Cardiff Studies in Archaeology - Specialist Report Number 18

THE FISH BONES

next page

By C. Ingrem

The initial examination of the fish bone assemblage concentrated on the lower two floor layers of the house on mound 3 and this is all that is discussed here. The assemblage recovered consisted of 2187 identifiable fragments and the range of species present is described in table 1.

The assemblage can be divided into two groups depending on the method of recovery and analysis. The residues were initially sieved through a 10 mm mesh to recover larger fragments of bone, stone and pot. The fish bones recovered from this sieving were all examined and identified where appropriate. The remaining residues were sub sampled and only a proportion of the total sorted and analysed. The estimated total quantity of fish bone in the samples examined is 195 from floor 1 (where very few samples were examined), 3941 from floor 2 and 3917 from floor 3. A striking feature of the assemblage is the difference in species composition between residues below 10 mm and the residues above 10 mm. The former are dominated by herring whereas the latter only provided a single herring bone. This yet again demonstrates that an accurate understanding of a fish bone assemblage is only possible by fine sieving (Colley 1987).

The herring bones present are almost entirely represented by vertebrae. The posterior abdominal and caudal vertebra are dominant whereas the anterior abdominal vertebra are poorly represented. This suggests that the head of herrings were removed prior to consumption in the house. In contrast the large cod family species were distinguished by low numbers of caudal vertebra and large numbers of the head bones. This suggests that gadid heads, with the upper body attached, were being cooked and consumed in the house. The horizontal distribution of fish bones shows specific concentrations in both floors, however, these are not related to particular features such as the hearth.

Almost all the herring bones were from individuals between 150 and 300 mm in length. This suggests that the majority of herring caught were adults over three years old (Wheeler 1968). Throughout most of the year adult herring occupy deep water and their movements are erratic which makes them difficult to hunt. However, in the coastal waters of the Hebrides, in spring and summer, they come together at specific spawning grounds and it is at these spawning grounds that they were traditionally caught. Spawning grounds exist on either side of South Uist but there are significant differences between the eastern Minch spawning grounds and those to the west at the edge of the Continental Shelf. After spawning the herring larvae lying to the west at the edge of South Uist are carried north towards Cape Wrath by the strong Atlantic currents. In contrast larvae in the east are protected by deep sea lochs, they hatch in the Minch where they remain for their early years of growth. This results in a different age structure in the two populations; to the east adults and juvenile herring shoal together but to the west only adults are present. It is therefore possible that the absence of juvenile herring in the Bornais sample indicates fishing in the deep water to the west, off the Continental Shelf. The size of the cod found in these floors supports the idea that deep water fishing was practised at Bornais. However, the other species present indicate that inshore fishing was also undertaken.

The social implications of these results are important. Herring fishing undertaken at night with nets and is an activity normally involving a number of families in their own boats. This investment in boats and nets implies a long term commitment to the task. The shoals have to be specifically located and this can involve long periods of waiting. If a shoal is successfully located the quantities of fish caught can be enormous and the processing and disposal of these fish require a large labour force and a well organised exchange network. The communal nature of the task contrasts with cod fishing which is generally believed to involve individuals using long lines (Barrett et al 1999).

These results are particularly important when compare with the pattern of fishing identified in the Northern Isles. In a recent study of this region (Barrett et al 1999) it was noted that evidence for the exploitation of herring was remarkably rare. Only 77 herring bones were identified from a total of 380,000 fish bones recorded and it was hypothesised that many of these could derive from the gut contents of large cod family fish (gadoids). The material recovered from Bornais represents the first Norse assemblage systematically recovered through fine sieving from the Western Isles. Since these discoveries were made an assemblage comparable in date has been examined from Bosta in Lewis (Ceron Carrasco pers comm) that is also dominated by herring. These discoveries were completely unexpected as the known assemblages of Late Iron Age date (Ceron Carrasco in Parker Pearson and Sharples 1999) were dominated by large gadoids, and are similar to assemblages in the Northern Isles. They suggest the Norse communities in the Western Isles made a conscious decision to exploit the herring and that this marked a break with traditional practices.

next page