Archaeology and Conservation : The Discovery of the Writing-Tablets and the Archaelogical Context

From Alan Bowman and David Thomas, Vindolanda: the Latin writing tablets London: Society for the Promotion of Roman Studies, 1983, pp. 19-24

I. Archaeology and Conservation

A. The Discovery of the Writing-Tablets and the Archaeological Context

The first of the Vindolanda writing-tablets was discovered in March, 1973, by Robin Birley in the course of his excavation in a deep trench at the south-west corner of the third-century fort at Vindolanda1. Birley describes the object as two small thin fragments of wood which looked rather like oily plane shavings. These slivers, when gently prised apart, proved to contain writing on their inner faces but this rapidly began to fade on exposure to the air2. This first example, published below as No. 38 (346), was the forerunner of over two hundred fragments (inventoried as Nos. 1-83) discovered during the excavation season of 1973. The curtailed excavation season of 1974 produced very few further finds but in the following year more extensive digging resulted in the discovery of several hundred more fragments, bringing the inventory numbers up to 202. Although this batch did contain some interesting texts, it is fair to say that the majority of our better tablets come from the finds of the first season3. Any statement about possible future discoveries can only be speculative, but it does seem very likely that there are considerable numbers of writing-tablets still in the ground in this area of the site. The great expense involved in this operation (both in excavating to the necessary depth and in conserving the finds) makes it difficult to predict when large-scale excavation might be resumed4.

The nature of the collection cannot properly be described without some analysis of the archaeological context5. The tablets were almost all6 found in a deep section which was cut, as has been said, at the south-west corner of the third-century fort (see FIG. 1). The material from this area (of various kinds, organic and artificial) belongs to a sequence of forts dating to the pre-Hadrianic period. It is probable that two separate forts should be identified, the second being much larger than the first and having at least two, perhaps three, phases of occupation. Table I shows how the phases are archaeologically distinguished as well as the approximate dates assigned to them7.

Table I

Table II shows the distribution of writing-tablets in the various archaeological layers. The leaf tablets are designated L and the stylus tablets S 8. The former category includes some blank fragments but excludes a number of pieces of wood which appear not to be from writing-tablets (this explains why the total is less than the original 202 inventory numbers).

Table II

It will be observed that writing-tablets are present in all phases of the pre-Hadrianic period except that which represents the first fort. The majority of tablets comes from Layers 10/8 (ca.A.D.95-105) and of these most originate in Layer 8. Most of the small number of stylus tablets, however, come from the later phase represented by Layer 6 (ca.A.D.105-15). No doubt this is pure chance and of no significance from the point of view of the type of writing-material used at different periods. We must, of course, take into account not only the vertical but also the lateral distribution of the tablets. As may be seen in FIGS. 1-3 the great majority of these tablets was found in a fairly restricted area, with the exception of two odd tablets which were some considerable distance away9. The clustering is suggestive but we cannot be sure of the significance of the strays until further excavation has taken place. As for the place concentration, it seems more than coincidence that the tablets were deposited in the same small area over several occupation-levels, but we find this very difficult to explain convincingly. Birley has suggested that the find-spots might be characterised as follows:

If this is correct, we might tentatively suggest that the tablets from Layers 10/8 come from a rubbish dump in an area near the fabrica (which itself might be close to the praetorium). This may account for the presence of urine and excreta in the deposit and also for the fact that some of the tablets show signs of burning10. The tablets from Layers 6 and 5 will then be simply the occupation-debris of what was possibly a barracks building.

It is therefore clear that the collection as a whole cannot be described as archival - we do not have the Romano-British counterpart of the Files of cohors XX Palmyrenorum 11. There are, however, three small archives in the collection: groups of letters connected with Flavius Cerialis (Nos. 21-29 (248) (250) (263) (270) (247) (280) (519) (290) (241)), with Crispinus (Nos. 30-33 (295) (297) (520) (298)) and with Flavius Genialis (Nos. 34-6 (218) (171) (223)). We think that the first two were praefecti of the units stationed at Vindolanda during the period represented by 10/8 (ca.A.D.95-105)12. We have no clue to the position of Flavius Genialis.

In order fully to understand the circumstances in which this collection of written material was preserved we have recourse to the expertise of the environmental scientist as well as to that of the archaeologist13. The timber buildings close to the south end of the pre-Hadrianic forts contained layers of flooring which were composed predominantly of compacted bracken. This flooring, belonging to a timber building of Period II (Layer 8), contained the first writing-tablets to be discovered in 1973. Apart from the bracken, the flooring contained straw and many other organic components (gorse-pods, heather, twigs) and the residue of human occupation, bones, oyster-shells, large quantities of leather, jewelry, cloth, wooden objects such as flesh-strippers and slickers and a considerable quantity of human excrement14. In addition, the layers were heavily saturated with urine, probably a legacy (like the flesh-strippers) of the tanning process which seems to have been carried out in the fabrica near this area. Pupae of a species of stable-fly were also found in a fine state of preservation15.

Several factors seem to have contributed to the survival of the writing-tablets in such a good state. First, there are clay compactions between the levels of occupation, and these created pockets of anaerobic conditions in which the tablets lay. Within these pockets, the presence of organic materials within the soil is also significant. The chemical conditions were created by the presence of bryophytes, whose green coloration was still apparent, straw, excreta and urine. Tannin produced by the organic materials (assisted by the presence of leather) enhanced the preservative factors. Finally, there are traces of vivianite or iron phosphate, which is partly caused by the accumulation of bones of horses, oxen, pig and duck16. There can be no doubt that these conditions are responsible for the relatively fine state of preservation of the tablets and it is probably worth making a comparison with conditions at Vindonissa, where considerable numbers of writing-tablets were found in the 1920s amongst the contents of a rubbish dump17.

We should note, however, that the conditions in which the tablets from Layers 6 and 5 were found do differ significantly from those in Layers 10 and 8. The floors of Layer 8 were composed almost exclusively of bracken and were thinner than those elsewhere, whereas those of Layers 6 and 5 were made principally of straw and contained no traces of urine or excreta. On the whole, we suppose that the anaerobic conditions in Layers 10 and 8 had the effect of preserving the writing which, in several cases, rapidly began to fade on prolonged exposure to air. But since this phenomenon was not consistent and since equally well-preserved examples were found in Layer 6, we cannot make any generally valid deductions18.

It may be that the failure to discover similar written material on a significant scale at other sites is due not so much to the fact that tablets have not survived in the ground as to the fact that excavators have probably failed to recognise them for what they are19. Recent discoveries at Ribchester and at Lechlade suggest that in future such finds might be more frequent20.