The History and Culture of Food and Drink: Northern South America

William F Keegan. Cambridge World History of Food. Editor: Kenneth F Kiple & Kriemhild Conee Ornelas. Volume 2. Cambridge, UK: Cambridge University Press, 2000.


In writing the history of culinary practices, there is a tendency to emphasize the ethnic character of diets (González 1988).Yet nowhere are historical entanglements more apparent than in the international character of modern cuisine, even if explicit ethnic territories are strongly defended. Foods are often defined with apparent regard to national origin: Indian corn, Irish potatoes, Italian tomatoes, Dutch chocolate, and Hawaiian pineapple, to name but a few. However, the plants that form the basis of many European cuisines in fact originated in the Americas (Keegan 1992), and American diets were transformed in what Alfred Crosby (1986) has described as the creation of the neo-Europes.

“You call it corn, we call it maize.” Contrary to the American television commercial in which a very Navaho-looking women makes that statement, the word is actually of Taino origin. Peter Martyr was among the first Europeans to describe this plant that the native West Indians called maíz (Zea mays) (Sauer 1966: 55). Other Taino words for plants and animals have also entered the English lexicon, including cazabi and yuca (Manihot esculenta Crantz), guayaba (Psidium guajava L.), bixa (Bixa orellana L.), iguana, and manati (Trichechus manatus) (Oviedo [1526] 1959: 13-16; Taylor 1977: 20-1).

Cultigens from the circum-Caribbean lowlands have also been of significant effect (Keegan et al. 1992). Tomatoes (Lycopersicon esculentum Mill.) were first encountered in coastal Mexico, where the Spanish were also treated to a drink called chocolatl, a blend of cacao (Theobroma cacao L.), peppers (Capsicum spp.), and other spices (including Bixa orellana L.). Cacao won immediate acceptance; together, it and vanilla (Vanillaspp.), a semidomesticated lowland orchid, have become the most important flavorings in the world. In contrast, the tomatl (tomato) languished under the specter of its membership in the “deadly” nightshade family of plants. First grown as an ornamental, and only much later as food, the tomato eventually reshaped Italian cuisine.

American foods have been reinterpreted in ever-changing dietary contexts since their “discovery” by Europeans. Maize, in particular, has been manipulated for use as a building material, lubricant, automobile fuel, and universal hardening additive. The objective of this review is to introduce a well-structured methodology for examining past diets, while at the same time providing comprehensive empirical coverage of the regions first reached by the Spaniards after 1492.

We begin with the native Tainos who occupied the Greater Antilles and Bahama Islands at the time of European contact (see Figure V.D.3.1). Their culinary practices are traced through the Saladoid expansion to their mainland South American origins. The earlier “preagricultural” Caribbean peoples, called Archaic in the regional parlance (Rouse 1986, 1992), are then briefly considered. Although little is known of Archaic diets, it has been suggested that the Tainos developed from the synergy of Archaic and Saladoid cultures (Chanlatte Baik and Narganes Storde 1990).The scope is then expanded to include pre-Columbian plant migrations into the circum-Caribbean.

Taino Diet and Nutrition

The study of prehistoric diet and nutrition in the Caribbean is a relatively recent phenomenon. Moreover, the investigators who have concerned themselves with these issues have approached the subject in a variety of ways. The earliest efforts involved the enumeration of plants and animals described in the contact-period chronicles (Fewkes 1907; Rouse 1948; Sturtevant 1961, 1969; Sauer 1966). More recently, attention has been drawn to the importance of diet and nutrition (Wing and Brown 1979).The analysis of animal remains in archaeological sites has been underway for some time (Wing and Reitz 1982; Wing and Scudder 1983). The investigation of archaeological plant remains has also been initiated. Although most of these studies are not yet published, they are investigating pollen, phytoliths, wood charcoal, and carbonized structures for the purposes of identifying cultigens, firewood, and modifications of the landscape (for example, Fortuna 1978; Pearsall 1985; Piperno, Bush, and Colinvaux 1990; Newsom 1993). Stable-isotope analysis has been used to investigate prehistoric diets (Keegan and DeNiro 1988; Klinken 1991), and studies of human skeletal remains have uncovered evidence of diet-related pathologies (Budinoff 1987).

Because different investigators approach questions of diet and nutrition from different perspectives, there has been a problem with the integration of their results. The author of this chapter has argued that diverse empirical studies need to be integrated in a well-structured research methodology (Keegan 1987; Keegan and DeNiro 1988). Such a methodology must include explanations for why people ate what they did, must predict what people should have eaten, and must incorporate techniques for determining the variety of foods people consumed.

The study of West Indian diets described here uses models developed in evolutionary biology under the title “Optimal Foraging Theory” (Winterhalder and Smith 1981; Smith 1983; Keegan 1986a; Smith and Winterhalder 1992). It was hypothesized that Caribbean diets would reflect the efficient capture of the aggregate nutritional currencies—calories and protein. The theory was examined with data derived from ethnological, ethnohistorical, zooarchaeological, paleobotanical, and other techniques, which specified what foods were eaten and the relative costs of obtaining these foods (Keegan 1985). Finally, osteochemical techniques were used to evaluate what food items were consumed from the record they left in the human skeleton (Keegan and DeNiro 1988). This three-tier approach provides a robust solution to questions of prehistoric diet and nutrition (see Bell 1992).

West Indian Food Items

The first task in reconstructing prehistoric diets involves compiling a list of the plants and animals that were consumed. Ethno-historic reports described the Taino in the Greater Antilles as practicing a mixed economy of root-crop horticulture and hunting-fishing-collecting (Fewkes 1907; Rouse 1948; Sauer 1966;Wing 1969;Wing and Reitz 1982).

The Spanish chroniclers reported that the Tainos cultivated manioc (Manihot esculenta Crantz),1 sweet potatoes (Ipomoea batatas [L.] Lam.), yautía or cocoyam (Xanthosoma sagittifolium Schott), llerén (Calathea allouia [Aubl.] Lindl.), peanuts (Arachis hypogaea L.), maize (Zea mays L.), beans (Phaseolus spp.), cucurbits (Cucurbita spp.), chilli peppers (Capsicum frutescens L. or Capsicum annuum L.), and fruit trees, including mamey (Mammea americana L.), jagua or genipop (Genipa americana L.), hicaco or cocoplum (Chrysobalanus icaco L.), guanábana or soursop (Annona muricata L.), bixa or annatto (Bixa orellana L.), guava (Psidium guajava L.), and cupey (Clusea rosea) (Fewkes 1907; Rouse 1948; Sturtevant 1961, 1969; Sauer 1966; Guarch 1974; Nadal 1981). Most of these cultigens were carried into the Antilles by the Island Arawak (see Ford 1984; Schultes 1984). A variety of wild or quasi-domesticated plants were also in use, including the stems of the cycad Zamia, primrose (Oenothera sp.), purslane (Portulaca sp.), mastic-bully (Mastichodendron foetidissimum ), sapodilla (Manilkara sp.), cockspur (Celtis iguanaea), palm fruits, sea grapes (Coccoloba uvifera), pigeon plum (Coccoloba diversifolia), and panicoid grasses (Setaria spp.) (Sturtevant 1969; Fortuna 1978; Newsom 1993). The carbonized remains of maize, chilli peppers, palm fruits, and at least two unidentified tubers (probably manioc and sweet potato) are among the plant remains identified in West Indian sites (Newsom 1993).

As is the case with modern tropical horticulturalists, life revolved around garden cycles (Malinowski 1978; Johnson 1983). Garden plots (called conucos by the Tainos) of about one to two hectares per household were cleared at regular, possibly annual, intervals. Clearing involved the use of stone axes or shell tools to slash brush, fell trees, and girdle large trees so that they would drop their leaves. After clearing, the brush was left to dry and was then burned, releasing the nutrients stored in the vegetation.

Planting came next and was done with a sharpened digging stick. Manioc was planted in small mounds of loose earth, called montones. After planting, the gardens were weeded and the mature crops harvested; replantings were made on a continuous basis until the garden was abandoned. (Unlike temperate gardens that die in the winter, tropical gardens can be maintained for years.) After a few years, garden production was reduced to certain tree crops, and new gardens had to be prepared. The old garden was then left fallow until covered by at least secondary forest growth, at which time it might again have been cleared. The importance of the agricultural cycle is reflected in stone alignments, which were used to chart the summer solstice and the rising and setting of stars that were important in Native American agricultural calendars. The most notable examples are site MC-6 on Middle Caicos in the Turks and Caicos Islands and La Plaza de Chacuey in the Dominican Republic (Castellanos 1981; Aveni and Urton 1982; Alegría 1983).

Taino cultigens grew and matured at different rates. Consequently, the diet of the inhabitants changed continuously with the seasons. Nonbitter maniocs and other root crops would have been available throughout the year. Boiling was the usual method of cooking. Bitter maniocs, so-called because they contain toxic levels of cyanogenic glucosides, must be grated and squeezed before consumption (Roosevelt 1980: 129). The juice releases its toxins when exposed to air or cooked; the boiled juice, called cassirepe, is a base for pepper pots and manioc beer. The pulp is dried for use as starch (flour); it is also toasted to make farina or for use as an ingredient in tapioca. The reward for so much additional processing effort is indefinite storability (Roosevelt 1980: 129). Water is added to the starch to make pancake-like cassava bread that is baked on large, round, pottery griddles called burénes.

Manioc was the staple crop intensively cultivated on mounds at the time of European contact (Sturtevant 1969). Maize is reported to have been grown for roasting ears and, to some extent, for bread grain, but it was not a major foodstuff (Sturtevant 1961; Sauer 1966). Maize was apparently a late introduction and may have been the focus of agricultural intensification in the Greater Antilles at the time of Spanish contact (Keegan 1987; compare Lathrap 1987).

In addition to “outfield” garden plots, there were house gardens closer to and around the dwellings. House gardens contained new varieties of cultigens, herbs and spices, medicinal and narcotic plants, vegetable dyes, fruit trees, and other cultigens that required special attention or were needed frequently in small quantities (Lathrap 1977, 1987).

A meal without meat would not have been considered complete. As is typical of most islands, the West Indies have a depauperate terrestrial fauna (Keegan and Diamond 1987). Indigenous land mammals are limited to bats, rodents of the tribe Oryzomyini, spiny rat (Echimyidae), and small rodents (Geocapromys sp., Isolobodon portoricensis, Plagiodontia sp.). Opossum (Didelphis virginiana), agouti (Dasyprocta aguti L.), and armadillo (Dasypus novemcinctus) were introduced from the South American mainland (Eisenberg 1989). In addition, a type of small deer called brocket (Mazama americana) was known in Trinidad, which was connected to the mainland until the end of the Pleistocene (Wing and Reitz 1982). Other land animals include iguanas (Iguanasp., Cyclura spp.), crocodiles (Crocodylus sp.), a variety of small reptiles (e.g., Anolis sp.), land crabs (Cardisoma guanhumi, Gecarchinus sp.), numerous birds, and a land snail (Caracolus sp.) (Wing 1969; Wing and Reitz 1982; Wing and Scudder 1983; Steadman et al. 1984;Watters et al. 1984; Morgan and Woods 1986; deFrance 1988, 1989; Fandrich 1990). The Tainos kept domesticated Muscovy ducks (Cairina moschata) and dogs (Canis familiaris) (Sauer 1966; Wing and Scudder 1983). It is also likely that a variety of grubs and insects were consumed (see Johnson and Baksh 1987).

A number of these animals, such as Isolobodon (Reitz 1986), crocodile (McKinnen 1804; deFrance 1991), and iguana (Iverson 1979), suffered local extinctions after the arrival of Europeans (Olsen 1982; Morgan and Woods 1986).

The main component of prehistoric Antillean vertebrate faunal assemblages is marine fishes (Wing and Reitz 1982). In the Bahamas, marine fishes account for more than 80 percent of the estimated maximum vertebrate biomass in the sites with analyzed assemblages (Wing 1969; Wing and Scudder 1983; Keegan 1986b; deFrance 1991).

In addition to fishes, several large aquatic animals have also been identified in Antillean sites. These include marine turtles (mostly Chelonidae), porpoises (Delphinidae), West Indian monk seals (Monachus tropicalis),and manatees (Trichechus manatus) (Wing and Reitz 1982;Watters et al. 1984).These large reptiles and mammals have been emphasized in the subsistence activities of peoples throughout the circum-Caribbean (Nietschmann 1973; Davidson 1974; Campbell 1978; Wing and Reitz 1982; McKillop 1985).

The final category of subsistence remains is marine mollusks. Molluscan shell is the most abundant type of refuse in prehistoric sites in terms of both volume and mass. Despite this abundance, molluscan shell represents relatively small edible packages. The most important mollusks were queen conch (Strombus gigas),West Indian top shell (Cittarium pica), tiger lucine clams (Codakia orbicularis), chitons (Chiton spp.), and nerites (Keegan 1985; deFrance 1988).

Meats were roasted over a fire or barbecued (“barbecue” is derived from a Taino word). It has been suggested that the Lucayans may have used another traditional form of tropical-forest cooking called the “pepper pot.” Pepper pots are stews, kept simmering over a low fire, to which meats and vegetables are added to replenish the pot. The large, thick clay pots made in the Bahamas and in the Lesser Antilles were well suited to this type of food preparation (Allaire 1984).

When viewed with regard to the number of different ways in which individuals could satisfy their hunger, the West Indies are noteworthy for the surfeit of options. It is difficult to imagine that anyone ever went hungry. Yet the fact that these people were selective in their food choices provides an important challenge. It is not sufficient simply to list the items that were, or may have been, eaten; rather, the criteria upon which dietary selections were based must be identified.

Because there is too much local variability to treat all West Indian diets as equivalent (deFrance 1988), the following study focuses on the subsistence decisions of the Lucayans. The Lucayans are of Taino ancestry. They occupied the Bahama archipelago between A.D. 600 and 1500 (Keegan 1992). Lucayan diet breadth is evaluated by comparison with an economic model of the diet that would result from the cost-efficient capture of nutritional currencies.

Lucayan Diet Breadth

Lucayan diet can be described as consisting of inputs from five general sources: cultivated roots and tubers, maize, terrestrial animals, marine fishes, and marine mollusks (Sears and Sullivan 1978; Keegan 1992). These foods were obtained from the forest, the coastal strip, and the sea. To evaluate the Lucayan’s behavior as cost-efficient foragers, the capture of foods in the three habitats is reviewed first. After identifying the foods and their associated costs, their return rates are compared by using a model derived from Optimal Foraging Theory. Finally, the diets predicted with the model are compared to the isotopic signatures in the bones of 17 individuals. The isotopic signatures reflect the foods that these individuals consumed.

Lucayan Food Procurement

The main component of the forest habitat was gardens. These were probably prepared in coastal accumulations of humic enriched sandy soil and would have followed the pattern of other tropical gardens (Ruddle 1974; Johnson 1983). Gardens were stable patches where output could be modified in response to changing needs. Root-crop horticulture provides high total and marginal rates of return. Average yields of manioc in Brazil are reported as 14.2 million calories per hectare (Roosevelt 1980). Using a generous estimate of human caloric needs (2,700 calories per day), 14 adults could have been supported on one hectare of land for an entire year. The availability of calories in other foods, which have higher net return rates, would have precluded the need for complete reliance on manioc production.

Because manioc does not require fertile soils for efficient tuber production, a single plot can be cultivated for many years. Long-term land use is not practiced in the tropical forest today because total human protein requirements also are met through garden production. The characteristics of maize make long-term, cost-efficient production on a single plot of land impossible (Keegan 1986a). Furthermore, yields from long-term manioc gardens can be increased by inter-cropping other cultigens (for example, sweet potatoes, cocoyams, fruit trees). These crops do not interfere with the growth of the manioc plants; rather, they increase yields per hectare for the additional investment of planting and harvesting, they aid in preventing weed growth, and they fill areas that would otherwise go unused (Ruddle 1974; Brierley 1985; Keegan 1986a). It is likely that the initial colonists planted small gardens with a diversity of crops, with garden size set by the caloric returns from manioc in relation to the caloric needs of the group. As population increased, garden size would have increased and other cultigens would have been added to the gardens.

The principal shortcoming of manioc is its low protein content (Roosevelt 1980). Since human nutritional requirements cannot be satisfied with manioc alone, other sources of protein had to be sought. In addition to animal protein sources, such cultigens as maize, beans, or ground nuts could have been added to the garden as protein supplements. These cultigens are more expensive in terms of harvesting and processing, contribute to a more rapid exhaustion of garden soils, and are available only during specific seasons.

Maize is the most cost-efficient of the high-protein cultigens. From Allen Johnson’s (1983) studies it was estimated that maize has a marginal return rate of 20 grams of protein per hour of labor (Keegan 1986a). This return rate is higher than that for many other Lucayan foods. This would have promoted its acceptance when it became available after about A.D. 1100 (Keegan 1987). The disadvantages of maize production are the plant’s need for fertile soils, its availability during only one season of the year, and the high cost of storing and processing maize when it is grown for use throughout the year. Thus, maize constituted only a partial solution to the problem of protein production.

Marginal return rates for animal species that inhabit the forest favor their use over horticulture. These animals (that is, hutia, iguana, and land crabs) could not, however, satisfy total needs. Their combined densities are equivalent to 1,861,950 calories per hectare, which could support only two individuals for a year if every animal was captured. But (in economic terms) intensive use would have rapidly reduced the frequency of encounters (that is, animals would not be seen very often), with the result a decline in the average return rate to a level below the marginal rate for other food sources (including horticulture).When this decline reached that level, those other foods would be added to the diet (Winterhalder and Smith 1981; Smith 1983). For the Lucayans, this other source of protein was the sea.

The inference from the economic model is that terrestrial animals were pursued whenever they were encountered. It is probable that game was taken in or near gardens (Linares 1976) and during visits to the coast. The forest is a difficult patch to traverse, and hunting trips in the forest were probably infrequent, especially after the decline of initially high prey densities. Since all of these small game were regular visitors to the coastal strip, the most efficient strategy would be to forage in this area. Travel along the coast is also less difficult, and other food sources would doubtless have been encountered (for example, wild plants, littoral mollusks).

The one exception to the forest-hunting proscription is land crabs, which congregate in the low areas that provide for moist burrows. These locations could be identified, and hunting could be accomplished with guaranteed results. But even land crabs can be taken on the coastal strip, especially when they congregate for mating (Gifford 1962; deFrance 1988).

The coastal strip is comprised of the beach patch and the rocky intertidal patch, and it provides access to marine habitats. The beach patch was the site of seasonal monk seal aggregations, turtle nesting grounds, and the accidental beaching of whales. In addition, terrestrial animals frequent this patch. Because settlements were located on the coast, the Lucayans were in a position to monitor activities on the beach.

The highest average return rate was available in the beach patch, although the highest return species were not available in all locations at all times (for instance, green turtles, monk seals, and whales). Turtles would have been available from April to July, and monk seals for about six weeks centered on December. Both would have been the focus of procurement efforts during their seasonal abundances.

The rocky intertidal patch supports dense aggregations of West Indian top shells, nerites, and chitons. Top shells have a relatively high net return rate, and of the three they are the only ones used today with any frequency. However, the use of top shells is limited by two factors. Rocky intertidal zones are small, averaging less than three meters wide, and they are irregularly distributed. These snails also are easily exploited and would have rapidly disappeared following the start of human predation. The other common littoral mollusks have lower return rates. These low values suggest that they were exploited during periods of food shortage. In any case, chitons and nerites should have been among the last items added to the diet.

The marine environment is comprised of a tidal-flat patch and a reef patch. The tidal-flat patch can be further divided by procurement strategies into infaunal mollusk collecting (that is, in the mud-sand substrate), epifaunal mollusk collecting (that is, on the substrate), and fishing. These strategies are discussed in turn.

The tiger lucine clam, Codakia orbicularis, is the most commonly used infaunal mollusk species. They occur at high densities beneath shallow grass flats and would have provided a stable resource supply (Jackson 1972, 1973). In terms of weight, tiger lucine clams provided a more significant source of food than intertidal mollusks (Rose 1982). Yet, tiger lucine has low net return rates of both calories and protein. It is, however, possible that these rates have been underestimated in the present study. A study of pelecypod collecting in Australia indicated a caloric return rate that is twice that estimated for Codakia orbicularis (Meehan 1977). Historical evidence suggests that even if the return rates were underestimated, the ranking relative to other food items is accurate. Tiger lucine clams should therefore have also been one of the last items added to the diet.

The epifaunal gastropod Strombus gigas is the highest ranked marine resource. It is available at high densities on shallow grass flats (Doran 1958; Hesse and Hesse 1977) and was a significant component of the diet. The high return rates place Strombus gigas in the initial optimal diet. Foraging trips over the shallow grass flats would also have led to encounters with marine fishes, such as bonefish (Albula vulpes).

The net return rate for fishes has been calculated as an average for all fishing strategies. Higher returns could have been obtained by pursuing particular strategies, such as the capture of fishes encountered during Strombus gigas collecting trips, but present evidence is not sufficient to discriminate the return rates for alternative strategies (Johannes 1981; Kirch 1982; Keegan 1986a).The average values for all fishing strategies places fishing just below rock iguana in the ranking of protein returns.

Optimal Horticulturalists

Having discussed the major habitats and their patches, attention is next focused on the food quest. In which patch should food be sought first? How much time should be devoted to the food quest in each patch? When should new foods be added to the diet? More complete information on time allocation and predation rates are needed before a quantitative solution for patch selection decisions can be calculated.

A qualitative solution can be proposed in the Lucayan case because the net return rates for the second through eighth ranked resources are sufficiently similar to analyze dietary change in a diet-breadth framework.This approach is based on the identification of horticultural production as part of the original subsistence endowment, and by the distribution of higher ranked resources in forest, coastal, and marine patches. Furthermore, the location of permanent settlements on the coast would eliminate significant differences in the time invested in traveling between patches.

The logic behind the use of the diet-breadth model, rather than the patch-selection model, is as follows. Although the patch-selection model compares average returns from different patches (including transportation time), while the diet-breadth model compares marginal returns to the overall foraging efficiency, the marginal return rates calculated for Lucayan food sources include some time investments that are better considered as search time. (Search time is a component of average return calculations.) This conflation of average and marginal returns results from the character of the information available to calculate those rates.

A second factor is that foraging from a central place should require similar investments in the time required to travel between patches. The high-ranked items in the Lucayan diet all have population densities that would have been rapidly depleted after the start of human predation. This reduction of animal densities would produce a reduction in the average return rates. Since the marginal rates for high-ranked items are similar, as is the time invested in traveling between patches, the marginal return rates should approximate the long-term decision-making problem. In other words, foraging decisions can be modeled as reflecting habitat selection based solely on short-term differences in resource distributions in each of the patches. This type of patch selection is so “fine-grained” (homogeneous) that it closely resembles and even operates like the diet-breadth model.

The diet-breadth model predicts that diet breadth will be expanded (in other words, items will be added to the diet) when the marginal return rate for a resource is equal to the average return rate for all higher-ranked resources. Because manioc cultivation was practiced when the Bahama archipelago was colonized, the higher-ranked resources would also have been in the original optimal set. These high-ranked resources are hutia, land crabs, queen conch, and rock iguana. This analysis suggests that despite their current absence in archaeological samples, green turtles and monk seals would have been captured during their seasonal availabilities. The food items mentioned should have provided a diet sufficient to preclude the need to eat any other foods. One qualification to this is that high-ranked fishes, such as bonefish, were probably pursued when encountered during foraging on the tidal flats.

Terrestrial animals in the Bahamas are susceptible to overexploitation. Their availability would have rapidly declined after a short period of intense predation and as human population growth increased the demand for these foods (Iverson 1979; Jordan 1989). In addition, land clearance and other changes to the landscape would effect the survival of other species (Olsen 1982).The first response to a decline in high-ranked resources would be to migrate to previously unexploited habitats. Thus, the rapid decline in high-ranked terrestrial animals would have encouraged the rapid migration of people to unoccupied islands. When new areas were no longer available, then the intensification of foraging in the marine habitat should have occurred.

The redistribution of the human population was no longer a cost-efficient option when the presence of other settlements prevented people from moving into pristine areas. When such social circumscription occurred, the currency demands had to be satisfied in the vicinity of the village over a longer period of time. In response to the combined effects of increased demand due to population growth and declining returns due to long-term exploitation, the intensification of production could only focus on two options. These options were increased use of marine habitats and changes in garden breadth. The total contribution of land animals would have continued to decline as human population numbers increased. In marine habitats, sea turtles, Strombus gigas, and certain species of fish would have been the initial focus, with other fishes added to the diet as other fishing techniques were introduced, and ending with the highest-ranked littoral mollusk, Cittarium pica, the West Indian top shells. But this species would also have been rapidly exhausted in areas of human settlement.

During the final phase, the lowest-ranked foods (that is, nerites, chitons, and tiger lucine) would have been added in turn. Strombus gigas would have been sought at more distant locations, and a variety of more intensive fishing strategies would have been introduced. It is likely that fish traps were introduced early in this phase as the availability of such visible diurnal species as bonefish declined and as fishes had to be sought at more distant locations, such as along the barrier coral reef. Horticultural production would have been intensified with the addition of beans, groundnuts, and maize, all of which are high carbohydrate- and protein-producing cultigens. Fields might have been fallowed for shorter lengths of time, although the sandy soils may have supported production for longer periods than was possible on other tropical soils (as in the Yucatan Peninsula, Mexico; Roosevelt 1980). Nitrogen-fixing legumes could have helped to maintain soil fertility. In the final phase, agriculture was intensified by the terracing of hillsides, the irrigating of fields, and the building of raised and drained fields (Zucchi and Denevan 1979).

The preceding discussion has served to identify three discrete diets. An initial diet was composed of root crops, land animals, and a few high-ranked marine species. This was followed by a second diet that included the consumption of more marine foods, a continuing contribution from root crops, and a precipitous decline in the contribution from land animals. In the third diet, land animals were reduced to a very minor level of use, marine production was further intensified, and horticulture was expanded to include higher-cost cultigens, such as maize and beans, which increased total protein and carbohydrate returns from the garden. It is this final pattern that is evident in the Lucayan faunal samples that have been analyzed to date (Wing and Reitz 1982; Wing and Scudder 1983; deFrance 1991).

Stable Isotope Analysis

The three diets just described were proposed on the basis of data gathered from ethnographic analogy, ethnohistoric reports, ethnobiological analyses, and formal economic models. To determine how well these diets reflect the actual subsistence practices of the Lucayans, a different analytical technique must be used to avoid circular reasoning. Stable-isotope analysis provides such a technique (Schoeninger and DeNiro 1984; Sealey 1986; Ambrose 1987, 1993; Keegan 1989c; Schoeninger et al. 1989; Sillen et al. 1989; see Stokes 1995 and de France et al. 1996 for new perspectives on this question.)

The Lucayan diet was evaluated by first measuring or estimating the carbon- and nitrogen-isotope compositions of food items and then measuring these in bone collagen extracted from 17 Lucayan skeletons. The results were then compared in order to determine the most likely components of the Lucayan diet.

An immediate division into marine and terrestrial food groups was observed in the carbon-isotope ratios. The marine group has an average delta13C of -11 per mil, whereas the terrestrial group has an average delta13C value of -25 per mil. Using these averages, the diet delta13C values calculated for the Lucayan individuals can be interpreted. For instance, a diet delta13C value of -11 per mil would suggest complete reliance on marine foods, a diet delta13C of -25 per mil would suggest complete reliance on terrestrial foods, and a diet delta13C value of -18 per mil would suggest equal contributions from both. Based on an estimated ±1 per mil uncertainty in the fractionation factor between diet and bone collagen, the diets of Lucayan individuals range from an estimated 71 ± 7 percent reliance on terrestrial foods (-21 per mil) to an estimated maximum of 74 ± 7 percent reliance on marine foods (-14.6 per mil).

If this interpretation is correct, the range in delta13C values estimated for the diets of Lucayan individuals can be interpreted as reflecting a shift in consumption practices through time. The three most negative delta13C values (all around -20 per mil) match the first of the proposed diets in which land animals were abundant and only the highest-ranked marine organisms were consumed. The second dietary pattern would account for the majority of Lucayan individuals (n = 11), whose diet delta13C values of -18 ± 1 per mil reflects almost equal contributions from marine and terrestrial sources. The remaining three individuals exhibit ideal consumption patterns in the 66 to 74 ± 7 percent marine range (diet delta13C values of -15.2 ±.6 per mil). Such a strong reliance on marine foods is unlikely due to the relatively higher costs of marine fishing and collecting in relation to horticulture. An alternative interpretation is that the higher delta13C values indicate that maize was being consumed in substantial quantities during at least part of the year.

The carbon isotopes confirm the presence of three distinct diets. One explanation for these differences is that they represent changes in diet breadth through time. Unfortunately, the skeletons came from burials in caves and were not associated with materials that could be used to date them (Keegan 1982; Keegan and DeNiro 1988). Other explanations, such as the influence of more localized factors, must be entertained. Further investigations are needed to relate the isotopic study to expectations based on empirical studies and theoretical projections.

A second part of the isotopic study examined nitrogen-isotope distributions. Nitrogen-isotope values for marine and terrestrial food sources overlap, and so they cannot be used to distinguish among the contributions of land animals, cultivated roots and tubers, maize, and reef fishes in the Lucayan diet (Keegan and DeNiro 1988). When the isotopic signatures of Lucayan individuals are compared to this average, 76 percent of the individuals (n = 13) fall within ±1 per mil of this range.

Deviations from that average diet can be explained with reference to the lower values of mollusks and the higher values of pelagic fishes and marine mammals relative to other dietary components. One individual from Crooked Island had a higher dietary value, which suggests that pelagic fishes and/or marine mammals comprised a larger component of his diet and that mollusks made a relatively minor contribution. The location of this burial in an area where the barrier coral reef approaches the shore and the reef flat has a restricted range places this individual in the vicinity of marine habitats that are the most likely sources of such a dietary combination (Keegan 1982, 1986b).

The three individuals’ diet values that are lower than the average probably reflect a stronger reliance on mollusks and perhaps a corresponding reduction in the consumption of higher-order carnivorous fishes. All three of these individuals are from islands with extensive Thalassia seagrass meadows (Grand Bahama Island and Great Abaco on the Little Bahama Bank, and Providenciales on the Caicos Bank). These shallow banks provide access to extensive mollusk populations while restricting access to reef-associated fishes (Wing and Reitz 1982;Wing and Scudder 1983; Keegan 1986b). Two of the individuals are less than

0.6 per mil below the average range, which reflects an increase in mollusk consumption to a level of less than 10 percent of total consumption. The other is 2.4 per mil below the average range, which suggests a reliance on mollusks and possibly other marine invertebrates approaching 40 percent of the diet.

The carbon- and nitrogen-isotope ratios are significant in their unequivocal rejection of commonsense interpretations of archaeological deposits. These deposits are largely composed of marine mollusk shells, the size and durability of which have led some investigators to propose that mollusks comprised as much as 95 percent of the meat protein in Lucayan diets. It is clear in the isotopic signatures of the 17 Lucayans that mollusks played a much more modest role in the diet. The same is true for the slave populations that followed them (Armstrong 1983).

The physical examination of the skeletons from the Bahamas indicates that the Lucayans enjoyed good health and nutrition (Keegan 1982). They certainly did not suffer from the nutritional and diet-related disorders of other prehistoric horticulturalists (Cohen and Armelagos 1984).They also lack the dental pathologies observed in Saladoid and Ostionoid burials at the Maisabel site, Puerto Rico. Linda Budinoff (1987), who analyzed the Maisabel skeletons, concluded that sand adhering to foods, exoskeleton and shell adhering to invertebrates, and a high percentage of carbohydrates in the diet conspired to destroy the teeth. One result was that middle-aged people had mouths more typical of old people.

In sum, the preceding examination of Lucayan diet has drawn together evidence from a variety of sources. On a general level, the empirical findings are consistent with the expectations derived from the formal model. There is every reason to believe that the Lucayans, and by extension other Tainos, were efficient, even optimal, horticulturalists.

Culinary Origins of the Tainos

The origins of the Tainos are conveniently traced to the banks of the Orinoco River in Venezuela (Rouse 1989a).As early as 2100 B.C., villages of horticulturalists who used pottery vessels to cook their food had been established along the Middle Orinoco. During the ensuing two millennia, their population increased in numbers, and they expanded downriver and outward along the Orinoco’s tributaries (Lathrap 1977, 1987; Roosevelt 1980; Sanoja Obediente and Vargas 1983). One path of expansion led these people to the coast of the Guianas (Rouse 1992). From the Guianas, the opening of the West Indies awaited only the discovery of Grenada, which is separated from Trinidad by the widest gap in the chain of islands leading to Puerto Rico.

The movement of these people down the Orinoco River and through the Lesser Antilles to Puerto Rico is well documented (Roosevelt 1980; Sanoja Obediente and Vargas 1983; Zucchi, Tarble, and Vaz 1984; Rouse 1986). It is easily traced because these people manufactured a characteristic type of pottery known as Saladoid after the archaeological site of Saladero, Venezuela, at which it was first described (Rouse and Allaire 1978). In particular, the use of white-on-red painted decorations has facilitated the identification of the path along which this population expanded. Another indicator of migration is the presence of certain animals in the island archaeological remains.

The Saladoid expansion into the Antilles occurred at a rapid pace. The earliest Ceramic-Age settlements in the West Indies date to about 400 B.C. (Rouse 1989b; Haviser 1991; Siegel 1991). Saladoid settlements appear simultaneously on Puerto Rico and the islands of the Lesser Antilles (Rouse 1992). Given what is known of human reproductive potentials and the time that elapsed between departure from the mainland and colonization of Puerto Rico, the inescapable conclusion is that only the very best locations on a few of the Lesser Antilles were settled at this time. The most likely scenario is that most islands were settled temporarily and were then abandoned in favor of more abundant food resources on other islands. Only those few locations with superior resource concentrations were settled for a period that could be considered permanent (Watters 1982; Keegan and Diamond 1987; Haviser 1991). This practice of establishing temporary settlements that were moved in response to resource availability is typical of extensive horticulturalists, which the Saladoid peoples are believed to have been (Conklin 1968; Ruddle 1974; Johnson 1983).

Since paleobotanical studies have only recently been attempted, the composition of the Saladoids peoples’ gardens has been hypothesized from ethno-historic and ethnological descriptions of cultivations (Sturtevant 1961; Roosevelt 1980), and from the presence of certain food-processing artifacts. It has been proposed that manioc was the staple, but that a variety of other cultigens were also grown. The hypothesized importance of root crops is consistent with the results obtained from stable isotope analysis.

The earliest known villages in the Lesser Antilles follow the riverine settlement pattern of the mainland. On Grenada, Antigua, St. Martin, Vieques, St. Croix, and St. Kitts, the villages were located inland on river terraces, which provided access to the best setting for gardens (Haviser 1991; compare Siegel 1991). The shifting, extensive character of gardening practices is evident from the absence of deeply stratified sites and from settlement patterns in which different components are arranged in horizontal and sometimes overlapping relationships (Watters 1982).

Shortly after the initial colonization of the Antilles, there was a rapid and almost complete shift from inland to coastal settlement locations. Although horticulture continued as the primary source of foodstuffs, the change in settlement patterns was accompanied by a shift in midden deposits from terrestrial to marine animal remains. At the earliest sites, the remains of land crabs predominate in archaeological deposits. However, following the shift to coastal villages, the shells of marine mollusks and bones of marine fishes are the main components of archaeo-logical deposits (Carbone 1980a; Goodwin 1980; Jones 1985; deFrance 1988, 1989; Keegan 1989b; Fandrich 1990). Stable isotope analysis has confirmed this strong initial reliance on terrestrial foods. One individual from the Hacienda Grande site (Rouse and Alegría 1990), an initial-period Saladoid settlement on Puerto Rico (around A.D. 100), has been analyzed. This individual exhibits a 93 ± 7 percent reliance on terrestrial foods (Keegan and DeNiro 1988).

Two explanations have been suggested to account for this shift from an emphasis on land resources to one on marine resources. The first proposed that a growing human population soon depleted the availability of terrestrial resources, which resulted in the shift to marine resources (Goodwin 1979). The second explanation proposed that changes in climate resulted in drier conditions, which acted to reduce population densities of the humidity-sensitive land crabs (Carbone 1980a, 1980b). Both are possible, but the first seems to have had the larger effect.

The examination of subsistence change on the mainland is instructive. In her study of Saladoid peoples on the Orinoco River of Venezuela, Anna Roosevelt (1980: 230-3) projected a population growth rate in Parmana that is exactly the same as the rate estimated for St. Kitts (Goodwin 1979). This coincidence suggested the question: As a shift in protein sources occurred when the population density doubled from about 1.5 to 3.0 persons per square kilometer in Parmana, did anything similar happen on St. Kitts? In Parmana, there was a shift in protein sources to an emphasis on maize in the diet (Roosevelt 1980; Merwe, Roosevelt and Vogel 1981); on St. Kitts, it has been estimated that the shift from land crabs to marine mollusks occurred at an equivalent population doubling point (Keegan 1989b).

In sum, Saladoid peoples expanded from northeastern Venezuela and the coast of the Guianas through the Lesser Antilles and Puerto Rico to establish a frontier in eastern Hispaniola. The initial migration through the Lesser Antilles to Puerto Rico took place in less than one century (Keegan 1995), a period that was insufficient for the establishment of permanent communities on every island in the Lesser Antilles. One stimulus to this rapid expansion was the small size of these islands and their limited terrestrial resource bases (Harris 1965). These constraints are apparent in the rapid and almost complete shift from terrestrial to marine sources of animal protein at the same time that the shift to coastal settlement locations provided ready access to the marine environment. The completion of this transformation of a riverine people to an island people was the economic foundation on which Taino societies developed.

Archaic West Indians

Archaeological investigations have documented the presence of human groups in the Greater Antilles by 4000 B.C. (Rouse and Allaire 1978; Veloz Maggiolo 1971-72). By most accounts, these groups were aceramic, preagricultural hunter-fisher-gatherers whose lifeways emphasized the gathering of wild plants and the exploitation of marine resources (Rouse 1948; Veloz Maggiolo and Vega 1982; Sanoja Obediente and Vargas 1983; Keegan 1994). G. J. van Klinken’s (1991) stable-isotope study of 24 individuals from Aruba, Bonaire, and Curaçao, in the Netherlands Antilles, indicated a diet of C4 grasses and marine animals from seagrass and coral-reef habitats. Dave Davis (1988: 181) has pointed out that the assumed lack of cultigens is based on the weakest of evidence. Lee Ann Newsom’s (1993) study of plant remains from archaeological sites has led her to propose that “preagricultural” West Indians (Archaic peoples) were managing, and possibly cultivating, a suite of indigenous seed-bearing plants. Such incipient cultivation is now well documented among foragers (Ford 1985; Vaquer et al. 1986).

In reconstructing Archaic subsistence, it is usually assumed that a population of hunter-gatherers called Guanahatabey (or erroneously, Ciboney) survived in western Cuba until European contact (Rouse 1948). The author of this chapter has recently shown that there is not sufficient evidence to support that belief (Keegan 1989a; compare Rouse 1992).Therefore, ethnohistoric reports of Guanahatabey subsistence cannot be used to illuminate Archaic diets.

Paleobotanical evidence for the Archaic is limited. Only the starchy stem of the wild cycad Zamia and a fruit known as cupey (Clusea rosea) have been recovered in excavations (Nadal 1981; Veloz Maggiolo and Vega 1982). Both of these plants were cultivated by the Tainos, and it is possible that other Taino cultigens (for example, soursop, sweetsop, hogplum, guava, pineapple) were also eaten (Davis 1988: 181).The most common fauna in Archaic sites are mollusks and reef fishes from shallow offshore habitats.

Marcio Veloz Maggiolo and Bernardo Vega (1982) propose an initial “adaptation” model of archaic subsistence in the Caribbean commencing around 9000 B.C. on the western coast of the United States and extending through Panama around to Trinidad. Trinidad was connected to the mainland until about 6000 B.C., with the Banwari-type occupation dating between 5500 to 3500 B.C. The diet was based on gathered plants and mollusks, with an emphasis on mangrove swamps, along with the hunting of small-to-middle-size game. Archaeological sites in the Dominican Republic date from between 2000 and 1500 B.C. Zamia and cupey were already present by this time, and land snails (Polydontes sp., Caracolus sp.), oyster (Crassostrea rizhoporae), intertidal gastropods (Cittarium pica), land crabs (Geocarcinus lateralis), marine cockroach (Acanthopleurasp.), parrot fishes (Scaridae), iguana (Cyclura sp.), and medium-size rodents (Heteropsomys sp., Isolobodonsp., and Nesophontes sp.) were also eaten.

The second adaptation, called Barrera-Mordan, dates to around 3000 B.C. in Cuba, Puerto Rico, and the Dominican Republic, and shows affinities with sites in the Colombian lowlands and Venezuela. This occupation was characterized by a more specialized emphasis on mollusks from sandy beaches (for instance, Arca spp.,Codakia orbicularis), and the disappearance of mortars and pestles, suggesting a decline in the use of seeds, roots, and berries.

The third adaptation was based on fishing and mollusk gathering from beaches and mangrove swamps. It is distinguished by the high frequency of shell tools and occurs in Sambaqui-type shell mounds in northern Venezuela, including the offshore islands of Cubagua and Manicuare, Trinidad, Cuba, and the Dominican Republic (Sanoja Obediente and Vargas 1983). These sites date to around 4000 B.C. in Venezuela and 2000 to 1200 B.C. in Cuba and the Dominican Republic.

These relationships among the peoples whose material remains have been used to characterize these adaptations await further specification. It is interesting to note that in contrast to the horticultural Saladoid peoples, whose earliest adaptation seems to have emphasized terrestrial resources, the Archaic peoples had a more pronounced emphasis on the marine environment. Davis (1988) has commented on the absence of land crabs in Archaic sites, given the importance of land crabs in early Saladoid sites. It is likely that the availability of an efficient calorie producer (manioc) allowed the early Saladoids to pursue the higher-ranked but less abundant animals of the interior, whereas the Archaic peoples had to emphasize calorie capture in coastal habitats. The shift to marine resources by Saladoid peoples shows a convergence in diet with regard to protein capture that would reflect the decline in terrestrial animals.

When Old Met New

The arrival of the Spanish brought dramatic changes to the circum-Caribbean (Keegan 1996). The most notable change was the rapid decline in the native population due to warfare, disease, and abuse. Prior to their demise, the native peoples were the major suppliers of food for the European colonists. By 1497 the five major cacicazgos (regions ruled by a paramount cacique or chief) were providing tribute in the form of food and labor, a practice that mirrored tribute made to Taino caciques (Moscoso 1986). Manuel García-Arevalo (1990: 272) has noted that “dietary patterns and foodways were among the most important of the Taino contributions to be integrated into colonial culture” and that “cassava bread came to be known as the ‘bread of the conquest.’“

The Spanish also introduced a variety of animals into these depauperate islands. Cattle (Bos taurus), which were brought from the Canary Islands multiplied in unprecedented fashion on Hispaniola (Reitz 1986: 319, 1988), where cattle ranching became a significant occupation (Ewen 1990). In addition, pigs (Sus scrofa),goats and sheep, horses (Equus sp.), chickens (Gallus gallus), and rats and mice (Rattus norvegicus, R. rattus, Mus musculus) were introduced, the latter as unintended stowaways (Crosby 1972, 1986; Reitz 1986). It is interesting to note that the Spanish presence in early contact-period sites is more often identified by the occurrence of pig and rat bones than by European objects (Deagan 1988). Cattle, pigs, goats, and chickens adjusted to the climate. Pigs were so well adjusted that they were released on islands to form feral herds that could be hunted when needed. As Charles Ewen (1991: 108-9) expressed it,”[T]he diet of the colonists [showed] a mixture of the Iberian barnyard complex of peninsular Spain and the mixed hunting-fishing strategies of the indigenous peoples.”

In colonizing the circum-Caribbean, the initial Spanish objective was to recreate their Iberian home-land (Crosby 1972, 1986).The colonists who founded St. Augustine in Florida brought seed stock for planting wheat and other cereals, cuttings for vineyards, and animals for breeding stock. They soon learned that the climate and soils were unsuited to this economy. Cereals withered on the stock, olives and grapes failed to grow, and their preferred meat source, sheep, did not thrive (Scarry and Reitz 1990: 344). In addition to raising cattle, the Spaniards introduced sugar-cane as a cash crop. By 1545, there were 29 sugar mills belonging to prominent people on Hispaniola (García-Arevalo 1990: 275). Sugar was the magnet for the next invasion (Mintz 1985).

Subsistence practices in the sixteenth-century Spanish colonies mirrored those of the native economies that preceded them. The major change was the use of domesticated animals as a meat source, comprising from 20 to 50 percent of the vertebrate taxa (Reitz 1986: 319). In addition, peaches, melons, and watermelons produced well (Scarry and Reitz 1990:350).Lastly, the Spanish continued to import Old World foodstuffs that would not grow in the colonies, including wheat, olives, and wine, but these supplies were rare and unreliable (Sauer 1966).According to Margaret Scarry and Elizabeth Reitz (1990: 344-5): “One soldier testified [in 1573] that rations were often short and that ‘when there was nothing they ate herbs, fish and other scum and vermin.’“

British and African Arrivals

The British colonies pursued two very different types of plantation economies. In Jamaica and the Lesser Antilles, sugarcane was grown by slave labor for export, whereas in the Bahamas, cotton, utilizing far fewer slaves, was the main commodity (Craton 1978; Riley 1983; Saunders 1985; Craton and Saunders 1992). In both areas the vegetable portion of the meal was grown largely on provision grounds, although wheat flour and locally unavailable foodstuffs were imported. Indian corn (Zea mays) was the staple. In addition, true yams (Dioscorea sp.), sweet potatoes, eddoes (Colocasia sp.), okras (Hibiscus esculentus), pigeon peas (Cajanus cajan), red peas or cowpeas, black-eyed peas, snap beans, cabbage, pumpkins, castor oil, and Guinea corn (sorghum) were also grown (Farquharson 1831-2; Handler and Lange 1978). The new additions to this list were imported from Africa and Asia in the course of the slave trade. Kimber (1988) provides a remarkable discussion of plant introductions and use on the island of Martinique.

Some livestock and fowl were raised on the plantations, but most of the meats that were consumed were imported. Imports included salt beef and salt pork. It is ironic that salt fish, imported from North America, was a staple. In the study of faunal remains at Drax Hall, Jamaica, Douglas Armstrong (1983) detected little in the way of fresh fish, mollusks, or other sorts of wild game. However, following emancipation, there was a strong reliance on rocky intertidal mollusks.

Although foodstuffs were usually distributed to slaves for cooking, a main meal was sometimes served from a central kitchen on Barbadian plantations in the nineteenth century (Handler and Lange 1978). Slaves prepared their meals on an open fire by roasting or boiling. Jerome Handler and F. W. Lange suggest that meals were nutritious and monotonous, and that food allowances were often of insufficient quantity. As a result, the theft of food was common, and it was not considered wrong to steal from the master.

Slaves were also given allowances of rum, one effect of which was a condition called “dry bellyache.” Handler and colleagues (1986) have shown that dry bellyache was caused by lead poisoning. Lead was the “demon” in rum; it was used in the equipment in which rum was distilled.

D. Gail Saunders (1985: 166) and others have noted that African foods such as accara, foo-foo, agedi, and my-my are remembered and still cooked in the Bahamas today. She describes accara as being a patty made with black-eyed peas, okras, onion, red peppers, flour, thyme leaves, tomato, and salt. With emancipation, what was remembered of African foodways was mapped onto the local environment.

Garifuna Diet

All of the preceding examples find their ultimate expression in the Garifuna. The Garifuna, often referred to as Black Caribs, presently occupy the coast of Central America from Belize to Nicaragua. They developed from intermixture of the Island Caribs and Africans in a social environment that was manipulated by white Europeans. By 1700, a new society that was racially and culturally distinct from the Island Caribs had developed on St. Vincent (Kirby and Martin 1972). What is striking about the Garifuna is that their diet reflects “the various exotic cultural influences experienced over the past 400 years” (González 1988: 98).

The best-known item in the Island Carib diet is human flesh. It was in reference to the Island Caribs that the term “cannibal” originated. More recently, investigators have questioned whether the Island Caribs did, in fact, consume human flesh or whether this practice was ascribed to them in order to permit their capture as slaves under Spanish law (Myers 1984; Davis and Goodwin 1990; Wilson 1990). The present consensus is that anthropophagy was practiced in ritual settings, perhaps even as a display of fierceness, but that it never served as a source of dietary protein as has been suggested for the Aztecs (Harner 1977; compare Garn 1979).

It is likely that many of the cultigens attributed to the Tainos were also cultivated by the Island Caribs. The staples of their diet were cassava bread made from bitter manioc, and fish. They also cultivated sweet potatoes, chilli peppers, peanuts, beans, guava, soursop, and mamey. Pineapple (Ananas comosus [L.] Merr.) is described as having been fermented into wine (Rouse 1948; Sturtevant 1969), an activity still practiced by the Garifuna.

Oranges, citrons, grapefruits, figs, rice, bananas, and plantains were introduced from the Mediterranean by the Spanish (González 1988: 101). Cacao was introduced from the mainland because of its importance as a European trade item, an importance it maintains today. The introduction of coconut, the most important source of oil for the Caribs, is not recorded. The Spanish also introduced fowl and pigs. Feral pig herds achieved substantial numbers following their release. Okra, akee (Blighia sapida Konig), pigeon peas, marijuana, senna (Cassia italica), yams (Dioscorea sp.), sorghum, and plantains accompanied slaves from Africa (Sturtevant 1961; Grimé 1979; González 1988: 101). Finally, mangoes, sugarcane, coffee, and arrow-root (Maranta arundinacea) were introduced at an early date, the latter by the Tainos (Sturtevant 1969; Handler 1971; González 1988: 101; Newsom 1993). Nancie González’s (1988) summary of modern ceremonial foods by probable date of introduction neatly summarizes the successive waves of culinary influence.

Through the eyes of the Garifuna, Nancie González (1988) describes the progressive homogenization of tropical diets among peoples of African ancestry. The pepper pot of the Island Caribs has today been replaced by falmou, a concoction of fish, tubers, and coconut milk. Both Capsicum (chilli peppers) and cassirepe (boiled juice of bitter manioc) have fallen out of favor. Fish stews similar to falmou are also popular with the Miskitos and other Belizean creoles (Nietschmann 1973: 37). Bitter manioc continues as an important source of food, and it joins maize, rice, and wheat as ubiquitous starches, along with mango, papaya, and watermelon as ubiquitous fruits on both sides of the Atlantic (González 1988: 105-6). Yet despite this convergence in food items, the menus maintain their separate ethnic dimensions.


Caribbean diets were strongly influenced by the diffusion of domesticated plants and animals. The initial conditions were established more than 10,000 years ago as biogeographic processes shaped the indigenous vegetation and fauna (Watts 1987). With the arrival of humans, this landscape was irrevocably modified to serve human needs. It is possible that the earliest immigrants did not bring new plants or animals with them, and that their procurement strategies had little impact on the islands. They were, however, followed by extensive horticulturalists who slashed and burned the forests and introduced a suite of new plants. Every wave of immigrants brought new cultigens and new ways to process or prepare those that were already present. In the end, the more than 40 different cultigens have been mentioned in this review.

Animals were also introduced. The earliest introduction was the dog, and others, such as hutia, agouti, and guinea pig, had their insular distributions enlarged. The Spanish made the biggest impact, bringing cattle, chickens, goats, rats, sheep, horses, and donkeys. These animals helped transform luxuriant tropical vegetation into scrub forage. As some animals were introduced, others became extinct (Olsen 1982; Morgan and Woods 1986). Manatee, monk seal, and sea turtles are either extinct or nearly so, and even the ubiquitous queen conch is now in short supply (Hesse and Hesse 1977).

Today, most West Indians have forsaken agriculture and instead rely on imported foods. For instance, Ifill noted in his study of Grenada that “in 1974 the estimated daily calorie intake per capita was 1,958.4 of which 1,535.9 [was imported].… With respect to

protein per capita, daily intake was 46.03g, of which imported food supplied 31.72g” (quoted in Brierley 1985: 52). Despite this national reliance on imported foods, rural peoples remain dependent on local production on both provision grounds and kitchen gardens.

This infield (kitchen garden)-outfield (provision grounds) division finds its recent origins in the slave plantation economy, but it mirrors practices that can be traced to South America and Africa. Present-day kitchen gardens reflect both continuity with and a synthesis of the past. For example, J. S. Brierley’s (1985) list of 20 common plants in Grenadian kitchen gardens includes crops from South America, Africa, Asia, Europe, and Oceania, as well as the Caribbean (1985). As Brierley (1985: 55) points out, these are not simply a random selection of available cultigens; rather, the “nutritional balance of [the] crops must be attributed … to traditional knowledge and a process of selection governed by the dietary needs and ecological potential of the region.”

Despite repeated waves of new peoples and new dietary items, the climate and ecology have shaped culinary practices. Until the very recent reliance on a cash economy, diets have been shaped by the low-cost caloric productivity of root crops (manioc, sweet potato, yams, dasheen), the paucity or expense of terrestrial animals (hutias, iguanas, cattle, sheep, and goats), and abundant but labor-intensive marine organisms (fishes, turtles, queen conch). In some ways, the more things changed, the more they remained the same. One result is a convergence of diets throughout the tropics as cultigens and animals are shared around the world.