Understanding the pattern and process of the colonisation of the eastern Caribbean is still a work in progress. It has long been widely agreed that the first horticultural, pottery-making Amerindian settlers (usually referred to as Arawaks, or, archaeologically, as Saladoid cultures) originated from the general area of the mouth of the Orinoco River in Venezuela, with settlement established on almost every island by c. 2200 BP. But there are also indications of an earlier Archaic occupation of the Caribbean (Figure 1).
In the Greater Antilles, where the evidence is best, the dates reach back as early as 6000 BP. Once referred to as the "lithic age" (Keegan 1994), and conceived as involving hunter-foragers using stone and shell tools, these pre-Arawak cultures have now yielded evidence of more complex subsistence regimes and even ceramic production; colonisation from Central America has been proposed (Wilson et al. 1998) but remains unconfirmed. The earliest Archaic sites in the Lesser Antilles have somewhat later radiocarbon dates (Fitzpatrick 2006), mostly not before c. 4000 BP. These earliest settlers too probably arrived from South America. The island with the largest number of pre-Arawak sites is Antigua, where several dozen have been located (Davis 2000: 82)—perhaps unsurprisingly, since it has by far the best sources of flint and chert in the eastern Caribbean (Knippenberg 2007). Yet Archaic material has been identified on only five other islands in the Lesser Antilles. Thus, understanding this earliest phase of settlement requires more and better data, and any additional evidence is welcome.
Montserrat is a small island near the southern end of the Leeward Island chain, its nearest neighbour being Antigua. Since 2010 Montserrat has been the focus of an intensive survey conducted by Brown University and Wayne State University, the Survey and Landscape Archaeology on Montserrat (SLAM) project (Cherry et al. 2012). Over 30 years ago, Watters (1980) located half a dozen sites of both the Early and Late Ceramic periods, and SLAM's fieldwork has now doubled that figure (Figure 2). The most thoroughly investigated location, Trants, yielded radiocarbon dates around 2500 BP, placing the site among the earliest known Saladoid settlements in the eastern Caribbean. Yet, despite its proximity to Antigua, and that island's abundance of evidence for an Archaic phase, nothing earlier had been securely documented on Montserrat.
Lithic materials donated to the collections of the Montserrat National Trust, with the vague provenance "from Centre Hills", hinted that there might be earlier material to be found on the island. Intensive survey on the northern flanks of Centre Hills in 2012 indeed located such finds. It should be understood that Saladoid-period sites throughout the Lesser Antilles are characterised by a lithic technology aimed at the expedient production of usable flakes, either by freehand hard-hammer percussion of flint nodules, or by bi-polar, "hammer-and-anvil" techniques. Earlier lithic technologies, by contrast, are dominated by a blade-core reduction system, which stands out very markedly, even in surface-collected materials. At the site of Upper Blake's (Site 35), SLAM encountered a rich and spatially concentrated flint assemblage, which is the product of a blade-core technology that predates anything typical of the Ceramic Age (Figures 3 & 4). It includes percussion blade-cores, blades, backed blades and blade-flakes of various sizes, and associated débitage, including flakes, micro-flakes and shatter, all found in close proximity, suggesting that one or more knapping events are preserved on the surface. Some of the blades are as long as 16cm (Figure 5) and imply a macro-core/blade technology akin to that known from the earliest Archaic sites in the Greater Antilles.
This new evidence, inevitably, leaves some questions unresolved. The site's inland location, at an altitude of c. 1000 feet (300m asl), is unusual compared to the generally coastal settings of Archaic sites on other islands (although it may be noted that the seasonally-occupied site at Plum Piece on Saba is even higher, at over 1300 feet, 390m asl). Furthermore, the Upper Blake's site lacks evidence of shell tools or subsistence remains, perhaps unsurprising on a purely surface site that lies amidst intensively cultivated fields. Most critically, we can offer no absolute date for the site, since no stratified deposits were encountered from which radiocarbon samples could be taken. Nonetheless, we are confident that the techno-typological characteristics of the lithic materials provide very solid grounds for asserting the existence of an Archaic population on Montserrat, most likely between c. 4000 and 2500 BP. This is a significant addition to our knowledge of the earliest stage in the colonisation of the eastern Caribbean islands.
*Author for correspondence