the other confounding ecological characteristics, we minimize the risk of missing an important factor.\u003c/p\u003e\n\u003cp\u003eThe specific question we ask here is: To what extent does the availability of terrestrial resources vs. near-shore marine resources drive estimates of island population density on Polynesian islands? To answer the question, we selected four islands in Eastern Polynesia in relatively close proximity, thus having similar climate, and with a range of agricultural resources and varying degrees of access to near-shore marine resources. In addition, each of the islands has archaeologically excavated samples of residential or food-preparation sites, providing island-specific information on the relative distribution of terrestrial and marine animal food sources. We estimated agricultural production by assigning island area to either non-agricultural use, or to the inferred agricultural use based on slope and substrate, as well as on ethnohistoric accounts of land use. We then used multiple published sources to estimate the nutritional value and expected yields of each agricultural zone. We estimated near-shore marine resources by quantifying the surface area of shallow reef-enclosed water. Then, given the uncertainty of fish yields in early Polynesia, we used three scenarios of reef productivity, spanning the range observed in modern times using various fishing tools and methodologies. The reef fish yield scenarios were used to convert the relative faunal contribution data from the excavations into absolute biomass numbers.\u003c/p\u003e"},{"header":"METHODS","content":"\u003ch2\u003eLand Classification\u003c/h2\u003e\n\u003cp\u003eWe imported topographical data for the islands of Moʻorea, Maupiti (Society Islands, French Polynesia), and Mangareva and Taravai (Gambier Islands, French Polynesia) and soil maps of Moʻorea and Mangareva into a GIS format. We extended the Moʻorea soil classes to Maupiti and the Mangareva soil classes to Taravai following Kirch et al. (2022), who verified the existing soil maps and the extrapolations with a lab analysis of soil samples and confirmed the expected nutrient availability as indicated by base saturation. We then characterized the land area by suitability for agriculture, and if suitable, the most likely use according to a rubric using slope, aspect, and substrate as variables. These productivity zones were then corrected, where necessary, by Kahn and Kirch, who have familiarity with landscape use and agricultural site distribution specific to particular ecotones on each island. Small associated coral islands (\u003cem\u003emotu\u003c/em\u003e) were excluded from calculations except where evidence existed for raised-bed agriculture. (See the note in \u0026ldquo;Sources of potential error\u0026rdquo; in the Results and Discussion section, below.)\u003c/p\u003e\n\u003cp\u003eOn Mangareva lithosols and lateritic soils (26.2% and 33.1% of area, respectively, per Tercinier (1974)) were deemed unproductive. On Moʻorea, ferralitic soils (5.6% of area, per Buffard-Morel and Thuilier (1994)) were also eliminated from our productivity estimates. Additionally, on Moʻorea and Maupiti, perennial streams, which are not found on the other two islands, were buffered by 15 meters and removed from the productivity estimates (with one exception for wet taro, described below). On all four islands, the remaining land with a slope in excess of 50% was deemed only suitable for upland semi-cultivation of low-yield foods, while zones with slope \u0026gt;30% and \u0026le;50% were classed for shifting cultivation. While some of the upland terrain was certainly too steep even for famine foods, we argue that the broad classification is a useful approximation, even if it contributes to what is probably an overestimate of this relatively minor contribution to terrestrial food resources (see below). Land with slope \u0026gt;5% and \u0026le;30% was designated for arboriculture and shifting cultivation, except on Mangareva, which assigned eutrophic soils with slope 0% to \u0026le;30% to this category. Land with slope \u0026le;5% and calcareous soils on Mangareva were assigned to breadfruit and coconut cultivation. Exceptions occur on Mangareva and Moʻorea, where hydromorphic soils (0.9% of Mangareva\u0026rsquo;s area) and \u0026le;5% slope land with evidence of raised beds (0.5% of Moʻorea\u0026rsquo;s area) were assigned to wet taro production. On Maupiti, a small portion of the \u0026le;5% slope area abutting Haranai Stream (0.4% of main island) was classed as suitable for wet taro and the island\u0026rsquo;s motu included 0.2 km\u003csup\u003e2\u003c/sup\u003e for raised beds of wet taro. See Tables 1 and 2 for a summary.\u003c/p\u003e\n\u003ch2\u003eEstimating Terrestrial Agricultural Productivity\u003c/h2\u003e\n\u003cp\u003eBased on our analysis of ethnohistoric and ethnographic sources, we defined a total of six different land use classes. Each was characterized by the dominant precontact agricultural products which could be cultivated there. In the absence of complete information on soil characteristics and the climatic data required to parameterize crop models, and given the paucity of such models for many of the traditional Polynesian food crops, we relied on best estimates of typical crop yields from the literature in similar contexts where traditional subsistence agriculture was observed. The land use classes are:\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eWet taro:\u003c/strong\u003e Hamilton and Kahn (2007) drew on Kirch (1994) and Spriggs (1981, 1984) to estimate that wet taro cultivation on Moʻorea yielded 25 U.S. tons of wet corms per hectare per year. To estimate caloric content of taro, we averaged six sources to estimate a yield of 56,950 kcal/ha/day, of which 89.8% come from carbohydrates and 4.8% from protein (Supplemental materials, Table S1). These energy estimates account for a 20% fallow span in wetland taro cultivation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eArboriculture:\u003c/strong\u003e We drew from Hamilton and Kahn (2007, Table 8.5, p. 149), who for Moʻorea adapted Kirch\u0026rsquo;s (1994) description of Futuna (a Polynesian Outlier) arboriculture. The system relies on breadfruit, coconut, banana, \u003cem\u003eAlocasia\u003c/em\u003e aroids, \u003cem\u003eColocasia\u003c/em\u003e taro, \u003cem\u003eDioscorea\u003c/em\u003e yams and sweet potato. We added coconut to the food crops, as it was not included in the previous starch-based analyses. For Moʻorea and Maupiti we assumed a breadfruit to coconut tree ratio of 40:60 and for Mangareva and Taravai we used a ratio of 75:25. Breadfruit occurs at an average of 45 trees/ha on the former two islands and 50 trees/ha on the latter, and the difference in the breadfruit:coconut ratio is achieved mostly through a reduction in the prevalence of coconut. Planting density of yams and sweet potato was assumed to be equal. Using the nutritional data in Supplemental materials Table S1 we estimate this zone yields 37,712 kcal/ha/day on Moʻorea and Maupiti, and 37,421 kcal/ha/day on Mangareva and Taravai (Table 3). Carbohydrates contributed 32,974 and 34,692 kcal/ha/day, respectively, and protein contributed 1,737 and 1,856 kcal/ha/day, respectively.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eShifting cultivation:\u003c/strong\u003e Kirch (1994, Table 12, p. 183) describes a shifting agricultural system on Futuna, which we modified to include sweet potato in place of 50% of the yam yield (yams not being as important in Eastern Polynesia). The dryland aroids were assumed to be exclusively dryland taro. We assume 3 years of cultivation and 10 years of fallow, following Hamilton and Kahn (2007, p. 146). This zone yields an average of 12,031 kcal/ha/day after accounting for fallow, of which 10,809 derive from carbohydrates and 584 derive from protein (Table 4).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eArboriculture and shifting cultivation:\u003c/strong\u003e We assume the area in this zone is divided evenly between the two agricultural types. After accounting for fallow, as above, this zone contributes on average 24,872 kcal/ha/day on Moʻorea and Maupiti, of which 21,891 derive from carbohydrates and 1,160 from protein. On Managareva and Taravai the yield is 24,726 kcal/da/day, with 22,751 and 1,220 coming from carbohydrates and protein, respectively.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eBreadfruit and coconut:\u003c/strong\u003e While breadfruit was an important source of starch in the Polynesian diet, coconut was used primarily for oil and milk in cooking, providing little in the way of starch. Duke (2001) estimates that coconut trees provide 200 kg of copra (dried meat from which oil is extracted) annually per ha in Polynesia. At 6% water content, this would provide 3,440 total kcal/ha/day in monoculture, of which 82.6% come from fat (Haytowitz et al. 2005). If only the fat is consumed it yields 2,842 kcal/ha/day. We assume this zone is dedicated to 40% breadfruit and 60% coconut by number of trees, following Kirch (1994, Table 11, p. 182), except on Mangareva and Taravai, where breadfruit makes up 75% of the trees. On Moʻorea and Maupiti breadfruit interspersed with coconut provides an estimated 19,471 kcal/ha/day, with 15,470 and 1,067 coming from carbohydrates and protein, respectively. On Mangareva and Taravai the yield is 19,180 kcal/ha/day, with 17,189 and 1,186 from carbohydrates and protein, respectively.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eUpland semi-cultivation:\u003c/strong\u003e This zone is made up of wild and semi-cultivated foods in the regions less accessible and otherwise unsuitable for intensive agriculture. Hamilton and Kahn (2007, p. 149) follow Massal and Barrau (1956, p. 17) to estimate 1.25 U.S. tons of starchy foods per ha per year. After accounting for a fallow of three years for each year of harvesting, the average yield is 979 kcal/ha/day, with 830 and 50 coming from carbohydrates and protein, respectively.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;To estimate the contribution of agricultural production to the diet of the islanders we scale the agricultural zone productivity (Table 5) by the zone-specific area on each island (Table 2). The potential agricultural energetic output by island is found in Table 6.\u003c/p\u003e\n\u003ch2\u003eMarine and Reef Classification\u003c/h2\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eOur study is the first to include productivity estimates of the near-shore marine ecology while modeling island productivity in relation to precontact human populations. To estimate near-shore contributions of fish and shellfish to the Polynesian diet on these islands, we examined maps and satellite images to divide the reef-associated area into \u0026ldquo;productive\u0026rdquo; and \u0026ldquo;non-productive\u0026rdquo; zones. Submerged area was designated as productive if it was shallower than 30m of depth, including the outer edge of the reef. Areas with a substrate identified as mud were excluded. The 30m limit is common in studies of reef productivity (see Craig et al. 2008, p. 238) and corresponds roughly to the depth at which light penetration declines sharply and primary marine productivity approaches zero. Where depth charts were not available we used visible-light satellite images imported into ArcMap to estimate the 30m contour. We found that where we had both visible-light imagery and bathymetric data (Institut G\u0026eacute;ographique National 2015), for Mangareva and Taravai only, the 30m contour corresponded well to the line at which the visible lagoon substrate graded into the color of deep water. For consistency we used the same visual spectrum method of lagoon and reef classification for all four islands. We manually corrected any errors and removed modern structures which extended into the marine zones, such as airport runways or docks. The productive reef zone area for each island is shown in Table 7. As Mangareva and Taravai share a fringing reef and lagoon, we allocated this productive marine area proportional to their land area for our calculations.\u003c/p\u003e\n\u003ch2\u003eEstimating Marine and Reef Productivity\u003c/h2\u003e\n\u003cp\u003eEstimates of nearshore yields of fish and shellfish derive mainly from two sources of information. The first includes studies of observed yields in populations practicing subsistence fishing and foraging using a range of methods and equipment, from those similar to what precontact Polynesians used to those that more resemble modern small-scale commercial fishing. The second set of sources seeks to estimate sustainable or maximum harvests using whatever methods might be available, including modern highly extractive commercial means. Both sources are informative, the first providing evidence of what people actually do, and the second summarizing data from multiple locations and fishing methods to estimate potential yield. We divide marine inputs into five categories: nearshore finfish, shellfish (mollusks and crustaceans), pelagic (open-water) fish, turtles, and marine mammals.\u003c/p\u003e\n\u003cp\u003eDalzell and Adams (1997) drew on data from 43 locations in the South Pacific to determine that contemporary observed yields range from 0.3 to 64 metric tons live weight of reef finfish (excludes shellfish) per km\u003csup\u003e2\u0026nbsp;\u003c/sup\u003eper year, with a mean of 7.7 mt. They estimate the maximum sustainable yield (MSY) varies by location between 6-20 mt/km\u003csup\u003e2\u003c/sup\u003e/year, with a typical value of about 16. Newton et al. (2007) expanded on Dalzell and Adams (1997) and included mollusks and crustaceans in a study of 79 data sets from coral islands around the world compiled by the U.N. FAO. The yields ranged from 0.2 to 40 mt/km\u003csup\u003e2\u003c/sup\u003e/year, with a median of approximately 3. For their analyses the authors assumed a \u0026ldquo;more realistic\u0026rdquo; MSY of 5 mt/km\u003csup\u003e2\u003c/sup\u003e/year, but bracketed that value with scenarios of MSY at 1 and 10 mt/km\u003csup\u003e2\u003c/sup\u003e/year. Leenhardt et al. (2012) reviewed six studies of Moʻorea finfish catches and argued that observational studies of fishers and of roadside sales significantly underestimate the true rate of removal because these fish rarely get tallied. They estimated the yield at the time to be about 25 mt/km\u003csup\u003e2\u003c/sup\u003e/year, exceeding the MSY of 23 mt/km\u003csup\u003e2\u003c/sup\u003e/year for Moʻorea calculated by Galzin (1987).\u003c/p\u003e\n\u003cp\u003eDalzell and Adams (1997) predict higher maximum sustainable yields than Newton et al. (2007), despite the fact that they exclude mollusks and crustaceans from their analysis. However, their data come exclusively from the South Pacific. We use Dalzell and Adams\u0026rsquo; (1997) estimate of 16 mt/km\u003csup\u003e2\u003c/sup\u003e/year for reef-associated fish in the South Pacific to represent the region\u0026rsquo;s maximum sustainable yield, an ecological measure of fish carrying capacity. We also examine the values of 5 and 1 mt/km\u003csup\u003e2\u003c/sup\u003e/year as moderate and low estimates of what subsistence fishing economies in remote parts of Polynesia may yield.\u003c/p\u003e\n\u003cp\u003eUsing modern yield data to estimate the potential contribution of mollusks and crustaceans is more problematic than is the case with finfish. The mix of mollusks and crustacean species targeted for precontact subsistence might be different than those that appear in yield records recorded in the FAO database. In a study of the outer islands of American Samoa, Craig et al. (2008) used conversions for estimating edible mass from live weight (including shells), determining that invertebrates made up approximately 17% of the annual harvest of reef-associated live weight (after excluding shell weights). The species captured included octopus (49.7% of invertebrate mass), polychaete worms (26.6%), spiny lobster (13.0%, including shell), turban snail (6.5%, without shell), giant clam (3.6%, without shell) and sea urchin (0.7%, without shell). In a comparison of fishing practices and yields between 1982 and 2002 in a remote subsistence fishery in Fiji, Kuster et al. (2005) found that in both years finfish made up 76% of the live weight caught and invertebrates were 24%. While we have estimates of marine contributions to diet from archaeological excavations, the data from our four islands excludes shellfish and mollusks (other than crustaceans from Moʻorea and Maupiti) due to excavation methodologies. Therefore, we average these two results and assume that shellfish and mollusks make up 20.5% of the combined live weight of the nearshore catch, equivalent to 25.8% of the finfish biomass (20.5%/79.5%=25.8%).\u003c/p\u003e\n\u003cp\u003eFinally, open-water, or pelagic, fish and sharks were a culturally important food in many locations across Polynesia. Their contribution to the diet, however, varied across Polynesia (Allen 2017). Fraser (2001, p. 127) argued that this variation \u0026ldquo;reflect[s] cultural-historical, rather than natural, processes.\u0026rdquo; Fraser found in a study of data from 21 excavated Polynesian sites that the bones of offshore species were rare, with three notable exceptions, all from Eastern Polynesia (see also Allen 2017). Following this, we used biomass estimates from excavated faunal remains from three archaeological sites on Maupiti and two on Moʻorea. We limited our analysis to data from layers dated from AD 1400-1800, and eliminated data from one location on each island (MAU-11 on Maupiti and ScMo-350 Block 3 on Moʻorea) that appeared to be high-status or ceremonial. We focus on the later phase of precontact occupation, when some of the foods favored by the earliest arrivals were no longer abundant and human populations were larger. We used these biomass estimates to reconstruct the relative contributions of faunal inputs to the traditional diet, first averaging across time within each location and then across locations to get an island-wide estimate. For Mangareva and Taravai we used biomass estimates from excavations on Agakauitai (Kirch et al. 2015), a small island which shares the lagoon with Mangareva and Taravai (Kirch et al. 2010). As before, we limited our analysis to layers dated AD 1400-1800, which consisted of two on Agakauitai and one from Taravai. Biomass distributions were averaged across the two layers at Agakauitai and the result was averaged with the Taravai data. We used this distribution to characterize both Mangareva and Taravai\u0026rsquo;s relative faunal inputs.\u003c/p\u003e\n\u003cp\u003eThe data show evidence of modest pelagic input on Moʻorea and Maupiti, despite numerous bones of nearshore fish. Requiem sharks were recovered in almost all precontact cultural deposits, accounting for the largest fraction of estimated biomass from pelagic species. On Maupiti requiem sharks made up 17.0% of the fish biomass, and 5.6% on Moʻorea. Pelagic mackerel and tuna, which like the requiem shark can sometimes be caught close to shore, contributed 1.3% of estimated fish biomass at Maupiti and 0.7% at Moʻorea. Reef-associated fish made up the largest part of the marine-derived diet on both islands (Table 8). At Mangareva and smaller islands in the lagoon pelagic fish are sparse in the archaeological record (Weisler and Green 2013; Rurua 2015). Family Scombridae, which includes pelagic tuna and mackerel, represented \u0026lt; 1% of the number of identified specimens (NISP) at Mangareva, and was the 13\u003csup\u003eth\u003c/sup\u003e-most common family observed (Weisler and Green 2013). Biomass estimates from Kirch and associates\u0026rsquo; excavations on Agakauitai and Taravai do not include differentiation below the level of class. For these islands we adopt an estimated pelagic fish fraction of 5% of total fish biomass.\u003c/p\u003e\n\u003ch2\u003eOther Animal Inputs\u003c/h2\u003e\n\u003cp\u003eThe distribution of estimated faunal biomass in the excavation data provides insight into the relative importance of the various food sources. Here we use the same subset of data from Kahn (Kahn, In press) for Maupiti and Moʻorea, and from the excavations on Agakauitai (Kirch et al. 2015) and Taravai (Kirch et al. 2010). In all cases the biomass estimates do not include shellfish or mollusks, which we instead estimate at 25.8% of the fish biomass, as described above. The two Maupiti commoner residential sites (MAU-1 and MAU-2) and intermediate-status (MAU-5) site differ somewhat from each other, and across time to an extent, but in aggregate biomass is dominated by finfish (46.5%) and mammals (45.3%), with turtles contributing 6.6% and birds 1.7%. Shellfish and mollusks were not included in the biomass analysis. The relative proportions of rat, dog and pig biomass are fairly stable, even across the status gradient, meaning that although the higher-status sites had greater biomass of identified mammals, the ratio of pig to dog biomass, for example, was more or less constant. Rat biomass was 5.8% of mammal biomass, dog was 8.1% and pig predominated with 86.1% of the estimated mammal biomass. Marine mammals contributed 0.1%.\u003c/p\u003e\n\u003cp\u003eAt the non-elite residential sites on Moʻorea there is a general trend over time toward less mammal and more fish biomass in pre-contact cultural deposits; this reflects \u003cem\u003etapu\u003c/em\u003e restrictions in the later period which excluded commoners largely from eating pig and dog (Kahn 2024; Oliver 1974, 276). However, averaging across the period 1400-1800 among commoner sites we find the biomass contribution of mammals was 37.8% and of finfish was 57.0%, with turtle contributing 4.5% and birds 0.8%. Of the estimated mammal biomass that could be identified from the Moʻorea sites, 5.3% came from rats, 28.0% from dogs, 54.8% from pigs and 11.9% from marine mammals.\u003c/p\u003e\n\u003cp\u003eAveraged across the two layers on Agakauitai and one layer on Taravai the biomass estimates from the Mangareva island group are 79.5% finfish, 0.8% turtle, 8.4% bird and 11.3% mammal, identified remains of which were made up mostly of pig (9.0% of faunal biomass) and rat (2.3%), while dog represented 0.1%. Turtle and marine mammal remains are absent from the samples.\u003c/p\u003e\n\u003ch2\u003eEstimates of Faunal Dietary Inputs\u003c/h2\u003e\n\u003cp\u003eThe estimated relative faunal contributions to late-precontact diet on the four islands are presented in Table 8, after including the shellfish contribution as a fixed 25.8% of the total fish biomass (see above). Being a function of reef-associated surface area, estimates of nearshore marine productivity are the only biomass inputs available that can be scaled across the four islands. We adopt values of 1, 5 and 16 mt/km\u003csup\u003e2\u003c/sup\u003e/year of nearshore finfish as our expected subsistence yields, representing the range of likely values.\u003c/p\u003e\n\u003ch2\u003eConverting Faunal Biomass to Food\u003c/h2\u003e\n\u003cp\u003eIn estimating the fraction of live weight that made its way into the local diet, what Lyman (1979) called \u0026ldquo;consumable meat,\u0026rdquo; we assume that the bulk of muscle (flesh) is available, while viscera, hide and bones are not consumed. Edible fat is assumed to be limited to that found in the meat, with the addition of blubber. We estimate the edible fraction of reef fish to be 50% of live weight, an approximation of fish yields of similar species (Buchanan 2023). Yield fractions of moderate to small sized fish vary widely, from as low as 30% to as much as 90% of live weight, but in general, smaller fish and fish with large, bony heads yield less. Pelagic fish, characterized by tuna, are assumed to yield 70% of live weight as edible meat (Smith 2011), and marine invertebrates are assumed to be entirely edible, as live weight excludes shells and exoskeletons. The edible portion (meat) of pigs is estimated to be 70% of live weight (White 1953). There are few measurements on the meat yields of dogs and rats, but Smith (2011) estimates these at 60% and 70% of live weight, respectively. These dogs were usually specially bred for eating and fattened on human foods, including breadfruit and taro (Titcomb and Pukui 1969; Bay-Petersen 1983). Lockyer (1991) includes an exhaustive study of 206 dissected whales from 9 species and all parts of the globe. The study found that 40% of whale live weight came from muscle and 26% came from blubber and we adopt these fractions for all marine mammals. The live weight of turtles includes the mass of the carapace and the edible fraction is estimated to be 50%, following a calculation in Frazier (1980, Table 1). Birds, like fish, vary widely in their particulars, but we follow Smith (2011) in the estimate of 70% edible fraction for small to medium birds, including terrestrial and wetland/marine species, in precontact New Zealand.\u003c/p\u003e\n\u003cp\u003eThe two nutritional components of greatest importance are the energy these foods provided and their protein content. Meat carbohydrates exist in small quantities, but are often recorded as zero in nutritional tables, and we will assume their contribution is negligible. The data we used to make nutritional component calculations, along with their sources, are included in the supplemental materials and are summarized in Table 9. The values reflect the nutrient content of multiple samples of meat (generally muscle tissue), which include incorporated fat. The blubber of cetaceans is an exception, and the values for birds includes the skin.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;To calculate the expected faunal energy available to the population we scale the faunal biomass distributions (Table 8) to the reef finfish amount, then multiply by the island-specific productive reef zone area (Table 7) to give metric tons of faunal biomass per metric ton of reef fish. We then multiply through by the edible fraction and caloric value of each food type (Table 9) to give food energy per metric ton of reef fish. Finally, we multiply these values by the scenario-appropriate value for reef fish catch per km\u003csup\u003e2\u003c/sup\u003e/year and then convert this to a daily value (kcal/day). The faunal energy yields under the three reef productivity scenarios are shown in Tables 10-12.\u003c/p\u003e"},{"header":"RESULTS AND DISCUSSION","content":"\u003ch2\u003eCombining Marine and Terrestrial Productivity\u003c/h2\u003e\n\u003cp\u003eTo better understand the estimated marine and terrestrial food yields we converted them into the maximum human population that might have been supported on those amounts. We adopted the simplifying assumptions that each person consumed 2700 kcal/day, representing the caloric requirement of a well-fed, active individual, and that 30% of the yield of plants and animals was unrealized, being either lost or diverted from the human diet. The 30% deduction represents a compromise between Bayliss-Smith (1974), who estimated populations at 70-80% of carrying capacity, and Spriggs and Kirch (1992), who estimated realized food production at approximately 50% of potential production, both in Pacific subsistence contexts. The results by reef fishing scenario, including a breakdown of food contributions by plant vs. animal sources, are summarized in Table 13. The same results are broken down by marine vs. terrestrial food sources in Table 14. The scenarios in the tables range from a low reef productivity level of 1 mt/km\u003csup\u003e2\u003c/sup\u003e/year of reef fish caught, through 5 and 16 mt/km\u003csup\u003e2\u003c/sup\u003e/year. These scenarios are labeled A-C, respectively. Recall that the archaeologically excavated faunal remains provide relative abundance data, which we convert to absolute values by pegging them to reef fish catch rate by scenario. As the estimate of fish catch increases, so does the mass of all other faunal food sources.\u003c/p\u003e\n\u003cp\u003ePutting the pieces together allows us to compare human population estimates across the varying islands. We have broken down the initial estimates of maximum population by the contributions of animal vs. plant resources (Table 13), and also by marine vs. terrestrial resources (Table 14). We find Moʻorea\u0026rsquo;s population maximum is many times larger than that of any of the other islands under all scenarios. But Moʻorea\u0026rsquo;s population ceiling does not increase much with estimates of fish abundance. The estimate of the island\u0026rsquo;s maximum population increases only from 43,575 to 44,368 even as reef fish inputs increase 16-fold. Population estimates among the other islands increase by almost 50% (Maupiti; from 1,489 to 2,210), 74% (Mangareva; from 2,954 to 5,139), or 94% (Taravai; from 701 to 1,364) across the scenarios. In the breakdown of faunal vs. agricultural food resources (Table 13) Moʻorea\u0026rsquo;s mix remains heavily agricultural in all cases, while the other three islands stand to gain much more \u0026mdash; in terms of maximum population size \u0026mdash; from better fishing success.\u003c/p\u003e\n\u003cp\u003eZooming out to examine the importance of ecological variables and food sources on population density, here we provide two measures. The first is whole-island population density (Table 15), which is population estimate by scenario divided by total land area, excluding motus. The second is density as a function of agricultural land area (Table 16), in which we divide by all land suitable for agriculture, excluding semicultivation zones, following Table 2.\u003c/p\u003e\n\u003cp\u003eIf questioning whether agricultural land is more or less important to food availability than marine resources, the answer is clear in Table 17. This table captures the tremendous value of terrestrial resources on all four islands under the most likely fish yield scenario, 5 mt/km\u003csup\u003e2\u003c/sup\u003e/y. Agricultural land is 150-200 times as productive as the near-shore marine area in terms of caloric availability. In all cases, fish and reef resources are important, but their potential contribution to the caloric budget is dwarfed by the terrestrial component.\u003c/p\u003e\n\u003cp\u003eTo this point we have calculated potential population purely on the basis of total calories, but the tremendous disparity in terrestrial and marine resources on Mo\u0026lsquo;orea (Table 14) raises the question of dietary balance. Human diets have a great deal of plasticity, but are ultimately constrained in their composition at the extremes. Fat and carbohydrates can be exchanged for one another by the body, within limits, but protein is different. Humans lack the ability to produce a number of essential amino acids, and dietary sources of protein are the only way to acquire them. Nutritionists suggest that the acceptable macronutrient distribution range (AMDR) is 45-65% of total calories from carbohydrates, 20-35% from fat (for ages 4+), and 10-35% from protein (for ages 18+) (Institute of Medicine 2005), and most human populations settle on diets that fall within these ranges. A number of Pacific subsistence/diet studies have observed starch percentages in excess of 65%, but even those include a protein fraction of 9.4% or greater (e.g., Bayliss-Smith 1974; Ross 1976:579; Lindeberg and Vessby 1995:48).\u003c/p\u003e\n\u003cp\u003eRecommended dietary allowances (RDAs) are a second way to assess macronutrient needs. These age- and sex-specific macronutrient mass values are assumed to capture the minimum requirements of 97 to 98% of the population (Institute of Medicine 2005). We used the UN Model Life Tables \u0026ndash; West to generate the age and sex structure of a hypothetical stable early Polynesian population and determined that the average RDA for protein in this population was almost exactly 45 g per person per day. This corresponds to 8% of a recommended 2,250 kcal daily diet. We have assumed that the subsistence Polynesian populations required 2,700 kcal per day (see above), and if protein makes up at least 8% of the energy in the diet, that comes to a minimum of 54g protein per day per person.\u003c/p\u003e\n\u003cp\u003eWe will consider both approaches to minimum protein requirements: 10% of the total kcal, and 54 g/person/day, corresponding to 8% of the total kcal. First, we calculate the number of people each island\u0026rsquo;s protein supply could support at the 10% minimum, independent of other macronutrients or calories, given the fish yields in Scenario B (Table 18). For example, Mangareva is expected to be able to supply enough protein to feed 6,632 people, but the maximum population based on total calories is only 3,537, so we assume protein is not limiting on Mangareva and the maximum population remains at the lower number. This is the case on all the islands, with the notable exception of Mo\u0026lsquo;orea, whose marine resources are far outstripped by its terrestrial ones, leading to the high likelihood of protein limitation. We estimate that the maximum population, fed with food procured on or nearby the island, would be 22,906, or 52.3% of the calorie-only estimate. In the second approach we relax the protein requirement to a minimum of 8% of the food energy (equivalent to 54 g/d/person), resulting in a maximum population on Mo\u0026lsquo;orea of 28,632 under the assumptions of fishing Scenario B. This is 65.4% of the na\u0026iuml;ve population estimate. While there is an argument to be made for either the 8% or 10% minimum protein requirement, seeing that the dietary studies support a protein minimum closer to 10% of total kcals, we regard 22,906 as more likely.\u003c/p\u003e\n\u003ch2\u003eSources of Potential Error\u003c/h2\u003e\n\u003cp\u003eIn comparison to estimated population densities on other islands in Polynesia before European contact, our values in Tables 15 and 16 are quite high. Puleston and Ladefoged (2022) concluded that estimates of precontact Polynesian population densities with respect to agricultural area ranged from about 50 to 250 p/km\u003csup\u003e2\u003c/sup\u003e. The estimates in Table 16 are several times greater than even that upper bound. There are several sources of variation, or error, that should be considered relating to our case studies. These include: (1) overestimates of agricultural productivity in one or more important agricultural types, (2) the mis-designation of less-productive land into a higher-productivity class, (3) underestimation of food diversion and losses due to ceremonial use, waste and spoilage, (4) the assumption that labor supply never limits productivity, and (5) the assumption that the populations were always working to maximize total production.\u003c/p\u003e\n\u003cp\u003eFirst, it is possible our methodology overestimates productive land area, but we have been careful to avoid doing so. We defined our land classifications precisely, but the suitability of land at the local scale is heterogeneous so perhaps only some fraction of the actual area was used in the manner we designated. However, our estimates of productivity by agricultural activity come from similar non-industrial systems in similar contexts, so the effect of heterogeneity is already incorporated into those calculations. One exception might be a modest overestimate of semicultivation zone area (see methods), but the productivity in these areas is so low that we do not expect it to make a significant difference to our results. Finally, density of population as a function of agricultural land should be less vulnerable to this form of error because overestimation of agricultural land should increase the denominator (agricultural land area) faster than the numerator (population size) in the calculation, particularly as our estimates of population include non-agricultural (marine) resources. It should be noted that our decision to include only known motu agricultural areas introduces a modest potential bias toward an \u003cem\u003eunderestimation\u003c/em\u003e of agricultural area by ignoring motu locations which might have hosted agriculture in the past, but are not currently known. We argue, however, that the likelihood of unrecognized raised-bed areas is small, and that the contribution of any such unincluded zones is unlikely to alter our results.\u003c/p\u003e\n\u003cp\u003eIn reference to the second possible source of error, it is possible that we overestimated potential production by classifying agricultural area as more productive than it might actually have been. This effect would be most pronounced in the zones we designated as appropriate for wet taro and mixed arboriculture and shifting.\u003c/p\u003e\n\u003cp\u003eIn terms of the third potential source of error, food crops are subject to stunting and also loss before and after harvest from diverse sources including damage from domestic and wild animals and insect pests, rainfall variability and drought, and rot after harvest. In addition, across Polynesia food offerings for ceremonial purposes diverted calories to the gods that might have otherwise been available to people. Furthermore, the cost of feeding agricultural calories to domestic animals convert one type of food into a more socially valuable one, but this is inefficient in terms of food energy. Obligatory tribute in the form of food resources was also common in Polynesia before European contact; the end result is a diversion of food from the people producing it into the hands and mouths of others. All of these play a role in decreasing food available to the agriculturalist population from a theoretical maximum. In most cases the yield figures we use account for losses in the fields, but the losses and diversions after that point need to be estimated separately. We have assumed 30% of potential production never makes it to the mouths of the human population, but some have argued that this is an underestimate. This correction factor also includes the effect of the fifth source of error, below. It should also be noted that a variable food supply, which includes the production of rain-fed systems, can have complex effects on population dynamics (Lee et al. 2009). Rare droughts and crop failures can have immediate catastrophic effects that may take generations to recover from, leading to a significant depression of mean population size vs. the theoretical maximum under the assumption of consistent yields. Societal features, including the requirement of tribute or labor diversion can exacerbate this effect (Winterhalder and Puleston 2018; Puleston and Winterhalder 2019).\u003c/p\u003e\n\u003cp\u003eThe fourth source of error centers around a potential shortage of laborers limiting the productivity of an agricultural system. Work by demographer Shripad Tuljapurkar and others, dubbed food-limited demography, (Lee and Tuljapurkar 2008; Puleston and Tuljapurkar 2008; Puleston and Winterhalder 2019) considered the role of labor in a rain-fed preindustrial agricultural system, focusing on the dynamic characteristics of the human populations that rely on that food. Such modeling work has found that in a natural-fertility population working to maximize productivity, labor is limiting while the population is growing and is more likely to be found in excess in a population living near its equilibrium or carrying capacity (Puleston, et al. 2014; Winterhalder and Puleston 2018). It is of course possible that ongoing diversions of labor into societal features like monumental architecture, the maintenance of infrastructure, or the presence of a highly specialized labor force, or a standing army, would make field labor scarcer, but in an established population the effect is small.\u003c/p\u003e\n\u003cp\u003eThe fifth assumption may be the most important one. Anthropologist Tim Bayliss-Smith (1980) has considered the tradeoff between population size and the intensity of labor effort. He and others found that in agricultural systems like the ones that dominated these islands there is a fairly common expectation as to how much labor a person contributes to agricultural work in a week. People would work less in many cases if they could, and in some cases work considerably more than the expectation of 10-20 hours a week (Bayliss-Smith 1978) in subsistence agriculture. However, in many places we do not really know what the local expectations of hours of field labor were. We also expect that the population, as individuals and as a group, might shift their energy investments to the activities that provided the greatest return, which might be measured in calories or in the more subjective currency of food desirability. The conclusion across a variety of scenarios is that while the number of laborers might be not be a limiting factor, the hours they work might be. The result is a level of total production, and a population size, diminished from its theoretical maximum by something on the order of 30%, as mentioned above. The point of Bayliss-Smith\u0026rsquo;s and food-limited demography\u0026rsquo;s findings is that there are tradeoffs between quality and quantity of life, and if the agricultural population has the power to make its own choices it may choose a smaller population with a better quality of life, in terms of work-hours, lifespan, infant mortality and food availability.\u003c/p\u003e"},{"header":"CONCLUSIONS","content":"\u003cp\u003eOur estimates of the maximum sustainable populations for the four islands are summarized in Table 19. \u0026nbsp;Moʻorea’s fertile valleys and wetlands would potentially generate sufficient food energy to support a maximum population \u0026gt; 40,000 people, regardless of fish yield scenario. The other three islands are much more dependent on estimates of fish yields, and the population numbers are much smaller, ranging from less than 1,000 to more than 3,500 in the most likely fish yield scenario. Clearly, ecology does matter: Moʻorea gets only 0.4% of its caloric potential from marine sources, while Maupiti gets 7.3%, Mangareva gets 15.0% and Taravai gets 18.4%. These relative inputs reflect the vast difference in productivity (in terms of calories per km\u003csup\u003e2\u003c/sup\u003e) in agricultural vs near-shore marine zones. Agricultural land is some 150 times as productive as the marine area at all four islands. We conclude that while terrestrial resources were much more plentiful, and thus played a more significant role, marine inputs could be important, particularly on smaller land-poor islands, and should not be ignored. Further, while terrestrial production had greater potential, these crops were both more variable and more vulnerable than the more passively maintained pool of marine resources, pointing to a promising area of inquiry for future study. Available studies suggest that traditional subsistence-level exploitation of near-shore marine systems led to sustainable use, despite the significant initial impact of human arrival and some evidence of long-term decline in the size of fish and invertebrates (e.g., Allen 2017). In contrast, local extirpations of high-value pigs and dogs on small or isolated islands, thought to be a result of intentional choices to reduce trophic competition with human populations, hints at the vulnerability of the terrestrial agricultural systems on smaller islands (Kirch 2000). Consider also the natural variability of rain-fed agriculture, which can be highly unpredictable (Lee et al. 2006). That many Pacific human populations were able to simultaneously rely on these two independent food systems surely contributed to their own resilience.\u003c/p\u003e\n\u003cp\u003eIt is important to note that in cases of extreme imbalance in the terrestrial vs. marine potential, such as is observed on the large, but geologically young island of Mo‘orea, the abundance of valued carbohydrate resources cannot be fully realized by the local population without sufficient protein availability. Our estimate of the island’s maximum population drops by almost half when we consider diet composition. Such constraints might well be relevant to other regions, like the Marquesas Islands and Rapa Nui in Eastern Polynesia, which have small fringing reefs due to young geological age, or to other ecological factors (e.g., the Humboldt Current in the Marquesas Islands). Even the starch-loving Polynesians required some quantity of fish, or meat, to make a proper meal (Kirch and O’Day 2003). Put another way, one cannot live by breadfruit alone.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003eAcknowledgements\u003c/p\u003e\n\u003cp\u003eSusan Andrews is thanked for calculating the initial biomass estimates from Kahn and Kirch\u0026rsquo;s faunal materials. We thank Neil Davies for helping to acquire GIS data from the Service de l\u0026rsquo;urbanisme of French Polynesia, whom we also thank. We thank Diane Ragone for generously sharing advice and information regarding breadfruit in Polynesia.\u003c/p\u003e\n\u003cp\u003eCompeting Interests\u003c/p\u003e\n\u003cp\u003eThe authors declare that there are no competing interests.\u003c/p\u003e\n\u003cp\u003eEthical Approval\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThere are no applicable ethics declarations.\u003c/p\u003e\n\u003cp\u003eFunding\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFunding for this work was provided by a National Science Foundation Grant (CNH # 1642894), awarded to Kahn and Kirch.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003eAb-Lah, Roslizawati, Joshua Smith, Dale Savins, Ashley Dowell, Daniel Bucher, and Kirsten Benkendorff. 2017. \u0026ldquo;Investigation of Nutritional Properties of Three Species of Marine Turban Snails for Human Consumption.\u0026rdquo; \u003cem\u003eFood Science \u0026amp; Nutrition\u003c/em\u003e 5 (1): 14\u0026ndash;30. https://doi.org/10.1002/fsn3.360.\u003c/li\u003e\n \u003cli\u003eAllen, Melinda S. 2017. \u0026ldquo;Spatial Variability and Human Eco-Dynamics in Central East Polynesian Fisheries.\u0026rdquo; https://academic.oup.com/edited-volume/34661/chapter/295345817.\u003c/li\u003e\n \u003cli\u003eAnn, Yong-Geun. 1999. \u0026ldquo;Dog Meat Foods in Korea.\u0026rdquo; \u003cem\u003eKorean Journal of Food and Nutrition\u003c/em\u003e 12 (4): 397\u0026ndash;408.\u003c/li\u003e\n \u003cli\u003eArai, Yoro, and Shigeru Sakai. 1952. \u0026ldquo;Whale Meat in Nutrition.\u0026rdquo; \u003cem\u003eThe Scientific Reports of the Whales Research Institute\u003c/em\u003e 7:51\u0026ndash;67.\u003c/li\u003e\n \u003cli\u003eBarrau, Jacques. 1958. \u003cem\u003eSubsistence Agriculture in Melanesia\u003c/em\u003e. Bulletin 219. Honolulu: Bernice P. Bishop Museum. https://cir.nii.ac.jp/crid/1570009750817146624.\u003c/li\u003e\n \u003cli\u003eBayliss-Smith, T. P. 1980. \u0026ldquo;Population Pressure, Resources and Welfare: Towards a More Realistic Measure of Carrying Capacity.\u0026rdquo; In \u003cem\u003ePopulation-Environment Relations in Tropical Islands: The Case of Eastern Fiji\u003c/em\u003e, edited by H. C. Brookfield, 61\u0026ndash;93. Paris: UNESCO.\u003c/li\u003e\n \u003cli\u003eBayliss-Smith, Tim. 1974. \u0026ldquo;Constraints on Population Growth: The Case of the Polynesian Outlier Atolls in the Precontact Period.\u0026rdquo; \u003cem\u003eHuman Ecology\u003c/em\u003e 2 (4): 259\u0026ndash;95. https://doi.org/10.1007/BF01531318.\u003c/li\u003e\n \u003cli\u003eBayliss-Smith, Tim P. 1978. \u0026ldquo;Batiki in the 1970s: Satellite of Suva.\u0026rdquo; In \u003cem\u003eThe Small Islands and Reefs\u003c/em\u003e, edited by H. C. Brookfield, 67\u0026ndash;128. Canberra: Australian National University, Development Studies Centre for UNESCO.\u003c/li\u003e\n \u003cli\u003eBay-Petersen, Jan. 1983. \u0026ldquo;Competition for Resources: The Role of Pig and Dog in the Polynesian Agricultural Economy.\u0026rdquo; \u003cem\u003eJournal de La Soci\u0026eacute;t\u0026eacute; Des Oc\u0026eacute;anistes\u003c/em\u003e 39 (77): 121\u0026ndash;29.\u003c/li\u003e\n \u003cli\u003eBeckerman, Stephen. 1977. \u0026ldquo;Protein and Population in Tropical Polynesia.\u0026rdquo; \u003cem\u003eThe Journal of the Polynesian Society\u003c/em\u003e 86 (1): 73\u0026ndash;79.\u003c/li\u003e\n \u003cli\u003eBradbury, J. Howard, and Warren D. Holloway. 1988. \u003cem\u003eChemistry of Tropical Root Crops: Significance for Nutrition and Agriculture in the Pacific\u003c/em\u003e. ACIAR Monographs 6. Canberra: Australian Centre for International Agricultural Research. https://ageconsearch.umn.edu/record/118050/files/6.pdf.\u003c/li\u003e\n \u003cli\u003eBuchanan, David. 2023. \u0026ldquo;Fish Fillet Butchering Yields.\u0026rdquo; Fish Fillet Butchering Yields. 2023. https://www.chefs-resources.com/seafood/seafood-yields/.\u003c/li\u003e\n \u003cli\u003eBurtenshaw, Mike, and Graham Harris. 2007. \u0026ldquo;Experimental Archaeology Gardens Assessing the Productivity of Ancient Māori Cultivars of Sweet Potato, Ipomoea Batatas [L.] Lam. in New Zealand.\u0026rdquo; \u003cem\u003eEconomic Botany\u003c/em\u003e 61 (3): 235\u0026ndash;45. https://doi.org/10.1663/0013-0001(2007)61[235:EAGATP]2.0.CO;2.\u003c/li\u003e\n \u003cli\u003eCraig, P., A. Green, and F. Tuilagi. 2008. \u0026ldquo;Subsistence Harvest of Coral Reef Resources in the Outer Islands of American Samoa: Modern, Historic and Prehistoric Catches.\u0026rdquo; \u003cem\u003eFisheries Research\u003c/em\u003e 89 (3): 230\u0026ndash;40.\u003c/li\u003e\n \u003cli\u003eDalzell, P., and T. J. H. Adams. 1997. \u0026ldquo;Sustainability and Management of Reef Fisheries in the Pacific Islands.\u0026rdquo; In \u003cem\u003eProc. 8th Int. Coral Reef Symp\u003c/em\u003e, 2:2027\u0026ndash;32. https://www.spc.int/digitallibrary/doc/fame/reports/adams_96_panama.pdf.\u003c/li\u003e\n \u003cli\u003eDignan, Cecily, Barbara Burlingame, Shailesh Kumar, and William Aalbersberg. 2004. \u003cem\u003eThe Pacific Islands Food Composition Tables.\u003c/em\u003e Ed. 2. Food and Agriculture Organization of the United Nations (FAO). https://www.cabdirect.org/cabdirect/abstract/20043191618.\u003c/li\u003e\n \u003cli\u003eDuke, James A. 2001. \u0026ldquo;\u003cem\u003eCocos Nucifera L.\u003c/em\u003e\u0026rdquo;\u0026nbsp;In \u003cem\u003eHandbook of Nuts\u003c/em\u003e, 1st ed., 100\u0026ndash;109. Boca Raton, FL: CRC Press.\u003c/li\u003e\n \u003cli\u003eFAO. 2016. \u0026ldquo;INFOODS Global Food Composition Database for Fish and Shellfish Version 1.0-uFiSh1. 0.\u0026rdquo; Rome, Italy. https://www.fao.org/3/a-i6655e.pdf.\u003c/li\u003e\n \u003cli\u003eFraser, K.L. 2001. \u0026ldquo;Variation in Tuna Fish Catches in Pacific Prehistory.\u0026rdquo; \u003cem\u003eInternational Journal of Osteoarchaeology\u003c/em\u003e 11 (1\u0026ndash;2): 127\u0026ndash;35. https://doi.org/10.1002/oa.551.\u003c/li\u003e\n \u003cli\u003eFrazier, J. 1980. \u0026ldquo;Exploitation of Marine Turtles in the Indian Ocean.\u0026rdquo; \u003cem\u003eHuman Ecology\u003c/em\u003e 8 (4): 329\u0026ndash;70. https://doi.org/10.1007/BF01560999.\u003c/li\u003e\n \u003cli\u003eGalzin, Rene. 1987. \u0026ldquo;Potential Fisheries Yield of a Moorea Fringing Reef (French Polynesia) by the Analysis of Three Dominant Fishes.\u0026rdquo; \u003cem\u003eAtoll Research Bulletin\u003c/em\u003e. https://repository.si.edu/bitstream/handle/10088/5853/00305.pdf.\u003c/li\u003e\n \u003cli\u003eHamilton, Brenda K., and Jennifer G. Kahn. 2007. \u0026ldquo;Pre-Contact Population in the \u0026rsquo;Opunohu Valley, Mo\u0026rsquo;orea: An Integrated Archaeological and Ethnohistorical Approach.\u0026rdquo; In \u003cem\u003eThe Growth and Collapse of Pacific Island Societies: Archaeological and Demographic Perspectives\u003c/em\u003e, edited by Patrick V. Kirch and J.L. Rallu. University of Hawaii Press.\u003c/li\u003e\n \u003cli\u003eHaytowitz, David, Jaspreet Ahuja, Xianli Wu, Meena Somanchi, Melissa Nickle, Quyen Nguyen, Janet Roseland, et al. 2005. \u0026ldquo;USDA National Nutrient Database for Standard Reference, Legacy Release.\u0026rdquo; Nutrient Data Laboratory, Beltsville Human Nutrition Research Center, ARS, USDA. https://doi.org/10.15482/USDA.ADC/1529216.\u003c/li\u003e\n \u003cli\u003eHipsley, Eben H., and Nancy E. Kirk. 1965. \u003cem\u003eStudies of Dietary Intake and the Expenditure of Energy by New Guineans.\u003c/em\u003e Technical Paper 147. Noumea, New Caledonia: South Pacific Commission. https://www.cabidigitallibrary.org/doi/full/10.5555/19662900948.\u003c/li\u003e\n \u003cli\u003eHuang, A. S., C. A. Titchenal, and B. A. Meilleur. 2000. \u0026ldquo;Nutrient Composition of Taro Corms and Breadfruit.\u0026rdquo; \u003cem\u003eJournal of Food Composition and Analysis\u003c/em\u003e 13 (5): 859\u0026ndash;64.\u003c/li\u003e\n \u003cli\u003eHunter-Anderson, Rosalind L., and Yigal Zan. 1985. \u0026ldquo;\u0026lsquo;Where\u0026rsquo;s the Fish?\u0026rsquo; Or, What Limited the Size of Aboriginal Populations in Small Tropical High Islands?\u0026rdquo; \u003cem\u003eThe Journal of the Polynesian Society\u003c/em\u003e 94 (1): 61\u0026ndash;65.\u003c/li\u003e\n \u003cli\u003eInstitut G\u0026eacute;ographique National. 2015. \u0026ldquo;French Polynesia Topographic Maps.\u0026rdquo; Paris: Polyn\u0026eacute;sie Fran\u0026ccedil;aise, Service de l\u0026rsquo;urbanisme, Section topographie.\u003c/li\u003e\n \u003cli\u003eInstitute of Medicine. 2005. \u003cem\u003eDietary Reference Intakes for Energy, Carbohydrate, Fiber, Fat, Fatty Acids, Cholesterol, Protein, and Amino Acids\u003c/em\u003e. Washington, D.C.: National Academies Press. https://doi.org/10.17226/10490.\u003c/li\u003e\n \u003cli\u003eJones, Andrew Maxwell Phineas, Diane Ragone, NuG Tavana, D. W. Bernotas, and Susan J. Murch. 2011. \u0026ldquo;Beyond the Bounty: Breadfruit (\u003cem\u003eArtocarpus Altilis\u003c/em\u003e) for Food Security and Novel Foods in the 21st Century.\u0026rdquo; https://scholarspace.manoa.hawaii.edu/handle/10125/21022.\u003c/li\u003e\n \u003cli\u003eJones, Kevin L., and R. Garry Law. 1987. \u0026ldquo;Prehistoric Population Estimates for the Tolaga Bay Vicinity, East Coast, North Island, New Zealand.\u0026rdquo; \u003cem\u003eNew Zealand Journal of Archaeology\u003c/em\u003e 9:81\u0026ndash;114.\u003c/li\u003e\n \u003cli\u003eKahn, Jennifer G. 2024. \u0026ldquo;Pig and Dog Use in the Pre‐contact Society Island Chiefdoms: Integrated Ethnohistoric, Archaeological and Use‐web Analyses.\u0026rdquo; \u003cem\u003eArchaeology in Oceania\u003c/em\u003e, April, arco.5314. https://doi.org/10.1002/arco.5314.\u003c/li\u003e\n \u003cli\u003e\u0026mdash;\u0026mdash;\u0026mdash;. In press. \u003cem\u003eFenua and Fare, Marae and Mana: The Archaeology of Ancient Tahiti and the Society Islands\u003c/em\u003e. Honolulu, HI: University of Hawaiʻi Press.\u003c/li\u003e\n \u003cli\u003eKirch, Patrick, and Sharyn Jones O\u0026rsquo;Day. 2003. \u0026ldquo;New Archaeological Insights into Food and Status: A Case Study from Pre-Contact Hawaii.\u0026rdquo; \u003cem\u003eWorld Archaeology\u003c/em\u003e 34 (3): 484\u0026ndash;97. https://doi.org/10.1080/0043824021000026468.\u003c/li\u003e\n \u003cli\u003eKirch, Patrick V. 1980. \u0026ldquo;Polynesian Prehistory: Cultural Adaptation in Island Ecosystems: Oceanic Islands Serve as Archaeological Laboratories for Studying the Complex Dialectic between Human Populations and Their Environments.\u0026rdquo; \u003cem\u003eAmerican Scientist\u003c/em\u003e 68 (1): 39\u0026ndash;48.\u003c/li\u003e\n \u003cli\u003eKirch, Patrick V., Eric Conte, Warren Sharp, and Cordelia Nickelsen. 2010. \u0026ldquo;The Onemea Site (Taravai Island, Mangareva) and the Human Colonization of Southeastern Polynesia.\u0026rdquo; \u003cem\u003eArchaeology in Oceania\u003c/em\u003e 45 (2): 66\u0026ndash;79. https://doi.org/10.1002/j.1834-4453.2010.tb00081.x.\u003c/li\u003e\n \u003cli\u003eKirch, Patrick V., Jennifer G. Kahn, and Oliver A. Chadwick. 2022. \u0026ldquo;Soils, Agriculture, and Land Use in Island Socio‐ecosystems: Three Case Studies from Southeastern Polynesia.\u0026rdquo; \u003cem\u003eGeoarchaeology\u003c/em\u003e 38 (1): 20\u0026ndash;34. https://doi.org/10.1002/gea.21934.\u003c/li\u003e\n \u003cli\u003eKirch, Patrick V., Guillaume Molle, Cordelia Nickelsen, Peter Mills, Emilie Dotte‐Sarout, Jillian Swift, Allison Wolfe, and Mark Horrocks. 2015. \u0026ldquo;Human ecodynamics in the Mangareva Islands: a stratified sequence from Nenega-Iti Rock Shelter (site AGA-3, Agakauitai Island).\u0026rdquo; \u003cem\u003eArchaeology in Oceania\u003c/em\u003e 50 (1): 23\u0026ndash;42. https://doi.org/10.1002/arco.5050.\u003c/li\u003e\n \u003cli\u003eKirch, Patrick Vinton. 1994. \u003cem\u003eThe Wet and the Dry: Irrigation and Agricultural Intensification in Polynesia\u003c/em\u003e. University of Chicago Press.\u003c/li\u003e\n \u003cli\u003e\u0026mdash;\u0026mdash;\u0026mdash;. 2000. \u0026ldquo;Pigs, Humans, and Trophic Competition on Small Oceanic Islands.\u0026rdquo; In \u003cem\u003eAustralian Archaeologist: Collected Papers in Honour of Jim Allen\u003c/em\u003e, edited by A. Anderson and T. Murray, 427\u0026ndash;39. Canberra: Australian National University, Centre for Archaeological Research and Department of Archaeology and Natural History.\u003c/li\u003e\n \u003cli\u003eKuster, C., V. C. Vuki, and L. P. Zann. 2005. \u0026ldquo;Long-Term Trends in Subsistence Fishing Patterns and Coral Reef Fisheries Yield from a Remote Fijian Island.\u0026rdquo; \u003cem\u003eFisheries Research\u003c/em\u003e 76 (2): 221\u0026ndash;28.\u003c/li\u003e\n \u003cli\u003eLee, Charlotte T., Cedric O. Puleston, and Shripad Tuljapurkar. 2009. \u0026ldquo;Population and Prehistory III: Food-Dependent Demography in Variable Environments.\u0026rdquo; \u003cem\u003eTheoretical Population Biology\u003c/em\u003e 76 (3): 179\u0026ndash;88.\u003c/li\u003e\n \u003cli\u003eLee, Charlotte T., and Shripad Tuljapurkar. 2008. \u0026ldquo;Population and Prehistory I: Food-Dependent Population Growth in Constant Environments.\u0026rdquo; \u003cem\u003eTheoretical Population Biology\u003c/em\u003e 73 (4): 473\u0026ndash;82.\u003c/li\u003e\n \u003cli\u003eLee, Charlotte T., Shripad Tuljapurkar, and Peter M. Vitousek. 2006. \u0026ldquo;Risky Business: Temporal and Spatial Variation in Preindustrial Dryland Agriculture.\u0026rdquo; \u003cem\u003eHuman Ecology\u003c/em\u003e 34 (6): 739\u0026ndash;63. https://doi.org/10.1007/s10745-006-9037-x.\u003c/li\u003e\n \u003cli\u003eLeenhardt, Pierre, R. Madi Moussa, and Ren\u0026eacute; Galzin. 2012. \u0026ldquo;Reef and Lagoon Fisheries Yields in Moorea: A Summary of Data Collected.\u0026rdquo; http://www.spc.int/DigitalLibrary/Doc/FAME/InfoBull/FishNews/137/FishNews137_27_Leenhardt.pdf.\u003c/li\u003e\n \u003cli\u003eLindeberg, Staffan, and Bengt Vessby. 1995. \u0026ldquo;Fatty Acid Composition of Cholesterol Esters and Serum Tocopherols in Melanesians Apparently Free from Cardiovascular Disease-the Kitava Study.\u0026rdquo; \u003cem\u003eNutr Metab Cardiovasc Dis\u003c/em\u003e, 45.\u003c/li\u003e\n \u003cli\u003eLockyer, Christina. 1991. \u0026ldquo;Body Composition of the Sperm Whale, \u003cem\u003ePhyseter Catodon\u003c/em\u003e, with Special Reference to the Possible Functions of Fat Depots.\u0026rdquo; \u003cem\u003eRit Fiskideilda\u003c/em\u003e 12 (January).\u003c/li\u003e\n \u003cli\u003eLyman, R. Lee. 1979. \u0026ldquo;Available Meat from Faunal Remains: A Consideration of Techniques.\u0026rdquo; \u003cem\u003eAmerican Antiquity\u003c/em\u003e 44 (3): 536\u0026ndash;46.\u003c/li\u003e\n \u003cli\u003eMassal, Emile, and Jacques Barrau. 1955. \u0026ldquo;Sweet Potato.\u0026rdquo; In \u003cem\u003ePacific Subsistence Crops\u003c/em\u003e, 10\u0026ndash;14. Bulletin 5. South Pacific Commission.\u003c/li\u003e\n \u003cli\u003e\u0026mdash;\u0026mdash;\u0026mdash;. 1956. \u003cem\u003eFood Plants of the South Sea Islands\u003c/em\u003e. Vol. 94. Noumea, New Caledonia: South Pacific Commission. https://library.wur.nl/WebQuery/titel/446787.\u003c/li\u003e\n \u003cli\u003eMcKee, Hugh Shaw. 1957. \u003cem\u003eSome Food Problems in the Pacific Islands\u003c/em\u003e. Technical Paper 106. Noumea, New Caledonia: South Pacific Commission. https://library.wur.nl/WebQuery/titel/401355.\u003c/li\u003e\n \u003cli\u003eMiller, Carey D., and Florence Pen. 1959. \u0026ldquo;Composition and Nutritive Value of Palolo (\u003cem\u003ePalola Siciliensis Grube\u003c/em\u003e).\u0026rdquo; \u003cem\u003ePacific Science\u003c/em\u003e 13:191\u0026ndash;94.\u003c/li\u003e\n \u003cli\u003eMurai, Mary, Carey D. Miller, and Florence Pen. 1958. \u003cem\u003eSome Tropical South Pacific Island Foods; Description, History, Use, Composition, and Nutritive Value\u003c/em\u003e. Honolulu: Univ. of Hawaii Press.\u003c/li\u003e\n \u003cli\u003eNewton, Katie, Isabelle M. C\u0026ocirc;t\u0026eacute;, Graham M. Pilling, Simon Jennings, and Nicholas K. Dulvy. 2007. \u0026ldquo;Current and Future Sustainability of Island Coral Reef Fisheries.\u0026rdquo; \u003cem\u003eCurrent Biology\u003c/em\u003e 17 (7): 655\u0026ndash;58.\u003c/li\u003e\n \u003cli\u003eO\u0026rsquo;Hara, Todd M., Paul Hoekstra, Cyd Hanns, Derek Muir, Dana Wetzel, and John Reynolds. 2004. \u0026ldquo;A Preliminary Assessment of the Nutritive Value of Select Tissues from the Bowhead Whale Based on Suggested Nutrient Daily Intakes.\u0026rdquo; \u003cem\u003eInternational Whaling Commission Scientific Committee Report SC/56 E\u003c/em\u003e 2:1\u0026ndash;19.\u003c/li\u003e\n \u003cli\u003eOliver, Douglas L. 1974. \u003cem\u003eAncient Tahitian Society\u003c/em\u003e. University of Hawaii Press. https://books.google.com/books?hl=en\u0026amp;lr=\u0026amp;id=kVrGDwAAQBAJ\u0026amp;oi=fnd\u0026amp;pg=PR2\u0026amp;dq=oliver+ancient+tahitian+society\u0026amp;ots=yL2sgRocjC\u0026amp;sig=lFqDzxvwN4-DkEZdvunxsnDbGTo.\u003c/li\u003e\n \u003cli\u003ePeters, F. E. 1957. \u003cem\u003eChemical Composition of South Pacific Foods: An Annotated Bibliography\u003c/em\u003e. Technical Paper 100. Noumea, New Caledonia: South Pacific Commission.\u003c/li\u003e\n \u003cli\u003ePuleston, Cedric O., and Thegn N. Ladefoged. 2022. \u0026ldquo;Ecology Limits Population, But Interaction with Culture Defines It: Carrying Capacity on Rapa Nui.\u0026rdquo; In \u003cem\u003eThe Prehistory of Rapa Nui (Easter Island)\u003c/em\u003e, edited by Valent\u0026iacute; Rull and Christopher Stevenson, 22:521\u0026ndash;51. Developments in Paleoenvironmental Research. Cham: Springer International Publishing. https://doi.org/10.1007/978-3-030-91127-0_20.\u003c/li\u003e\n \u003cli\u003ePuleston, Cedric O., and Shripad Tuljapurkar. 2008. \u0026ldquo;Population and Prehistory II: Space-Limited Human Populations in Constant Environments.\u0026rdquo; \u003cem\u003eTheoretical Population Biology\u003c/em\u003e 74 (2): 147\u0026ndash;60.\u003c/li\u003e\n \u003cli\u003ePuleston, Cedric, Shripad Tuljapurkar, and Bruce Winterhalder. 2014. \u0026ldquo;The Invisible Cliff: Abrupt Imposition of Malthusian Equilibrium in a Natural-Fertility, Agrarian Society.\u0026rdquo; \u003cem\u003ePLoS One\u003c/em\u003e 9 (1): e87541.\u003c/li\u003e\n \u003cli\u003ePuleston, Cedric, and Bruce Winterhalder. 2019. \u0026ldquo;Demography, Environment, and Human Behavior.\u0026rdquo; In \u003cem\u003eHandbook of Evolutionary Research in Archaeology\u003c/em\u003e, edited by Anna Marie Prentiss, 311\u0026ndash;35. New York: Springer International Publishing. https://doi.org/10.1007/978-3-030-11117-5_16.\u003c/li\u003e\n \u003cli\u003eRagone, Diane, and Catherine G. Cavaletto. 2006. \u0026ldquo;Sensory Evaluation of Fruit Quality and Nutritional Composition of 20 Breadfruit (\u003cem\u003eArtocarpus, Moraceae\u003c/em\u003e) Cultivars.\u0026rdquo; \u003cem\u003eEconomic Botany\u003c/em\u003e 60 (4): 335\u0026ndash;46. https://doi.org/10.1663/0013-0001(2006)60[335:SEOFQA]2.0.CO;2.\u003c/li\u003e\n \u003cli\u003eRoss, Harold M. 1976. \u0026ldquo;Bush Fallow Farming, Diet and Nutrition: A Melanesian Example of Successful Adaptation.\u0026rdquo; In \u003cem\u003eThe Measures of Man: Methodologies in Biological Anthropology\u003c/em\u003e, edited by E. Giles and J.S. Friedlander, 550\u0026ndash;615. Cambridge, Massachusetts: Peabody Museum Press.\u003c/li\u003e\n \u003cli\u003eRurua, Vahine. 2015. \u0026ldquo;Biodiversit\u0026eacute; et Exploitation Des Ressources Marines En Polyn\u0026eacute;sie Fran\u0026ccedil;aise Sur La Longue Dur\u0026eacute;e : \u0026Eacute;tude Comparative Des Archipels Des Marquises et Des Gambier.\u0026rdquo; These en pr\u0026eacute;paration, Polyn\u0026eacute;sie fran\u0026ccedil;aise. https://www.theses.fr/s140157.\u003c/li\u003e\n \u003cli\u003eSmith, Ian. 2011. \u0026ldquo;Meat Weight, Nutritional and Energy Yield Values for New Zealand Archaeofauna.\u0026rdquo; Department of Anthropology \u0026amp; Archaeology; University of Otago. https://ourarchive.otago.ac.nz/bitstream/handle/10523/5942/OALR_No8.pdf?sequence=1.\u003c/li\u003e\n \u003cli\u003eSpriggs, Matthew. 1984. \u0026ldquo;Taro Irrigation Techniques in the Pacific.\u0026rdquo; In \u003cem\u003eEdible Aroids\u003c/em\u003e, edited by S. Chandra, 123\u0026ndash;35. Oxford: Clarendon Press.\u003c/li\u003e\n \u003cli\u003eSpriggs, Matthew James Thomas. 1981. \u0026ldquo;Vegetable Kingdoms: Taro Irrigation and Pacific Prehistory.\u0026rdquo; Canberra: The Australian National University. https://www.researchgate.net/profile/Matthew-Spriggs-2/publication/35933028_Vegetable_kingdoms_taro_irrigation_and_Pacific_prehistory/links/55ff430108aeafc8ac8b9a21/Vegetable-kingdoms-taro-irrigation-and-Pacific-prehistory.pdf.\u003c/li\u003e\n \u003cli\u003eSpriggs, Matthew, and Patrick V. Kirch. 1992. \u0026ldquo;\u0026lsquo;Auwai, Kanawai, and Waiwai: Irrigation in Kawailoa-Uka.\u0026rdquo; In \u003cem\u003eAnahulu: The Anthropology of History in the Kingdom of Hawaii\u003c/em\u003e, edited by M. Sahlins and P.V. Kirch, 2:118\u0026ndash;64. Chicago: University of Chicago Press.\u003c/li\u003e\n \u003cli\u003eTercinier, G. 1974. \u0026ldquo;Les sols de l\u0026rsquo;\u0026icirc;le de Mangareva (Gambier): \u0026eacute;tude p\u0026eacute;dologique t\u0026eacute;moin d\u0026rsquo;une \u0026icirc;le haute de la Polyn\u0026eacute;sie Fran\u0026ccedil;aise.\u0026rdquo; In \u003cem\u003eCahiers du Pacifique\u003c/em\u003e, 18:341\u0026ndash;57. Paris: Fondation Singer-Polignac.\u003c/li\u003e\n \u003cli\u003eThomas, Frank R. 2003. \u0026ldquo;Shellfish Gathering in Kiribati,Micronesia: Nutritional,Microbiological, and Toxicological Aspects.\u0026rdquo; \u003cem\u003eEcology of Food and Nutrition\u003c/em\u003e 42 (2): 91\u0026ndash;127. https://doi.org/10.1080/03670240390202246.\u003c/li\u003e\n \u003cli\u003eTitcomb, Margaret, and Mary Kawena Pukui. 1969. \u003cem\u003eDog and Man in the Ancient Pacific, with Special Attention to Hawaii\u003c/em\u003e. Bishop Museum Special Publication 59. Honolulu, HI: Bernice P. Bishop Museum. https://cir.nii.ac.jp/crid/1130000796938777600.\u003c/li\u003e\n \u003cli\u003eWeisler, Marshall I., and Roger C. Green. 2013. \u0026ldquo;Mangareva Fishing Strategies in Regional Context: An Analysis of Fish Bones from Five Sites Excavated in 1959.\u0026rdquo; \u003cem\u003eJournal of Pacific Archaeology\u003c/em\u003e 4 (1): 73\u0026ndash;89.\u003c/li\u003e\n \u003cli\u003eWhite, Theodore E. 1953. \u0026ldquo;A Method of Calculating the Dietary Percentage of Various Food Animals Utilized by Aboriginal Peoples.\u0026rdquo; \u003cem\u003eAmerican Antiquity\u003c/em\u003e 18 (4): 396\u0026ndash;98.\u003c/li\u003e\n \u003cli\u003eWills, Ron B. H., Jessie S. K. Lim, Heather Greenfield, and Tim Bayliss‐Smith. 1983. \u0026ldquo;Nutrient Composition of Taro (\u003cem\u003eColocasia Esculenta\u003c/em\u003e) Cultivars from the Papua New Guinea Highlands.\u0026rdquo; \u003cem\u003eJournal of the Science of Food and Agriculture\u003c/em\u003e 34 (10): 1137\u0026ndash;42. https://doi.org/10.1002/jsfa.2740341015.\u003c/li\u003e\n \u003cli\u003eWinterhalder, Bruce P., and Cedric O. Puleston. 2018. \u0026ldquo;The Exchequer\u0026rsquo;s Guide to Population Ecology and Resource Exploitation in the Agrarian State.\u0026rdquo; \u003cem\u003eCliodynamics, 9 (2)\u003c/em\u003e. https://www.academia.edu/download/58096716/Winterhalder_and_Puleston_2018.pdf.\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp\u003eTables 1 to 19 are available in the Supplementary Files section\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[{"identity":"fcad4a02-94f5-4887-a0ef-4e22bd6fcaa5","identifier":"10.13039/100000001","name":"National Science Foundation","awardNumber":"CNH #1642894","order_by":0}],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"National Science Foundation","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"French Polynesia, marine productivity, subsistence agriculture, pre-contact population","lastPublishedDoi":"10.21203/rs.3.rs-4797211/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4797211/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eMost studies of the food resource potential of early Polynesian populations focus exclusively on agricultural potential, and specifically starchy staples, despite the importance of the marine world to the Polynesians. In an attempt to more accurately estimate total precontact food resource availability, we characterized the terrestrial and near-shore marine environments of four Polynesian islands: Moʻorea, Maupiti, Mangareva and Taravai. We estimate the agricultural potential of each island after a consideration of the ecological factors related to productivity. We also estimate the productivity of the near-shore marine environment as a function of surface area. Using a range of fish yield scenarios from Pacific subsistence systems we scale relative measures of faunal food biomass derived from local archaeological excavations to generate absolute biomass estimates. We convert our estimates of agricultural and fished/foraged food potential into a maximum population size based on calorie availability. Moʻorea’s fertile valleys and wetlands would potentially generate sufficient food energy, mostly from starchy staples, to support a considerable population, many times larger than the other islands. In the most likely fish yield scenario Moʻorea gets only 0.4% of its calories from marine sources, while the others range from 7-18%. These relative inputs reflect the vast superiority in productivity (in terms of calories per km\u003csup\u003e2\u003c/sup\u003e) of agricultural vs near-shore marine zones. We also find that population size on islands with smaller fringing reefs, as illustrated by Moʻorea, may have been limited by a lack of fish protein. Moʻorea’s maximum population is approximately halved when diet breadth is considered.\u003c/p\u003e","manuscriptTitle":"People of the sea, or of the soil? How the balance of marine and terrestrial resource availability informs maximum population on four Polynesian islands","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-07-29 08:43:14","doi":"10.21203/rs.3.rs-4797211/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"e09cd77c-3b3e-4100-8e4d-be682a066b8d","owner":[],"postedDate":"July 29th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":35085204,"name":"Agroecology"},{"id":35085205,"name":"Ecological Modeling"},{"id":35085206,"name":"Population Biology"},{"id":35085207,"name":"Marine and Freshwater Ecology"},{"id":35085208,"name":"Terrestrial Ecology"},{"id":35085209,"name":"Anthropology"},{"id":35085210,"name":"Archaeology"},{"id":35085211,"name":"Geographic Information Systems"}],"tags":[],"updatedAt":"2024-07-29T08:43:15+00:00","versionOfRecord":[],"versionCreatedAt":"2024-07-29 08:43:14","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-4797211","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4797211","identity":"rs-4797211","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}