4 Metal detecting in fieldwork

4.1 Introduction

Many professional archaeologists who have used detectors in fieldwork hold them in high esteem. Their benefits go far beyond the mere recovery of larger numbers of artefacts. At some sites they have provided the only means of dating specific periods of occupation. They can also recover types of object almost invariably missed by the naked eye - minute droplets from metalworking, for instance, which can provide the only evidence that the process had taken place. Metal detectors have also made impressive contributions to the detailed understanding of site layout.

Yet few of these benefits have yet been fully harnessed by archaeology. A fifth of the 50 professional units in our survey do not use metal detectors at all. Of the rest, few use detectors consistently on all sites likely to produce metal. Fewer than half the units in our survey possess their own metal detector, and in some cases the machine is old, obsolete and seldom used. Few units seem up to date with recent developments in detector technology, or aware of the potential of different types of machine - a position not helped by an extraordinary shortage of literature on the use of detectors in excavation and on their technical capabilities.

Part of the reason for this uncertainty about metal detectors may be that many units 'sub­contract' metal detecting to amateur detectorists, recognising that they generally possess more experience and skill at detecting than professional archaeologists. Some units work in regular liaison with local detecting clubs, while others employ individual detectorists as site specialists.

The benefits that hobbyists have brought to archaeology should not be underestimated. They not only bring valued and sophisticated equipment to an excavation free of charge, but many even work unpaid themselves. Many of the archaeological advances recorded in this chapter could not have been made without them. A case study of especially fruitful co­operation between hobby detectorists and archaeologists, at the annual South of England Rally, can be found in the following chapter.

4.2 Metal detecting in archaeology
4.2.1 The technology adopted

Archaeologists were slow to adopt the metal detector as an archaeological tool. Despite the advocacy of R J C Atkinson, who described the predictive applications of army­surplus mine detectors in his book Field Archaeology (Atkinson 1953), metal detectors were not embraced to any significant extent in the 1950s and 1960s. Post­war mine detectors were cumbersome machines and, in any case, by the late 1960s had declined in availablity. By the time that a range of relatively cheap, efficient and manageable machines appeared on the market, the metal detector was already beginning to attract a contentious reputation.

Few archaeological units, perhaps for this reason, chose to employ metal detectors during the 1970s, and even towards the end of the decade many remained openly resistant. In Crowther's 1978 survey of 26 units, 17 (65%) claimed they knew of no instances where detectors had been used in an archaeological application. Eleven units (42%) asserted that they would not use the equipment under any circumstances (Crowther 1983, 11-12; Hall 1992, 24-5).

Our survey reinforces these findings. Of the 50 units canvassed, only nine were using the technology by about 1980, whilst a further 24 began to use metal detectors between 1981-85 - the period that followed the zenith of the hobby's popularity. During this period, the first (inconclusive) attempt at subjecting the metal detector to rigorous field trials under controlled conditions appeared in the archaeological press (Crowther 1981). Far more importantly, in 1984 a seminal article by Tony Gregory and Andrew Rogerson of the Norfolk Unit appeared in Antiquity (Gregory and Rogerson 1984). The paper convincingly demonstrated that scanning by club detectorists in preliminary work at the Gallows Hill (Fison Way) site at Thetford produced meaningful distributions of topsoil metal finds, which subsequently informed the strategy adopted for the excavation.

But Gregory and Rogerson's work seems not to have convinced everyone immediately. From 1986 to 1990, only ten more units within our sample of 50 adopted metal­detecting equipment, while six more took up the practice between 1990-94.

Ten units (20%) out of 50 still do not use metal detectors today, most of them based in the south west and the north, where sites rich in non­ferrous metal are less common than elsewhere in the country. The absence of archaeological detecting in the far south west, and particularly in Cornwall, also parallels the apparent lack of illicit activity in the region (Table XVII).

4.2.2 Machines

Only 21 (42%) of the units in our survey own a metal detector. Of these, 18 own only one machine - the exceptions being the Central Archaeology Service at English Heritage which owns two, the Essex Unit which owns three, and the Museum of London Archaeology Service (MoLAS) which owns six. Experienced operators, however, suggest that an ideal technical kit comprises several different machines with a range of capabilities and applications (eg Barber 1990, 25). A full list of the machines owned by these units is given in Appendix K.

Although not a point addressed in detail within the survey, our impression is that few units feel obliged to keep up with developments in detecting technology. Some of the units' instruments are old, obsolete, and rarely used. Elsewhere, those in use represent a cross­section of those on the market, and it is unclear on what basis the selection of particular makes and models has been made. The Rimatron Viking 5D Series 2 owned by Trent and Peak Archaeological Trust, for example, retails at only £109.00 (dealers' lists, Treasure Hunting, February 1994). The Central Archaeology Service's Eagle Spectrum, however, would cost £599.00 if bought today.

Units using detectors usually have several machines at their disposal in addition to those they own; but nonetheless, the range of machines that are owned may suggest uncertainty among archaeologists over which are most appropriate for archaeological detecting - a situation perpetuated by the virtual absence of recent technical literature on detectors within archaeological publishing. Given the huge literature generated over several decades on the technicalities of such things as aerial photography, field survey, geophysics, sampling, sieving, and statistics, the absence of any comparable studies on the metal detector is surprising.

4.2.3 Operators

The dominance of volunteer detectorists in archaeological excavations, working in effect as site specialists, is probably the main reason why professional archaeologists have mostly failed to get to grips with the technical aspects - and implications - of the metal detector.

Of the 40 units using detectors within our sample, 33 (83%) work with independent volunteer (unpaid) detectorists. Of these, 14 units work exclusively with such operators, whilst the remainder use them in combination with their own staff and/or paid detectorists. In rare cases, such as MoLAS or the Leicestershire Unit, operators have been taken on to the unit staff specifically as detectorists.

The initiative for co­operation seems usually to have come from the detectorists. The Avon Archaeological Unit, for example, began detecting in 1992 following an approach from a detectorist who offered assistance on a specific site. The relationship proved worthwhile, and all of the unit's detecting experience has been with the same operator (A Young: information, 28.2.94). Occasionally the arrangement is ad hoc: the Birmingham University Archaeological Field Unit periodically receives offers of help from detectorists, and accepts them when mutually convenient (S Buteux and P Leach: interview, 18.2.94). The Humberside Unit began using volunteer detectorists at the Flixborough excavations in 1989 primarily as a security measure, as the site had suffered frequent raiding from illicit operators (B Whitwell and D Tomlinson: interview, 16.2.94). Occasionally volunteer detectorists have joined a unit's staff, as in the City of Lincoln Archaeological Unit, one of whose project officers first assisted the unit as a detectorist in 1988 (J Hockley: interview, 10.2.94). South Eastern Archaeological Services are in a similar position. Most of their detecting is in the hands of Luke Barber, whose expertise as a detectorist and trained archaeologist is regularly deployed in the unit's projects (L Barber: interview, 8.2.94).

Ten (25%) of the units using detectors work in regular liaison with detecting clubs. Here too, the initial contacts often came about on the initiative of the detectorists themselves. The Trust for Thanet Archaeology, for instance, began liaison with a local club as early as 1978, following an invitation to attend a club meeting to speak in defence of archaeology (D Perkins: interview, 16.2.94). Liaison between the Society of Thames Mudlarks and the Museum of London began in 1981, when two members were invited to detect on the Swan Lane site for a day, and so impressed museum staff that the experiment was extended (G Egan: interview, 28.1.94; Egan 1986). Thames Valley Archaeological Services have worked regularly with the Wessex Metal Detecting Association for some years, after the club volunteered its services, and now contacts the club for assistance where appropriate. If working outside its base area, the unit also attempts to make contact with a local club (S Ford: interview, 28.2.94). The same is true of the Central Archaeology Service of English Heritage, whose national role has led them to attempt contacts with clubs in several parts of the country.

4.3 Metal detecting in fieldwork
4.3.1 Literature

Among the 40 units in our sample with metal­detecting experience, few operate a consistent policy in which the impact on finds recovery is realistically assessed, let alone quantified. Consequently few papers dealing with the use of detectors in fieldwork have been published. Publication lag will account for some of the shortfall. Reticence may explain yet more. Information on the potential benefits of metal detecting, and on how best to use the machines, has undoubtedly been impeded as a result.

A search of literature published between 1988-93 has revealed few rigorous assessments of the use of detectors in survey work. The most frequently cited use of metal detectors is in evaluation and survey, which often involves teams of detectorists and archaeologists jointly fieldwalking. Such programmes are reported by Didsbury (1990), Malim (1990), Nenk et al (1991), Perkins (1989), Newman (1992; forthcoming) and in Buckinghamshire (MEF and ARP 1989). Occasionally, however, the use of detectors or the help of detectorists is acknowledged in a survey or excavation report but no other information is given (Chapman and Jackson 1992; Williams 1992a). The knowledge that metal­detector users have made finds in an area is sometimes noted as a reason for conducting an archaeological evaluation in advance of development (Hearne 1991; Leach 1991). There are also a limited number of occasions when a metal detector survey is reported as taking place at the same time as excavation in a neighbouring area (Lawson and Farwell 1990; Mackreth 1988; Neal 1989).

Rigorous assessments of the use of detectors in excavation are even rarer. Currie (1993), Eccles et al (1989) and Williams (1992b) mention detecting for salvage of material from construction­industry spoilheaps; but only one reference was found to the use of detectors in a predictive role, namely David Perkins' work in recovery of a bronze age hoard in Thanet (Perkins 1991). Largely dismissive but cursory remarks about metal detecting have been made by A Clark (1990, 121-2) and Scollar (1990, 570-5).

Indications, drawn from the literature, of the gains made to archaeological knowledge as a result of the use of the metal detector can be found in Appendix G.

4.3.2 Methods of use

Few units appear to be using metal detectors in a consistent way or to their full capacity. Nor is there much evidence to suggest that rigorous sampling methods are in general use. We asked staff at the 40 units using metal detectors to categorize their use of detectors in one or more of four possible ways. These were topsoil scanning prior to stripping (prospection); scanning stratification during excavation (prediction); spoilheap scanning (salvage); or some other way, to be specified.

In response, we found that one unit (3%) used detectors solely for topsoil scanning; five (13%) used them solely for spoilheap scanning; four (10%) used detectors for topsoil and spoilheap scanning; five (13%) used them for scanning stratification and spoilheaps; and 24 (60%) used them for topsoil, stratification and spoilheap scanning.

Behind these simple figures lies a complex array of practical arrangements. Few units consistently employ detecting in the reported modes on all sites likely to produce metal, though some acknowledge this as an ideal and are attempting to work towards it. One difficulty is raised by the preference of many units for working with voluntary detectorists, who are not always available when needed. This applies especially in the continuous scanning of features during excavation, in contrast to comparatively brief pre­stripping topsoil surveys, or to spoilheap scans, which can be done at the detectorists' convenience.

In cases where units use full­time staff with conventional archaeological training as detector operators, it is sometimes conceded that their limited experience with the machines makes for poor results. Some units committed to detecting have solved these problems by employing experienced detectorists, often drawn initially from volunteer ranks, as full­time unit staff. On the whole, these units alone are able to establish detecting on a highly systematic basis.

4.3.3 Positive gains: topsoil scanning

One of the most frequent accusations that detectorists make of the archaeological profession is that the mechanical stripping of topsoil prior to excavation involves a massive loss of metallic objects. Painful though it may be to admit, this accusation is largely true. The information to be gained from the horizontal context of material in ploughsoil - including metal - is considerable. Many archaeologists remain slow - or reluctant? - to appreciate this point: only 29 of the 50 units polled in our survey have used detectors to scan the topsoil on their sites.

Gregory and Rogerson's paper on the topsoil detector surveys at Thetford and Middle Harling remains a baseline for any discussion of the use of the metal detector in prospection (Gregory and Rogerson, 1984). The Thetford survey, conducted in 1980 in advance of extensive excavations of a late iron age to Roman complex by the Norfolk Unit, demonstrated beyond doubt that the gains accrued in material quantities and spatial information by controlled detecting were enormous.

In quantities, a total of 126 metal objects intrinsically datable to the period of occupation of the site were found, both in the area excavated and in the surrounding field. Of these, 104 objects were found in the topsoil as a result of metal detecting.

From metal­detected topsoil and excavated contexts combined, it was shown that material of the 1st century AD was largely confined to the area within and to the north of the main site enclosure: 45 objects were found by detecting, and 18 by excavation of features. Moreover, 3rd and 4th century metalwork, was found almost exclusively by detecting in the topsoil: of 63 objects found, only one was excavated from a feature. The detected topsoil finds were present across the whole site; additionally, a tight concentration was centred some 150 metres outside the main enclosure. None of this vital information for later Roman activity on the site would have been obtained without the detector survey.

The chronological information gained from detecting at Thetford greatly enhanced the data available on the latest periods of occupation. Cases are also documented where detecting provides more or less the only means of dating specific periods of site occupation. Newman's work in Suffolk draws attention to the difficulties in certain areas in distinguishing between early Anglo­Saxon pottery types and those of the later iron age. There is thus no simple way to distinguish between surface scatters of one or the other type of material. Metalwork in the ploughsoil represents the only means of defining the period of occupancy (Newman, forthcoming).

Didsbury's (1990) study of a fieldwalking and detecting survey of an Anglo­Saxon site on the Yorkshire Wolds provides a similar object lesson. Here, Anglo­Saxon metalwork is reportedly co­extensive with a cropmark which on morphological grounds would normally be considered to be of Roman date. Fieldwalking here produced a sparse distribution of Roman pottery, which superficially would have been accepted as confirmation of a Roman dating. Anglo­Saxon sites in this region are generally aceramic, and in the absence of the metal topsoil finds the site would have been misattributed. Similar principles may apply in areas where occupation sites, if not aceramic, have softer­fired pottery types which survive poorly in ploughsoil.

4.3.4 Positive gains: stratification scanning

The use of metal detectors in excavation can enormously increase the volume of finds made, yet only 29 of the 50 units in our survey use them in this way. One unit that does is the Museum of London Archaeology Service. Geoff Egan, in his introduction to the Museum's recent volume dealing with medieval dress accessories from London, writes that the use of metal detectors in the 'skilled hands' of the Society of Thames Mudlarks was the principal factor in the recovery of 'virtually all of the metal items' included in the volume from Swan Lane [Swan Lane 1981] and Billingsgate [Billingsgate Lorry Park watching brief 1983], accounting for some 68% of the objects listed in the volume. Several parts of the study, he wrote, including 'some of the most outstanding individual items as well as entire categories of objects (eg, shield­shaped strap­ends) ... could not have been written without this fruitful cooperation' (Egan 1991, x).

Egan also reports that detectorists can also increase the quantities of non­metallic finds made on an excavation. Simply working closely over deposits with the detector increases the general level of observation (Egan op cit; interview, 28.1.94) - even though most hobby metal detectorists, working independently of archaeologists, appear on the whole not to be making non­metallic finds (Appendix I).

Those experienced in using metal detectors for stratification scanning suggest in addition that they can greatly enhance the recovery of certain categories of material - in particular finds that are very small, and objects made of lead.

Luke Barber reports that detecting, when introduced at Beddingham Roman villa, dramatically increased the rate of coin finds (interview: 8.2.94). During the 1987 season, working without detector support, four coins were recovered. With the introduction of the detector in 1988, 42 coins were found. In particular, one scattered hoard of barbarous radiates - diminutive late Roman issues - would have been impossible to recover in full without the detector. Barber reports having scanned excavated spoil which was thought to be sterile of finds, only to discover more coins. Similarly, minute droplets from metalworking, perhaps the only evidence for the process in certain contexts, have been recovered with the detector when otherwise they would have been missed.

Barber reports that the lead finds rate generally increases when the detector is used. Lead is one of the most difficult metal types to recover by observation, since unlike many other metals it produces no staining through corrosion product in its surroundings. This finding is also supported by records from the Norfolk SMR. Gurney (1993) has calculated the proportions of metal of different types, of all periods, from Norfolk surface finds records (fig 25). Though no control data from 'conventional', or non­detector excavations have been gathered to set against this graph, the proportion of lead objects - higher than any other non­ferrous metal type - seems exaggerated in relation to general site experience.

Barber reports that the detector's usefulness can be diminished, however, on some types of site. First, when the site is extremely rich in ferrous items, such as nails, which impair the detector's ability to discriminate. Second, when material is severely corroded, and will not give a proper signal.

David Perkins of the Trust for Thanet Archaeology, one of the first field archaeologists in Britain to use metal detectors, recommends using them as a predictive tool, especially for the excavation of accompanied inhumations. Perkins points out that the grave fill in the immediate area of the body is often much looser than the surrounding matrix. Thus the trowel often displaces finds where the detector locates them (D Perkins: interview, 16.2.94).

Furthermore, the detector can make an impressive contribution to the level of intra­site spatial information. A detector survey by Thames Valley Archaeological Services on excavations at Charnham Lane, Hungerford, targeted both ferrous and non­ferrous material (Ford 1989), and showed that iron objects tended to concentrate on the yard surfaces external to the habitation structures, which seemed to represent an activity area. In addition, non­ferrous material of diagnostic function was located around the site. This type of survey, in which almost all the metalwork in the stratification was presumably located, provides a firm basis for enhanced understanding of intra­site variability in area function and activity.

4.3.5 Variety in application

The site practices of units using detectors in excavation vary enormously. Given the potential for finds recovery enhancement provided by the instruments, this variety must lead to some inconsistencies in sampling. At a national level it seems clear that greater uniformity is necessary.

The Avon Archaeological Unit, for example, uses detectors for selective feature­scanning on site prior to excavation (A Young: telephone interview, 28.2.94). The Central Archaeology Service tends to use detectorists for marking, though not lifting, finds from stratified contexts (D Batchelor: interview, 21.2.94). Surrey Archaeological Unit reports the use of detectors to target sectioning in linear features, in much the way recommended by Gregory and Rogerson (R Poulton: interview, 23.2.94), whilst the Test Valley Archaeological Trust uses detectors for context­controlled scanning of spoil during excavation (F Green: interview, 3.3.94).

Moreover, 38 units in our sample used detectors for spoilheap scanning, among whom five (10%) used them solely in this way. It seems unlikely that these units would continue to use the machines on spoilheaps if no finds were ever made there. Yet the application of the detector to scanning topsoil and stratification offers the opportunity to recover finds from controlled contexts, rather than under salvage conditions. One factor limiting the ability to do this will be the availability of detecting equipment, when skilled volunteers only are used. But there seems little doubt that spoilheap scanning should be something of a last resort.

4.3.6 Reservations

Of the units in our survey not using metal detectors, some question the efficacy of the machines, either on the types of site with which they regularly deal, or more generally. Others have a broadly ethical difficulty over the danger of being seen to condone metal­detecting as a whole. Some have had unhappy experiences with volunteer detectorists, and now prefer to keep a distance from the practice, or use it sparingly with their own field staff as operators.

The City Archaeologist for Chester Archaeology Service, for instance, has experience of detecting on projects elsewhere in the country, but in those cases found that the instruments did not produce a significant increase in finds rate (M Morris: interview, 1.3.94). The Colchester Archaeological Trust has used a detector once, at a gravel­pit site in 1991. The instrument was operated by a trusted volunteer detectorist, but the finds rate did not convince archaeological staff that the exercise had been worthwhile (P Crummy: interview, 17.2.94).

Some units have tried detecting but abandoned it. The Exeter Museums Archaeological Field Unit gave up detecting more than five years ago, owing to dissatisfaction with the liaison they had developed with local detectorists. On one occasion in 1984 the detectorist discovered two bronze age gold objects and, despite being asked to respect confidentiality, reportedly passed the story to the local press immediately (C Henderson: interview, 1.2.94).

Similar arguments are used by units that employ detectors only rarely and in the hands of their own staff. The Hertfordshire Archaeological Trust has only used a detector once, in this way, and the results in recovery enhancement were not impressive (A Havercroft: interview, 16.2.94). Havercroft stresses, however, that the Trust's limited use of detectors is partly a result of working several prehistoric sites with small metalwork assemblages. Periodic approaches from volunteer detectorists offering assistance are presently declined. The Trust feels nervous over the involvement of an unfamiliar third party, and considers that this may give rise to difficulties with the ownership of finds.

The Test Valley Archaeological Trust has a similar experience. The unit bought its own detecting equipment in 1986-7. Before this the unit did try to work with individual detectorists, but found them unwilling to operate within the disciplined framework of archaeological recording. All detecting is done in­house (F Green: interview, 3.3.94).

The Surrey Archaeological Unit has used detectors rarely and with caution, since staff there are wary that operating the machines themselves will encourage undesirable activity. The unit's main experience was at the Wanborough temple, where such activity was indeed rife, though not as a result of the unit's application of the technology (R Poulton: interview, 23.2.94).

Without questioning the authority behind these reservations, two answers may be tentatively offered. First, sufficient evidence is available that an appropriate machine, on the right type of site, in skilled hands, can enhance finds rates significantly. Single contrary experiences may only suggest that the detectorist is inadequately skilled, or is using an instrument of insufficient quality or suitability for the site. To employ a metal detector once or twice and then discontinue the practice because of poor results is, surely, altogether too cursory a trial. How the metal detector might compare to other forms of intensive recovery, such as selective sieving, remains to be assessed through rigorous field trials and it may ultimately prove to be the case that an integrated approach using both techniques will prove to be the most beneficial. The main difficulty is a lack of systematic detecting, informed by a thorough grasp of sampling methodologies. The (usually) irregular availablilty of volunteer detectorists simply compounds this problem.