DESCRIPTION OF THE BLUEFISH CAVESLocation and environmentGeomorphological contextDepositPalynological dataPalaeontological dataChronological context
Location and environment
The Bluefish Caves (Note 2) are located 54 km southwest of the Vuntut Gwichin village of Old Crow, overlooking the middle course of the Bluefish River, a tributary of the Porcupine River (Fig.1). This is a region of Devonian limestone hills (Norris 1985) at the northern end of the Keele Range, which in turn forms the foothills of the great massif of the Ogilvie Mountains that mark the centre of the Yukon. These hills also mark the southwestern edge of the enormous network of lacustrine basins, which during the upper Pleistocene were inundated by the waters of the Bluefish, Old Crow and Bell glacial lakes (Hughes 1972). The environment is characteristic of boreal forest in mountainous regions (Ritchie et al., 1982), with spruce trees (Picea glauca and Picea mariana) on the pediments and more or less continuous tundra zones along ridges over 750 m high. The climate is typical of this boreal type of environment, and the landscape has been shaped by multiple periglacial processes characteristic of continuous permafrost regions.
The caves are found at the western extremity of a ridge dominating a narrowing of the Bluefish River. They are nestled at the foot of a series of limestone outcrops standing about 250 m above the river valley. There are three caves (I, II and III), actually small cavities whose volume ranges from about 10 m3 to 30 m3 (Fig. 2). Contrary to what had been suggested earlier (Cinq-Mars 1979), they appear to have been formed in a very broad geographical and chronological context namely, that of a regional karst landscape now under study (Cinq-Mars and Lauriol 1985; Lauriol et al. 1989; Roberge et al. 1986). In other words, these are not cavities formed by congelifraction, but rather the remains of a former, greatly reduced karst network uncovered by the erosion of the slopes.
Sampling and excavations both inside and outside the three caves, as well as subsequent sedimentary analyses, have revealed a depositional sequence characterised by a relative uniformity both in terms of the sedimentary sequence and the mode of deposition (Fig. 3). This sequence can be summarised as follows:
- The floor is a rocky surface (bedrock) studded with cryoclastic fragments, with some residual areas of highly altered sediments in fissures and depressions.
- The overlying sediment, of eolian origin, is relatively homogeneous loess. It may reach more than a metre deep and contains variable quantities of cryoclastic elements, mainly from the walls and ceiling of the cavities and from the cliff faces which rise about 10 m over the entrances.
- Results of granulometric and sedimentological analyses on samples collected under the dripline of Cave II suggest that, despite its apparent homogeneity, the loess can be divided into three facies which reflect both the transport conditions of the sediments and their place of origin (Note 3)
- Overlying the loess, gradually or discomformably, is a layer of humus-rich cryoclastic rubble which, outside of the caves, can be slightly more than a metre deep. The thickness of this unit declines rapidly at the cave entrances where it becomes, on the interior of the cavities, a simple organic enrichment of the upper portion of the loess.
- Finally, the surface of the fill is characterised by herbaceous and shrubby vegetation on the exterior of the caves, and by a discontinuous cover of ferns, mosses and lichens on the interior.
action of periglacial phenomena such as cryoturbation and congelifluction, and which can sometimes make interpretation difficult. These deposits do not lend themselves readily to a precise stratigraphic reading, and more than in most locations, decoding them requires the contribution of other types of data.
The analysis of sediments from the interior of Cave I and the exterior of Cave II has provided pollen diagrams which, despite the nature of the deposits, are indicative of a certain degree of depositional integrity. These diagrams have already been published (Cinq-Mars 1979; Ritchie et al. 1982) and so we will outline only the most significant details here:
- First, corresponding stratigraphically to the lower level of the loess, we find a pollen assemblage presenting the characteristics of tundra rich in herbaceous species.
- Above this, in the upper level of the loess, is a zone characterised mainly by an increase in, and predominance of birch (Betula).
- Finally, in the portion of the deposit which has been described as humus-rich rubble, a third assemblage is characterised by a discernible decrease of herbaceous species and a significant increase in spruce (Picea) and alder (Alnus).
This pollen biostratigraphy, though only weakly defined, holds up well in relation to the findings of other palynological research carried out in the region (Ritchie and Cwynar 1982; Ritchie 1984). It is therefore apparent that the lower portion of the loess contains traces of a xeric herbaceous tundra environment. Following these environmental conditions, which differ vastly from what we find today, there is a series of changes, starting with the development of shrubby birch tundra and leading eventually to the rapid establishment of boreal forest conditions quite similar to the ones we see at present time.
This correspondence provides several useful chronological markers. It would appear that the development of shrubby tundra, with increasing birch, dates to around 14,000-13,500 BP; and that conditions suitable for the development of boreal spruce forest in the region appeared around 10,000 BP. Finally, it must be emphasised that the correspondence between the Bluefish pollen diagrams and the regional ones supports the idea noted above of the relative integrity of the deposits, and this in spite of the inevitable periglacial scrambling.
All three caves have yielded thousands of bone remains, which are extremely well preserved due to the very favourable sedimentological and taphonomic conditions. The fauna comprises both large and small mammals, including a significant series of microtines, as well as birds and fish (Table 1). It can be divided into two series, found in the loess and in the humus-rich rubble respectively.
The loess contains a wealth of fauna, in terms of quantity, complexity and diversity, which correspond to Guthrie’s (1982, 1985) description of Beringian late Pleistocene mammoth steppe fauna. The megafauna includes horse (Equus lambei), caribou or reindeer(Rangifer tarandus), sheep (Ovis dalli), bison (Bison priscus), moose (cf. Alces alces), wapiti or elk (Cervus elaphus) and mammoth (Mammuthus primigenius). There are also saiga (Saiga tatarica), muskox (Ovibos moschatus), bear (Ursus), wolf (Canis lupus) and lion (Panthera).
In contrast, both quantitatively and qualitatively, the bone remains found in the humus-rich rubble show characteristic signs of a significant impoverishment of the regional megafauna, undeniable indication of mass extinction and extirpation on a regional and continental scale. In fact, this unit yielded only a few sparse specimens of caribou and sheep, a faint echo of regional Holocene megafauna that also includes moose, bear and wolf.
This contrast in the fauna, like that seen in the palynological record, is concordant with a definite chronostratigraphic break (loess/humus-rich rubble) (Note 4). Its significance, which remains to be fully understood, seems to be related to global ecological changes in Beringia and many other parts of the world during the last millennia of the Late Glacial (Lundelius 1989).
As we have seen, the sedimentological, palynological and palaeontological data enable us to place the cave deposits on a chronological scale that definitely includes the Holocene (humus-rich rubble) and the end of the late Pleistocene, or Late Glacial (loess). Several 14C measurements made early in the research (Cinq-Mars 1979; Morlan and Cinq-Mars 1982) roughly confirmed this chronostratigraphic estimate and even gave some indication of precision. We were able to date an episode of forest fire, definitely Holocene, in the (cryoturbated) upper sediments of Cave I. We also obtained a date of 12,900 BP from the femur of a horse, collected in the upper level of the lower loess of Cave I; and a date of 15,500 BP from a mammoth scapula found in the lower loess of Cave II, where the palynological signs of the herbaceous tundra were identified.
A new series of 14C dates obtained recently from specimens of megafauna (Note 5; Cinq-Mars and Nelson 1989) recovered from the three caves have enabled us to put forward a somewhat more detailed chronostratigraphy which sheds new light on a number of palaeontological, palaeoecological and archaeological issues.
For example, some of these findings can help us identify periods of extinction and extirpation and, in retrospect, furnish valuable information on the evolution of the environment. The example of the saiga is particularly instructive. Its ecological requirements (Vereshchagin and Baryshnikov 1982) make it an unusually fine indicator of a xeric environment (semi-arid steppe), relatively rich in herbaceous species and characterised by only slight accumulations of snow in winter. Until recently, the presence of saiga in eastern Beringia was dated prior to the Glacial Maximum (Harington 1980; Matthews 1982). However, its discovery in the sediments of Cave III indicate that populations of saiga, apparently living in an environment which was not altogether unfavourable, survived in far-eastern Beringia as late as about 13,400 BP. The overlap between this date and the age of the (palynological) birch zone which, as we have seen, is an indicator of a major ecological change, probably means that the Bluefish saiga is one of the last to have lived in these regions.
Besides providing clear confirmation of the loess/humus-rich rubble dichotomy and of its dating to about 10,000 years ago, many of these new chronometric results have led to an unexpected increase in the range of our chronostratigraphy. As we shall see below, some results indicate that the history of the deposit goes as far back as about 25,000 BP. More importantly, these results demonstrate that the mammoth steppe fauna mentioned above, in a combination that remains to be precisely determined, constituted an essential element of the Glacial Maximum biotope of eastern Beringia (between about 17,000 and 25,000 BP). Without elaborating further on this information, which sheds light on an important Beringian controversy (Note 6), we must emphasise that this demonstration of the viability of a Glacial Maximum environment is of major significance for our understanding of several aspects of Beringian archaeology.