Corrected Proofs

Corrected Proofs

Radiocarbon, Vol 57, Nr 5, 2015, p 1–9 DOI: 10.2458/azu_rc.57.18313 © 2015 by the Arizona Board of Regents on behalf of the University of Arizona

RADIOCARBON AND STRATIGRAPHIC CHRONOLOGY OF CANÍMAR ABAJO, MATANZAS, CUBA Mirjana Roksandic1,2 • William Mark Buhay1 • Yadira Chinique de Armas3 • Roberto Rodríguez Suárez3 • Matthew C Peros4 • Ivan Roksandic1 • Stephanie Mowat5 • Luis M Viera3 • Carlos Arredondo3 • Antonio Martínez Fuentes3 • David Gray Smith6 ABSTRACT. Twelve accelerator mass spectrometry (AMS) radiocarbon dates from the shell-matrix site of Canímar Abajo (Matanzas, Cuba) are reported. Eleven were obtained directly from human bone collagen in burials and one was obtained from charcoal recovered from a burial context. The site stratigraphy presents two episodes of burial activity separated by a shell midden layer. The AMS dates fall into two compact clusters that correlate remarkably well with the stratigraphy. The older burial dates to between 1380–800 cal BC (2σ) and the younger one to between cal AD 360–950 (2σ). The AMS dates are compared to eight conventional 14C dates previously obtained on shell and charcoal. One of the conventional dates on charcoal (5480–5380 cal BC; 2σ) has been reported as the oldest 14C date in the Caribbean region; its context and reliability are clarified. The suite of AMS dates provides one of the most reliable chronometric dating of a cultural context during this timeframe in Cuba. The correlation of 14C and stratigraphy establishes a solid chronology for investigating the important economic and ritual features of Canímar Abajo.

INTRODUCTION

Canímar Abajo is a shell-matrix site in Matanzas Province, on the north-central coast of Cuba. The stratigraphy of the site presents two burial contexts separated by shell midden deposits. Among pre-contact burial sites in Cuba, Canímar Abajo is a rare example of a site that was subject to a large-scale excavation, with 36 continuous 1 × 1 m excavation units. As such, it offers a unique opportunity to pursue several important problems concerning the early inhabitants of the island, especially for establishing chronology and clarifying the subsistence regime. The site is located near the city of Matanzas (23°2′15.5″N; 81°29′49.1″E), on the western bank of the Canímar River, which discharges into the Bay of Matanzas (Figure 1). It is situated near the ancient beach of the estuary of the Canímar River, which is navigable to over 11 km from the estuary, reaching 12 m of depth in some areas. The estuary represents a very diverse ecosystem, characterized by intertidal change between freshwater and seawater, mangrove forest, evergreen forest—where trees have small leaves and superficial roots that brake and seep rocks—and abundant thorns and climbing plants (Martínez et al. 1993). Among animal resources, there are abundant sea, estuarine, and reef fauna including mollusks, fish, and turtles, as well as birds and manatees. Of terrestrial mammals, only different species of hutia were so far identified. The site of Canímar Abajo was partially destroyed by construction of a campsite in the early 1980s and subsequently excavated by Ramón Dacal Moure from 1984–1987. Rivero de la Calle (1987) reported 51 burials from a limited excavation in the northwest corner of the undisturbed portion of the site. Since 2004, the site has been more intensively excavated by the Montane Museum of the University of Havana, directed by Roberto Rodríguez Suárez. The site can be divided into five stratigraphic layers. From top to bottom, these are Layer 1, a superficial layer that contains recent soil intermixed with rock face; Layer 2, or the younger cemetery (YC), which is characterized by a mixture of soil and shell with human burials; Layer 3, represented by several strata of shell concentrations with charcoal lenses and ash, with sporadic and fragmented 1. University of Winnipeg, 515 Portage Avenue R3B2E9, Canada. 2. Corresponding author. Email: [email protected]. 3. University of Havana, Calle 25 #4553ntre JeI, Vedado Habana, C.P. 10400, Cuba. 4. Bishop’s University, 2600 Rue College, Sherbrooke, QC, J1M 0C8, Canada. 5. University of Manitoba, 727 McDermot Ave, MB, R3E 3T5, Canada. 6. University of Toronto, Mississauga, 3359 Mississauga Road, Mississauga, ON, L5L 1C6, Canada.

2

M Roksandic et al.

Corrected Proofs

Figure 1 Map of the study region

human remains without formal burials (the shell midden); Layer 4, or the older cemetery (OC), represented by a mixture of soil, charcoal, and sporadic shells with human burials; and Layer 5, sterile soil, possibly resulting from an earlier mangrove forest (Figure 2). To date, at least 213 individuals (83 adults and 130 subadults) have been excavated in 50 OC and 92 YC burials. There is no readily identifiable stratigraphic separation of burials within either the OC or the YC.

Figure 2 Stratigraphic profile indicating relative positions of samples for AMS 14C dating

Corrected Proofs 14

C and Stratigraphic Chronology of Canímar Abajo, Cuba

3

A set of eight conventional 14C dates were obtained by Ramón Dacal Moure on charcoal and shell recovered in the 1980s. Canímar Abajo has elicited substantial attention among the Caribbean archaeologists because one of these assays yielded a date of 6460 ± 15 BP (UNAM-0715: 5480– 5380 cal BC; 2σ). This is the oldest 14C date yet reported for the Caribbean archipelago (see Cooper 2010). It is evident, however, that the Canímar Abajo site area has been subject to significant bioturbation from tree roots and crab burrows. Such disturbance lessens the reliability of any 14C dates on material, such as charcoal, that does not have a secure cultural association. The set of AMS 14 C assays directly from burial contexts reported in this paper was submitted to provide a secure chronology for the stratigraphy and to see if the date from UNAM-0715 could be replicated. MATERIALS AND METHODS

A total of 14 samples from Canímar Abajo, nine from the older cemetery (OC) and five from the younger cemetery (YC), were submitted for dating at the University of Arizona Accelerator Mass Spectrometry Laboratory (lab code AA). Bone samples were ultrasonicated in water, dried, surface milled, and then crushed with a porcelain mortar and pestle to 0.5- to 1.0-mm fragments. These were loaded into continuous flow cells through which a programmed sequence of washes containing deionized water, 0.1N HCl, 0.5N NaOH, and 0.01N HCl was pumped, over the course of 16 hr. The demineralized fragments were gelatinized in test tubes at 70°C overnight and then filtered through 0.45-μm glass fiber filters. Collagen was recovered from the filtrates by lyophylization. Sample combustions, carbon stable isotope measurements, graphitizations, and AMS measurements were carried out according to standard AMS laboratory protocols. The 14C background inherent in bone collagen processing was determined from 14C measurements from multiple isolates of collagen from a 14C-free bovine bone obtained from a permafrost environment. The averages of these control measurements and their uncertainties were subtracted from all routine bone collagen measurements. Of the 14 submitted samples, 12 with sufficiently preserved collagen produced reliable dates. The choice of individuals submitted for dating was guided by their potential value for confirming the stratigraphic relationship of the two inferred cemetery occupations. While stratigraphic superposition is a basic tool in archaeology for establishing relative chronology, interpretation is complicated in a burial context where younger burials may intrude into older layers. To add to these general problems in reconstructing the chronology of cemetery usage, shell-matrix sediments often prevent observations relating to the original surface or the surface from which the burial pit was dug into the soil. Interpretation is further complicated by the bioturbation noted above. Direct dating of a representative number of burials ensures that stratigraphic observations are correctly interpreted. More individuals were selected from the lower than upper stratigraphic layers (nine as opposed to five, respectively) to counteract poorer preservation in the lower layer. One charcoal sample from a closed burial context, in direct association with a dated individual (the charcoal is found directly associated with the skull), was taken to ascertain comparability of results obtained on charcoal and human collagen. CALIBRATION

The AMS and conventional 14C dates were calibrated to calendar years BC/AD using the CALIB 7.0.2 mixed Marine and NH Atmosphere, IntCal13 and Marine13 (including a 400-yr offset; Stuiver and Reimer 1993) age calibration data sets for the collagen, charcoal, and shell samples, respectively (Reimer et al. 2013). Previous studies record localized (circum-Caribbean) marine reservoir age offsets that differ, sometimes significantly, from the average marine reservoir. The estimated contemporary correction values (∆R) for the three most proximal sites to Canímar Abajo range between 33 and 146 yr (∆R weighted mean 46 ± 40 yr; Druffel 1982; Lightly et al. 1982). Adding to

4

M Roksandic et al.

Corrected Proofs

this uncertainty, Wagner et al. (2009) more recently recorded an average ∆R value of –32 ± 25 yr from modern corals derived from four sites within the Gulf of Mexico and southern Caribbean Sea. There is also a reasonably distinct possibility that past ∆R values differed from the modern estimates as a result of local or regional (circum-Caribbean) changes in the supply of 14C (stored as dissolved inorganic 14C, DI14C) to the upper ocean mixing layer from the atmosphere (recent 14C) and the deeper ocean (older stored 14C). Evidence suggesting an ITCZ position that was more proximal to the Caribbean region than at present, and the associated reduction in prevailing easterlies, increased precipitation and consequent runoff regime during the mid-Holocene could have also resulted in a reduction in the (DI14C) reservoir exchanged between the Caribbean mixed ocean layer and the deep ocean (Haug et al. 2001; Poore et al. 2004; Greer and Swart 2006). This, in turn, could have resulted in marine reservoir age offsets of less than the standard 400 yr and subsequent ∆R values that are significantly more negative. In light of the different hydroclimatic regime that are proposed to have existed during the midHolocene it is reasonable to assume that the ∆R values used to calibrate the OC dates would also differ from the modern circum-Caribbean ∆R estimates. Druffel et al. (2008) reported a reservoir age correction of 292 ± 30 yr from mid-Holocene coral sequences (Biscayne Park Florida ~150 km NE from Canímar Abajo) dated to 3040 and 4930 cal yr BP. Accordingly, a ∆R of –108 ± 30 yr is used for the OC collagen calibrated ages. Cooper and Thomas (2011) found that a ∆R= –70 ± 40 yr value (Islas Tortugas, Marine Reservoir Correction Database, Reimer and Reimer 2001) provided the best overlap between site congruent wood and shell age ranges (2σ) collected at the archaeological site of Los Buchillones (22°22′20″N; 78°48′10″E) on the northern coast of Cuba, 308 km east of Canímar Abajo (Figure 1). The Los Buchillones ages ranged between cal AD 1280–1480 (2σ), chronologically younger but close to the Canímar Abajo YC ages. Therefore, it also seems reasonable to employ the more contemporary and established ∆R = –70 ± 40 yr correction for the YC calibrated dates relative to the Caribbean hydroclimatic changes that have likely occurred since the mid-Holocene (see above). The Canímar Abajo age model calculations require marine diet intake estimates, which were calculated using adjusted carbon and nitrogen isotopic foodweb compositions of adult [both male and female) Canímar Abajo residents (Chinique de Armas et al. 2015)] in conjunction with the stable isotope analysis in R (SIAR) Bayesian multisource stable isotope-mixing (Moore and Semmens 2008; Jackson et al. 2009; Parnell et al. 2010). The δ15Ncol and δ13Ccol isotope values for the OC and YC Canímar Abajo residents (Chinique de Armas et al. 2015) were used in conjunction with probable diet protein source nitrogen and carbon isotopic compositions, nitrogen and carbon concentration dependencies, and variable δ15N and δ13C diet-collagen fractionations, to estimate (using SIAR 4.2) marine diet intakes of approximately 30% (mode values) for both the OC and YC Canímar Abajo adult residents. RESULTS

Twelve dates were obtained: seven for the older cemetery (OC) and five for the younger cemetery (YC). These data, with the results of the calibrations generated by the method described earlier, are presented in Table 1. The calibration plots are presented in Figure 3. Interpretation of these results is based on the 2σ standard deviations, which provide a confidence level of 95% that the calendar date of the death of the organism—human individuals in the case of the assays on bone collagen, and a tree in the case of the charcoal—falls within the date range. It is tempting, but unwise with a statistically small sample, to attempt to use the 1σ standard deviations

Corrected Proofs 14


5

(only 68% confidence level) or the peaks in the graphs to infer a narrower timespan within which the burials occurred. Table 1 AMS (collagen) and conventional (charcoal, shell) 14C dates (BP) calibrated to calendar years BC/AD using the CALIB 7.0.2 mixed Marine & NH Atmosphere, IntCal13 and Marine13 (including a 400-yr offset; Stuiver and Reimer 1993; Reimer et al. 2013) age calibration data sets for the collagen (C:N atomic weight ratios all within the range of 2.9–3.5; van Klinken 1999; Chinique de Armas et al. 2015), charcoal, and shell samples, respectively. Sampling levels (m) are provided for the conventional (charcoal and shell) dates; na = δ13C values not available. Canimar Abajo sample name

Lab code AMS dates AA101056 E-77 AA89060 E-92 AA89062 E-79 AA89064 E-72 AA101055 E-7 AA101059 E-105 AA89063 E-19 AA101052 E-120 AA89061 E-15 AA101057 E-113 AA101054 E-119 AA101053 E-120 Conventional dates UNAM-0714a C-119, 0.2 m UNAM-0717 C-118, 0.40 m UNAM-0716 C-153, 0.45 m UNAM-0715 C-119, 0.6–0.7 m A-14315 C-96, 0.9–1.0 m A-14316 C-96, 1.8–1.9 m UBAR-170 C-157, 1.6–1.7 m UBAR-171 C-157, 1.8–1.9 m

Sample type

δ13C (‰ VPDB) Date (BP)

Cal AD/BC (2σ)

collagen collagen collagen collagen collagen collagen collagen collagen collagen collagen collagen charcoal

–19.7 –18.1 –16.1 –14.0 –19.1 –20.0 –16.3 –15.0 –14.1 –15.6 –15.3 –25.6

1289 ± 46 1420 ± 59 1536 ± 51 1617 ± 46 1661 ± 52 2791 ± 51 2922 ± 34 2946 ± 57 2960 ± 33 2996 ± 53 2999 ± 61 3057 ± 39

cal AD 690–950 cal AD 600–860 cal AD 440–670 cal AD 420–620 cal AD 360–600 cal BC 990–800 cal BC 1130–910 cal BC 1220–910 cal BC 1210–980 cal BC 1290–950 cal BC 1370–940 cal BC 1380–1090

charcoal charcoal charcoal charcoal charcoal charcoal charcoal shell

–25.8 –27.3 –26.2 –26.9 –28.2 –26.3 na na

800 ± 50 2520 ± 60 3460 ± 60 6460 ± 15 2515 ± 75 2845 ± 90 4200 ± 79 4700 ± 70

cal AD 1060–1290 cal BC 800–430 cal BC 1920–1630 cal BC 5480–5380 cal BC 800–420 cal BC 1260–820 cal BC 3010–2500 cal BC 3330–2920

The pattern of the 2σ date ranges is readily apparent and remarkably consistent. All dates correspond without exception to stratigraphic context. The dates separate into two independent clusters associated with the OC and YC, respectively, with no chronological overlap. Within each cluster, there are no statistical outliers (2σ standard deviations that do not overlap with at least one other 2σ date range). The 2σ standard deviations of AA101052 (bone collagen) and AA101053 (charcoal), taken from the same individual burial context, overlap significantly. With a high degree of confidence (95%), we can state that the OC burials that were assayed were interred between 1380 and 800 cal BC (2σ), and that the YC burials that were assayed were interred between cal AD 360 and 950 (2σ). Within the OC cluster, six of the seven dates (AA101052, AA101053, AA101054, AA101057, AA89061, AA89063) are statistically identical (all have significant 2σ overlap). These may indicate that the OC was used for a relatively short period, and possibly for a very brief period, sometime between 1380–910 cal BC (2σ). Within a small statistical sample, however, AA101059

6

M Roksandic et al.

Corrected Proofs

Figure 3 Radiocarbon dating results from Canimar Abajo

(990–800 cal BC; 2σ) cannot be considered an outlier. As noted above, there is no stratigraphic differentiation of burials with the OC to confirm or disconfirm chronological position. The standard deviations of the five dates corresponding to the YC show some statistical overlap, but not the same pattern that occurs in the OC. At face value, dates indicate continuous use of the cemetery over the cal AD 360–950 (2σ) timespan, but this inference must be tested with further 14C dates. We cannot estimate the period of time within which the midden deposits accumulated with any degree of confidence. We can only argue that the midden was formed sometime between 990– 800 cal BC (2σ) and cal AD 360–600 (2σ). While the obtained dates are consistent with stratigraphy, we have no evidence to infer that we assayed the latest burial interred in the OC or the earliest burial interred in the YC. Until dates from the midden context become available, we are unable to narrow this timeframe any further, for either the beginning or the end of midden deposition. DISCUSSION

The chronological pattern indicated by the AMS 14C dates is clear and unequivocal, but the limitations on chronological inference from these results need to be discussed. It bears repeating that the dates reported here provide a chronological framework only for the 11 burials for which we were able to obtain reliable assays. Eleven individual interments out of a total of 213 is not a statistically reliable sample. We cannot claim to know when either of the burial contexts was first or last used. In addition, we do not know when deposition of the intermediate shell midden began or ended.

Corrected Proofs 14


7

That being said, there are some parameters that can be placed on the timeframe within which these contexts occurred. One important implication of the study is that there appears to have been little (or no) displacement of human bone between the OC and YC as the result of bioturbation or other agencies. This evidence suggests that no YC interments extended into the OC. It follows from this that interment in the OC ended before cal AD 360–600 (2σ) at the latest, and most likely several hundred years before this to allow sufficient time for the midden deposits to build up. Using the same rationale, the earliest burial in the YC was interred by 990–800 cal BC (2σ) at the earliest and probably several hundred years after. Although the eight conventional 14C dates on charcoal and/or shell assayed previously are far less reliable than the AMS 14C dates on human bone collagen, they will be briefly discussed to assess their reliability. The results of the conventional dates are presented in Table 1, with associated stratigraphic context. As can be seen, all but two of the conventional dates correlate at least generally with the stratigraphy and provide support for the chronological pattern discussed above (Figure 2). UNAM-0714a (cal AD 1060–1290; 2σ) was assayed on charcoal recovered 20 cm below site surface. Its stratigraphic and/or cultural context is unclear; thus, it provides no additional evidence. Four of the conventional 14C dates were assayed on charcoal recovered from midden deposits. A-14315 (800–420 cal BC; 2σ) was taken from the bottom of the midden and UNAM-0717 (800– 430 cal BC; 2σ) from the top of the midden layer. The calibrated dates fall within the earlier half of the timespan between the OC and the YC. Because these dates are virtual duplicates, they may indicate that the midden was laid down immediately after use of the OC, with a hiatus of several hundred years before the YC burials began. Although tentative, this interpretation reinforces the observation that the OC and YC are independent burial contexts separated by a chronological hiatus of several hundred years. The other two assays were taken on charcoal from the top of the midden. They both produced dates much older than any of the AMS dates associated with the OC; their possible significance is discussed below. A-14316 (1260–820 cal BC; 2σ) was assayed on charcoal from an OC context in association with the same burial from which bone collagen was assayed for AA89061. Its 2σ standard deviation falls completely within the overall date range for the OC; it may be considered reliable and lends support to the AMS date ranges for the OC. The final four conventional dates yielded dates older than the earliest standard deviation of the AMS dates for the OC. Two of these, UBAR-170 (3010–2500 cal BC; 2σ) and UBAR-171 (3330– 2920 cal BC; 2σ) were yielded by charcoal and shell, respectively, from OC deposits, and are, therefore, stratigraphically tenable. The other two, UNAM-0716 (1920–1630 cal BC; 2σ) and UNAM0715 (5480–5380 cal BC; 2σ), were returned from charcoal assays recovered from the top of the shell midden deposits (as noted above). Both are clearly out of stratigraphic context, although the position of either one (or both) could be the result of bioturbation. UNAM-0715, with a range of 5480–5380 cal BC (2σ), is a statistical outlier; there is a 2050-yr gap between the standard deviations of UNAM-0715 and that of UBAR-171, the next oldest date. That being said, the results of the AMS dating neither confirm nor disconfirm any of the older conventional dates. Because we cannot argue that any of the AMS assays date the earliest OC interment, we cannot argue with certainty they do not date older OC contexts. It is also possible that they date a cultural context that predates the OC by up to 4000 yr. 14C dates from other sites in Cuba, albeit conventional, also do not preclude the age of these dates, including UNAM-0715. Prior to further AMS 14C dating, inference of deposits earlier than 1380–1090 cal BC (2σ) must remain tentative and unconfirmed.

8

M Roksandic et al.

Corrected Proofs

CONCLUSIONS

The combination of stratigraphy and AMS dates on human bone collagen reported in this paper provides one of the most reliable chronometric dating of any cultural context during this timeframe in the Caribbean (Fitzpatrick 2006; Rodríguez Ramos 2010; Fitzpatrick and Giovas 2011). There are, however, a number of important questions that remain to be answered. We do not know the actual temporal parameters of either the OC or YC. Further AMS dates from human bone collagen will provide the larger sample necessary to refine the chronology of these burial contexts. We need AMS dates from the midden deposits in Level 3. This layer is not, in fact, uniform, but consists of numerous substrata and lenses composed of shells, charcoal, and ash. We need to investigate the possibility of cultural deposits at Canímar Abajo that predate 1380–1090 cal BC (2σ) and, if such deposits are identified, determine if they relate to the OC or an earlier occupation of the site. The chronological framework thus far established for Canímar Abajo will allow us to pursue a number of important issues related to the site and to the community to which it relates. Mortuary practices from this time period in the Caribbean are not well understood. Subsistence regime, diet, and nutrition are all topics that can be fruitfully investigated at Canímar Abajo. Cultural practices including dental modification are in evidence in the OC. One of the individuals buried in the OC, from which bone collagen was recovered, yielded starch grains in dental calculus that are typical of bean (Fabacaea), some of which are consistent with common bean (Phaseolus vulgaris). If substantiated, this starch provides the earliest direct evidence of domestic plant utilization in the Caribbean (Chinique de Armas et al. 2015). Dental calculus from an individual from the YC yielded three starch grains that are possibly consistent with maize (Zea mays). Common bean and maize are important cultigens that are not native to the Caribbean. Stable isotope analysis of human bone samples recovered from burials in the both the OC and YC and reported by Buhay et al. (2013) indicate that the people buried at Canímar Abajo may have consumed a diet that included significant amounts of domesticated plant species. Pursuit of these and other questions may now be conducted on a firmer chronological basis. ACKNOWLEDGMENTS

We acknowledge research support by Social Sciences and Humanities Research Council of Canada Standard Research Grant. We thank Weldon Hiebert (the University of Winnipeg) for his diligence and expertise in producing the Figure 1, map of the study area, and Figure 3, the 14C dating results; and Dr Greg Hodgins (NSF-Arizona AMS Facility at the University of Arizona) for providing the bone collagen extraction protocol. REFERENCES Buhay WM, Chinique de Armas Y, Rodriguez Suarez R, Arredondo C, Smith DG, Armstrong SD, Roksandic M. 2013. A preliminary carbon and nitrogen isotopic investigation of bone collagen from skeletal remains recovered from a Pre-Columbian burial site, Matanzas Province, Cuba. Applied Geochemistry 32:76–84. Chinique de Armas Y, Buhay WM, Rodríguez Suárez R, Bestel S, Smith D, Mowat SD, Roksandic M. 2015. Starch analysis and isotopic evidence of consumption of cultigens among fisher-gatherers in Cuba: the archaeological site of Canímar Abajo, Matanzas. Journal of Archaeological Science 58:121–32. Cooper J. 2010. Pre-Columbian archaeology of Cuba: a

study of site distribution patterns and radiocarbon chronologies. In: Fitzpatrick S, Ross A, editors. Island Shores, Distant Pasts: Archaeological and Biological Approaches to the Pre-Columbian Settlement of the Caribbean. Gainesville: University of Florida Press. p 81–107. Cooper J, Thomas KD. 2011. Constructing Caribbean chronologies: comparative radiocarbon dating of shell and wood artefacts from Pre-Colombian sites in Cuba. Achaeometry 54(2):401–25. Druffel EM. 1982. Banded corals: changes in oceanic carbon-14 during the Little Ice Age. Science 218(4567):13–9. Druffel ER, Robinson LF, Griffin S, Halley RB, South-

Corrected Proofs 14


on JR, Adkins JF. 2008. Low reservoir ages for the surface ocean from mid-Holocene Florida corals. Paleoceanography 23:PA2209. Fitzpatrick SM. 2006. A critical approach to 14C dating in the Caribbean: using chronometric hygiene to evaluate chronological control and prehistoric settlement. Latin American Antiquity 17(4):389–418. Fitzpatrick SM, Giovas CM. 2011. New radiocarbon dates from the Grenadine Islands, West Indies. Radiocarbon 53(3):451–60. Greer L, Swart PK. 2006. Decadal cyclicity of regional mid-Holocene precipitation: evidence from Dominican coral proxies. Paleoceanography 21:PA2020. Haug GH, Hughen KA, Sigman DM, Peterson LC, Rohl U. 2001. Southward migration of the Intertropical Convergence Zone through the Holocene. Science 293(5533):1304–8. Jackson AL, Inger R, Bearhop S, Parnell A. 2009. Erroneous behaviour of MixSIR, a recently published Bayesian isotope mixing model: a discussion of Moore and Semmens, Ecology Letters, 2008. Ecology Letters 12:E1–E5. Lighty RG, Macintyre IG, Stuckenrath R. 1982. Acropora palmate reef framework: a reliable indicator of sea-level in the western Atlantic for the past 10,000 years. Coral Reefs 1(2):125–30. Martínez A, Rives A, Baena G. 1993. Área arqueológica Canimar-Morato-Yaití, provincia de Matanzas. La Habana: Editorial Academia. Moore JW, Semmens BX. 2008. Incorporating uncertainty and prior information into stable isotope mixing models. Ecology Letters 11(5):470–80. Parnell AC, Inger R, Bearhop S, Jackson AL. 2010. Source partioning using stable isotopes: coping with too much variation. PLoS ONE 5(3):e9672. Poore RZ, Quinn TM, Verardo S. 2004. Century-scale movement of the Atlantic Intertropical Convergence

9

Zone linked to solar variability. Geophysical Research Letters 31(12):L12214. Reimer PJ, Reimer RW. 2001. A marine reservoir correction database and on-line interface. Radiocarbon 43(2):461–4. Reimer PJ, Bard E, Bayliss A, Beck JW, Blackwell PG, Bronk Ramsey C, Buck CE, Cheng H, Edwards RL, Friedrich M, Grootes PM, Guilderson TP, Haflidason H, Hajdas I, Hatté C, Heaton TJ, Hoffmann DL, Hogg AG, Hughen KA, Kaiser KF, Kromer B, Manning SW, Niu M, Reimer RW, Richards DA, Scott EM, Southon JR, Staff RA, Turney SM, van der Plicht J. 2013. IntCal13 and Marine13 radiocarbon age calibration curves 0–50,000 years cal BP. Radiocarbon 55(4):1869–87. Rivero de la Calle M. 1987. Reporte de la presencia de sacros con espina bífida en el sitio aborigen de Canímar Abajo, Matanzas, Cuba. Revista Biologia 1:75–83. Rodríguez Ramos R. 2010. Rethinking time in Caribbean archaeology: the Puerto Rico case study. In: Fitzpatrick SM, Ross AH, editors. Island Shores, Distant Pasts: Archaeological and Biological Approaches to the Pre-Columbian Settlement of the Caribbean. Gainesville: University Press of Florida. p 21–53. Stuiver M, Reimer PJ. 1993. Extended 14C data base and revised CALIB 3.0 14C age calibration program. Radiocarbon 35(1):215–30. van Klinken GJ. 1999. Bone collagen quality indicators for palaeodietary and radiocarbon measurements. Journal of Archaeological Science 26(6):687–95. Wagner AJ, Guilderson TP, Slowey NC, Cole JE. 2009. Pre-bomb surface water radiocarbon of the Gulf of Mexico and Caribbean as recorded in hermatypic corals. Radiocarbon 51(3):947–54.