News

Cave of the Speckled Horses, Fiskavaig, Skye

Posted by steven on 10/04/2008 at 09:14 AM

Read on for more details about the rock shelter at Fiskavaig.....

Uamh an Eich Bhric (Cave of the Speckled Horses), Fiskavaig, Skye

Introduction

The rock shelter and midden known as Uamh an Eich Bhric, or Cave of the Speckled Horses, lies below west-facing sea cliffs some 3 kilometres south-west of the village of Fiskavaig and some 400 metres south of Geodh’ an Eich Bhric. The cliffs at this point are around 110 metres high and are composed of basalt lava’s, laid down over several successive volcanic events during the Tertiary Period between 60 and 50 million years ago. Access to the site by land is difficult and involves a 90 minute walk across wet undulating moorland, followed by a steep descent down a 100 metres high grass slope. This leads down to the boulder shore, where a 400 metres scramble along the high tide mark leads to the rock shelter. The alternative to the walk out to site would involve a rather hazardous landing by boat on the rock-strewn and pebble beach, which is only accessible during the high tide. Waterfalls cascade from the cliff-tops to the beach on both sides of the rock shelter, creating a dynamic setting for the site.

The rock shelter at the base of the basalt cliffs - note the two figures (one in red) for scale

Another view of the rock shelter with the towering cliffs above dwarfing the site - spot the people inside the shelter

Discovery

In the later summer of 2006 Mr. Rod Richard of Strathpeffer informed John Phillips (Highland Council Ranger for Skye & Lochalsh) that he had seen bones exposed in a steep sediment and boulder bank at the base of the cliffs, while walking on the shore near Fiskavaig. John Phillips in turn passed on these details to Martin Wildgoose of Archaeological and Ancient Landscape Survey and along with Steven Birch and George Kozikowski, visited the site in October 2006. Our first impressions were of a massive midden within the rock shelter, in an incredible location in the landscape, on the limits of land and sea.

The visible remains comprised a sea-washed section of what was once the fill of a large rock shelter located just above the High Water Spring Tide level. The eroding section within the shelter measured 30 metres long by 3 metres high and visible in the face of this section, under more than a metre of accumulated roof collapse, was a layer of cultural deposits averaging between 0.4m and 0.6m thick.

Martin Wildgoose standing in front of the exposed section with the rock shelter beyond (scale=1m)

This cultural layer contains well-preserved animal bone, fish bone, shellfish, heat-cracked pebbles, charcoal, ash deposits and other organic residues. A closer inspection of the eroding section and the debris at its base, revealed pebble tools, several pottery sherds, bone tools, fragments of bronze (copper-alloy) and the complete upper stone of a rotary quern manufactured from Torridonian Sandstone. Distinct activity areas could also be recognised within the eroding face showing discete depsotional events; areas containing material potentially related to copper-alloy smelting or smithing, limpet and fish bone middens, and dumps of ash.

Martin standing in front of the eroding face of the section. The cultural midden is the orange to black-coloured band near the base of the scale

Storm action had already eroded at least 4 metres of the midden face at the south end of the site, with between 1 and 2 metres removed at the north end, while debris from the collpased section lay across several metres of the foreshore along the entire length of the midden face. During this first visit to the site the exposed and eroded cultural deposits, including the quern stone, were removed as the next big storm would sweep them away.

A closer view of a section of the midden in the eroding section showing bone, fire-cracked pebbles and a sherd of pottery (Scale=0.2m sections)

The upper stone of the rotary quern exposed in the section and ready to be claimed by the sea (Scale=1m)

This image shows the eroding section from above, with the partially exposed quern stone before removal (Scale=1m)

The recovered upper rotary quern stone (Scale=1m)

Site Recording and Monitoring

Several visits have now been made to the site. During the second visit, a series of base-line erosion markers were established so that we could monitor the loss of material from the site through time. A plan was also made of the rock shelter and associated features and profiles recorded through the site, while detailed section drawings have also been made of the eroding midden face. Photographic and video records have also been taken during the visits to site, while the inspection of the eroding midden face has produced a steady stream of small finds. 

A view from inside the huge rock shelter looking down onto the eroding section. The large incut shown in this image indicates where most of the cultural material has been lost

Martin Wildgoose recording the eroded section of the midden in October 2007

We knew from previous monitoring trips that the site suffered from increased erosion rates due to the combination of winter storms and high tides. Therefore, during a trip to the site in October 2007 we recovered samples from three vertical columns in the eroding face of the midden. By carrying out this task, we knew that we could at least salvage important deposits from the site before they were lost to the sea. Extracting the samples also allowed us to make a more detailed assessment of the midden, especially when combined with the detailed recording of the eroding section.

A close view of one part of the eroding section showing the midden containing the animal bone (Scale=1m)

In particular, this exercise allowed us to recover a large amount of animal bone from the midden, much of which appeared to be from cattle and sheep-goat, although we did find some bone from red deer and pig. All parts of individual animals appear to be represented including cranial fragments and teeth, and bones from the extremities such as hooves and lower limb bones. We also noticed that some of the bone appeared to be semi-articulated, almost as if animal carcasses had been deposited in the midden. During the removal of the samples of animal bone we also identified some clusters of hatched fly pupae cases, which were embedded between the layers of bone.

Semi-articulated vertebrae from cattle skeleton within the midden (Scale=0.5m)

During the sampling, we also recovered deposits of fish bone, shellfish, ash and charcoal; while small finds included fragments of a copper-alloy pin, a bone needle and pottery. Two samples of animal bone have been collected from the base and top of the bone midden for radiocarbon dating, and we are most grateful for the financial support received from Mrs. Suddaby and Ian Suddaby (Ian, who works for CFA Archaeology Ltd in Edinburgh, has been assisting us with fieldwork at the site) in processing these samples. The results of the carbon 14 assays are as follows:

1870+/-35BP (upper bone midden)

1905+/-35BP (lower bone midden)

This suggests that the deposition of the bone midden was over a limited time range, some time between 50AD and 240AD. The dates place the use of the site during the middle Iron Age, while the dates are similar to those obtained during the excavation of Dun Ardtreck by Euan MacKie. Dun Adtreck is located around 4.5km to the northeast of the rock shelter at Fiskavaig, in Loch Bracadale. Excavated between 1964 and 1965, the earliest date for the construction of the dun came out at 2005+/-105BP, but the occupation of the site continued for at least 400 years after this. The excavation produced quite a range of Roman items, especially for an Iron Age site situated on the Scottish west coast. 

Ian Suddaby working at the rock shelter during October 2007

A second rock shelter is located around 400 metres to the north of Uamh an Eich Bhric, which has not been subjected to the erosive forces seen at the latter site. No midden is visible within this shelter, but its form and morphology allows us the opportunity to see what the Cave of the Speckled Horses probably looked like before it was breached and partially destroyed by the storms in January 2005.

The image above shows George Kozikowski at the south end of the entrance leading into the intact rock shelter. Note the talus slope of grass-covered sediments and boulders at the front edge of the site protecting the interior

The image above shows Martin Wildgoose looking south out of the entrance of the rock shelter. The interior is quite dry and protected from the prevailing winds. The image below shows the Cave of the Speckled Horses from the south for comparison. The grass-covered talus at this end of the site is still complete, but the north end has been completely washed away leaving behind the remains of the interior of the rock shelter including the cultural deposits

The Cave of the Speckled Horses is located in a very unusual location at the foot of the west-facing cliffs near Fiskavaig. Seclusion was apparently more important than ease of access and the function of the site at present remains a mystery.....However, we now have a unique opportunity to unravel the story of the site and place it in its wider landscape setting. Excavations will commence at the site on Monday 14th April, so visit the ‘News Page’ again soon for regular updates on our progress.

Looking into the rock shelter during our latest visit to the site in March 2008

For details of the excavations at the Cave of the Speckled Horses see the ‘News Page’

Next entry: Excavations start at the Cave of the Speckled Horses at Fiskavaig....

Previous entry: Radio Carbon Dates and phasing at the High Pasture Cave Site

Add Your Comment Here

Commenting is not available in this section entry.