Internet Archaeol. 10. Dye.

6.0 Discussion and conclusions

Sinkhole site 9574 is the type site for the hypothesis that Polynesian activities led to extirpation of native, extinct land snail taxa at Kalaeloa. There is a decline in relative proportion of native, extinct taxa observed at site 9574 and this proves to be a regional phenomenon, which is present at all sites. The decline at site 9574 appears suddenly and relatively late in the sequence after a long period of relative stasis, and it led to extirpation of native, extinct taxa from the sinkhole. In this regard site 9574 is unique. The pattern of sudden, late change leading to extirpation discovered there is unlike that of any of the region's sinkholes excavated subsequently. These features of the site's land snail sequence, which are crucial for the argument that Polynesians were responsible for extirpation of native, extinct taxa, are not found elsewhere.

Our analyses indicate a regional pattern of long-term decline in the relative proportion of native, extinct taxa. (Curiously, L. gracilis is absent in the leaf litter along with native, extinct taxa. The other introduced taxon, G. servilis, is the dominant component of the leaf litter collection.) At all other sinkhole sites, except site 1710-1 whose basal deposits were laid down in a flooded environment, this decline begins at the base of the stratigraphic column and continues, as a first-order trend, through to the top of the column. Dating the onset of this long-term, first-order decline is critical to understanding its causes. On the basis of ¹⁴C dates from sinkhole deposits, the decline began before structural collapse deposits began to form, sometime before 50 BC-AD 950. This is perhaps 1,000 years before significant human settlement of the 'Ewa Plain, which is argued to be the 14th century AD at the earliest (Athens et al. 1997; Tuggle 1997; Davis 1990), although temporary base shelters for fishing expeditions might have been established somewhat earlier. The archaeological community is divided over the date Polynesians colonised Hawai`i, but if the estimate for a 'late' colonisation at the end of the first millennium AD proves correct, then the long-term decline probably began before humans set foot on the island. In either case, any human role in this decline is extremely unlikely.

We propose that lowering of the water table, probably caused by a drop in local sea level from its mid-Holocene high stand, is the explanation for the long-term decline in relative proportion of native, extinct taxa in Kalaeloa sinkholes (lowering of the water table has been frequently mentioned in regard to environmental change on the 'Ewa Plain (Allen 1995; Davis 1990; Christensen 1995), although this is usually attributed to causes other than sea level change). The mid-Holocene high sea stand on O`ahu is believed to have culminated 6,000-4,000 years ago at a height of 1.6±0.45m, with a drop to present levels about 2,000 years ago (Athens and Ward 1991, 1993; Fletcher and Jones 1996), a period that overlaps deposition of basal diagenetic sediments in sinkholes. Effects of this process show most clearly in the land snail sequence at sinkhole site 1710-1, in which basal diagenetic deposits are dominated by the aquatic taxon A. nitida. These deposits today stand a bit higher than 1.5m above mean sea level and only 1m above modern high water table (Davis 1995, 355), and would have been underwater during most of the mid-Holocene high stand. The proportion of A. nitida declines markedly in succeeding structural collapse and transported sediment deposits, as the water table dropped and the sinkhole environment became drier. Wetter conditions early in the regional sequence are also indicated by wet forest snail taxa recovered from basal diagenetic deposits at site 9647-2. Similar effects would have been felt at sinkholes throughout the region. Thus, one component of decline in relative proportion of native, extinct taxa does not indicate drastic ecological disturbance, and we interpret this trend as indicative of ecological succession in the context of natural environmental change. (Sea-level decline is unlikely to have been the only environmental change underway in the region. Other evidence indicates a pattern of increased rainfall, and possibly onset of ENSO conditions (Athens et al. 1997). If this were the case, it might have mitigated the effect of a falling water table on the land snails, producing a more gradual change than would otherwise have been the case.) Too litle is known of land snail ecology in Hawai`i to develop this hypothesis further.

The relatively great diversity of land snail collections from Kalaeloa sinkholes, compared to other site types, leads us to characterise the `Ewa plain as a patchy environment for land snails, where moist conditions inside sinkholes provided a relatively congenial habitat compared to harsh, dry conditions on the surface of the plain. Our analysis of change over time in diversity of native, extinct land snail taxa indicates that stability, rather than population stress, characterises most of the stratigraphic sequence. Sinkhole habitats might have provided a buffer against changes on the surface of the plain. Recent evidence from coring in a wetland at Kalaeloa is interpreted as indicating a drastic change in the dry forest cover of the region around AD 1000 unrelated to Hawai'ian farming or settlement of the region (Athens et al. 1997). If this change did, in fact, occur, it had no measurable effect on land snail faunas in sinkholes. Similarly, our analysis indicates that environmental consequences of Polynesian settlement at Kalaeloa, with the exception of the introduction of L. gracilis, were not felt by land snail populations of sinkholes. This does not necessarily indicate that Polynesian settlement was without environmental consequence, however. It does indicate that Polynesian use of this environmentally marginal corner of O`ahu did not result in ecological disturbance sufficient to extinguish land snail populations in sinkholes.

Diversity indices indicate a complex pattern of decline near the top of several stratigraphic columns, and we interpret this as indicating some stress in the population of native, extinct land snails in sinkholes. In two dated sinkhole sites, decline in diversity is associated with transported sediment deposits, and thus might date either to the end of the pre-Contact period or to the post-Contact period. We prefer the latter alternative because massive environmental changes to the region brought on by sugarcane cultivation in the 19th century and by the introduction of plants and animals that now dominate the flora and fauna of the region should have had an observable effect on sinkhole land snail faunas. If we are right about the importance of the water table to the sinkhole habitat, then water wells drilled to irrigate sugarcane fields after 1880, which lowered the water table significantly (Macdonald and Abbott 1970, 247), might have had an appreciable effect on land snail faunas of sinkholes.

The apparent simplicity of the karst setting of the 'Ewa Plain belies a complex environmental and human history. There have been a series of changes associated with sea level, rainfall patterns, Polynesian introductions and activities, and post-Contact land uses. We conclude that land snail extinctions at Kalaeloa resulted from the first and last of these factors, and that activities of Polynesians prior to the advent of European influence in AD 1778 contributed little, if anything, to the extinction process.