Earliest Settlement in the Marianas-A Response
Hung, Hsiao-chun, Carson, Mike T., Bellwood, Peter, Antiquity
Some initial corrections
Winter et al. refer to the colonisation of the northern Marianas at a latitude of 18 [degrees] N. It is well known that the earliest Marianas sites appeared in Guam, Tinian and Saipan (Russell 1998). Guam, the largest and southernmost of the Mariana Islands, is positioned at 13-14[degrees] N. Saipan, the northernmost of the earliest inhabited islands, is 15[degrees] N. Winter et al. also propose a date of 3400-3200 BP for the Unai Bapot site, but it is actually dated to 3500 BP (Carson 2005, 2008; Carson & Kurashina 2012). The earliest Marianas settlement is securely dated to 3500-3300 BP at two sites in Guam, a further two in Tinian, and three in Saipan (Craib 1993; Butler 1994; Carson 2005, 2008, 2010, 2011, 2012a; Spriggs 2007; Carson & Kurashina 2012).
The early colonisation of the Marianas must have been difficult from any point of departure, given their location more than 2000km from any contemporary populated area at 3500-3300 BP. The other far western Micronesian islands of Palau and Yap were settled evidently no earlier than 3100 BP (Intoh 1997; Liston 2005). Present-day and historical records of winds and currents are of some interest, but we would caution against using such data to adjudicate ancient migration routes. Such logic led Thor Heyerdahl more than 50 years ago to claim that Polynesia had been settled via the Americas. Surely Winter et al. do not intend to convince us that the Remote Oceanic islands were populated entirely by unmediated drifting at sea? To date, no preserved Neolithic canoes have been excavated in either the Marianas or the Philippines, and as such, ancient sailing conditions cannot be known.
While an exact node of origin may never be known, the earliest Marianas pottery resembles a sub-set of findings in the Philippines, with some localised modifications, as expected in a classic founder-effect scenario. The Philippines pottery is best documented at Nagsabaran (Tsang et al. 2002; Hung 2008) and Magapit (Aoyagi et al. 1993) in northern Luzon, but other examples occur in the central Philippines, in the Batungan Caves on Masbate Island (Solheim 1968; Hung 2008), so travel from here or the eastern Visayas is also quite likely. We cannot yet accept possible sources south of the Philippines, unless some convincing evidence can be presented.
The ceramic analysis by Winter et al. validates some of our own findings, but it is incomplete. The authors confirm thin-walled pottery made with local clays and fine beachsand tempers at Unai Bapot, already reported previously (Carson 2005, 2008). They discern coil-building versus paddle-beating, but these are parts of a continuous construction and finishing sequence (Rye 1981; Rice 1987). Winter et al. do not tell us the primary-forming techniques of Unai Bapot potters, nor do they tell us the secondary-forming and finishing techniques of Nagsabaran potters. Our wider research on the earliest Marianas pottery traces the full process of initial slab-building and coil-building, followed by paddle-beating and trimming. We further found diagnostic paddle-impression marks in both the early northern Philippines and Marianas pottery (Carson et al. 2012).
Prehistoric populations will not have followed only one method of pottery manufacture, regardless of raw materials and specific cultural knowledge of the potters concerned. Ethnographic and ethnoarchaeological studies show that more than one type of manufacturing method could exist among potters, even within a single community or settlement (Longacre 1991; Stark 1999; Stark et al. 2000). Given the expected technical variation in pottery manufacture, the cross-regional consistency in decorative system is most remarkable. In this regard, we maintain our view that the earliest Marianas pottery was related more closely to contemporary traditions in the northern and central Philippines than to anywhere else. …