Number of Incidents

We turn first to the numbers of known incidents of illicit metal detecting in England and Wales.^{Footnote 44} The HE list, the FLO survey, and FOI responses give us different elements of the picture, with the first providing us with numbers that have come to the attention of Historic England, the second the number of law enforcement cases involving illicit metal detecting with which FLOs have been engaged, and the third giving us the (incomplete) numbers that the police could extract from their databases. One of the conclusions of the Nighthawking Survey was that illicit metal detecting has increased slowly since 1995. The survey found a gradual increase in incidents between 1995 and 2006, with a dramatic increase for 2007; this increase was thought to be a result of the data collection rather than an actual increase of incidents.^{Footnote 45}

The HE list demonstrates that knowledge of illicit metal detecting has increased over the period 2010–17, though at a greater level than previously known (Figure 1). This is demonstrated by two statistics: first, the HE list contains information on 276 incidents over a seven-year period, while, in contrast, the Nighthawking Survey contains information on 274 incidents over a 13-year period.

Figure 1. Illicit metal detecting on SMs and non-designated sites by year (HE list).

However, we should be cautious about the extent of incidents and trends identified via the HE list as this does not necessarily represent the true extent of illicit metal detecting over this time period; the true figures are likely to be much higher as many attempts and successful illicit metal detecting events are likely to go unnoticed due to the remote/rural locations of many suitable sites.^{Footnote 46} Yet the differing methodology between the previous survey and this dataset mean that they are not directly comparable. Indeed, both datasets are of limited use given that an “incident” is simply a report of illicit activity rather than a conviction of illicit activity.

That said, we have a second observation; although there is a two-year gap in knowledge between the end of the Nighthawking Survey data (2008) and the beginning of the HE list (2010), the number of incidents recorded in 2010 was more than twice that in 2008.^{Footnote 47} Notably, this figure has now doubled from just over 20 incidents in 2010 to just over 40 in 2017 (Figure 1). Again, whilst this should be treated with caution due to the differing methodologies of the two sources, this is a dramatic difference and suggests at the very least a greater policing of the problem of illicit metal detecting.

Although the general trend is one of greater numbers of recorded incidents, the HE list recorded a decline in 2016 when just 25 incidents were noted. This may well be artificially low; indeed, the survey of FLOs noted that staff were involved in 27 cases in 2016 and, collectively, had information on at least 60 further alleged incidents.^{Footnote 48} While some of the cases on the HE list rest on anecdotal evidence provided by FLOs and others, the majority of incidents are those that came to the attention of the police. It is likely that some, but not all, of the incidents identified by FLOs in the survey (Figure 2) are a subset of those shown in Figure 1. The data does not allow for a direct comparison.

Figure 2. Number of incidents per year where FLOs assisted law enforcement authorities (FLO survey).

The overall trend noted by the FLO survey is one of increasing involvement with law enforcement authorities,^{Footnote 49} with 19 cases in 2015, 27 in 2016, and 31 in 2017 (Figure 2). The types of incidents that FLOs advised on during the period 2015–17 were dominated by reports of illicit metal detecting, closely followed by reports of items that potentially qualify under the Treasure Act 1996 but which had not been reported to the coroner (Figure 3). FLOs also assisted law enforcement authorities with their enquiries into the illicit sale of potential treasure and, less commonly, of archaeological material that does not fall under the auspices of the Treasure Act 1996. Unfortunately, the data is such that it is not possible to cross-reference individual cases with the HE list, but there is likely to be considerable overlap given the close working relationship that FLOs have with Mark Harrison and Operation Chronos.

Figure 3. Types of incidents that FLOs advised on during the period 2015–17 (FLO survey).

A similar problem of cross-referencing cases occurs when we examine the scarce data provided in response to the FOI requests. In total, 48 cases of illicit metal detecting were highlighted in response to the request. However, after examining the detail provided, a number of these responses were irrelevant, including knife crime and cable theft. This left us with an estimated 44 recorded illicit metal detecting crimes. Of these, 29 were identified within just three police forces: Lincolnshire, Norfolk, and Suffolk. There is clearly much difficulty with identifying relevant cases: 18 of the 39 responding police forces refused the FOI request on the basis that manual searches would have to be conducted of the tens of thousands of records (in most cases) to identify cases of illicit metal detecting. Those forces that were able to provide data were careful to caveat their response with the limitations of the data searches they were able to conduct. The method of searching also varied by police force area, meaning that the numbers of incidents are not reliable nor directly comparable—either to other forces or to the other sources of data we discuss herein. For example, the FOI response from the Northumbria Police contains information on one incident, though the HE list contains information on eight incidents. The incident recorded on the FOI response occurred in September 2017; however, in 2015, the Northumbria Police issued a joint press release about incidents occurring on scheduled ancient monuments along Hadrian’s Wall.^{Footnote 50} These incidents were not caught in the FOI trawl.

Similar instances of “missing data” were also noted in the FOI responses for other police force areas. The FOI response from the West Midlands Police returned no incidents for illicit metal detecting, but it did include one incident of metal theft (presumably theft of lead from a church roof). However, the HE list contains 18 incidents of illicit metal detecting for this region. Two of the 18 incidents of illicit metal detecting listed on the HE list are accompanied by crime numbers; nonetheless, neither was caught by the FOI search. The joint FOI response issued by the Norfolk and Suffolk Police includes information on 20 incidents of illicit metal detecting and is by far the most comprehensively recorded response in comparison to the other police forces. The large number of incidents is also reflected in the 33 incidents recorded on the HE list. On face value, there appears to be much overlap between the Norfolk/Suffolk response and the HE list, though it is difficult to be absolutely sure in most cases.

The FOI response exercise demonstrates that there is much variation in the way that heritage crime, particularly with regard to illicit metal detecting, is recorded (or not, as the case may be). Indeed, “known” incidents on the HE list, which contains crime reference numbers, sometimes do not appear on FOI responses even though they do appear on the HE list. The HE list therefore arguably gives the most reliable picture of illicit metal detecting across England, though the survey of FLOs suggests that the occurrence of incidents might be somewhat greater.

The FLO survey also captured information on the type of assistance offered to law enforcement authorities using the following categories: (1) reporting first-hand information; (2) reporting third party information; (3) advising on policy/legislation; (4) advising on whom to contact; and (5) appearing in court as an expert witness (Figure 4). Broadly, the statistics for each year compare well with one another and show that FLOs contact law enforcement authorities with both first and third party information. FLOs were also instrumental in advising on policy and legislation surrounding archaeological antiquities and sites; these include the Treasure Act 1996, the Ancient Monuments and Archaeological Areas Act 1979, the Dealing in Cultural Objects (Offences) Act 2003, and the legislation regarding export licenses.^{Footnote 51} It might be notable that there was a decline in the proportion of incidents in which FLOs advised on whom to contact, which probably reflects both the greater awareness of the issue of illicit metal detecting and the greater networking between agencies that has arisen from Operation Chronos and the efforts of Historic England and others.^{Footnote 52}

Figure 4. Capacities in which FLOs assisted law enforcement authorities. Figure shows the percentage of incidents (FLO survey).

A decline in the proportion of occasions where FLOs reported first-hand information was also noted over the period, though third party information remained consistent. This statistic is difficult to interpret, but it does stress the important role that responsible metal detectorists have in reporting potentially illicit activity.

Illicit Activity at Legally Protected and Non-Protected Sites

The HE list provides information on the designation of the site that has been attacked—that is, whether the site is legally protected or not. We have previously commented that SMs under cultivation may be perceived by illicit metal detectorists to be prime targets, in part owing to their status as SMs and also because of the information and publicity about a particular SM that often is found online.^{Footnote 53} The number of incidents recorded on the HE list that affected SMs is significant: 151 incidents reported across 121 designated sites. The numbers range from between 12 and 30 incidents, giving an average of 15 incidents per year over the period. By way of comparison, the Nighthawking Survey, which covers the period 1995–2008, identified 88 SMs that had been attacked.^{Footnote 54} This latter statistic, however, was thought to be an underestimate of the problem.^{Footnote 55} Indeed, it is much lower than the findings of a survey published by the Council for British Archaeology in 1995, which identified 188 SMs attacked between 1989 and 1993.^{Footnote 56}

The HE list demonstrates that non-designated sites were also frequently targeted— indeed, slightly more so than SMs. The number of reported incidents of illicit metal detecting on non-designated sites between 2010 and 2017 is somewhat variable, but there is an overall increase. The dramatic rise seen after 2013 might be partly attributable to the greater liaison between HE and the stakeholders, which naturally resulted in an increase in the reporting of both incidents and intelligence for non-designated sites. Post-2013 statistics, therefore, do appear to give a more reliable view of the situation in England and Wales, with the levels of reporting being approximately more than 40 cases per year. Overall, the HE list indicates that the number of incidents affecting SMs is rather constant, while incidents affecting non-designated sites is increasing.

Accordingly, it might well be the case that SMs are increasingly not seen as the beacons as they once were; 20 years of recording by the PAS has demonstrated that the rural landscape contains many more sites than were previously known.^{Footnote 57} These “new” and non-designated sites may now be perceived as more fruitful targets. It may also be the case that a greater awareness of the law has resulted in areas adjacent to SMs being targeted rather than the protected site itself.^{Footnote 58}

The HE list indicates that of the total number of incidents recorded each year, around 40–48 percent will affect a SM. Naturally, this statistic should be of interest to any police officer who apprehends an individual on or leaving an archaeological site; there is a good chance that if the individual has committed theft, there will also be offences under the 1997 Archaeological Areas and Ancient Monuments Act. Indeed, SMs make up a very small proportion of the total archaeological resource, perhaps as little as 2–4 percent. Even with biases in reporting, SMs still appear to receive a disproportionate amount of attention from illicit metal detecting.

Distribution

The distribution of the 276 incidents recorded on the HE list demonstrates the same spatial trend in offences that was noted by the Nighthawking Survey—that is, there is a general background level of illicit metal detecting across the country. We can view general trends in the reporting and recording of illicit activity in more nuanced ways by classifying data according to the geographical regions defined by Historic England. These regions are useful on two accounts: first, they divide the country according to broad landscape types and, second, they make it easier for Historic England to tailor historic environment advocacy to a particular area.^{Footnote 59} Classifying illicit activity by these regions reveals greater (reported) activity occurring in the southeastern and eastern counties of England, with relatively large numbers of incidents also noted in the southwest and East Midlands (Figures 5 and 6).^{Footnote 60} These “prime” regions, of course, do contain large areas of land under arable cultivation—conditions that make sites more vulnerable to illicit metal detecting owing to the ease of recovering artifacts.^{Footnote 61}

Figure 5. Incidents in England and Wales 2010–17 (HE list). Note that data from Scotland are not included.

Figure 6. Number of incidents by region (HE list).

The incidents identified from the HE list are expressed more clearly when the HE list is viewed by county, with the highest reported incidents occurring in Hampshire, Kent, Suffolk, Norfolk, Cambridgeshire, Essex, and North Yorkshire respectively (Figure 7). Yet a degree of caution needs to be exercised when interpreting county statistics; the dramatic rise in incidents in Hampshire when compared with the Nighthawking Survey is an anomaly but is one that can be explained by the recent efforts of the Police Country Watch Team to increase reporting.^{Footnote 62} Indeed, there is a broad correlation between hot spots and those areas where a greater degree of police engagement has taken place. As with all crime statistics, these data are sensitive to a range of nuances, not least local policing initiatives.

Figure 7. Number of incidents by county (HE list).

It is not surprising that we see clusters of illegal metal detecting in certain counties, such as Hampshire, Kent, Norfolk, and Suffolk. We know that once a place is targeted by offenders on one occasion, it is far more likely to be targeted again. This is known as repeat victimization, and it is observed across all crime types (with the notable exception of murder, which even then is often preceded by a number of attempts or violent offences). The reasons for this are usually a combination of state heterogeneity and event dependence factors. In this case, an area may be particularly targeted because of features that make it attractive for illicit metal detecting: first, a rural, remote area with many sites that may be perceived as having potential for a relatively high return on effort (state heterogeneity) or, second, an offender who has successfully retrieved artifacts from the ground who may return to the area to identify further items (event dependence). We also see a proportion of individual sites being repeatedly targeted according to the HE list, with an average of 1.25 incidents per site.^{Footnote 63}

There are other reasons that we may see a clustering of reports that bear no relation to the actual incidence of illicit metal detecting. Landowners, visitors, residents, and caretakers in some areas may be more aware of abnormal behavior in an area and, thus, more likely to report their observations. There may be greater membership in schemes such as Heritage Watch, which formalize awareness and encourage reporting. There may be police officers or volunteers with a particular interest in heritage or archaeology, resulting in a greater likelihood of reports being forwarded for inclusion in the HE list. Regardless, it is likely that the incidents identified are an understatement of the problem, and we may also see a greater concentration of crime events with increased reporting.

Character of Site

The HE list can also be interrogated for information regarding the character of the site—that is, what period it belongs to and what “type” of site it is (for example, castle, fort, temple, farmstead). Naturally, these broad categories must be treated with due caution; many archaeological sites were used over long periods of time and cannot be assigned to any one particular period. Likewise, the nature of a site may have changed over time; land that was used as a farmstead in the Roman period may be a cemetery in the Anglo-Saxon period, for instance. These complexities aside, the HE list can be gleaned for broad trends.

Regarding the period of site, when examined on a national level, there is a clear trend toward incidents occurring on Roman and medieval sites (Figure 8). In fact, this is a trend that was established by both the Council for British Archaeology survey^{Footnote 64} and the Nighthawking Survey.^{Footnote 65} However, it is difficult to know whether illicit metal detectorists specifically target Roman and medieval sites or whether it is a bias resulting from sites of these periods being more common.^{Footnote 66} Closer analysis of incidents within these two archaeological periods reveals the general nature of the sites being attacked (Figure 9). In total, 64 percent of incidents on Roman sites involved settlements (42 incidents), while 23 percent of incidents occurred on forts, many of which are SMs (15 incidents). Damage to sites of medieval date concentrated on castles (17 incidents [40 percent] of sites within the period), and abbeys or priories (19 incidents [45 percent] of sites within the period). Again, the majority of these being SMs.

Figure 8. Period to which site is dated (HE list).

Figure 9. General type of site (HE list).

A slight increase in attacks on Iron Age sites is notable in the HE list; incidents were recorded on 21 hill forts, all of which are SMs. Curiously, a number of prehistoric stone circles were also targeted in spite of these being sites that predate the introduction of metalwork. In these instances, it might be that illicit metal detectorists with a poor understanding of archaeology are simply drawn to sites because of their perceived status.^{Footnote 67}

Time of Year

As well as giving us an annual overview of illicit metal detecting, the HE list also records the month in which incidents took place (Figure 10). Incidents of illicit metal detecting were reported fairly evenly across the agricultural year, though there was a notable reduction during June and July probably owing to the height of crops reducing access to arable land. The rise in August may be associated with crop harvesting, which results in large areas of land becoming accessible once again. Far fewer incidents were reported for September, however. The reasons for this are difficult to ascertain, but it could be due to fields being rough plowed after harvest in preparation for the next crop. Rough plowing creates deep furrows that make metal detecting very difficult.

Figure 10. Number of incidents in 2010–17 by month (HE list).

The high level of incidents experienced in the summer months was also a feature noted by the Nighthawking Survey, but the reduction in incidents in July and September revealed by the HE list is a trend that was masked in the Nighthawking Survey, in which incidents were grouped according to season (for example, “summer”).^{Footnote 68}

Of course, by grouping all of the incidents of illicit metal detecting together by month we inevitably mask trends on finer scales of time and place. Perhaps the most visible of these trends emerges when we explore the relationship between land use and site type—in this instance, Iron Age hill forts and multi-period settlements. The majority of Iron Age hill forts in England comprised large areas enclosed by multiple ditches and banks. Owing to the grand scale on which most were constructed, many survive today as earthworks. Those that are protected are often set to permanent pasture and are therefore more accessible throughout the year. This is in contrast to settlements, which for the most part are undesignated and under cultivation. The strong association between settlements and modern arable cultivation, however, does mean that they are only accessible during the months when there are no crops on the field or when the crops are still very young. These differing cycles of access, of course, do have implications for their risk from illicit metal detecting.

We can see traces of this trend within Figure 11, which shows the percentage of incidents affecting hill forts and settlements. For both types of sites, there are peaks in May, August, and November, broadly complementing the national picture. However, there are no, or few, incidents on hill forts during September and October—months for which there are incidents on settlements, notably so in October. The situation flips for November, where hill forts see a dramatic upsurge in detecting activities, while settlements see a dramatic decline. This might indicate that the same individuals are targeting both types of site, but are diverting their attention according to what is immediately favorable; illicit metal detectorists target fields containing ancient settlements when they are freshly rolled and then divert their attention to hill forts when the nights get longer and arable fields become less accessible. Of course, these are (over)generalizations, but it is important to stress that it is the fine detail—that which derives from local context—that will help us to best understand how to deter illicit metal detecting.

Figure 11. Proportion of incidents of illicit metal detecting per month on Iron Age hill forts and multi-period settlements (HE list).

Incongruencies in Data

These three sets of data have provided us with information that is not only distinctly different from any previous published data in its scope but also not comparable across the three sources. There are four areas where the data differs: purpose, accessibility, recording, and reporting. The data sources reflect datasets that have been gathered for different purposes. The FOI requests were reliant on police recording systems, which vary between forces. There are strict rules in terms of what is recorded as a crime, and the purpose of police data is to reflect recordable offences according to Home Office counting rules. The database is therefore massive, as it records all crime types. In contrast, the HE list was collated by an individual, largely in response to a specific police operation—Operation Chronos—but with the aim of getting a sense of the national scale of illicit metal detecting. The FLO surveys were undertaken to reveal the extent to which the PAS as a whole was engaging with local law enforcement authorities and how this varied across counties. With different aims and objectives, it is not surprising that the content of each dataset differs significantly.

There are differences also in the accessibility of data on the respective systems. Both the FLO and the HE list were small-scale datasets that were collated by an individual and were specific to metal detecting and finds. We were able to access these because the former was collated by the first author and the latter was collated by Mark Harrison, with whom we have an excellent working relationship. In contrast, relying on the FOI responses to access police data relied on phrasing the request in such a way as to facilitate police staff understanding and compliance. Even so, it became apparent that the police databases are not fit for this purpose; details about modus operandi and other attributes of the crime event are often not available except by manual searches of thousands of potentially relevant cases. This seems in part due to poor database design and in part due to the aforementioned counting rules that provide limitations within which the police must record crimes.

This leads us to the third area, which demonstrates the distinct differences in data—namely, recording practices. The FLO survey addressed all issues that had come to the attention of the relevant officers, regardless of whether they had come to the attention of the police. The HE list was mostly built by keeping records on police activity, which makes the differences between this data source and the police responses to the FOI request an interesting reflection of the ways in which the recorded details are accessible at a later date. The police data reflect both reporting and recording practices. These issues are well documented in policing generally.^{Footnote 69} Individuals and organizations may lack awareness of how and in what circumstances illicit metal detecting may be reported to police. Police must decide whether a reported incident is deemed recordable as crime under Home Office counting rules.^{Footnote 70} These two issues combine to result in an undercount in official crime statistics, known as the “dark figure” of crime.^{Footnote 71} This will doubtless affect both the information available from the HE list and the FOI requests, whereas the FLO survey will likely be less affected by this problem, being a survey of already engaged professionals.

In practice, these differences have meant that, although each dataset has been able to provide a small part of the picture of illicit metal detecting, they cannot be used to cross-reference cases or to confirm overall figures. This once more draws our attention to the need for a consistent and regular recording of illicit metal detecting incidents—and, indeed, heritage crime more broadly.^{Footnote 72}