All dates in the text have been converted to calibrated calendar years (cal BP; Table

1). A median basal date of 10,057 cal BP[1] was obtained for the 210-200-centimetre interval in Bass Pond through bulk radiocarbon dating (Table 1). This date has an unusual ¹²C/¹³C ratio (R. McNeely pers. comm.). It is possible that “radioactively dead” carbon in the underlying carbonate rocks may have been incorporated into the sediment, producing a date that appears older than the actual age of the sediment. This seems likely since the Bass Pond sediment core contained high quantities of carbonates. Similar contamination effects have been noted at other sites in Atlantic Canada (Wolfe and Butler 1994; Stea and Mott 1998; Andrews et al. 1999) and elsewhere.

Radiocarbon dating using the atomic mass spectrometry method was carried out on terrestrial plant fossils from five other levels in the core (Table 1). Ages for other depths in the core were estimated by linear interpolation.

In BPC-1, an increase in warm-adapted midges (e.g., Chironomus and Dicrotendipes) was observed (Figure 4). Chironomus is predominant in temperate lakes with summer bottom-water oxygen depletion (Walker 1987). Dicrotendipes is a midge widely distributed south of treeline (Oliver and Roussel 1983). These data may relate to a warming trend often seen in the early Holocene.

A decline in temperate midge taxa (e.g., Chironomus and Dicrotendipes) was noted in BPC-2 (Figure 4). An apparent lack of cold-adapted species is generally noted in this zone, although towards the end of the zone a few fossils of the cold-water midge Heterotrissocladius were noted. Heterotrissocladius is abundant in arctic lakes as well as cold bottom waters of deep, pristine lakes in southern Canada (Walker et al. 1997). Since this pond is shallow, there is no summer refuge for cold-water insects in this lake. Consequently, most of the climate-inferred assessments in this study are based on the presence/absence of warm-adapted midge taxa which are mostly characteristic of shallow, near-shore environments.

The significant reduction in overall midge abundance at the beginning of zone BPC-3 marks this zone change (Figure 4). The head capsule concentration decreased at the 110-centimetre interval, approximately 6500 cal BP (Figure 5); the average count below 100 centimetres was 325 head capsules per cubic centimetre, and the average above 100 centimetres was 86 head capsules per cubic centimetre. An increase in the abundance of several warm-adapted midge taxa (e.g., Microtendipes, Polypedilum, and Stempellinella/Zavrelia) was also observed. Both Microtendipes and Polypedilum are midges widely distributed south of treeline (Oliver and Roussel 1983).

In Zone BPC-4, Cladopelma, Microtendipes, and Stempellinella/Zavrelia increased, at least initially, whereas Chironomus, Dicrotendipes, and Polypedilum decreased (Figure 4). All of these taxa are warm-adapted (Walker et al. 1997). Towards the end of zone BPC-4, several taxa ( Stempellinella/Zavrelia, Endochironomus, Glyptotendipes, and Microtendipes) declined. All of these taxa are warm-adapted with distributions extending little, if at all, into arctic or alpine environments (Walker et al. 1997).

In BPC-5, a resurgence of temperate midge taxa (e.g., Chironomus, Dicrotendipes, and Microtendipes) was observed (Figure 4). Dicrotendipes and Microtendipes are characteristic of warm, shallow-water habitats (Walker et al. 1997). Walker et al. (1997) have estimated the temperature optimum for Chironomus at 24.7℃; however, the genus is widely distributed, with some Chironomus species’ distributions extending even into the high arctic (Walker 1988).

A midge-temperature inference model developed by Walker et al. (1997) was used to reconstruct temperature changes from Bass Pond’s midge fossil record (Figure5). Due to a lack of good analogues for the data point at 170 centimetres, it has been removed from the reconstructed palaeotemperatures. Apart from the basal sample (170-165 cm), early and mid-Holocene inferred summer water temperatures (Zones BPC 1 to 3) are generally warm, varying between 19 and 23℃. A distinct cooling trend is indicated beginning in BPC-4, approximately 4900-2400 cal BP. The lowest inferred temperature, about 16℃, was reconstructed for about 2200 cal BP. This temperature minimum was followed by an inferred warming trend, with a maximum temperature of 22℃ being reconstructed at the 20-centimetre level in BPC-5.

It is important to note that the temperature inferences are for maximum summer surface-water temperatures, which are typically 4℃ higher than mean July air temperatures (Livingstone et al. 1999). Since Bass Pond is a very shallow lake, its surface-water temperature would tend to be slightly higher, at least briefly during the summer. The current mean July air temperature at Rocky Harbour, Newfoundland is 15.6℃; thus, a maximum summer surface-water temperature of $20℃ would not be unusual today.

The palaeosalinity reconstruction for Bass Pond (Figure 5) indicates a decrease in salinity from 2.7 grams/litre initially (subsaline) to less than 0.6 grams/litre by 7000 cal BP. The reconstruction indicates freshwater conditions thereafter, with the exception of a salinity peak of 1.9 grams/litre around 2000 cal BP. The peak salinity at 2000 cal BP could record an exceptional storm surge, for example, linked to a hurricane. This idea is speculative. The salinity reconstruction should be verified using additional indicators (e.g., diatoms).

The pattern of deglaciation and accompanying sea level changes has allowed many coastal archaeological sites to be well preserved in the Port au Choix region of Newfoundland (Tuck 1976; Grant 1994; Liverman 1994; Renouf 1999; Bell et al. this volume). The region offers researchers an opportunity to explore potential correlations between the long human cultural record and diverse other palaeoenvironmental data. Although it has been suggested that patterns of cultural migration cannot be inferred from vegetation history (Fitzhugh and Lamb 1985), previous palynological research (Macpherson 1995b) suggested that major climatic trends inferred at Bass Pond could be correlated with settlement patterns of both MAI and Dorset Palaeoeskimos.

The first evidence of the MAI (6290 cal BP; Bell and Renouf this volume) suggests their arrival at Port au Choix during a warm period in BPC-3 (Figure 5). Sustained MAI occupation (4800 cal BP; Bell and Renouf this volume) began during an inferred cooling period at the beginning of BPC-4. They disappeared more than 1500 years later with the beginning of a more drastic and prolonged inferred cooling trend (Figure 5).

At about 3000 cal BP, the Groswater Palaeoeskimos began to occupy the area (Bell and Renouf this volume). The midge-inferred temperature record suggests that the Groswater Palaeoeskimos occupied the region through a prolonged cool phase (Figure 5). The appearance of the Dorset Palaeoeskimos (around 2000 cal BP) occurred at a time when midge-inferred temperatures began to increase (Figure5). The Recent Indian culture appeared as this warming trend continued. Whether these cultures became extinct as the region was subsequently repopulated by other people, or whether their populations just migrated to other areas is still open to debate.

It thus appears that summer temperature may be a significant indicator for the succession of cultures as revealed in the Port au Choix archaeological record. Since chironomids are thought to be among the best indicators of local climate change (Battarbee 2000), midge stratigraphies may provide key evidence for interpretation of the prehistoric cultural record.

The initial occurrence of Cricotopus ornatus type chironomid remains at Bass Pond, and the midge-inferred palaeosalinity record both suggest a decreasing salinity trend at Bass Pond through the early Holocene, with freshwater conditions being established prior to around 7000 cal BP. Statistical tests suggest that the midge fauna at Bass Pond may portray a stronger palaeosalinity than palaeotemperature signal. The first axis of the midge fossil data in detrended correspondence analysis correlates more strongly with the palaeosalinity record (r = -0.88 on log (salinity)) than with the palaeotemperature record (r = 0.14).

The relatively great abundance of grasses (Poaceae) and other herbs in the pollen diagram from Bass Pond may be indicative of a salt marsh environment during Bass Pond’s initial stages (Macpherson 1995b). Macpherson (1995b) places the isolation of Bass Pond at approximately 7500 cal BP (~130 cm) due to the rapid increase in the freshwater alga, Pediastrum, above this level. This estimate is 1000 years earlier than proposed by Grant’s (1994) and Bell et al.’s (this volume) sea-level curves. As salinity decreased, so did the head capsule concentration in the sediments, indicating a transition from the more productive salt marsh to the less productive upland environment around 6500 cal BP (Figure 5). There is some variation in the emergence dates suggested by these indicators. The precise timing is difficult to resolve since the transition was likely very gradual, and different indicators will respond to different aspects of this transition. For example, the vegetation response may indicate when much residual salt had been leached from the marsh, and the resulting soil was sufficiently well drained to allow trees and shrubs to colonize the former marsh. The composition of the midge community was likely directly impacted by salinity, but the head capsule concentrations more likely reflect a slower decline in overall nutrient availability.

Interpretation of the later salinity peak around 2000 cal BP is more speculative, but may be due to a marine incursion (storm surge or tsunami) or the reworking of marine sediments. Intense tropical storms are common in Atlantic Canada. In contrast, little is known about tsunami frequency. Newfoundland was hit by a 12-metre tsunami in 1929 following an earthquake centred 350 kilometres south of the island (Liverman et al. 2001). Such events probably also occurred in the past. Bell et al. (this volume) link this salinity peak to anthropogenic factors.

The pattern of sea level regression and transgression along the west coast of Newfoundland varies markedly along a north-south axis. Since deglaciation, sea level has continuously fallen at the northern tip of the island (a type-A sea-level curve; Bell et al. this volume). However, farther south at St. George’s Bay (Bell et al. 2003), sea level reached a lowstand about 9700 cal BP, and sea level has been slowly rising for several millennia (producing a type-B relative sea-level curve ). Bell et al. (this volume) place the transition between these two patterns of sea-level change to the south of Port au Choix. The coastline at Port au Choix would therefore have emerged continuously since deglaciation. Emergence was likely very rapid initially, but may have been much slower after 7000 cal BP (Bell et al. this volume). This slow emergence of Bass Pond would be consistent with the gradual transition we perceive in the pollen and midge fossil data, thereafter allowing occasional salt-water incursions to have occurred.