Go to the Main Scottish Cave and Mine Database Search Page

Ulva Cave

Ulva, SW of A' Chrannag, Argyllshire.

A raised beach Sea Cave, Livingstone's Cave is on Ulva, a small island to the west of the Island of Mull. Excavations undertaken in this cave in the 1980s and 1990s by the University of Edinburgh uncovered Mesolithic shell middens and later material, providing evidence of the use of this cave extending back more than 8000 years ago. The cave faces south-southeast.

The cave is in a small cliff on the hillside. A dark hole in grey basalt, it is intriguing and curious, otherworldly even. Climbing the short hill slope and approaching from below, along a path through tall grass and bracken, drystone walls and tumbled stone are remnants of recent use of the cave. The dim interior is a surprisingly large space. With light streaming in from outside it is not gloomy, instead it is dry and spacious, liberally sprinkled with bird droppings. A couple of startled pigeons fly past out into the sunlight. The people who used this cave 7000 or 8000 years ago left behind their refuse to be picked apart by archaeologists. The same space was used by hunter-gatherer groups, sporadically perhaps, for a period of 2 or 3000 years who left deposits of marine molluscs, fish, animal and bird bone and burnt plant remains - the left-overs from meals. Flint and bone tools, and shell artefacts; the remains of lives lived. And occasional human bones. Reminders of real people. [KirstyMill]

Two periods of fieldwork totalling 7 weeks were undertaken in 1989 using teams of undergraduate students from Edinburgh University, the first in March-April, and the second in July. Two areas within the cave were investigated.

Area B

In the north-west part of the cave interior removal of the thin surface layer of animal dung exposed an area of stone pavement constructed of densely packed, angular fragments of basalt. This overlies clayey sediments of (?)Late Pleistocene age, and in one area was found to seal a pit containing burnt fragments of bone, shells and charcoal. Mammalian bones and flint artifacts were also found occasionally at the boundary between the dung layer and the clay. A detailed plan of the stone pavement and the associated features was made.

Entrance (Area C)

The deposits of the entrance area of the cave were investigated by a 10 x 1 metre sondage. This provided a section through the "wall" of basalt stones and boulders constructed across the cave entrance and through the midden deposit on the inner side of the "wall". The midden had a maximum thickness (in the excavated section) of 0.35 metres. It is composed largely of shells of limpets (Patella vulgata) and periwinkles (Littorina littorea), but contains bones of mammals and fish, fragments of crabs' claws, and flint artifacts (cores, flakes, bladelets). The midden and the rubble "wall" are underlain by a series of very stony sediments from which flint artefacts and faunal material were also recovered. Samples of limpet shells from the base and top of the midden were submitted for radiocarbon dating. When adjusted for the apparent age of sea water, the dates indicate that the midden was deposited between ca 7650 and 5680 BP. The date for the base of the midden establishes that the history of occupation of Ulva Cave extends back to at least the Mesolithic. The radiocarbon dating programme is being extended to date the deposits underlying the midden, and to establish the chronological relationships between the deposits of the entrance area and those in the cave interior.

The fieldwork on Ulva received support from the Department of Archaeology Fieldwork and Research Fund, the Carnegie Trust for the Universities of Scotland, the Munroe Fund, the Royal Archaeological Institute, the Society of Antiquaries of London, and the Society of Antiquaries of Scotland. The British Academy provided a grant towards the cost of post-excavation work and radiocarbon dating.

A total of six weeks fieldwork was undertaken in 1990 during the Easter and Summer vacations. Further work was undertaken in the entrance zone (Area C) and the cave interior (Areas A, B). The Mesolithic shell midden (Feature 3) was exposed over an area of 16 square metres in the SW part of the entrance zone, in preparation for systematic excavation in 1991. The overlying talus deposits proved to be quite complex stratigraphically; those directly overlying the midden produced a number of flint artifacts and several sherds of Early Neolithic Unstan Ware (identified by Ian Armit). Area B was extended to expose the stone pavement (Feature 34) over the entire NE part of the cave interior; much time was devoted to making a detailed plan and photographic record of this feature. Area A was extended horizontally as far as the SW wall of the cave. The major feature uncovered in this area was an extensive deposit of ash (Feature 26) up to 25cm thick, overlying a 'floor' of large basalt stones. As yet, this feature is undated but its stratigraphic position suggests that it post-dates the Mesolithic occupation of the cave. Four radiocarbon dates were received from the SURRC Radiocarbon Laboratory at East Kilbride, bringing the total so far to six. Two dates are of particular significance: 7800 +/- 160 BP (GU-2704) for an organic-ring layer in Area A, and 4990 +/- 60 BP (GU-2707) for charcoal from a pit in Area B containing burnt bone, shell, and carbonized cereal grains. GU-2704 dates the earliest remains of Postglacial occupation in the cave interior and is indistinguishable at the 1 sigma level from the date of 7660 +/- 60 BP (GU-2600) for the base of the midden. GU-2707 demonstrates Early Neolithic use of the cave, and accords with the presence of Unstan Ware in post-midden contexts in the entrance zone.

Excavation: (a) Area C: Midden (Contexts 3, 80): Excavation was begun of an area c.3.5 x 4.0m between the 1989 sondage and the north-west wall of the cave. The excavation was based on a 50cm-grid, and the deposits removed in approximate 5cm "spits". Three-dimensional recording was employed for major finds - artifacts, fragments of mammalian bone, concentration of fish bones, rounded/burnt stones and rarer species of shell; all excavated material was transported to Edinburgh for processing and analysis. Only the upper part of the midden was removed in 1991, the number of excavated spits varying from one at the north-east end of the trench to eight at the south-west end. Artifacts recovered from these upper levels include pottery, artifacts of flint and pitchstone (including two pieces with bifacial retouch), and a bevelled bone implement ("limpet scoop"). The latter is the first artifact of "Obanian" type to be recovered from the site. The presence of pottery and bifacial implements in spits 1-7 suggests that these upper levels of the midden contain material which relates to early Neolithic use of the cave.

(b) Area B: Excavation and recording of the stone pavement (context 34) were completed and investigation of the complex set of deposits around the Neolithic pit (Context 8) was begun. These deposits were overlain by the stone pavement and, in turn, overlie till-like sediments of presumed Late Devensian age. Finds to date include pockets of marine shells (mainly Littorina sp. and Patella sp.) often associated with irregular lenses of black soil, occasional animal bones and teeth, charcoal fragments and a single fragment of Neolithic pottery.

This cave continues to yield results of importance. It is now certain that the midden in the entrance zone accumulated over a very long period of time, from at least c.7800 BP. It is the clearest example so far recovered from Britain of a midden deposit relating to both Mesolithic and Neolithic activity in a single site. Early Neolithic use of the cave is also indicated by the pit in Area B, which has provided one of the earliest radiocarbon dates for the Neolithic in western Scotland (Bonsall et al, 1991), and the earliest evidence for cultivation of oats.

Grants to support the fieldwork were provided by The British Academy, The Carnegie Trust, The Department of Archaeology Fieldwork and Research Fund, The Society of Antiquaries of Scotland, and the University of Edinburgh. The British Academy also provided funds for post-excavation work. [C Bonsall, N J Russell and G M Coles 1991].

8/8/23. After five years, Edinburgh University continue excavation of the Mesolithic site in Ulva Cave (Livingstone's Cave). This is a nationally important site that has revealed evidence of human use from 8000 to 5000 BP (Before Present). Dr Catriona Pickard is leading the excavation with students from Edinburgh, and we hope to take some video of their excavation, along with an explanation of the site and post it here. Visitors are very welcome during the excavation - those who are able to walk to the site - and there are a limited number of places for volunteers. We also hope to give a talk about the site during the two weeks that Edinburgh are here, somewhere closer to the ferry than the Cave itself - details to follow about that.

Alternate Names: Livingstone's Cave

Notes: Pliestocene funeral remains and possible food processing area. [Bonsall]

Livingstone was a famous African explorer and missionary. The cave was named because his grandparents lived here for a short time.

'Logaga lwa ga Kobokwe' in Botswana is also called Livingstone's Cave.

Hydrographic Feeds: None

Hydrographic Resurgences: None