Comments on the Mainland Origins of the Preceramic Cultures of the Greater Antilles

COMMENTS ON THE MAINLAND ORIGINS OF THE PRECERAMIC CULTURES OF THE GREATER ANTILLES

RichardT. Callaghan

Computersimulations are used to shed light on the probable origins of the earliest Preceramic cultures of the Greater Antilles and to understand the navigation skills necessary for island colonization. These cultures, dating to between ca. 4000 B.C. and 2000 B.C., are found on Cuba, Hispaniola, and possibly Puerto Rico. Two areas, northern South America and northern Central America, have assemblages that bear resemblance to the assemblages of the Greater Antilles, but there are importantdifferences. Chance discovery of the GreaterAntilles is possible from three areas: northern South Amer- ica, northern Central America, and southern Florida. Directed voyages have a high degree of success from all three areas. However,voyages from northernSouth America requirethe least navigational skill, making it the most likely source of col- onization. From northern Central America, foreknowledge of the islands appears to be required, while directed voyages from southern Florida encounter considerable risk.

Se empleanmodelos de simulaci6npor computadorapara entenderlos origenesposibles de las primerasculturas precerdmicas de las Antillas Mayores, asi como las habilidadesnecesarias para realizar su colonizaci6n. Estas culturas, que tienenfechas entre4000 a. C. y 2000 a. C., se encuentranen la Cuba, la Espaiiola y posiblemente en Puerto Rico. Dos areas, la parte norte de Sudamericay la regi6n norte de Centroamerica,tienen conjuntosarqueoldgicos que presentansemejanzas con aquellos de las Antillas Mayores. Sin embargoexisten diferencias importantes,especialmente en las industrias liticas. Los resultados de la simulaci6n indican que hubo posibilidades del descubrimiento de las Antillas Mayores desde tres areas: el norte de Sudamerica,el norte de Centroamericay el sur de la Florida. Hay grandesposibilidades de exito en viajes intencionales desde cualquierade estas areas. Sin embargo,los viajesdesde Sudamericarequieren menos habilidadesparala navegaci6n, haciendo que esta sea lafuente mds probable de la colonizaci6n. Parece ser necesario tener un conocimientoprevio de la existencia de las islas para realizar el viaje desde America Central,y el viaje directo desde el sur de la Florida involucra riesgos consi- derables. Finalmente,este estudio sugiere que los conjuntosarqueoldgicos de la Centroamerica,de Sudamericay de las Anti- Ilas Mayoresrequieren de un andlisis comparativodetallado antes de poder sacar una conclusidnsolido con respecto al origen de las primeras culturas de las Antillas Mayores.

irtually every area of the Caribbeanand for the earliestpopulations in the GreaterAntilles Gulf mainland (Figure 1) has, at some (Keegan 2000; Wilson et al. 1998). The majority time, been hypothesizedto be the origin of the suggested origins for the GreaterAntillean of the early PreceramicAntillean cultures. Coe Preceramiccultures are based on comparisonsof (1957) suggested the Yucatan Peninsula, while a few artifacts,a few artifacttypes, or manufac- Hahn (1961:181-182) suggested the Gulf Coast turingtechniques. Despite calls from scholarssuch and the southwestern United States. Western as Rouse,no systematiccomparison has been made Venezuelawas suggestedby a numberof authors between the early Preceramicassemblages of the (Veloz and Martin 1983; Veloz and Ortega 1973; GreaterAntilles and any mainlandassemblages. Velozand Vega 1982).An originhas also beenpos- Even withinthe Antillean material, conflicting tax- tulated in the microblade-producingcultures of onomies exist (Rouse 1992:58). Louisiana(Febles 1982), and in Nicaragua(Crux- I would suggest that the early Preceramiccul- ent andRouse 1969). Currentlythe Yucatan Penin- turesof the GreaterAntilles probably had multiple sula seems to be favoredas the most likely source inputs from the surroundingmainland, but some

Richard T. Callaghan * Departmentof Archaeology,University of Calgary,Calgary, Alberta CanadaT2N 1N4

Latin AmericanAntiquity, 14(3), 2003, pp. 323-338 Copyright( 2003 by the Society for AmericanArchaeology

323

This content downloaded from 129.252.86.83 on Wed, 14 Oct 2015 20:47:15 UTC All use subject to JSTOR Terms and Conditions 324 LATIN AMERICAN ANTIQUITY [Vol. 14, No. 3, 2003]

Figure 1. Staging areas for voyages (letters indicate regions of departure considered in this study). group had to arrivefirst. This issue of primacyis cations have been proposed for the occupations investigatedthrough the use of computersimula- includingthose of Kozlowski(1974), Rouse(1992) tions of both accidentaland intentionaldiscovery andVeloz andVega (1982). AlthoughRouse's clas- of the islands from various mainland points of sificationhas been criticizedfor maskingvariabil- departure.The resultis to evaluatethe potentialfor ity in the assemblages,I use it hereas it is the most discovery of the islands by mainlandpeoples and widely known.Rouse (1992:51) classifiedthe arti- to assess the navigationalskills necessaryto reach factsof theearliest occupants of theGreater Antilles the islands.The simulationprogram is a morecom- intohis CasimiroidSeries, for a periodlasting from prehensiveversion of an earlierwork (Callaghan ca. 4000 B.C. to 400 B.C. The seriescontains three 1999, 2001). The currentversion has numerous subseries: Casimiran (4000 B.C.-2000 B.C.), advantages:it is more flexible with respect to the Courian (2660 B.C.-A.D. 240), and Redondan, questions that can be asked, it contains more beginning perhapsas early as 2050 B.C. (Rouse detailedwind andcurrent data, and it incorporates 1992:51-61). According to Rouse (1992:51), the archaeoclimaticmodeling factors in climatic vari- earlypeoples making these artifactsoccupied Cuba ation duringthe Preceramicperiod of the Greater andHispaniola and possibly PuertoRico; the early Antilles thatmight be significant. assemblagesfound on Antiguamay also represent Once likely placesof originon the mainlandare a manifestationof the CasimiroidSeries. No sim- identified,I presenta brief discussionof the simi- ilar assemblages are found south of the Guade- laritiesand differences between the relevantmain- loupe Passagein the LesserAntilles (Figure 1). land assemblagesand those of the GreaterAntilles These Casimiroidpeoples did not makeceram- based on previouslypublished work.' ics or practiceagriculture. Initially, their chipped stonetools were madeon blades,but these became The Preceramic Period Early less frequentin comparisonto tools madeon flakes of the Greater Antilles over time. Both blade and flake tools are over- The earliest human occupations known in the whelmingly unifacially retouched and shaped. Caribbeanislands date to approximately3000 B.C. Retouchon the dorsalsurface is usuallymarginal. for Cuba (Kozlowski 1974:67) and 4000 B.C. for Occasionally,a few retouchflakes are removed on Haiti (Moore 1991). A numberof artifactclassifi- the ventralsurface, often nearthe base (Kozlowski

This content downloaded from 129.252.86.83 on Wed, 14 Oct 2015 20:47:15 UTC All use subject to JSTOR Terms and Conditions REPORTS 325

1974). Extensivelychipped bifacial stone tools are the exceptionof two small sites, Boutbois and Le extremelyrare. Kozlowski (1974:70) reported three Godinot on Martinique (Allaire and Mattioni poorlyprovenienced bifacially chipped tools from 1983), and one very dubious site, Buccament,on Cubancollections, but doubtedthat one originated St. Vincent (Hackenberger1991). In fact, there is in Cuba. very little evidence of Preceramicpeoples in the The Old Harbor collection (Loven 1935: islands southof the GuadeloupePassage until one 219-222) of bifacialpoints from Jamaica has been reaches Trinidadand Tobago. Direct crossings, suggested as belonging to the first peoples occu- which bypassed these islands, were possible pying the GreaterAntilles (Harris1991:81). How- (Callaghan2001), but the Banwari Trace site is ever, the collection is highly problematic. Sven datedto approximately5000 B.C. It is in theRedon- Loven (1924, 1935) reportedthat the collection dan period, however, 3,000 years later, that shell was acquiredby CaptainA. F. Scholanderin 1920 gouges appearin Cuba.Even if shell gouges were from a Swedish sailor in Jamaica.The collection introduced from Trinidad 3,000 years after the consistedof 75 bifacialprojectile points. The sailor occupationof BanwariTrace, such an originwould claimedthat the pointscame from a moundat Old still not explainthe earliestPreceramic cultures of Harbour, Jamaica (Loven 1935:219). Rouse the GreaterAntilles. (1960:19) noted that the points are more in keep- Simulations of ing with the Florida Archaic period and do not Computer Voyaging resemble anything found in the Caribbean.Fur- Computer simulations have been successfully ther,Bullen (1976:595) was of the opinionthat this developedto studymaritime colonization, primar- collection represented three periods of Florida ily in the Pacific.Levison et al. (1973) designedthe points, going so far as to state that the material firstexperiment of this kind in the late 1960s. The appearedto be Floridachert. Given that these points simulationexamined the possibilitythat Polynesia fit well into a Floridatypology, the circumstances was colonized from Melanesia or from South of their acquisition, the absence of Preceramic America,as Heyerdahl(1952) espoused, by drift- periodsites on Jamaica,and a lack of similartools ing with the prevailingwinds and currents.The anywhereelse in the Caribbean,I feel it best not to resultsof 101,016simulated voyages demonstrated include the collection with Carsimiroidor other that, with only a few exceptions, undirecteddrift CaribbeanPreceramic assemblages. voyages could not accountfor the colonizationof Duringthe later partof the Preceramicperiod Polynesia. of the GreaterAntilles, the time of the Courianand A version of their programwas then run that RedondanSubseries, new types of tools appeared assumed the intention of the crew was to try to in the assemblages.These includeintricate ground maintaina general bearingeast. This experiment stone tools such as stone bowls, axes, beads, and ran 8,052 simulatedvoyages. It demonstratedthat items of unknownfunction (Rouse 1992:51-62). of all the EasternPolynesian Islands, the Marque- Many of these artifacts bear incised geometric sas Islands, would have had the best chance of designs.Shell gouges anddecorated wooden batons being colonized firstfrom West Polynesia,despite have also been recovered. the fact thatthe Marquesasare abouttwice the dis- The firstcultures of the GreaterAntilles do not tance from West Polynesia as the Society Islands, appearto have reachedthe islands via the Lesser also located in East Polynesia.This findingcorre- Antilles. Nor do they appearto be related to the spondsremarkably well withlinguistic evidence for BanwariTrace culturesof Trinidad.While some the patternof colonizationof EastPolynesia (Lev- researchers(Sanoja 1987; Veloz 1991) have sug- ison et al. 1973:48-50). The results of this model gesteda connectionbetween the two regionson the continueto be indirectlysupported by ethnographic basis of Strombusshell gouges, Rouse (1992) sees and experimentalobservations of the navigational the use of Strombusshell gouges as an independent skills and efficient voyaging canoes of the Pacific developmentin the GreaterAntilles. Further,as Islanders(Finney 1977, 1979, 1987, 1991; Glad- Keegan (1994:266) has noted, Preceramicsettle- win 1970; Goodenoughand Thomas 1987; Lewis ments are not found in the WindwardIslands with 1972).

This content downloaded from 129.252.86.83 on Wed, 14 Oct 2015 20:47:15 UTC All use subject to JSTOR Terms and Conditions 326 LATIN AMERICAN ANTIQUITY [Vol. 14, No. 3, 2003]

More recently,Geoffry Irwin of the University Structure of the Simulation of Aucklanddevised a second simulationprogram and Program Modifications (Irwin 1989, 1990, 1992; Irwin et al. 1990). The Aucklandteam utilizedthe same wind and current At issue in thepresent study are two points:(1) how data as Levison et al. 1973, but the canoes in their difficult would chance discovery of the Greater study were programmedfor interactive sailing Antilles have been from surroundingmainland strategiesthat permitted navigation by dead reck- areas;and (2) once the islands were known, how oning, thus providing the option of returntrips difficultwould it have been to returnto themusing eitherby deadreckoning or by following the Poly- the type of watercraftlikely to have been in exis- nesian traditionof latitudesailing. tence at the relevant time? Chance discovery The results(Irwin et al. 1990) seem to bearout includes such events as voyagersbeing lost at sea Irwin's earlierconclusions that humanexpansion during storms while fishing or conductingother across Remote Oceania was not a unidirectional activities. Dening (1963:138-153) noted that the process, despite being characterizedby a "bow- limitedempirical evidence of knowndrift voyages wave advance"from west to east, nor can it be jus- in Polynesia suggests a commonpattern of behav- tifiably regarded as the cumulative result of ior. People concludethey are lost earlyin the voy- countlesshaphazard, one-way voyages. Rather, the age and respond by allowing the vessel to drift simulationstrengthened the evidence for a safer, before the wind with no attemptto navigatein a but much more complicated, multidirectional particulardirection. In the secondscenario, the voy- processof navigationthat initially occurred against, agerseither know the locationof the islandsor they then across,and finally down the prevailingwinds. are deliberatelyexploring in a particulardirection. This complexpattern suggested that the overriding The fewer the numberof coursechanges required concernof earlycolonists was theirability to back- to reach the target, the simpler the navigational track to their point of departure,rather than the skills necessary. speed or rangeof exploration. While answeringthese questionsdoes not auto- A thirdsimulation (Thorne and Raymond 1989), matically tell us from which mainland area the developedat the AustralianNational University in GreaterAntilles were firstcolonized, it does tell us Canberra,utilized ethnographicdata for the per- the probabilityof discoveryfrom specific locations formance characteristicsof Indonesian bamboo andthe level of navigationalskill required to inten- rafts to plot driftvoyages from Timorto northern tionally reach the islands from locations on the Australia-parametersrelevant to the initial colo- mainland.An understandingof the navigational nization of Sahulland,the combinedlandmass of problemsencountered by prehistoricpeoples in the Australiaand New Guinea,in the LatePleistocene. Caribbeancan be used to evaluatepossible origins Givenknown wind andcurrent patterns, almost all of the Preceramicpeoples of the GreaterAntilles. raftsended up on theAustralian coast, most within The watercraftused in the simulation were a week to ten days. When Pleistocene sea levels canoes from the Upper Orinoco and Maya-style were considered,it was virtuallyimpossible for a canoes from CentralAmerica. Canoes were cho- raftto miss Australia. sen because they were in use in Floridaby 6000 A fourth,ongoing simulationstudy (Callaghan B. P. (Hartmann1996:90), the approximatetime of 1991a, 1991b, 1995, 1999, 2001; Callaghanand the earliestevidence for humanoccupation of the Fremont1992) deals with the initial colonization GreaterAntilles. Both canoe styles have a wide of the Caribbeanislands and subsequentinterac- distributionaround the Caribbean,and the Ye'k- tion patterns.The resultsshow thatthree mainland wana style has been recoveredfrom the Bahamas areashad the potentialfor chancediscovery of the (Callaghan and Schwabe 2001). Earlier work GreaterAntilles and that during the Ceramicperiod (Callaghan1991 a: 150-203) demonstratedthat rafts therewere no technologicalor environmentalbar- could not have made these requisite crossings riers requiringnavigation to follow a "stepping becauseof slow speedand excessive drifting before stone"pattern from island to island. the wind.Data regarding how thecanoes performed

This content downloaded from 129.252.86.83 on Wed, 14 Oct 2015 20:47:15 UTC All use subject to JSTOR Terms and Conditions REPORTS 327 were derived from both on-locationexperiments present such a reconstructionfor the last 10,500 and from analysis using naval architecturepro- years based on 180/160 ratios in Haitian Lake grams.Performance data for the Maya andYe'k- Miragoaneostracod shells that reflect changes in wana style canoes were entered into a dynamic precipitationfor the period.From about 6000 B.P. wind andcurrent simulation. The wind andcurrent to 4000 B.P., the 180/160 values and variation informationis the same as that used to compile (1991 :Figure2) were lower thanover the last 200 American Navy pilot charts (Defense Mapping years,the period during which pilot chartdata were Agency Hydrographic/TopographicCenter 1982). compiled.The values indicatehigher precipitation The simulationprogram in the following dis- levels than currentconditions (1991:792). There cussion is a much more advancedversion of the was a slightshift toward drier conditions at approx- one cited above (Callaghan1999, 2001). This sec- imately5200 B.P.than just priorto thattime, a con- ond generationprogram is based on the United dition that lasted until approximately3200 B.P. StatesNavy Marine Climatic Atlas (US Navy 1995) During this period, conditions resembled the andhas beenexpanded to includeall of the world's wettest conditionsof the last 200 years. Hodell et seas andoceans with the exception of Arcticwaters. al. (1995) published a similar study from Lake The dataare organized at the finerresolution of one Chichancanabin the centralYucatanPeninsula. For degree Marsdensquares (one degree of longitude the period from -7000 B.P. to 4000 B.P., the cli- by one degree of latitude)rather than two degree matic history of the two lakes is very similar Marsdensquares as in the previousstudies. In par- (1995:375). ticular,this change allows the effects of smallerand Hodell et al. (1991:792) note a correlation more variablecurrents to be accuratelyreflected in between precipitationanomalies and variationin the outcomes. The advancedprogram also auto- the annualclimatic cycle in the Caribbeanregion. matically shifts from the data for an initially They state:"(e)nhancement of the annualcycle led selectedmonth to thatfor thefollowing month after to yearsof anomalouslyhigh precipitation, whereas the originalinterval has expired.This featurebet- a reduction led to a deficient rainy season" ter reflectsthe reality of changingwind and cur- (1991:792).Thus reconstruction of variationin pre- rentconditions over long voyages.A conversionto cipitationshould accurately indicate variation in the sphericalcoordinates has been addedthat increases annualcycle. The annualcycle itself is controlled positionalaccuracy outside of thetropics. The oper- by the summerdisplacement of the NorthAtlantic atoralso has the option of definingsuccess in dif- subtropicalhigh by the northwardmovement of the ferentways. Success can be definedas sightingan intertropicalconvergence zone and the reverse island from a particulardistance or as the vessel movement in winter. Hodell et al. compare their actually making landfall. Finally, the program datawith annualcycle intensitychanges estimated allows the operatorto changethe bearingof a ves- from "theseasonal insolation difference at the top sel duringa voyage to reflect decisions made by of the atmosphereat 10? N between August and the crew.This last featureis importantwhen assess- February"(1991:792). Changes in bothrecords for ing the level of navigationalskill requiredto reach the 10,500-yearperiod are similar (1991 :Figure2). a selected target. This agreementreinforces the conclusionthat while variationfrom present climatic conditions, includ- Past Climate and Water Levels ing surface wind patternscontrolled by the sub- Although climatic conditions during the period tropicalhigh, existed duringthe early Preceramic from 6000 B.P. to 4000 B.P.differed from present period,they were not substantial. conditions,it is possible to estimatethe natureof Further,data from sea cores and other sources these differencesand includethem in the interpre- (Bryson1987:2) indicate that, on a global scale, the tationof the results.It is also importantto consider boundarybetween the prevailing westerlies and the variationin sea levels. tradewinds has been at about20 degreesnorth lat- High-resolution reconstructions of the itudethroughout the Holocene.This meansthat the Caribbeanclimate are available. Hodell et al. (1991) NorthAtlantic subtropical high pressurecell would

This content downloaded from 129.252.86.83 on Wed, 14 Oct 2015 20:47:15 UTC All use subject to JSTOR Terms and Conditions 328 LATINAMERICAN ANTIQUITY [Vol. 14, No. 3, 2003]

0.55 2

0.5 1.8 V) UJ z 2 0.45 1.6

(-- ? 0.4 1.4 O 0 1.2 OC u I 0.3 u 0.3 1 03 LU co 0.25 0.8

0.2 0 0.6 -14000 -12000 -10000 -8000 -6000 -4000 -2000 0 RADIOCARBONDATE BP

ANNUAL * SEPT

Figure 2. Tropical cyclone history, South Florida (provided by Reid Bryson, Center for Climatic Research, University of Wisconsin). have been furthernorth in terms of annualmean, Tropicalcyclone activity is also an important butprobably not northof its presentnorthern (sum- issue when consideringnavigation. Figures 2 and mer)limit (Lamb1987:Figure 9.3). Therefore,con- 3 (providedby Reid Bryson, Centerfor Climatic ditions similar to present summer may have Research,University of Wisconsin)model tropical prevailedyear roundin the study area duringthe cyclone activity for the LesserAntilles and south periodcentered around 6000 B.P. to 4000 B.P. Florida.For the LesserAntilles, activityis slightly Bryson and Bryson's (1998) archaeoclimatic less from6000 B.P.-5000 B.P.than from about 500 modeling indicates very similar patternsof pre- B.P.-100 B.P., and is significantlyless from 5000 cipitation for a number of areas around the B.P.to 4000 B.P.Between 6000 B.P.and 4000 B.P., Caribbeanand Gulf region.The modelingis based tropicalstorm activity for southFlorida was roughly on Milankovitchcycles, geographicinformation, half what it was between 500 B.P. and 100 B.P. and modern atmosphericcalibration data. Such Overall,the availabledata suggest a 6000-4000 modeling effectively demonstratesthe magnitude B.P. climate very similar to that reflected in the of the changes and is used to calculate past posi- United States Navy Marine Climatic Atlas (US tions of the intertropicalconvergence zone, the sub- Navy 1995), based on databeginning in the early tropicalhigh pressuresystems, the jet stream,and nineteenthcentury. The main climatic difference other major features of atmosphericcirculation. appearsto havebeen a significantreduction in trop- Fromthis modeling,precipitation and temperature ical stormactivity in the northwesternCaribbean can be derived.The resultsof the modeling show and easternGulf region duringthe periodof inter- precipitationbetween -6000 B.P.and 4000 B.P.for est, perhapsmaking voyaging safer. the coast of Venezuela,the CaymanIslands, Key Many sea areas relevantto this study are very West in Florida,and Isla de Cozumel to be equiv- shallow.If anysignificant land was exposedaround alent to, or slightly greaterthan, that of the period 6000 B.P. it might have providedvoyagers with from -500 B.P.to 100 B.P.(Reid Bryson, personal intermediatestopping places on their way to the communication2002). GreaterAntilles. Hendry (1993:Figure7.3) pre-

This content downloaded from 129.252.86.83 on Wed, 14 Oct 2015 20:47:15 UTC All use subject to JSTOR Terms and Conditions REPORTS 329

1.5 3

2.5 0 -J 1.25 2 - 1

o 0.75 1.5 I- cc co 0.5 ...... 1 E 0. UJ 0.25 0.5

o l 0 -14000 -12000 -10000 -8000 -6000 -4000 -2000 0 RADIOCARBONDATE BP ANNUAL * SEPT

Figure 3. Tropical cyclone history, Lesser Antilles (provided by Reid Bryson, Center for Climatic Research, University of Wisconsin). sents sea-levelcurves for a numberof areaswithin of threeregions-south Florida,theYucatan Penin- the Gulfand the Caribbeanregion. Those pertinent sula, and Venezuela-there was virtuallyno pos- to thisdiscussion include Surinam, Jamaica, Belize, sibility of either chance discovery of the Greater and Florida. Sea level reached present levels at Antilles or of directed(paddled) voyages. This is about6000 B.P. in Surinamand was about30 cen- not surprisingfor the western Caribbeanwhere timeterslower in Belize, 40 centimeterslower in navigatingto the islands is to windward.Colum- Jamaica,and 50 centimeterslower in Floridaby that bus, in a letterto Ferdinandand Isabellaconcern- time. Hendry(1993:115) also points out that tec- ing his disastrous fourth voyage, commented on the tonic movementgenerally cannot explain the vari- problemsof sailing to windwardin the Indies: ationin relativesea level withinthe Caribbeanand In the Indies,if ships do not sail except with Gulf region. Rates of verticaltectonic movement the wind abaft, it is not becausethey are ill are at least one orderof magnitudeless thanrela- built or becausethey are clumsy.The strong tive sea-level rise, thus giving them little signifi- currentsthat are there, together with the wind, cance in most localities over the last few thousand bringit aboutthat none can sail with the bow- for in one wouldlose as much An examinationof nauticalcharts reveals that line, day they years. way as they might have made in seven, nor no significantnew land would have been exposed does a carvelserve, even if it be a Portuguese in areassuch as the BahamaBanks, the Nicaragua latteenrigged vessel [Columbus1932:98]. Rise, or southwestFlorida. Winds and currentsin the Gulf of Mexico are Simulation Results extremelyvariable for periodsextending voyages Figure 1 shows the 19 areas of departureconsid- beyond the five-monthlimit used in the model. A ered (Callaghan1991a and 1991b) and then rein- five-monthlimit was used becauseit representsthe vestigated with the current second-generation duration of some of the longest drift voyages program.Three thousand voyages were simulated recorded(Levison et al. 1973:20). fromeach areafor both driftand directed voyages, As the Maya-style canoe and the Ye'kwana- totaling 114,000 simulations.With the exception style canoeresults are nearly identical, Table 1 pre-

This content downloaded from 129.252.86.83 on Wed, 14 Oct 2015 20:47:15 UTC All use subject to JSTOR Terms and Conditions 330 LATINAMERICAN ANTIQUITY [Vol. 14, No. 3, 2003]

Table 1. Drift VersusIntentional Voyage Landfall Success Rates for Specific Source Points.

Geographic Drift Voyage Drift Voyage IntentionalVoyage IntentionalVoyage Source Pointa Success Rate Avg. Duration Success Rateb Avg. Duration A 0.00% 0.00% B 0.10% 15 days 100.00% C 0.00% 0.00% D 0.00% 0.00% E 0.00% 0.00% F 0.00% 0.00% G 0.00% 0.00% H 0.00% 0.00% I 0.00% 0.00% J 0.10% 4 days 0.20% 2 days K 0.00% 100.00% 5 days L 0.00% 100.00% 6 days M 0.00% 0.00% N 0.00% 0.00% 0 0.00% 0.00% P 0.00% 0.00% Q 3.00% 39 days 100.00% 7 days R 5.00% 34 days 100.00% 6 days S 10.00% 30 days 100.00% 6 days aSee Figure 1. bAt 3.4 knots per day. sents only the results of the Maya-style vessel. For Florida. The results from the second generation more than a 300 nautical mile stretch of the program show low rates of .1 percent in Septem- Venezuelan coast, points Q, R, and S, there is a rea- ber and April comparable to those for Isla Mujeres. sonable chance of an undirected vessel successfully Similar to the case for Isla Mujeres, moving even reaching the Greater Antilles given fall environ- 5 nautical miles off this starting point eliminated mental conditions starting in September. Success any chance of success. was defined in the simulations as making landfall In comparing the likelihood of successful drift on the islands. The success rate ranges from 3 per- voyages to the Greater Antilles from the three stag- cent in the western part of the staging area to 10 ing areas, a departure from the Venezuelan coast percent in the east. From the west, landfalls are pri- has a considerably greater chance of success than marily in western Cuba. As one moves east, land- from northernCentral America or southern Florida. falls are overwhelmingly on the southwest coast of However, there are other factors that suggest suc- Hispaniola. Interestingly, because no Preceramic cess from South America was much more likely. sites have been found on Jamaica, only one of the First, given the length of the staging areas for the 9,000 simulated drift voyages made landfall on three locations, there is approximately 15 times Jamaica. more coastal area from which success is possible For voyages originating from the northern Cen- for South America than for northern Central Amer- tral American coast, successful drift voyages were ica or southern Florida. Given the presence of limited to a small area east of Isla Mujeres off the dugout canoes along the Venezuelan coast and northeastern coast of the Yucatan Peninsula. The some time depth for their use, it appears almost success rate was .1 percent and was limited to April inevitable that some people would have found and September conditions. Moving the starting themselves in the Greater Antilles. position as little as ten nautical miles north or south The program gives success rates for vessels, but of Isla Mujeres eliminated any chance of success. does not include calculations for crew survival. The Results for drift voyages from the southern tip of means of estimating risk to the crew is provided by Florida in the eastern Gulf of Mexico show that suc- Levison et al. (1973:Figure 6). Their estimations cess was possible from a point near Key West, are derived from the work ofMcCance et al. (1956),

This content downloaded from 129.252.86.83 on Wed, 14 Oct 2015 20:47:15 UTC All use subject to JSTOR Terms and Conditions REPORTS 331 wheresurvival rates were calculated from data gath- intentionalvoyages from all three areas,the prob- ered on approximately27,000 persons forced to abilityof crew loss was less than 1 percent. abandon ship during World War II. In tropical The resultspresented in the currentstudy differ waters,risk is determinedlargely by lengthof time from those of the first generationprogram for a at sea. The datashow dangerto crew for driftvoy- numberof reasons.As noted above, the improved ages as follows: successful driftvoyages from the programuses wind and currentdata at a finerres- SouthAmerican coast took from4 to 6 weeks, with olution, thus allowing more accuratereflection of a probablecrew loss of 10 to 12percent. This means smallercurrents. Its greatestimproved effect is in that 1 or 2 members(out of a totalof 8 to 10) could watersoff southwestFlorida and along the south- be expectedto die en route.Such a loss could still erncoasts of the GreaterAntilles wheremost of the result in a viable foundingpopulation, according variationbetween the two programsoccurs. Hav- to Birdsell (1985), or allow for a return trip ing the program automatically shift to the next (Callaghan1999). The probabilityof crew loss in month'sdatabase after 30 days also has a signifi- successful drift voyages originating near Isla cant effect. Because the datawere not availablein Mujereswas about1 percentand therefore insignif- electronicform at the time the initialprogram was icant, while the probabilityfrom Key West was 5 written, that version only contained data for the to 7 percent. months of January,April, July, and October.It is There is considerabledifferentiation between the Septemberdatabase, however, that reveals the mainlandstaging areas when we considerthe ques- most variation. tion of intentionalvoyages to the GreaterAntilles. Voyagersfrom the coast of Venezuelahad the First, the configurationof the target area when greatest likelihood of chance discovery of the approachedfrom South America offers a greater GreaterAntilles. The probabilityof such a discov- widththan from the othertwo areas.There is a tar- ery throughintentional exploration and for subse- get windowof 80 degreesfrom that coast compared quenttrips was also veryhigh. Suchvoyages would to 50 degrees from northernCentral America and have been well within the technicalcapabilities of 75 degreesfrom the easternGulf Coast. people with dugout canoes and would not have More importantis the variationin the paddling been hamperedby the environmentof the early speedthat had to be maintainedto reachthe Greater Preceramicperiod. The level of knowledge and Antilles. Success was possiblefrom all threeareas navigationalskill requiredwas minimal,requiring if a speed of 3.4 knots was maintainedas a daily only thata northwardbearing be kept. Unforeseen average.Nevertheless, from SouthAmerica, voy- events that reducedthe paddlingspeed below 3.4 agers simply maintaininga northerlybearing and knots usually would not have resultedin missing an averagedaily speed of .5 knotsstill would have the GreaterAntilles. usually made landfallin the GreaterAntilles, and Chance discovery of the GreaterAntilles was maintaininga speedof 3.4 knotscould have resulted unlikelyfrom northern Central America, given the in landfallin about6 days (Callaghan2001). Only very limitedstaging area from which it was possi- .2 percentof attemptsto reachCuba from nearIsla ble. Intentionaldiscovery also seems somewhat Mujereswould have been successfuleven if a speed unlikely given the foreknowledge required to of 3.4 knots were maintained without course change course in the open sea. However, once changesin the open sea. Success also was possible knowledgeof theGreater Antilles was acquired,the from furthersouth (areas K and L), but voyagers islands could have been reachedrelatively easily wouldhave needed to maintaina speedof 3.4 knots from CentralAmerica with little risk, as therewas andhave had the knowledge and skill to makecourse a strongtendency for vessels to be blown back to changesin the open sea. If voyagerspaddled south the mainlandif they were unsuccessful.The results fromFlorida and maintained 3.4 knots,they would heresuggest that initial contact between theYucatan haveno problemreaching Cuba. However, if speeds Peninsulaand Cuba was likelyto havefirst occurred droppedbelow 2.5 knots the voyagerswould have from Cuba, with the necessary knowledge being been at considerablerisk of being swept east and transmittedto CentralAmerica by people already then northinto the AtlanticOcean. For successful in the islands.

This content downloaded from 129.252.86.83 on Wed, 14 Oct 2015 20:47:15 UTC All use subject to JSTOR Terms and Conditions 332 LATIN AMERICAN ANTIQUITY [Vol. 14, No. 3, 2003]

Figure 4. Sandhill phase artifacts, Belize (courtesy of Richard S. MacNeish).

The possibilityof chancediscovery from south- ern Floridacoast, continuingout into the Atlantic ernFlorida was limitedin the samemanner as from trappedin the Gulf Stream,or makinglandfall on northernCentral America. While intentionaldis- eitherGrand Bahama Island orAndros Island. The coveryand subsequent voyages did not require nav- Bahamasdo not appearto havebeen occupieduntil igationalskill beyond keeping a southwardbearing, well into the Caribbeanceramic period (Keegan this trip would have involvedthe highest risk.An 1992:48; see Seidmann [2001] for an in-depth inability to maintain a relatively high paddling analysisof the problemsof canoe navigationalong speedwould have resulted in landingfar up theeast- the southwestcoast of Florida).Simulation results

This content downloaded from 129.252.86.83 on Wed, 14 Oct 2015 20:47:15 UTC All use subject to JSTOR Terms and Conditions REPORTS 333

Figure 5. Orange Walk phase artifacts, Belize (courtesy of Richard S. MacNeish). do not exclude some sporadic contact between occupied. Returntrips from PuertoRico to north- Florida and Cuba during the Preceramicperiod. ern SouthAmerica would have been no more dif- Kozlowski (1974:71) suggested that one of the ficultthan the precedingones from SouthAmerica threebifaces noted in Cubancollections might rep- to the north(Callaghan 2001). This also holds true resentsuch a contactduring the "Neoindian"period for returntrips from Hispaniolaand easternCuba. of Florida. From western Cuba, no voyages directly to the Previous work (Callaghan 1995) investigated South Americanmainland were successful in the the possibilitiesof reachingvarious points on the simulationdue to the need to proceedinto the pre- mainlandfrom the GreaterAntilles once they were vailing winds and currents. However, as noted

This content downloaded from 129.252.86.83 on Wed, 14 Oct 2015 20:47:15 UTC All use subject to JSTOR Terms and Conditions 334 LATIN AMERICAN ANTIQUITY [Vol. 14, No. 3, 2003] above andin Callaghan(1995:Figure 3), therewas bladetechnology arrived in theAntilles fully devel- little navigationalproblem in reachingthe Yucatan oped. However,pure blade cores arepoorly repre- Peninsula from western Cuba. Reaching Florida sented in some sites (Febles 1982). In other sites, from centralCuba (Callaghan 1995:Figure 2) was thereis a progressionin core formfrom blade cores also a relativelyeasy proposition. to flake cores (Kozlowski 1974: 40). In Levisa I, Levels VII and VI, dating between roughly 5000 Comparison of Assemblages B.P. and 4000 B.P., tools include end-scrapers, from Identified Staging Areas burins, retouched blades, side-scrapers,backed implements, notched implements, and triangular Simulationprograms can be usedto directresearch flakes resembling Levallois points (1974:55). in the most profitabledirection by identifyingthe Kozlowski named other unifacial tools made on most likely possibilities. The results suggest that blades "tanged,""shouldered," and "leaf shaped" voyages from northern South America and the points. While I do not question that at least some YucatanPeninsula offered the greatestprobability of these artifactsare indeed points, the term may of chance discovery of the Greater Antilles. be misleading, and the reader is referred to Althoughno detailedcomparative analysis of mate- Kozlowski'sFigures 10 and 11 (1974:54,56). Arti- rial culturehas been conductedfor any region to facts otherthan chippedstone do not appearuntil evaluate early contact, some comparisons have Level V. been noted for the two most probablemainland A number of northernSouth American areas regions. have been suggested as sources of early Antilles populationson the basis of similar technologies, NorthernSouth America includingColombia and the Stateof Falconin west- Most authors who have suggested similarities ernVenezuela. Other localities, such as Louisiana between northernSouth American assemblages in the United States, were also consideredon the and those of the early Preceramicperiod of the basis of microbladetechnology. These possibilities GreaterAntilles have worked in Cuba and the were eventuallyrejected, although some of these Dominican Republic (Febles 1982; Kozlowski previouslylisted authors(Febles 1982;Kozlowski 1974;Ortega and Guererro 1981; Veloz andMartin 1974) refer to a perforatortype as being a Poverty 1983; Veloz and Ortega 1973; Veloz and Vega Pointform. The termmay have been used descrip- 1982). Their work tends to give a more complete tively,however, rather than to suggesta connection, pictureof the earlyAntillean assemblages through as it was introducedby Kozlowski who specifi- drawings and photographsthan sources such as cally rejectedFlorida or the southerncoast of the Rouse (1992), in which only a very limited selec- North American Southeast as an origin for the tion of artifactsis illustrated.The more extensive Antillean culturesin his conclusions (Kozlowski illustrations show greater variation, both geo- 1974:68). Virtuallyall of these authorsacknowl- graphicallyand temporally, in the assemblages.Of edge that the South American mainlandassem- interest is their finding that the earliest cultural blagesreferenced in comparisonare poorly known. manifestationsin the GreaterAntilles are concen- Probablythe best discussionof the earlylithics tratedin easternCuba and southernand western of the GreaterAntilles andtheir possible origins in Hispaniola(Kozlowski 1974:59; Ortega and Guer- northernSouth America is outlinedin the work of rero 1981:218). Veloz andMartin (1983). They suggestthat assem- Duringthe first 2,000 years of occupation,the blages primarilyin the State of Falconof western assemblagesinclude blade, microblade,flake, and Venezuelaare the predecessors. These assemblages Levellois-like technologies. Chippingtechniques contain both unifacial and bifacial tools and are are largely directpercussion and pressureflaking linked by the authorsto El Jobo industries.Ini- (Febles 1982:19).Beginning before 5000 B.P.,the tially,the unifacialtools were made on flakes,but earliestsites, such as Levisa I in Cuba (Figure 1), over time a blade technology developed,particu- primarily contain blades with marginal retouch larly in the coastal sites. Kozlowskistates that the (Kozlowski 1974). Kozlowski believed that this bladetechnology developed sometime around 6000

This content downloaded from 129.252.86.83 on Wed, 14 Oct 2015 20:47:15 UTC All use subject to JSTOR Terms and Conditions REPORTS 335

B.C.-5000 B.C. (1974:24). Veloz and Martin localized in the New World, the nine industries (1983:14) list a numberof unifacialtools charac- exhibiting it that he studied "appearto be inde- teristicof these assemblages,including plano-con- pendentdevelopments with no historicalconnec- vex perforators,notched scrapers,plano-convex tions between them" (1994:87). We should be blades with retouch on the dorsal surface, semi- cautious about making a connection between the lunatescrapers with dorsal retouch or backed scrap- YucatanPeninsula and the GreaterAntilles on the ers, bilateral semi-lunate scrapers with dorsal basis of manufacturingtechnology and chronology retouch,blades or knives, and scraperswith mar- alone. ginal retouch.The methodof manufactureis direct However,if such a connectionproved to be cor- percussionwith pressureflaked retouch. rect, it would be necessaryto question why other The authorsconnect these westernVenezuelan aspects of the Belizean assemblagesare not found assemblages to the early Preceramicera of the in the GreaterAntilles. There are significantdif- GreaterAntilles while at the same time notingthat ferences. Belizean Preceramicsites discussed by there is not a complete correspondence.Signifi- Wilson et al. (1998) containpointed unifaces that cantly,bifaces are present in theVenezuelan assem- resemble some of the unifaciallyworked macrob- blages while they are virtually absent from the lades from the GreaterAntilles (1998:Figures 4 GreaterAntilles. The correspondencebetween the and 5) but differin the degree of working.As Pan- unifacial assemblagesof the State of Falcon and tel states, "It has been clearly demonstratedthat the GreaterAntilles, as shown in their figures, is secondary flaking is mostly absent from West more convincing. However, their analysis lacks Indian assemblages" (1988:165). Those from such basic informationas frequenciesthat would Belize are usuallychipped over most, if not all, of be neededto discriminateamong similarly general the dorsalsurface. matcheswith assemblagesfrom manyparts of the The other unifacialtool from Belize discussed world. by Wilsonet al. (1998:Figure4B), the "constricted uniface,"has no counterpartin the Antilles. These The YucatanPeninsula constrictedunifaces were originallytermed "snow- Wilsonet al. (1998) presenteddating for Belize that shoe shapedend scrapers"by MacNeish(1982:40). broughtthe Preceramic Sand Hill andOrange Walk Usewear analysis conducted by George Odell phases (MacNeish 1982; MacNeish et al. 1980; (MacNeishand Nelken-Tumer 1983:15-24) show MacNeish and Nelken-Turner 1983) into line these tools to havebeen used for "adzing,gouging, chronologicallywith the earlyPreceramic cultures wedging and chopping hard and soft wood" of the GreaterAntilles. Workin Belize duringthe (1983:46).The only Antilleanartifact that could be early 1980s (MacNeish 1982; MacNeish et al. considereda "constricteduniface" might be a Couri 1980; MacNeish and Nelken-Turer 1983) had macrobladeshown in Wilson et al. (1998:Figure recoveredmaterials bearing some similarityto early 5A). It is difficultto imagine this tool being used Preceramic artifacts in Cuba and Hispaniola for the above tasks. (Callaghan 1990). The most striking similarity Both of the chipped stone tool types discussed between the two regions is the presence of mac- above were considereddiagnostic of the Sand Hill roblades with little or no retouch. Wilson et al. and OrangeWalk phases in Belize (MacNeish and (1998) presentedmore recentlyrecovered mater- Nelken-Turer 1983:Figure 4). While the dating ial of this type and showed that there is a strong of these phases was problematicat the time of the correlationin manufacturingtechnology. However, study,the associations with other artifact types were it shouldbe notedPantel (1988) demonstratedthat not. In the three years I was associated with the differencesin chipped stone technologythrough- Belize Archaic Archaeological Reconnaissance out the GreaterAntilles are due primarilyto varia- projectdirected by MacNeish,I observedSand Hill tions in the qualityof the availableraw materials. and OrangeWalk phase living floors and activity In short, macrobladesoccur in conjunctionwith areas in primarycontext. There were many tool large nodulesof high-qualitychert. Further, Parry types with no counterpartsin the Antilles (Mac- notesthat although this technology is fairlyrare and Neish and Nelken-Tuner1983:Figures 4 and 5),

This content downloaded from 129.252.86.83 on Wed, 14 Oct 2015 20:47:15 UTC All use subject to JSTOR Terms and Conditions 336 LATIN AMERICAN ANTIQUITY [Vol. 14, No. 3, 2003] includingbifacial projectile points anda varietyof for canoes and have had the knowledgeto change distinctivescrapers, choppers, and scraping planes coursein theopen sea. Assuming these criteria, con- (MacNeishand Nelken-Turer 1983:43-51). Bifa- tact between northernCentral America and the cial tools are not common in the Belizean assem- GreaterAntilles may then have followed two sce- blages, but those that do exist show considerable narios.One posits a determinedeffort to discover skill in manufacture(MacNeish and Nelken-Turer the islands. This scenario would not be too haz- 1983:Figures6 and 7). Bifacial tools other than ardous, as the majority of unsuccessful vessels blades showinga few flakescars on the ventralsur- would have returnedto the mainland.The second face are not found in the early Antillean assem- scenario posits that the necessary navigational blages. If thereis a directconnection between the knowledgewas obtainedfrom people already in the Archaicperiod in Belize andthe early Preceramic islands who would have had little problemreach- culturesof the Antilles, it is importantto explain ing northernCentral America, either accidentally why all the distinctivechipped stone tools of the or intentionally.From southernFlorida, a 100 per- Belizean cultureswere deleted, leaving primarily cent success rate was possible if a simple south lightly retouchedmacroblades. bearingand high paddlingspeeds were maintained. The risks were high in thatlower paddlingspeeds Conclusions could have resultedin canoes being swept into the Simulationsof navigationproblems can be used to open Atlantic.For successful intentionalvoyages rankmainland regions as possiblesources for island from all three areas, the probabilityof crew loss colonization.Analysis of chance discovery of the was minimal. GreaterAntilles via undirectedvoyaging shows Of the three potential staging areas, northern that trips with any chance of success could have SouthAmerica and northern Central America have originatedfrom only three areas of the Gulf and assemblagesthat share more similarities with those Caribbeanmainland: northern South America, the of the early Preceramicin the GreaterAntilles. YucatanPeninsula, and south Florida. Vessels drift- Assemblagesfrom both mainland areas also exhibit ing off the coast of northernSouth America had a significantdifferences that will only be understood reasonablechance of reachingCuba and Hispan- when detailedcomparative analyses have been con- iola, while those from the lattertwo areas had a ducted. Florida also may have contributedto the much lower probabilityof success. Further,the early Preceramiccultures of the GreaterAntilles lengthof the SouthAmerican coast from which suc- but,given the hazardsof the voyage, it is likely that cessful voyages were possible is about 300 nauti- contactwas extremelysporadic. cal miles, as comparedto 5 to 10 nauticalmiles for the YucatanPeninsula and for West off south- Key Acknowledgments.Funding for this work was providedby a ern Florida.The rateof crew survivalwas high for University of Calgary Research Services Pilot Grant and a all three areas.Interestingly, successful driftvoy- Social Sciences and Humanities Research Council of ages from northernSouth America usually made CanadaResearch Grant. Donna Fremontof the Universityof landfallin southwestern CalgaryComputer Science Departmentcarried out computer Hispaniola,which, along In I would like to the with eastern has the earliestdated programming. particular acknowledge Cuba, occupa- contributionsof a large numberof Belizeans who helped to tions in the GreaterAntilles. construct rafts and test canoes, most especially Morris Analysis of directedvoyaging where the crew Lesley of Sand Hill along with the Ye'kwana people of La paddleson a bearingto the islands showed a land- Esmerelda,Venezuela, who sharedsome of their knowledge fall success rateof 100 from all threeareas of canoe design with me. I would like to thank Lesley percent Nicholls for the and Cristobal within a reasonabletime frame.While even proofreading manuscript very Gnecco and Santiago Mora for the Spanish translationsof low paddling speeds and a simple north bearing the abstract.Finally, I would like to acknowledge Dr. R. S. resulted in success when departingfrom South MacNeish for slides of artifactsfrom the BARR project. America, similar from northernCentral voyages References Cited America and southernFlorida demonstrated only and Mario Mattioni a qualifiedsuccess. Fornorthern Central America, Allaire, Louis, 1983 BoutboisetLe Godinot:Deux gisementsaceramiques voyagers would have had to maintaintop speeds de la Martinique.Proceedings of theIX International Con-

This content downloaded from 129.252.86.83 on Wed, 14 Oct 2015 20:47:15 UTC All use subject to JSTOR Terms and Conditions REPORTS 337

gressfor the Studyof the Pre-ColumbianCultures of the Waters. US Defense Mapping Agency Hydrographic/ LesserAntilles,pp. 27-38. Centrede RecherchesCaraibes, TopographicCenter, Washington, D.C. Montreal,Quebec. Dening, GregM. Birdsell,James B. 1963 The GeographicalKnowledge of the Polynesiansand 1985 Biological Dimensions of Small, Human Founding the Natureof Inter-IslandContact. In PolynesianNaviga- Populations. In InterstellarMigration and the Human tion,edited by JackGolson, pp. 138-153. PolynesianSoci- Experience,edited by Ben R. Finney and Eric M. Jones, ety MemoirNo. 34, Wellington. pp. 110-119. Universityof CaliforniaPress, Berkeley. Febles D., Jorge Bryson,Reid A. 1982 Estudio tipologico y tecnologico del material de 1987 On Climatesof the Holocene.Proceedings of the 17th piedra tallada del sitio arqueologico Canimar1, Matan- Annual Chacmool Conference 1984:1-13, Calgary, zas, Cuba.Academia de Ciencias, Habana. Alberta. Finney,Ben R. Bryson,Reid A., and RobertU. Bryson 1977 VoyagingCanoes and the Settlementof Polynesia.Sci- 1998 AnArchaeoclimatologyWorkbook: High-Resolution, ence 196:1277-1285. Site-SpecificClimate Modeling for Field Scientists.Cen- 1979 Voyaging.In The Prehistoryof Polynesia, edited by terfor ClimaticResearch, University of Wisconsin-Madi- Jesse David Jennings,pp. 323-350. HarvardUniversity son and Department of Geology and Geophysics, Press, Cambridge. Universityof Minnesota.University of Wisconsin,Madi- 1987 AnomalousWesterlies,ElNiio, andthe Colonization son. of Polynesia.American Anthropologist 87:9-26. Bullen, Ripley P. 1991 Myth,Experiment, and the Reinventionof Polynesian 1976 Did Palaeolithic,Archaic, and FormativeMan Enter Voyaging.American Anthropologist 93:383-404. the Antilles from Florida?XLI International Congress of Gladwin,Thomas Americanists1974 3:592-599, Mexico City. 1970 Eastis a Big Bird.Harvard University Press, Cambridge Callaghan,Richard T. Goodenough,Ward H., and StephenD. Thomas 1990 Possible PreceramicConnections between Central 1987 Traditional Navigation in the Western Pacific: A Americaand the GreaterAntilles. Proceedingsof the XI Searchfor Pattern.Expedition 29(3):3-14. International Congress for Caribbean Archaeology Hackenberger,Steven 1985:65-67. Universityof Puerto Rico, Recinto de Rio 1991 ArchaeologicalTest Excavation of BuccamentValley Piedras. Rockshelter,St. Vincent:Preceramic Stone Tools in the 1991a MainlandOrigins of the PreceramicCultures of the WindwardIslands, and the EarlyPeopling of the Eastern GreaterAntilles.Ph.D. dissertation, University of Calgary. Caribbean.Proceedings of theXIII International Congress UniversityMicrofilms, Ann Arbor. for CaribbeanArchaeology, pp. 86-91. Curacao. 1991b Passages to the GreaterAntilles: An Analysis of Hahn,Paul G. Watercraftand the MarineEnvironment. Proceedings of 1961 The Cayo RedondoCulture and its Chronology.Ph.D. the XIV Congress of the InternationalAssociation for dissertation,Yale University.University Microfilms, Ann CaribbeanArchaeology 1991:33-39. Barbados. Arbor. 1995 Antillean CulturalContacts with MainlandRegions Harris,Peter O. as a Navigation Problem. Proceedings of the Fifteenth 1991 A Paleo-IndianStemmed Point from Trinidad,West Congressof the InternationalAssociation for Caribbean Indies. Proceedings of the XIV Congress of the Interna- Archaeology 1993:181-189. San Juan,Puerto Rico. tional Association for Caribbean Archaeology, 1999 Computer Simulations of Ancient Voyaging. The 1991:33-39. Barbados. NorthernMariner/Le Marin Du Nord 1:24-33. Hartmann,Mark J. 2001 Analysis of CeramicAge Seafaringand Interaction 1996 The Development of Watercraftin the Prehistoric Potential in the Antilles. Current Anthropology SoutheasternUnited States. Ph.D. dissertation,Texas A & 42(2):308-313. M University.University Microfilms, Ann Arbor. Callaghan,Richard T., and Donna Fremont Hendry,Malcolm 1992 A ComputerSimulation of Routes of Passage in the 1993 Sea-LevelMovements and ShorelineChange. In Cli- Caribbeanand Gulf of Mexico. Paperpresented at the 57th mate Changein the Intra-Americas Sea, editedby George AnnualMeeting of the Societyfor American Archaeology, A. Maul, pp. 115-161. EdwardArnold, London. Pittsburgh. Heyerdahl,Thor Callaghan,Richard T., and StephanieJ. Schwabe 1952 AmericanIndians in the Pacific.Allen & Unwin, Lon- 2001 Watercraftof the Islands.Proceedings of the 18th Con- don. gress of the International Association for Caribbean Hodell, David A., Jason H. Curtis, Glen A. Jones, Antonia Archaeology1999: 231-242. Grenada. Higuera-Gundy,Mark Brenner, Michael W. Binford,and Coe, WilliamR., II KathleenT Dorsey 1957 A DistinctiveArtifact Common to Haiti and Central 1991 Reconstructionof CaribbeanClimate Change over America.American Antiquity 22:280-282. the past 10,500 Years.Nature 352:790-793. Columbus,Christopher Hodell, David A., Jason H. Curtis,and MarkBrenner 1932 FourthVoyage of Columbus. In Select Documents 1995 Possible Role of Climate in the Collapse of Classic Illustratingthe FourVoyages of Columbus.Vol. II, second Maya Civilization.Nature 375:391-394. series, No. LXX. HakluytSociety, London. Irwin,Geoffry Cruxent,James M., and IrvingRouse 1989 Against,Across, and Down the Wind:A Case for the 1969 Early Man in the West Indies. ScientificAmerican Systematic Colonization of the Remote Pacific Islands. 221(5):42-52. Journalof the PolynesianSociety 98:167-206. Defence MappingAgency Hydrographic/TopographicCenter 1990 Human Colonization and Change in the Remote 1982 Publication 106. Pilot Charts of CentralAmerican Pacific. CurrentAnthropology 31:90-94.

This content downloaded from 129.252.86.83 on Wed, 14 Oct 2015 20:47:15 UTC All use subject to JSTOR Terms and Conditions 338 LATIN AMERICAN ANTIQUITY [Vol. 14, No. 3, 2003]

1992 ThePrehistoric Exploration and Colonizationof the Pantel,Agamemnon Gus Pacific. CambridgeUniversity Press, Cambridge. 1988 PrecolumbianFlaked StoneAssemblages in the West Irwin, Geoffry,Simon Bickler,and Philip Quirke Indies. Ph.D. dissertation,University of Tennessee.Uni- 1990 Voyagingby Canoe and Computer:Experiments in versity Microfilms,Ann Arbor. the Settlementof the Pacific Ocean.Antiquity 64:34-50. Parry,William J. Keegan,William F 1994 PrismaticBlade Technologiesin NorthAmerica. In 1992 The People WhoDiscovered Columbus:The Prehis- TheOrganization ofNorthAmerican Prehistoric Chipped tory of the Bahamas. University Press of Florida, Stone Tool Technologies, edited by Philip J. Carr,pp. Gainesville. 87-98. ArchaeologicalSeries 7. InternationalMonographs 1994 WestIndian Archaeology. 1. Overviewand Foragers. in Prehistory,Ann Arbor. Journalof ArchaeologicalResearch 2: 255-284. Rouse, Irving 2000 West IndianArchaeology. 3. CeramicAge. Journal 1960 TheEntry of Man Into TheWest Indies. Yale Univer- of ArchaeologicalResearch 8(2):135-167. sity Publicationsin Anthropology,No. 61. New Haven. Kozlowski,Janusz K. 1992 The Tainos: Rise and Decline of the People Who 1974 Preceramic Cultures in the Caribbean. Zeszyty GreetedColumbus. Yale UniversityPress, New Haven. naukoweUniwersytetu Jagiellonskiego. Prace archeolog- SanojaO., Mario icyne; zesz. 20. PanstwoweWydawn Naukowe, Krakow. 1987 Ecologia y asentamientoshumanos en el noreste de Lalueza-Fox, C., M. T. P. Gilbert,A. J. Martiez-Fuentes,F. Venezuela.In Actas del TercerSimposio de la Fundation Calafell, and J. Bertranpetit Arqueologiadel Caribe,edited by O. M. Sanoja,pp. 1-10. 2003 MitochondrialDNA from Pre-ColumbianCiboneys Washington,D.C. from Cuba and the Prehistoric Colonization of the Seidmann,Rayan. M. Caribbean.American Journal of Physical Anthropology 2001 The Bahamian Problem in Florida Archaeology: 121:97-108. OceanographicPerspectives on the Issueof Pre-Columbian Lamb,Hubert H. Contact.Florida Anthropologist 54:9-23. 1987 Weather,Climate and HumanAffairs. Methuen, Lon- Thorne,Alan G., and RobertRaymond don. 1989 Man on the Rim: The Peopling of the Pacific.Angus Levison, Michael, R. GerardWard, and JohnW. Webb and Robertson,Sydney. 1973 TheSettlement ofPolynesia: A ComputerSimulation. US Navy Universityof MinnesotaPress, Minneapolis. 1995 MarineClimaticAtlas of the World.National Climatic Lewis, David Data Center,Asheville, North Carolina. 1972 We,the Navigators. University Press of Hawaii,Hon- Veloz, Marcio,M. olulu. 1991 Panoramahistorico del Caribeprecolombino.Banco Lov6n, Sven Centralde la RepublicaDominicana, Santo Domingo. 1924 Uber Die WurzelnDer TainischenKulture. Elanders Veloz, Marcio,M., and CarlosA. Martin BoktryckeriAktiebolag, Goteborg. 1973 Las technicasunifaciales de los yacimientosel Jobo, 1935 Originsof the TainanCulture, West Indies. Elanders sus similitudescon el paleo-arcaicoanillano. In Boletin BoktryckeriAktiebolag,Goteborg. del Museodel HombreDominicano 11 (18):13-39. Museo McCance, RobertA., Colin C. Ungley, James W. L. Crosfill, del HombreDominicano, Santo Domingo. and Elsie M. Widdowson Veloz, Marcio,M. and Elipidio Ortega 1956 TheHazards to Men Lost at Sea, 1940-44. Medical 1983 El preceramicode Santo Domingo, nuevos lugares, ResearchCouncil, Special Report No. 291. London,Med- y su posible relacioncon otrospuntos del area antillana. ical ResearchCouncil. Museo Del Hombre Dominicano, Papeles Ocasionales, MacNeish, RichardS. No. 1. 1982 ThirdAnnualReport of the Belize ArchaicArchaeo- Veloz Marcio,M., and BernardoVega logical Reconnaissance.Center for ArchaeologicalStud- 1982 The Antillean Preceramic:A New Approximation. ies, Boston University,Boston. Journalof New WorldArchaeology:33-44. MacNeish,Richard S., S. JeffreyK. Wilkerson,and Antoinette Wilson, SamuelM., HarryB. Iceland,and ThomasR. Hester Nelken-Turer 1998 PreceramicConnections between the Caribbeanand 1980 First Annual Report of the Belize ArchaicArchaeo- the Yucatdn Peninsula. Latin American Antiquity logical Reconnaissance.Philips Academy, Andover. 9:342-352. MacNeish, RichardS., andAntoinette Nelkin-Turer 1983 Final Report of the Belize Archaic Archaeological Notes Reconnaissance. Center for Archaeological Studies, Boston University,Boston. 1. DNA analysis of early human skeletal remains from Moore, Clark Cuba with comparisons to DNA from North, South, and 1991 Cabaret:Lithic Workshop Sites in Haiti. In Proceed- CentralAmerican have recently been published(Lalueza-Fox ings of the ThirteenthCongress of the InternationalAsso- et al. 2003). The results are significantto this work but were ciationfor CaribbeanArchaeology, edited by Edwin N. not availablein time to include in the discussion. Ayubi and Jay B. Haviser, pp. 92-104. Reports of the Archaeological-AnthropologicalInstitute of the Nether- landsAntilles, No. 9. Willemstad,Curagao. Ortega,Elipidio, and Jos6 Guerrero 1981 Estudiode 4 nuevossitiospaleoarcaicos en la isla de Santo Domingo. Museo del HombreDominicano, Santo SubmittedJuly 10, 2002; Accepted April 27, 2003; Revised Domingo. May 8, 2003.

This content downloaded from 129.252.86.83 on Wed, 14 Oct 2015 20:47:15 UTC All use subject to JSTOR Terms and Conditions