Caribbean Archaeology


The Archaeology of St. Eustatius


St. Eustatius (or Statia) differs from the larger islands around it (Fig. S-1 and EUX-3): the smallness of the island probably made the whole island exploitation area of the inhabitants of one settlement in one period, and did not permit two groups of any size to live on the island together and exploit it.
During the early preceramic period one group lived in the Corre Corre Bay area, leaving us two archaeological sites, of which one site is relatively large and the other small (no 4 in Fig. EUX-3) relatively small. During the Saladoid period one group lived at Golden Rock (site 1 Fig. EUX-3). During the post-Saladoid period one group lived at the Godet-Smoke Alley area which may be considered as two sub-sites left us by one post-Saladoid group, in view of the nearness of both (sub-)sites and similarity of artefacts (sites 2 and 3 Fig. EUX-3).
The following text is a summary of two publications: the Golden Rock monograph by Versteeg & Schinkel (eds.), 1992 and the paper in Analecta 26 by Versteeg, Schinkel & Wilson. Some data are presented here for the first time, and some data stem from the three chapters on St. Eustatius (by Versteeg, Louwe Kooijmans, and Delvoye) in Versteeg (ed.), 1994.


The earliest inhabitants

The earliest, preceramic inhabitants of St. Eustatius left the marks of their activities near the only large coral reef of the island, at Corre Corre Bay. These two sites (no 4 Fig. EUX-3) manifest themselves as shell, shell tool, coral and stone tool deposits. Only test pits have been made in them. The larger one of the two sites (Corre Corre-2) has artefacts to a depth of ca 70 cm below the surface. Some shells from these layers were dated at 2740 ▒ 40 BP (GrN-17071); a sample collected at a depth of ca 40 cm in the other Corre Corre site (Corre Corre-1; Fig. EUX-4) was dated at 2400 ▒ 50 BP (GrN-17073). This means that, after calibration and after correction for marine reservoir effect of 430 years, the former sample dates to 410-208 BC and the latter to 98 BC - 120 AD. The inhabitants obviously exploited the coastal resources nearby, especially shellfish. These sites still have to be excavated in detail in the future.


Large-scale excavations at Golden Rock and Smoke Alley

The only way to get an idea of the structure of a settlement is by using an excavation technique with large-size pits. Such methods were used at the Golden Rock and Smoke Alley sites. An essentially uninterrupted area of 3300 m2 was excavated in the former site (Fig. EUX-5); only less than 10% had to remain unexcavated between the large excavation pits.
It is evident that large-scale excavations, and a detailed analysis of all features (discolorations that are the result of Indian activities), are necessary to obtain basic data for the archaeology of the Caribbean at the lowest level: that of the individual structure, and that of the settlement. Such data yield detailed information on the prehistoric Caribbean villages, and that means on the communities that lived in these villages.


The Golden Rock site

Golden Rock is the largest and most important Saladoid period archaeological site of St. Eustatius. Valuable information on this site was gathered when sets of large and deep features were met in the area around the midden. These proved to be sets of postholes from which the shape of structures could be reconstructed.
These floor plans of Saladoid period houses range from small one-family compounds (8 m in diameter) to malocas 19 m in diameter (Fig. EUX-2, EUX-5). Dating of the well-preserved large charcoal samples from several of the postholes resulted in a tight framework of chronological control for the site (both the midden and the houses) and a chronological sequence for the houses, ranging from ca. 600 AD to ca. 900 AD (Versteeg & Schinkel, 1992:202-209). A plaza with ceremonial features (caches) related to burials was situated to the north of the houses, and the midden was situated south of the houses, slightly uphill from the flattest part of the plain.
There is no doubt that midden, plaza and houses are structural elements of one settlement. Several phases in the subsequent use of the houses could be distinguished.

The GR-1 excavations have provided many new, and partially unique, data on the structure of a Saladoid settlement. Proof was found that middens have, in a geographical (spatial) perspective, a direct relationship with houses (living areas). The houses are characterized by a round shape, and strong, vertical supporting posts which were founded several meters deep in the ground (Fig. EUX-6). The shape and construction of such houses were not known from the pre-Columbian archaeological record of the Caribbean and Northern South-America, but they are well-known from the ethnographic record of the South-American mainland (Fig. S-6a, Fig. S-6b; Fig. S-7; Fig. S-11).
The Golden Rock site is situated more or less precisely in the middle of the island (Fig. EUX-3), at 1-1.5 km from both coasts, and at ca 4 km from the north and the south point of the island. The site is also situated in the middle of the Cultuurvlakte, and this 560 ha plain was completely available for cultivation of cassava, and of other crops. It offered space for ca 860 slash-and-burn plots of ca 0.65 ha size. Allowing for a 30-year interval between use and reuse, 25 - 30 plots could be exploited simultaneously. Estimates from the Amazonian ethnographic record suggest that this number and size of plots provides the cake of the starchy cassava root daily to 75 - 100 individuals.
There is no doubt that protein was primarily procured from marine resources: the archaeological record and the isotope analyses of human bone are in agreement here. Meat from iguanas, agutis, rice-rats, birds, and landcrabs offered, without doubt, a welcome change to the usual meat intake, which consisted of fish and shellfish. A probable deficiency in the archaeological record (the lack of a representative number of conch, Strombus gigas) prevents an estimation of the relative importance of the latter food sources: fish may or may not have been a more important protein supplier than shellfish. The isotope analyses favor an important position for shellfish, but as yet it is impossible to come to pertinent conclusions. Shellfish in large quantities were probably available from the sandy flats in front of the beaches (Strombus gigas), and from the rocky shores (Cittarium pica), on both the windward and leeward sides of the island.
Apart from the Cultuurvlakte-exploitation for cassava cultivation, and the exploitation of the shores and beaches on the leeward and windward sides of the island, the Northern Hills provided clay within 1.5 km of Golden Rock.
The most probable freshwater extraction location is near the cliffs on the leeward side of the island. Coral is available at several locations around the island: the largest coral occurrence at present being the Corre Corre Bay on the leeward side of the island at ca 4 km from the site. There are no indications that this was different during Indian times.
These data suggest that the Golden Rock site was located in a central position in respect to the locations that were exploited for the commodities which were preserved in the archaeological record. These data suggest that the whole island was an exploitation zone of the GR inhabitants.
Wood was only preserved in the form of charcoal. Nevertheless, several wood species could be identified. Van der Valk's studies suggest that a semi-decidious forest was growing in the Cultuurvlakte. So, most probably, wood was available in large quantities near the settlement. Completely lacking in the archaeological record are boats. The best landing and storage location is in the Godet/Smoke Alley area. Here no coastal cliffs are to be found, and the sea is calm. Interestingly, in the pottery repertoire of the Godet/Smoke Alley post-Saladoid site(s) a small percentage of Saladoid sherds occurs (Versteeg, Schinkel & Wilson, 1993). Combining these data, one may speculate that activities around a canoe landing place in the Godet valley are the origins of this archaeological material. This is the only indication of a possible satellite site for special activities in the Saladoid period. An attractive interpretation of the Smoke Alley data associates the house structure excavated there in 1989 with a time frame of the early GR-1 settlement.
Another commodity of which we have no trace, is feathers. Amongst many present-day Indian groups in Northwest Amazonia, feathers are used for ceremonial purposes. Some of the bird bones in the GR-1 midden may well be the end result of feather procurement activities rather than of meat procurement activities. Most of the bird species identified for GR-1, such as the magnificent frigatebird and flamingo, have beautiful feathers indeed.
Not many finds suggest commodities of which the raw material does not exist on St. Eustatius. The GR archaeological record suggests a primarily self-supporting community, but, in view of the important place of trade-networks between tribal groups in the Amazonian ethnographical record, it is possible that the GR community participated in a comparable network. We found only a few indications: Van der Valk did not find the flint of GR-1 on the island. Probably it originates from St. Martin. The only items excavated at Golden Rock which we can interpret with confidence as having been prestige items are 81 quartz beads, found in a child's grave (Burial 11, Chapter 4). This type of clear quartz does not occur on St. Eustatius; it is unknown where the nearest provenance is for such quartz as a raw material. The scarcity of such status-indicators supports the interpretation (which is common in the ethnographical record of comparable Amazonian societies) that the Golden Rock society did not feature important status differences. Combining all evidence, the best interpretation of the excavated material seems to be that GR functioned on the level of a Complex Tribe. This also is supported by the technological level of the GR-1 inhabitants: the quality and complexity of the houses, especially the malocas was high. The (ceremonial) pottery was of equally high quality.
Golden Rock is a Late Saladoid site, inhabited in the 7th, 8th, 9th century AD; earlier dating results can only been explained as the results of incipient activities on the Golden Rock location, c.q. as a terminus post quem for the dating of the settlement (the hearth datings below the midden). The settlement was abandoned in the 9th century for unknown reasons. The last maloca did not burn down. The fact that posts were removed from this structure suggests that inhabitation of the island continued.
The GR pottery repertoire (Fig. EUX-1a) is uniform, including zoned-incised-crosshatched decorations in all layers of the midden. On the basis of the extensive Golden Rock evidence, for sure this type of Saladoid decoration cannot be considered to belong exclusively to the early phase of the Caribbean Saladoid. Unfortunately it is difficult to compare Golden Rock results to those of other Saladoid sites on the neighbouring islands. Not much except pottery and food items and some datings have been reported from these sites. The pottery repertoire seems to be a stable factor in Saladoid times and, on the basis of the insights we have now, it probably is not the best instrument to come to decisions on contemporaneity of Saladoid sites.

One of the concepts that resulted from these Golden Rock investigations is cultural continuity. On the basis of the uniformity between the Golden Rock houses and the relatively recent Indian houses in northern South America, it is evident that the basic ground plan of Indian houses is a concept hardly subject to changes through time. Even nowadays, modern building materials being available, the ground plan and the shape of the heavy post-frame remain the same. Only the most labor-consuming parts, like the roof and wall, are constructed of modern building materials (Kloos, pers. comm.). This concept has a long history, going back at least 1200 years. This also remains true for the lay-out of the settlement, still found today in the tropical lowlands, indicating a remarkable degree of cultural continuity through time. The similar lay-out of houses and village, then and now, points to shared ideas about the organization of the physical world. These ideas are structured by cosmology, which represents an organizing principle for Lowland societies across Amazonia and perhaps for the Caribbean as well. It appears that the basic cosmological notion, that is circularity in time and space, did not change between Saladoid times and the 20th century (Siegel, 1992:388-9). This circular lay-out of village/settlement and houses, and of the concepts of the makers, such as this is convincingly argued by Siegel for the Maisabel Saladoid and post-Saladoid settlement in Puerto Rico, is strikingly supported by the GR-1 houses and settlement lay-out.


The Smoke Alley Site (no 2 in Fig. EUX-3)

A post-Saladoid midden was found by Van der Valk on a small terrace (ca. 60 x 20 meters) some 8 meters above the sea at a distance of 50 meters from the sea (Fig. EUX-12). It is situated 250 meters south of the coastal Godet Site (Haviser, 1985; Versteeg et al., 1986), discovered by A.E. Figueredo in l975. Both sites share a number of characteristics: both are predominantly post-Saladoid, judging from the ceramics, but both have a small number of typical Saladoid sherds. At least 90% of Godet is covered under a l.5 - 2 meter overburden (Fig. EUX-11)dating from the early Colonial period when large-scale deforestation activities caused landslides and other erosion effects.
The Smoke Alley site, less prone to colluviation, only received some 40 cm of soil overburden. More 18th century buildings are near the site; the terrain was still used in this century for agriculture. The archaeological context of the artefacts in the site has been greatly disturbed by this agricultural activity. Insight into the total sub-surface area of the Smoke Alley terrace suggests that these large-scale disturbances destroyed all archaeological information that was not protected by the midden. South of the midden 18th century building made prehistoric remains invisible, and north of the midden a whitish-colored fill predominated.
Excavation of five 1 m▓ test pits established that the midden was approximately 10 x 10 meters in size. Excavation of four large machine-made (255 m▓, Fig. EUX-7) pits resulted in the discovery of 80 features. Sixty-five of them proved to be postholes. Twelve postholes belong to one slightly oval 10-meter diameter house, 7 others belong to a round 8-meter diameter house (Fig. EUX-8).
The Smoke Alley pottery from the test pits and from the features of the machine-made pits was studied by C.L. Hofman and T. Hamburg (Hamburg, 1992). This study was on 2077 pottery sherds from the excavation pits and 187 from the features. Of these sherds, 6 are certainly decorated Saladoid ones. If the average of the Golden Rock relation between decorated - undecorated Saladoid sherds (ca. 21%) is used (established on a sample of more than 40,000 sherds, see Versteeg and Schinkel, 1992:64) for this sample, then some 30 sherds of the total were Saladoid. Even allowing for the uncertainty of the low number of six, it is improbable that more than 5% of the Smoke Alley ceramics are of Saladoid origin.
Thirty-three decorated post-Saladoid sherds were found, representing about l.5% of the total pottery recovered. This figure is supported by field observations of the Godet excavations, where, among thousands of post-Saladoid sherds, only a very small percentage was decorated.
Typical post-Saladoid decorative modes such as wide-line incisions and wide-line curved patterns ending in punctations are characteristic of the Smoke Alley decorated sherds (Fig. EUX-9). Some of these patterns are reminiscent of Esperanza pottery (Hamburg, 1992:4). Three test pits (Nos. 1, 3, 5) had many West Indian Top Shells (Cittarium pica) and some conches (Strombus gigas). The same three pits featured most pottery sherds.
Two features in the margin of the midden contained a Hawksbill turtle (Eretmochelys imbricata), and a zemi. Several layers of West Indian Top Shells (Cittarium pica) and crab pincers were found. This latter complex is interpreted as a cache.
It is striking that both houses and also two burials are situated exactly at the location of the midden. The refuse of the inhabitants of the houses was probably thrown down the slope, and from there subsequently was washed into the sea or into areas lower on the slope where erosional material was accumulating. The Smoke Alley midden itself probably originates from one or more houses higher on the slope. Dating the houses and associated archaeological features has presented some problems. There are three 14C samples: one is recent (160 ▒ 70 BP; GrN-18448) and obviously not associated with Indian activities. One is bone material of the human burial F. 50. This sample reflects the age of the post-Saladoid activities on Statia: 1105 ▒ 30 BP (GrN-17070), which, after adjustment for Marine Reservoir Effect (- 130 y) and calibration, results in an age between ca 1000 - 1160 AD. A third sample, charcoal from the lowest level of the midden, was dated at 1720 ▒ 30 BP (GrN-17072), which, after calibration results in an age between ca 250 - 400 AD. This result reflects the age of the Saladoid, and not post-Saladoid activities. The latter date corresponds very well with the earliest dates (samples GrN-11512 and 11513, cf Versteeg & Schinkel, 1992:204) from Golden Rock. These relatively old charcoal samples may be the result of the first Saladoid activities on Statia, such as clearing the vegetation of settlement locations.
The postholes of both houses contain no charcoal, and little pottery or shell. The houses obviously did not burn down. They were built before the midden, and their life cycle ended before the midden accumulated. This implies that they are either Saladoid (but up to now no Saladoid refuse of any extent has been found in this part of the island), or post-Saladoid, but older than the Smoke Alley post-Saladoid midden.


Back to top of the page

A total view of all the pictures

Back to main page

Back to main page

To contact the author send an E-mail


Hamburg, T., 1992. Het aardewerk van Smoke Alley, St. Eustatius. Unpublished manuscript, Leiden University. Leiden.

Haviser, J.B., 1985. An Inventory of Prehistoric Resources on St. Eustatius, Netherlands Antilles. Proceedings of the 10th International Congress for the Study of the pre-Columbian Cultures of the Lesser Antilles:61-81.

Siegel, P.E., 1992. Ideology, Power, and Social Complexity in Prehistoric Puerto Rico, (2 vols.). Thesis State University of New York. Binghamton. 805 pp.

Versteeg, A.H. & C. Schinkel (eds.), 1992. The Archaeology of St. Eustatius: the Golden Rock site. Publication of the St. Eustatius Historical Foundation 2, Publication of the Foundation for Scientific Research in the Caribbean Region 131. 284 pp.

Versteeg, A.H., C. Schinkel & S.M. Wilson, 1993. Large-scale excavation versus surveys: examples from Nevis, St. Eustatius and St. Kitts in the Northern Caribbean. Analecta Praehistoria Leidensia 26:139-161.

Versteeg, A.H. (ed.), 1994. Between St. Eustatius and the Guianas. Publications of the St. Eustatius Historical Foundation 3, 99 pp.

Back to main page

Copyrightę 1998, Archeologisch Museum Aruba, St. Eustatius Historical Foundation, Stichting Surinaams Museum, A.H.Versteeg

Last update July 2000