In search of magnetic anomalies associated with haunt-type experiences: pulses and patterns in dual time-synchronized measurements

Jason J. Braithwaite

Clearly not all buildings have a reputation for being haunted. For those buildings that do, such anomalous experiences and events do not take place all of the time. Furthermore, when such instances do occur, not all persons present report them. These observations imply that there may be some critical dimensions or factors that distinguish such properties from other neutral locations and differentiate certain observers from other individuals. Certainly one effective method for a contemporary field-based investigation of a haunting would be to evaluate (1) environmental factors specific to the location, (2) individual factors specific to the observer, and (3) factors specific to an interaction between the location and the observer. In the case of specific locations associated with numerous instances of anomalous haunt-type experiences, an evaluation of how the surrounding microenvironment could be responsible for inducing such an experience would seem crucial. The idea that a given location may be influencing the perceptions of observers is nothing new. Many studies have carried out detailed surveys of such locations and revealed potential contributing influences from (1) contextual and situational specific factors, (2) diverse lighting levels, (3) drafts, (4) infrasound levels, (5) the localized distribution and changes in geomagnetic fields (GMFs), (6) time-varying electromagnetic fields (EMFs), and (7) transient tectonic events, to name but a few (Gearhart & Persinger, 1986; Houran, 2000; Lange & Houran, 1997, 2001; Persinger, 1974a, 1974b, 1985, 1988, Persinger & Cameron, 1986; Persinger, Ludwig, & Ossenkopp, 1973; Persinger & Koren, 2001; Persinger, Koren, & O'Connor, 2001; Persinger, Tiller, & Koren, 2000; Richards, Persinger, & Koren, 1993; Roll & Persinger, 2001; Tandy, 2000; Tandy & Lawrence, 1998; Wiseman, Watt, Greening, Stevens, & O'Keeffe, 2002; Wiseman, Watt, Stevens, Greening, & O'Keeffe, 2003; see also McCue, 2002, for a discussion). All of these factors, either collectively or individually, could either induce a direct experience or facilitate an experience-prone state in certain observers and under certain circumstances.

With respect to magnetic fields, researchers are proposing that perhaps some aspect of these fields have "experience-inducing properties"--even more so if observers have shown a degree of increased neuronal hypersensitivity and susceptibility to these fields (Cook & Persinger, 2001; Makarec & Persinger, 1987; Persinger, 1983, 1984, 1988, 1993a, 1993b; Persinger & Makarec, 1993; Persinger & Koren, 2001; Persinger & Roll, 1985; see also Fuller, Dobson, Wieser, & Moser, 1995, for an example of induced epileptiform activity). The general hypothesis from this is that such Experience Inducing Fields (EIFs: Braithwaite, 2004) could be present at reputedly haunted locations and may well underlie a number of reports ranging from nebulous and ambiguous sensations to extreme and complex hallucinations (Persinger & Koren, 2001; Persinger et al., 2001; Persinger & Richards, 1994; Persinger, Richards, & Koren, 1997; Roll & Persinger, 2001). (1) Within this view, what many haunt-type experiences could represent is, in essence, a spontaneously occurring magnetically induced hallucination. Here discrete shifts and changes in the localized magnetic field would correlate with sympathetic changes and shifts in the neurophysiology, perception, and behavior in observers. The net consequence of this process would be some degree of altered state for cognition and consciousness under certain circumstances. This proposal need not necessarily assume any degree of paranormality is involved in the experience or event. Here a magnetically remarkable environment may interact with an observer's brain, which may also show an increased susceptibility to such fields.

Clearly, a crucial step towards a proper field-test of the magnetically remarkable hypothesis would be to demonstrate that such magnetic anomalies are indeed present and available in such environments in the first place. The purpose of this would simply be to ascertain whether, at least in principle, there is some distinguishing component or factor that identifies certain regions of interest when compared to baseline areas. If this could be shown, then one side of the environment/observer equation may be supported. This could provide field-based physical evidence that potentially stimulatory magnetic fields (EIFs) that are available in the natural environment may underlie some instances of anomalous experience. Conversely, no magnetic anomalies may be detected, which may indicate that such components are not crucial for these specific reports, and if enough studies find similar null results, then the relevance of the idea applied to the natural setting could be legitimately questioned.

Crucial Aspects of Experience Inducing Fields (EIFs)

Growing evidence suggests that crucial EIFs are characterized primarily by their complexity rather than overall field strength/ amplitude (Persinger, 1999a, 1999b; see Persinger & Koren, 2001). In the laboratory, complexity has been implemented in a number of ways, including (1) increased varying amplitudes and amplitude modulation, (2) varying frequencies and frequency modulation, (3) using patterned amplitude-modulated fields, (4) using complex patterns of pulsed fields, and (5) using rotating fields. In all cases the amplitudes used are very low, usually between 100 nanoTesla (nT) and 1000 nT, and rarely above 5000 nT. Furthermore, only small windows of frequencies seem to have potent consequences for neural activity and anomalous consciousness, and these can generally be described as being within the spectrum of the human brain (i.e., between 0 and 50 Hz and typically below 30 Hz: Bell, Marino, & Chesson, 1992, 1994; Persinger & Koren, 2001). The low-amplitude, low-frequency, complex nature of these fields seems important in order for them to be integrated into, and alter, the overall current perceptual gestalt.

With reference to spontaneous cases of haunt-type reports, this evidence is interesting for a number of reasons. Firstly, high amplitudes themselves need not necessarily index areas associated with haunt-type experiences at all (i.e., a metaphorical hot spot). Such regions could well be indistinguishable from baselines in terms of actual field strength (see Wiseman et al., 2002, 2003). Secondly, the field strengths employed in the laboratory have not been excessive and would be readily available from the living environment. In this sense the proposal has some ecological validity, at least in terms of amplitude levels. Thirdly, complexity itself at such field frequencies could be created by complex distortions in the magnetic environment from natural sources (i.e., tectonic movement, structural building factors, and local geology), manmade sources (i.e., malfunctioning house wiring configurations and faulty appliances), or indeed an interaction of these factors. So it is entirely conceivable that such complex EIFs could be readily available in the natural setting under certain circumstances, at certain times. (2) Collectively, these findings have given considerable currency to the idea of a magnetic component underlying some instances of anomalous haunt-type experiences.

Magnetic Anomalies and Reputedly Haunted Locations

Field-based investigations have revealed a mixed set of findings in accurately characterizing the magnetic microenvironments of reputedly haunted locations. This may be due, in no small part, to inconsistent and inappropriate methodologies and the varying measuring technologies employed. Some research suggests that "higher" levels in either the localized ambient magnetic field or in EMF contributions may define haunted areas (Nichols & Roll, 1998, 1999; Roll & Nichols, 1999; Persinger et al., 2001; Roll, Maher & Brown, 1992; see Persinger & Koren, 2001; Roll & Persinger, 2001). Other studies suggest that it is not necessarily the overall ambient levels that are crucial but the way in which the localized fields are varying and changing (i.e., their complexity) that is the important factor (Braithwaite, 2004; Persinger & Koren, 2001; Wiseman et al., 2002, 2003). In addition, further studies have argued that large transient magnetic pulses and tectonic events could be associated with instantaneous experiences and events (Persinger & Cameron, 1986). To add even more confusion, some studies failed to find any noticeable magnetic signature in their investigations of spontaneous cases (Maher, 2000; Maher & Hanson, 1997). Despite these differences, a common emerging theme across a number of studies is that specific areas associated with haunt-type experiences do seem to contain EIFs that vary more in terms of amplitude than baseline areas (Braithwaite, 2004; Wiseman et al, 2002, 2003; see also Persinger & Koren, 2001).

What Is "Magnetically Remarkable" About Magnetically Remarkable Locations Associated With Haunt-Type Experiences?

Across a number of repeated investigations, the laboratory studies suggest that magnetic fields can induce potent hallucinatory perceptions in certain observers. However, what is not clear is how these findings are to be applied to the natural setting and the nature of spontaneous anomalous experience. If we accept the general case being made for magnetic anomalies, then the question becomes "How are these anomalies realized in the natural setting?" To put it another way, "How should we be thinking of EIFs in relation to reputedly haunted buildings?" These questions could be directed at potential sources such as geomagnetic field (GMF) and electromagnetic field (EMF) contributions summated into a distorted and complex magnetic microenvironment. However, these questions are more fundamental in the sense that what we really need to evaluate is the very nature of these "magnetically remarkable signatures" and what they can actually "look like." This should then allow researchers to evaluate what is both necessary and sufficient for a magnetic context to contain potential experience-inducing properties. This is important not just from the perspective of the environmental physics of such magnetic signatures themselves but also from the perspective of neuroscience and psychology, as establishing the crucial mechanism for interaction would greatly advance our current understanding of brain function and human-environment interaction.

There are a number of theoretical possibilities, all of which are supported to some degree by the existing literature (see Braithwaite, 2004, for a discussion of these). We suggest that there are three fundamental possibilities for how such anomalies could be realized in the natural setting. For instance, EIFs could be transient, volatile instances that may accompany an experience or event more or less instantaneously (such as a pulse or train of pulses, as documented by Gearhart & Persinger, 1986; Persinger & Cameron, 1986). In the absence of these events, the location in question may not be magnetically distinguishable in any way. What this implies for researchers is that simply measuring the area at any given time may actually miss the important characteristics that define that particular anomaly, in that particular case. In order to demonstrate its existence and detail its characteristics the researcher would need to "capture" the anomaly as and when it occurred. We refer to this general concept as a transient anomaly (see Persinger & Cameron, 1986; Roll & Persinger, 2001).

A similar idea is that of a more prolonged, enduring and sustained anomaly. Under this scenario magnetic shifts may come and go, but when present they may be around for some time. Such a shift would reflect changes in magnetic components from the more usual background behavior of such fields. These changes need not be linked exactly in time to a reported experience (as the experience may be the result of continued exposure to these anomalies), but the magnetic context around that general time period could be noticeably different from other periods in some way.

Finally, it could also be the case that such EIFs exist as a kind of constantly available distorted undercurrent that is more or less present all of the time (Braithwaite, 2004, Wiseman et al., 2002, 2003). Here, spontaneous experiences can occur when other important factors such as susceptible brains in observers, specific attentional biases, favorable levels of attention and arousal, other contextual variables, and prolonged exposure are all present and favorable. By this account the important distinguishing feature is seen more as a permanent magnetic characteristic of the locale (or specific areas within it). The implication for investigators is that the crucial distinguishing features are readily available to be measured freely at any time. Importantly, the magnetic fields measured during a reported event/ experience would not necessarily be distinguishable from the background fields measured at any other time, though this background may itself be a distinguishing factor relative to baseline areas and locations.

It is important to point out that the accounts outlined above for magnetic anomalies are all time-based accounts. Magnetic fields could also display unusual characteristics over space (i.e., across rooms/locations) that may also be a distinguishing factor of reputedly haunted locations. These anomalies could be thought of as within what could be termed the spatial uniformity of the fields (i.e., how the fields are distributed over space). These distinctions are important as they clearly indicate the legion of possibilities by which a magnetically remarkable signature could occur, for whatever reason. Perhaps all types of anomalies can occur across different cases and are associated with different experiences and events. Perhaps only one type of anomaly is crucial. Either way, it is important to be clear about the varieties of possibilities when these ideas are applied to the natural setting. These different theories also have implications for how measurements are taken and interpreted by researchers.

In many ways the discussion outlined above is analogous to that of searching for seizure-type patterns that may indicate neural storms in an EEG (electroencephalograph) scalp recording of a suspected epileptic patient. In some patients, seizures can only be distinguished around the time of the ictal event. In others, there can be a more constantly available abnormality in the EEG, which can be picked up during inter-ictal periods. In both cases the researcher is interested in the underlying mechanisms mediating how and where the anomalies occur, how they are sustained, how they propagate, and how they disappear--if indeed they do. The EEG has been an invaluable tool to the researcher interested in detailing cortical electrical anomalies and how these relate to neuro-cognitive processes. In essence the task is no different from that of detailing magnetic anomalies, which may exist as invisible thunderstorms in certain locations that may occasionally strike at vulnerable brains.

Appropriate Baseline Areas and Locations

In order to truly establish whether a magnetic environment or a particular signature within it is indeed remarkable in some way, such measurements must be compared to appropriate baseline measurements. The baseline provides estimates of variability contained within such signals where no apparent anomalous consequences are associated. What constitutes an appropriate baseline can be a matter of debate and contention dependent on the questions being asked, but a study that does not employ some form of baseline can be seriously compromised (see Houran & Brugger, 2000, for a fuller discussion of these issues). Also, it is not always explicit from field studies what attempts have been made to try to match the baselines (or not as the case might be) to the location of interest. There are a number of approaches to choosing appropriate baselines.

One approach has been to compare any target site with a selection of random other locations (see Houran & Brugger, 2000, for a discussion). The underlying logic is that most parameters should cancel themselves out across these randomly selected locations and build a representative template of natural variability of the important factors being studied, in the absence of anomalous reports. This approach often shows that what might look anomalous at the target site is actually part of the natural variability of a host of other "control" locations as well; therefore these factors are not particularly distinguishing as anomalous. However, one problem with using lots of random baseline locations is that many other dimensions are also free to vary. These include, room/area dimensions, location age, wiring configurations, electrical appliance demand, experiential context, architecture, lighting levels, sound levels, subterranean geology, person/ occupancy frequency, and so on. This should still be fine if it is being assumed that the magnetic signatures themselves are solely crucial (i.e., none of these other factors matter). However, if a complex interaction between magnetic signatures and the specific experiential context is assumed, the usefulness of such random locations as an appropriate baseline can be questioned.

Therefore, under some circumstances there is an argument that it might be prudent to try to match target and baseline locations along certain dimensions as much as possible. This is important for a number of reasons. All these other situational and contextual factors may indeed interact with magnetic signatures themselves, creating an experience-prone state in certain observers. Therefore a random baseline here may not share any geological, contextual, architectural, or magnetic properties at all with the target site, which could make it more difficult to identify the crucial components in any particular case as all factors are free to vary considerably. One possibility is to match sites along certain dimensions while ensuring that they differ considerably along the crucial dimensions of interest (e.g., geologically defined magnetic fields). These would provide more conservative estimates as all other variables are being held relatively constant.

Although searching for appropriate and separate baseline locations is one way to address this issue, another has been to take multiple measurements from the same location of interest but to compare areas that are associated with anomalous reports to areas that are not. (Wiseman et al., 2002, 2003; see Roll & Persinger, 2001). Note that all of the above possibilities involve comparing different spatial locations or areas at different times. A major methodological improvement has been to compare a specific and known area of interest to other proximal spatial areas at exactly the same time (Braithwaite, 2004). This method gets around the problem of trying to locate other separate locations that are matched in a number of ways. With these issues in mind, perhaps the most parsimonious approach would be to initially couple measuring the specific area of interest to more proximal time-linked baseline measurements in the same location. Both sets of measurements could then be further compared to other separate baseline location measurements for other estimations of magnetic variability. This approach should be appropriate for most purposes.

For the present study we needed to consider several factors when choosing an appropriate baseline. These were mainly that the baseline area should (1) share as many physical parameters (lighting levels, temperature, sound levels, and so on) as possible with the "haunt area" of interest, particularly those areas directly implicated in inducing haunt-type experiences; (2) share as many structural parameters as possible, such as room dimensions and subterranean geology; (3) share as many (potential) witness parameters as possible, such as occupancy level and frequency; and (4) be sufficiently distal so as not to share any specific localized anomaly possibly surrounding the haunt-area. These factors, taken together, should to some degree isolate magnetism and the presence of haunt-reports as the sole differentiating parameters between the baseline area and the haunt-area. Furthermore, this method is further improved by employing synchronized time-linked measurements at both the baseline and haunt-area simultaneously. This allows for a more direct comparison between the areas and has the potential to reveal anything special about both the spatial and temporal aspects of the magnetic signatures present. With these parameters in mind, the obvious choice for a baseline area in the present study was within the same room, several meters away from the main area of interest and any obvious artificial magnetic sources. We were lucky in that the actual haunt-area associated with the most striking and consistent experiences could also be estimated to within a few centimeters (as explained below).

Coupled to this time-linked baseline measure, we also provide some comparison data from other baseline locations. These baseline measurements have come from (1) modern living environments of both low and high dwelling loads, and (2) another stately building matched for structural and geological similarity (granite-based structure). The general normative data from these baseline locations are presented in the General Discussion section for further comparison.

Experiential Instances and Magnetic Anomalies at Muncaster Castle

Muncaster Castle has been the home of the Pennington family for over 800 years. It is situated on the tip of the Eskdale valley on the west coast of the Lake District, in Cumbria, England. The author (JJB) has been investigating haunt-type experiences and reports from the castle since 1992. For many reasons this location represents a classic case of an English castle haunting. Over the course of investigation numerous reports have been collected documenting first-hand eyewitness accounts. These reports currently span over 60 years with secondhand accounts reaching back even further, beyond living memory. This database receives additional entries on a yearly basis. Although a detailed discussion of these accounts is outside the scope of the present article, it is important to provide some of the experiential context that has motivated the present study.

Although there are numerous and repeated reports of strange experiences from inside and outside the castle, a clear epicentre has emerged. This epicentre is a bedroom situated on the first floor of the castle, called the Tapestry Room (TR). Reports from the TR pertain to experiences during both the day and night and have been provided from day tourists, staff, family guests, and family members. Some staff members staying overnight in the TR have also reported the old heavy door opening abruptly and apparently of its own accord. Reports that relate physiological components include (1) sudden headaches/migraines, (2) eyes watering, (3) runny nose, (4) ringing in the ears, and (5) bouts of dizziness. These reported sensations can occur alone or can be accompanied by other haunt-type components in some instances. Some observers have reported the strong feeling of a "sensed presence," being watched, hearing footsteps, and a periodic feeling of foreboding. As these experiences often happen while in the TR, observers typically attribute the sensations to the room, often with a paranormal interpretation concerning it.

However, the most distinct experiences have been reported by family guests or staff that have stayed overnight in the TR and occupied its bed (the TR was used as a guest room by the family for many years). In line with other aspects of the case at this location, all these witnesses have been interviewed over the years for their firsthand accounts. A number of the striking experiences reported by overnight TR occupants related to the distinct sounds of children crying/screaming in the night from inside the room. These particular reports are distinctive for a number of reasons. Firstly, they have been reported by a number of observers without any self-reported explicit prior knowledge of the haunting. Secondly, as a haunt-type report, these experiences are somewhat novel in that observers report that the sounds continue for some considerable time (starting off faint and becoming increasingly prominent and disturbing), long enough in some instances for the individual to investigate the room in an attempt to locate the sounds. All witnesses stated clearly that the noises were coming from within the room. Indeed, observers typically report not being able to sleep much during the night due to these sustained instances. Thirdly, of the observers who have reported these disturbing sounds/sensations, on most occasions they themselves were situated in the same location, settling down or falling asleep in the TR bed. The fact that the majority of the observers reporting these specific experiences were located in a relatively standardized place may not be a coincidence. It is these particular startling "children-crying" and other associated sustained bed reports that we are primarily interested in for the present investigation.

Contributions From Experiential Context, Expectation, and Magnetic Fields

There may be a number of reasons for these quite striking personal accounts from overnight occupants of the TR bed. Two important possibilities are (1) the role of the experiential and situational context at the time of the experience, and (2) prior knowledge or expectations on the part of the observers. The context and setting of an ancient castle is, of course, unavoidable in this instance. In some circumstances this may well, on its own, be sufficient to induce, influence, or shape many anomalous reports and interpretations (see Houran, 2000; Lange & Houran, 1997, 2001). Such an ancient and suggestive context may well predispose individuals to interpret otherwise ambiguous stimuli in a paranormal manner-particularly if they are unfamiliar with the surroundings (see Lange & Houran, 2001). Once an individual starts to perceive events as being "odd," this may initiate a cascade process whereby the experience becomes further distorted and embellished. Therefore the experiential context of the TR may well be conducive (visually, semantically, suggestively) to anomalous reports by inducing a subtle attentional bias in observers. This is certainly one possibility and may well underlie many accounts reported at the castle.

However, the role of context alone as being crucial can be questioned with regard to the specific, prolonged, and intense experiences reported by the overnight TR bed occupants. For instance, context alone does not explain why such reports have not been forthcoming from other guests staying in adjacent rooms that presumably are also equally contextually loaded as being "castle bedrooms," and the same also applies to the idea of expectation alone being crucial. Indeed, situated immediately next door to the TR is the West Dressing Room. This room is painted in the same color as the TR (a kind of distinctive turquoise); is also endowed with hanging tapestries, imposing family portraits, and antique furniture; and is of a similar size to the TR. In many ways this room is visually matched to the TR; indeed, until the 1860s both rooms were all one large area. Despite the relatively matched visual context of both rooms, anomalous reports from the West Dressing Room are surprisingly thin. Frequency of individual occupancy is also unlikely to be a major factor here as both the TR and West Dressing Room have been used roughly equally over the years by the family as guest rooms because both are proximal to the same private bathroom facilities. The same also holds for the King's Room on the other side of the TR, about which there are absolutely no anomalous reports to date despite its being a very dark and suggestive oak-paneled room.

Another problem for context alone being crucial for these specific experiences is that, in many cases, individuals have stayed at the castle at length, occupying a number of other bedrooms before or after staying in the TR. It is not clear why such general context alone could crucially influence perceptions only when these observers are in the TR, particularly when it is noted that it is not distinct along any obvious visual/semantic dimension relative to the other rooms occupied. Although as with any ancient haunting there will always be a degree of suggestive context (and in some instances this will be sufficient), this alone does not explain the bias in both the frequency and intensity of anomalous reports from the TR bed occupants, implying that other factors may also be contributing to the experiences reported from TR bed occupants.

One other important factor is that prior knowledge or expectation, either alone or coupled to other factors such as context, may be crucial (Lange & Houran, 2001). Such expectation could come in two forms, either general or specific. For instance, there may have been some general and nonspecific expectation on the part of the observers due to the suggestive nature of the ancient surroundings. Again, this is certainly an important factor and may well be crucial for many reports from this location and others like it. However, as with the role of context, it is difficult to see why any general expectation held about staying in a castle should be predominantly responsible for producing these bed experiences that are specific to the TR. Therefore, it is not clear why general expectation should be higher for the TR and why this alone should induce relatively specific reports of children crying, singing, and other voices from this area. In order for expectation to be crucial here, we would expect the observer to have some form of prior knowledge or a more specific expectation for a particular type of experience related to that particular area, or both. Although this explanation will have merit for future TR reports due to the fact that these experiences are now becoming known, the association of the room to these specific experiences and sensations was not generally known about when most of the striking TR-bed experiences occurred (between the late 1960s and mid 1990s). These particular experiences had never been openly reported in any media (local or national) at the time the individuals reported them. Therefore, although it is likely that general expectation and context are contributing to experiences, and indeed may be sufficient in some circumstances (i.e., reports from day tourists; see Houran & Lange, 2001), these factors alone do not obviously explain both the predominance and intensity of reports from overnight guests staying in the TR bed, at least as they are currently proposed.

It is also important to note that many of these observers were close family friends who lived in their own ancient houses and were quite used to being surrounded by such suggestive architecture. Furthermore, as the primary experiences were often unpleasant, the family did not openly discuss the reports from the TR, at least when these particular witnesses reported their experiences and this room was being used for overnight family guests. Although always aware and somewhat perturbed by TR reports, the family typically ignored them or viewed them as little more than a curiosity.

Instead, we suggest that the similarity of accounts to many reported physiological and experiential components from laboratory studies may implicate a potential magnetic stimulatory component to these particular reports. Preliminary evidence for this suggestion has been recently provided by Braithwaite (2004), who initially reported measuring a severe "undercurrent" magnetic anomaly in the TR. In that study, two magnetic sensors were placed in the TR, one in the pillow region of the bed (approximating the position of observers' heads during the reported experience). The other sensor was placed at some proximal distance from this at an estimated area from which observers had reported the phenomena (children crying) being emitted (Baseline sensor).

There were two important findings from that study. Firstly, a large and significant difference in magnetic field amplitude was measured between the sensors. This difference was in the region of 47,000 nT and is far outside what might be considered a usual difference for such a short distance between the sensors in the absence of any obvious artificial contributions (a large degree of spatial nonuniformity in terms of the accounts outlined earlier). Furthermore, the fields measured by the baseline sensor were far higher (in the region of 77,000 nT) than what would be predicted for the castle area (49,000 nT: British Geological Survey data), and the fields measured in the pillow region were lower (around 30,000 nT). To account for the increased fields measured in the baseline area, Braithwaite (2004) suggested contributions from both man-made and local geological sources combined. The reduction in amplitude in the pillow area was attributed primarily to a possible localized anomaly created by the heavy metal/iron lattice bed supports underneath the mattress of the TR bed. This lattice did not extend to the pillow area, but covered an area approximately from the ankles to the upper shoulders/chest area. The presence of such a magnetically permeable object may well have been distorting the background field away from the pillow area (which is supported by wood) and thus reducing the amplitudes in that area.

Secondly, there was a significant difference between the magnitudes of variance measured by the sensors. The variance in the crucial pillow area was far greater than that measured on the nearby baseline sensor placed in the same room a number of meters away. This difference occurred throughout the measuring period (4 hr) and appeared to be a constant component of the background variance. This variance is also in line with independent evidence gathered from other field studies (Wiseman et al., 2002, 2003) and the predictions from laboratory studies that suggest EIFs would need to be present for long enough to induce neurophysiological changes in observers (typically around 20-30 min: Persinger & Koren, 2001).

Aims of the Present Study

The purpose here was to further investigate the general claim that locations (and areas within them) associated with haunt-type experiences may be magnetically remarkable in some way. The present paper reports a more detailed follow-up study of the findings from Braithwaite (2004). More specifically, it explores further the temporal aspects of magnetic anomalies at Muncaster Castle by returning to investigate the nature of the TR undercurrent anomalies.

Over the years our investigation of witness reports has shown that the striking experience of children crying has been predominantly reported by observers when they were located in the TR bed. Knowing the approximate body/head position of a number of crucial eyewitnesses is helpful as it allows for the detailed investigation of a relatively specific region for the presence of magnetic anomalies. Braithwaite (2004) assumed that if some of the TR experiences (or some component of them) could be viewed as a magnetically induced hallucination, then perhaps such anomalies would be present around the head/pillow area of the bed in the TR. Arguably this represents a point in space where observers may have been exposed to stimulatory magnetic fields with experience-inducing properties. This means that a crucial sensor location was actually dictated by the experiential reports themselves (more specifically, the children crying experiences). In order to see if such magnetic fields were potentially interacting with neurocognitive processes, it was paramount to measure such fields in the exact region of the observers' heads when the experience occurred. Similarly, the choice of baseline sensor location was made on the pure premise of a location that was as close to the crucial sensor as possible but as far away as possible from all potential artificial sources of EMF, hence the middle of the room was the chosen location in the present study (see below).

As in the prior study, here simultaneous time-linked magnetic measurements were taken over a prolonged period of time. These included magnetic measurements taken from an area of high interest and a proximal synchronized baseline measurement from within the same room. However, there were a number of important differences between this study and that of Braithwaite (2004). Firstly, here we further tested the undercurrent anomaly account by placing one sensor in the same crucial location (pillow area) as that in Braithwaite's study, but for a more prolonged 6-hr continuous measuring period. If the fields measured in the prior study truly are a constantly available undercurrent, then we should replicate the findings of Braithwaite (2004) by also measuring them here. Furthermore, additional evidence of the constant nature of the fields would be provided by the use of a more prolonged measuring period here.

Secondly, a new baseline-sensor location was employed, which was more localized relative to the active sensor in the pillow area. The use of a new baseline area was very important to the investigation of the TR anomalies. For instance, in the initial study the difference in variance between the locations could be due either to excessive field variation in a crucial area (as argued), or to a somewhat reduced variance measured by the baseline (for whatever reason). Therefore, using a new baseline region would provide further important information as to whether the pillow area variation is indeed distinguishable from a number of other areas within the immediate vicinity. If time-based variances were still distinguishable in the pillow area compared to this new baseline location, then further evidence would be provided that the magnetic environment around the pillow area is somewhat remarkable relative to a number of other points in the immediate vicinity.

Thirdly, the original study revealed an unusual time-based varying component to the amplitudes of the fields measured by the active sensor placed in the pillow area. This return visit was designed specifically to investigate this time-based anomaly in the amplitude variance. To do so effectively, the gain-offset function of the magnetometer system (see below) was used. Basically, this procedure offsets the local ambient background fields present by canceling out (i.e., filtering) and zeroing the sensors relative to the available field. The sensors then measure all deviations that occur subsequently from this zero value (increments and decrements) over time. Although this procedure means that we do not know the "true" background field magnitude (i.e., a value of say 50,000 nT), it has the distinct advantage of showing all deviations in the same coordinate system (deviations from a zero point over time). It also allows for small transients to be detected even in the presence of a strong static (geomagnetic) field.

The magnetic measurements for this experiment were carried out using the dual sensor Magnetic Anomaly Detection System (MADS: Braithwaite, submitted). The MADS consists of two separate high-speed, 3-axes digital fluxgate magnetometers. The sensors are customized versions of the model 540 from Applied Physics Systems USA (see http:// www.appliedphysics.com). One sensor is labeled as the Active sensor (Sensor A) and is placed in regions of high interest. The other sensor is labeled as the Baseline sensor (Sensor B) and is placed at a baseline area or location. Each sensor provides a full 3-dimensional representation of the magnetic environment (axes x, y, z). These sensors are incredibly sensitive (down to 0.5 nT) and capable of measuring both the AC and DC components of the magnetic field. The MADS sensors are interfaced to their own individual dedicated laptop PCs (Dell computers) and are equipped with their own real-time data acquisition software. This provides a constant, real-time magnetic stream of time-series data measuring the environment.

Due to its high capabilities, the MADS can detail the magnetic microenvironment in a way previously not possible via other technologies. This is important not only when trying to grapple with the varying forms of magnetic complexity that could be crucial but also when trying to provide a legitimate and detailed magnetic representation of the natural setting at that time. It is highly likely that previous approaches have underestimated the level of detail needed and underrepresented the contributions from many magnetic factors. In light of this, the results section in the present paper details not only the overall ambient magnetic fields measured but also a formal comparison between the sensors in terms of (1) amplitude levels, (2) amplitude variation, and (3) a detailed decomposition of within-sensor directional contributions to the specific area (axes x, y, z), and between-sensor directional differences from the areas surveyed.

Based on standard geological and geomagnetic predictions, we would expect these separate components to be represented in a particular way. For instance, we would not predict large differences between sensors that were only a few meters apart unless this represented a severe localized anomaly in one area influencing only one of the sensors. Therefore, it is not an x, y, z difference per se that would be interesting (as this is normal), but a relative difference between the x, y, z contributions from both locations. Finally, a number of discrete pulse events measured in both sensor locations and a burst-pattern measured primarily in the pillow area were revealed in the data series. These were evaluated separately.

METHOD

Design and Procedure

The study was carried out over the course of one night from 11:45 p.m. on April 1, 2004 to 5:45 a.m. April 2, at Muncaster Castle, Ravenglass, Cumbria. Researchers present included the author (JJB) and fellow researcher Ian Topham (IT), who assisted in setting up the sensors and taking measurements (IT was blind to the motivation of the experiment). The Active MADS sensor (Sensor A) was placed just above the pillow area on the Tapestry Room bed at an approximate height of 110 cm from the floor to the middle of the sensor. The sensor was placed roughly in the middle of the pillow area, and it was approximately 120 cm from both side lamp fittings on either side of the bed. The Baseline MADS sensor (Sensor B) was placed at the same height from the ground as Sensor A but was located at an approximate midpoint in the room. The distance between the midpoints of both sensors was approximately 4 m 45 cm with Sensor B being placed diagonally south-southeast from Sensor A.

Each sensor was orientated (using a compass) so that the x-axis was pointing East/West, the y-axis was North/South, and the z-axis was Up/ Down and was fixed so that its orientation was parallel to the floor (assessed via a spirit level). The calibration of the sensors was checked before the experiment began following guidelines from the manufacturer. The sensors were configured to gather data at a rate of 33 samples per second (at 9600 baud). The docks on both laptop computers were synchronized (using the internet) and the data files configured so that they provided a time stamp with every reading. At the beginning of the measuring period, both sensors were configured so that they were zeroed into the ambient fields present and therefore responded only to deviations from this zero point as and when such events occurred. This process is done electronically by a command from the operating computer and takes approximately 10 s to complete. The zeroing process was carried out only once at the beginning of the study.

Before the beginning of the experiment, both sensors were placed together and a time-calibration test was carried out. This involved passing a low-strength magnet a few centimeters in front of the sensors, which produced a significant peak in the signals. These peaks were used in the data file for further time-based alignment calibrations. The measurement period lasted for 6 hr of continuous time-linked monitoring. All data gathered was recorded and stored automatically by the software on the laptop computers. In terms of nearby electrical devices, the room is equipped only with side lighting and these were left on. There were table lamps on either side of the Tapestry Room bed and two lamps situated on either side of a dressing table near the window area. With the exception of the sensors and a ceiling-mounted fire alarm, these were the only electrical devices in the room (and a similar arrangement is employed in all adjacent rooms). All other sources of man-made EMF would come from internal house wiring through the walls and floor. No individual entered the TR during the measurement period, and this was monitored by both the author (JJB) and IT.

RESULTS AND DISCUSSION

The results were analyzed in the following manner. Firstly, all measurements were actual deviations in magnetic amplitude from a zero point. This means that all values are a measure of variability (i.e., change) around a mean zero value. The data files from both sensors were checked and matched for the time calibration test and edited down so each sensor master file now contained 6-hr worth of raw magnetic time-series data. We then calculated overall descriptive statistics on both data sets for the full 6-hr worth of measurements. This included a mean total of the 3-axis combined (i.e., MagTotal which equates to energy from AC and DC fields summed together), a range, and a standard deviation for each of the axes (x, y, z). These values are given in Table 1. All values are given in nanoTesla (nT). We then calculated the mean variability and standard error for each time-linked 1-hr measuring session, from both sensors. These values are shown in Figure 1.

[FIGURE 1 OMITTED]

Analysis of Ambient Magnetic Field Variance

The background ambient magnetic field variability was much larger for Sensor A (pillow area) than for Sensor B (baseline area). This was indexed in a number of ways. For instance, we took the MagTotal values (overall combined field) and calculated an overall range (maximum minus minimum values) and standard deviation averaged across the 6-hr measuring period. This revealed an overall range of 195 nT (standard deviation = 33 nT) in the pillow area compared to 131 nT (standard deviation = 21 nT) from the baseline area. In terms of individual axes contributions, across both sensors, the overall highest variability in the magnetic fields (as indexed by standard deviation and range) came from the x axes (East/West). However, the range of values from the x axes was far higher overall for Sensor A (417 nT, standard deviation = 65 nT) relative to sensor B (213 nT, standard deviation = 34 nT).

These data were formally analyzed in the following way. Mean averages for the variability were calculated every 15 min (bins) for both sensors for the whole 6-hr measuring period. This process corrected the data for nonstationarity and parametric analysis (3). A 2 x 6 (Sensor x Session) mixed-subjects Analysis of Variance (ANOVA) was carried out with Sensor as the between-subjects factor and Session as the within-subjects factor. This revealed a significant main effect of Sensor, F (1, 6) = 1519.1, p <.001. The overall variability measured by Sensor A (pillow area) was significantly greater than that observed for Sensor B (baseline). The analysis also revealed a significant main effect of Session, F(5, 30) = 200.760, p<.001. The sessions produced reliably different field variability over the 6-hr measuring period. The Sensor x Session interaction was also significant, F (5, 30) = 19.320, p <.001. Figure 1 shows the data plotted for both regions over the six separate 1-hr sessions. From this figure, it is clear to see the large difference between the areas covered. The error bars represent 1 standard deviation above and below each mean for that session and clearly show the distribution of measurements. From this the greater overall variability in the pillow area can also be seen. The interaction seems to be mainly due to the increase in mean variability from Session 2 to Session 6 being greater in the baseline area relative to the pillow area (40 nT vs 12 nT, respectively).

Within-Sensor Analysis

The data gathered at each sensor location were then further assessed in relation to the individual axes contributions and analyzed further by dividing the variability from one axis into the variability from the other and calculating an F-ratio (with a Bonferroni comparison correction). This was done by averaging the sample series into 1-s mean (bins) for 120 s of time (2 min) at the beginning of Session 1. For Sensor A (pillow area), the variability measurements from the separate x, y, z axes were all significantly different from each other, (all F's > 4, all p's < .01). For Sensor B, the comparison between y (north/south) and z axes (up/down) did not approach significance, (F < 1.1, p > .05). All other comparisons were significant: (F's > 14, p's < .0; see Figure 2).

[FIGURE 2 OMITTED]

Between-Sensor Analysis

Data for all three axes were then compared to their counterparts across sensors in the way described above. An F-ratio was calculated in the same way and revealed a significant difference between the variability for both the x axis and y axis from the pillow area relative to the baseline area with x, F (119, 119) = 5.30, p <.001; and y = 9.36, p <.001. The difference between the z axis on both sensors failed to reach significance, z = 1.34, p >.05. These results demonstrate that on the whole, the variability in the magnetic field measured in the pillow area was much greater and statistically distinguishable from that obtained by the baseline sensor. However, these analyses also reveal what components and directions the highest variability contributions were coming from for both areas measured. Furthermore, by decomposing the magnetic measurements across the individual directional axes here, it can be seen that the relative differences between the axes for the pillow region are far greater (i.e., more disparate) when compared to the baseline area (see Figure 2).

Analysis of Transient Pulses Events

Over the 6-hr measuring period, a preliminary visual analysis of the data series revealed seven instances of what appeared to be large and transient pulses (large increases and decreases) in the magnetic data series. In all cases these pulses were measured in both sensor locations and in most instances could be seen to occur across all three axes (though they were predominantly represented in the x and y axes). These pulses were also generally symmetrical, with an approximate equal swing of increase and decrease in amplitude over the duration of the pulse. For illustration, Figure 3 shows one measured instance of a pulse across all three axes measured in the pillow area (Sensor A). All pulse instances across both sensors appeared similar to that in Figure 3, with the exception that the magnitude of the pulses was lower in the baseline area (see Table 2). These pulses were explored further here with respect to their (1) duration over time (how long they were present), (2) magnitude, and finally (3) regularity throughout the measuring period.

[FIGURE 3 OMITTED]

The average peak-to-peak swing in the pillow area was 236 nT compared to 111 nT in the baseline area (see Table 2). Put in perspective, for the pillow area these pulses were, on average, an increase in magnitude of approximately 120 nT (+/- 60 nT) over and above the background variability and approximately 60 nT (+/- 30 nT) for the baseline area. The average duration of all pulse events was approximately 450 ms (milliseconds), with the duration being 500 ms in the pillow area and 400 ms in the baseline area. This difference was not significant: t(12) = .420, p =.682 (see Table 3).

Analysis of Amplitude Patterns

In addition to revealing discrete pulses events, examining the data series also showed one instance of a peculiar magnetic pattern that we describe here as a burst-pattern. This pattern was measured in both areas (though more predominantly in the pillow area) and lasted for 2 min 50 s. The real time of pattern onset was 1:41 a.m. The characteristics of this pattern are not pulses in the sense of increased transient events relative to the background variability. Instead, this pattern is characterized as a reduction in background variability interspersed with periodic bursts back up to (but not above) the ambient level of variability. A section of this pattern is shown in Figure 4.

[FIGURE 4 OMITTED]

Starting at the left side of the data series, this figure shows the typical degree of ambient background variability (approximately and just over 150 nT at this time). Shortly after this the magnitude of variation drops considerably (to around 90 nT - 100 nT) and stays at this level for about 30 s. At this point there was a train of four bursts back up to the original ambient level of variability interspersed with periodic reduced levels of variability. The bursts were sustained for between 6 and 8 s (across the different instances) with the intermittent drop-off periods having a similar time period. After repeating this cycle four times, the variability then remained reduced for a further 30 s before this burst pattern was repeated again. This pattern was formally assessed in the following way. Firstly, all the magnetic bursts were separately removed from the series, then summed and trimmed to the nearest whole second. This provided 64 s' worth of raw data. These data were then averaged into 4-s mean bins (producing 16 separate mean values). To match this, a further 64 s worth of data was taken from the remaining period (taken randomly from the reduced period leading up to and in between the bursts) and averaged in the same way. We then calculated a variance statistic for each data series and divided the variances into each other to compute an F-ratio. We did this based on the total combined (MagTotal) magnetic values. This difference was significant, with F (15, 15) = 6.04, p <.01 in the pillow area. This analysis shows that the fast-changing differences in variability from the burst increases were significantly distinct from the reduced background variability. We repeated the analysis for the baseline area, where the same anomaly, though of a lesser amplitude, was measured, and this was also significant, F (15, 15) = 3.61, p <.01.

To summarize, the magnetic signatures measured in the TR pillow area were significantly distinguishable from those available in the immediate proximity. This is in terms of the overall background variability and both transient and sustained temporal anomalies. The fact that this area has also been associated with some striking instances of strange experience may also be no coincidence.

GENERAL DISCUSSION

The idea that some haunt-type experiences could be associated with magnetically remarkable environments has been gaining considerable currency over recent years (Braithwaite, 2004; Nichols & Roll, 1998, 1999; Persinger & Koren, 2001; Persinger et al., 2001; Roll & Nichols, 1999; Roll & Persinger, 2001; Wiseman et al., 2003; see Persinger & Koren, 2001; Roll & Persinger, 2001, for reviews). The tantalizing implication is that these magnetic microenvironments could influence and bias susceptible observers, inducing anomalous experiences, perceptions, and interpretations. However, few field studies have detailed the magnetic microenvironment in an appropriate way, and even fewer have reported detailed investigations of the anomalies themselves across studies.

The present paper reports a follow-up study of a recently documented magnetic anomaly (Braithwaite, 2004). As well as replicating the basic findings from the original study, the present investigation has revealed even more about the varieties of magnetic signatures that may distinguish reputedly haunted areas. The evidence provided here shows that a crucial region associated with numerous reports of striking haunttype experiences contains a number of magnetic components that may distinguish this area, both relative to other proximal regions and baseline locations. Based on the data provided so far these components include (1) significantly greater degrees of amplitude variability, (2) significantly more disparate magnetic contributions from all directions, and (3) increased burst patterns impinged upon and distorting the ambient background field.

It should be noted that these findings relate to a detailed investigation of one type of experience reported from a classic case of a reputedly haunted English castle. Future research will need to address whether these findings can be generalized to other experiences also reported from this location and whether they can be extended to other instances of similar experiences from elsewhere. These findings are next discussed in terms of the general anomaly accounts outlined earlier and their potential relevance to haunt-type reports in the natural setting.

Ambient Magnetic Variability

In line with previous findings, we found that the pillow area of the TR contained magnetic fields that varied significantly more than those measured at exactly the same time only a few meters away (see Braithwaite, 2004). This was the case throughout the measuring period. If we assume that the nature of this variability could be crucial, then it would appear that this particular anomaly exists as an undercurrent permanently available and present to be measured at any time. In many respects it is plausible to assume that this undercurrent could be sufficient to induce such experiential changes in observers.

Firstly, the range of variability encountered was not far from that used in laboratory studies of brain stimulation. They are also similar to, and above, the levels of variability measured in other field studies that were directly linked with questionnaire responses of strange perceptions and feelings (see Wiseman et al., 2003). Baseline measurements across both this and the original study have revealed a standard deviation of magnetic variability in the region of 15 nT to 20 nT, increasing to 30 nT to 50 nT in areas associated with anomalous reports. These values are comparable to other field studies. For instance, Wiseman et al. (2003) measured fields varying from around 11 nT, which were also linked to concurrent increases in anomalous interpretations given in questionnaire responses from individuals at that time. Therefore, it is not unprecedented that background variability of this magnitude (around 30 nT_to 50 nT) has been associated with experiential reports in the natural setting.

However, it is clearly the case that not everyone who has spent time in the TR bed has reported a strange experience. This may be due to the fact that such an undercurrent probably exerts its influence only on observers who also show increased signs of temporal-lobe instability (Makarec & Persinger, 1987; Persinger, 1983, 1984, 1988; Persinger & Makarec, 1993; Persinger & Roll, 1985; see Persinger & Koren, 2001 for a review) or particular forms of attentional biases (Houran, 2000; Lange & Houran, 1997, 2001). Note also that in the laboratory it is typical for participants to undergo at least 20 to 30 rain exposure before any experiential effects take place and are reported. This highlights a possible indirect mechanism that requires a more prolonged period of exposure before such energetic components are fully recruited into the experiential gestalt. In the natural setting, as long as the varying fields are readily available, it is likely that at some point favorable positions, level of arousal, and an appropriate degree of susceptibility could co-occur, the consequence of which could be some form of anomalous experience or interpretation. However, from the point of view of the magnetic field itself, it merely needs to be present and available over a sustained or prolonged period in order to exert its potential effects.

Magnetic Pulses and Patterns

We reported seven instances of large and transient pulses (large increases and decreases) in the magnetic data series over the 6-hr measuring period. In all cases these pulses were measured in both sensor locations but were higher in overall magnitude in the pillow area relative to the baseline region. In terms of the types of magnetic anomalies outlined earlier, these could certainly be classed as transient anomalies, existing, as they did, for less than 1 s. It is likely that these events have their basis in artificial sources. Evidence for this comes from the fact that there was a certain degree of regularity between the pulse instances (generally in the region of every 50 min) and the fact that the magnetic variability increased slightly after the pulse events. (4)

Although these pulse events are certainly interesting and are important to document, they are unlikely to contain experience-inducing properties on their own, as they occurred here. As such we do not believe they are themselves an index of any important EIF component. There are a number of reasons for not viewing these events as particularly important for haunt reports, the main one being that the time period between the pulses is quite large. This makes it very unlikely that these temporally disparate transient events alone could sustain any prolonged neurophysiological shifts that may have consequences for experience.

Perhaps more interesting was the detection of a specific period where a peculiar burst-pattern occurred. This pattern was measured in both areas, though to an increased degree in the pillow area. At nearly 3-min duration, the pattern was certainly more sustained than that measured for the pulses and can be characterized in terms of a more sustained transient account. Although the presence of this pattern was more prolonged, it still occurred on only one occasion throughout the 6-hr worth of measurements (and it was not present in the previous study at all). In this sense it fits the description outlined earlier. The burst-pattern appears to be best described as periodic reductions in the background variability, followed by equally periodic bursts or returns to more usual levels of variation.

In relation to potentially important components of EIFs, this magnetic pattern is interesting in a number of ways. Firstly, the magnitude of the variability was well above that documented elsewhere as being associated with haunt-type reports from the natural setting (around 11nT reported by Wiseman et al., 2003; and Stevens, personal communication). Here the total variability was often in excess of 60 nT. Secondly, this particular pattern does bear some similarity to the amplitude-modulated sequences used to stimulate brains artificially in the laboratory (see Persinger, 1999a, 1999b; Persinger & Koren, 2001; for discussions). Therefore, at least in principle, this spontaneously occurring incarnation of a complex burst-pattern has a certain degree of isomorphism to known experience-inducing parameters (at least in terms of amplitude variability). Thirdly, the magnitude of the pattern was substantially increased in the crucial pillow area (in a region where observers' heads would be located) relative to a baseline area. In addition, the existence of such a pattern can be taken as further indication of a much more complex magnetic environment existing in a crucial region. Here, the magnetic fields are not just varying more in the pillow area, but the patterns of variability themselves are changing over time in a way not occurring elsewhere in the room to that degree (at least as indexed by the areas so far surveyed across the studies).

Furthermore, the sustained duration of this pattern may also underlie its potential importance. Although the measured instance here (3 min) was somewhat sustained, it was considerably shorter than that known to be effective in the laboratory. Nevertheless, there may be other contributing factors here worth considering. For instance, for certain hypersensitive individuals, in certain conducive mental states, expectation and levels of arousal (i.e., sleeping/drowsy/resting), this period of complexity may be sufficient to initiate a process that could become integrated into neural temporal codes and propagate through sensitive areas, disrupting stable firing patterns. This process could certainly continue long after the magnetic field pattern has ceased by a process perhaps similar to that of cognitive-kindling (Persinger, 1993a). Therefore, in certain circumstances, it may not be necessary for the magnetic anomaly to be present to set up and sustain its experiential influence. Instead, merely being present for a period sufficient to initiate a process--which can then be continued, amplified, and propagated within neural structures--may be enough. This is indeed one possibility.

However, another possibility relates to the duration itself. For instance, it is not clear as to whether the duration of 3 min is a typical duration for this type of pattern. If this pattern is important and does occur periodically (for whatever reason), then it may be the case that the duration could vary across instances, dependent on other factors. Therefore it remains to be seen whether this particular magnetic signature will occur again, and if so, how it may differ from the current documented event. These tantalizing possibilities will be addressed in future studies as part of the ongoing detailed investigation of this location.

Irrespective of these possibilities, the present data clearly show further indexes for the magnetically complex nature of a crucial area associated with a number of haunt-type experiences. These anomalies have been presented here in some detail and have revealed characteristics similar to those used in the laboratory setting with known experience-inducing properties. (5) These findings provide further evidence for distinguishing the magnetically remarkable characteristics of crucial areas associated with haunt-type reports.

Baseline Locations

The data gathered by our present magnetic investigation further testify to the magnetically remarkable signatures available in the locale. It is clear from these findings that there is a case for taking a number of measurements from baseline locations with which to couple the present data. As well as taking a number of time-linked measurements from within the TR, and further baseline measurements from other rooms of the castle (Braithwaite, 2004), we have also taken some comparison measurements from a number of separate baseline locations. These measurements have come from (1) modern living environments and (2) another stately building matched for structural and geological similarity (granite-based structure). These baseline locations represent detailed studies in themselves, and a full description of findings is outside the scope of the present article.

However, some points are worth noting. Firstly, we have surveyed a number of standard modern living environments in low-density population areas (mainly in rural Cumbria) and high-density population areas (London and Birmingham). These give parameter estimates for differing levels of electromagnetic demand and contribution. In both cases we tend to find internal background variability ranging from between 5 nT and 50 nT, with variations within this range. An average standard deviation would be around 2 to 8 nT. Transient pulses very similar to those reported here have also been documented in the region of around 50 nT to 100 nT (which again supports our contention that these pulses are not a crucial component of EIFs). These findings seem relatively normal, at least for the locations covered so far.

This point is interesting for a number of reasons. For instance, the modern home environment is likely to contain more modern electrical appliances, distributed across a more compact area in a more frequent user-demand situation (i.e., demands from other local inhabitants such as neighbors) relative to those in a stand-alone stately home in which a majority of rooms are only period-furnished. In this sense, we might expect the magnetic fields to vary more in the modern home due to increased manmade sources present in a more condensed living space. However, this does not appear to be the case. The amplitudes and variances measured at Muncaster Castle are many times greater than those measured at a variety of other modern baseline locations. This may be a crucial distinguishing factor and suggests that the local geology may be important (in conjunction with any wiring configurations specific to the location). This may help to explain why the castle seems to contain a much-increased level of variability relative to baseline locations (as an overall feature), and this variability is increased significantly more in areas associated with haunt reports. Future studies are planned to assess both geological and artificial contributions to the magnetic signatures at this exciting and prominent case of a classic English haunting.

Our investigation of this location is longitudinal, and at this point we are more concerned with detailing the presence and nature of any anomalies, their potential experience-inducing properties, and their behavior in the natural spontaneous setting, rather than speculating about the potential contributing sources underlying them. We suggest that a full description of all contributory factors would likely consist of sources from (1) artificial man-made factors such as internal wiring configurations and appliances, (2) higher density of magnetic minerals in the granite stone surrounding the castle and potentially used in the castle structure, and (3) local geology and the presence of tectonic fault lines in the immediate vicinity (see Persinger & Koren, 2001; Roll & Persinger, 2001). Our initial geological and geochemical examination of the site has started to confirm that the castle is built on top of a localized geological granite intrusion that contains anomalous concentrations of specularite and possibly magnetite, both of which are naturally magnetic (see British Geological Survey, sheet 37; Akhurst, Chadwick, Holliday, McCormac, McMillian, Millward, & Young, 1997; Millward, Johnson, Beddoe-Stephens, Young, Kneller, Lee, & Fortey, 2000). Furthermore, the castle is placed in the immediate vicinity of two major tectonic fault lines: the Eskdale fault and the Windgate fault (Akhurst, personal communication; Akhurst et al, 1997; British Geological Survey, sheet 37). The Eskdale Fault (clearly visible on satellite images) passes approximately 10 m SSE (south/south-east) of the castle, whereas the Windgate Fault is around 100 m west of the building. We are currently conducting a more detailed examination of the geochemical and geological context as part of a further study currently being undertaken.

Conclusions

This study reports a further detailed investigation of magnetically remarkable signatures possibly underlying some reported haunt-type experiences. Further evidence has been provided that locations, and specific areas within them, can be differentiated in terms of their magnetic variability. It is perhaps no coincidence that these areas also appear to be associated with some striking anomalous reports. The level of magnetic variability measured in crucial areas here are comparable to other field studies that were also linked to concurrent increases in anomalous perceptions given in questionnaire responses from individuals (Wiseman et al. 2002, 2003). Therefore, it is not unprecedented that background variability of less than those documented here have been associated with experiential reports.

As well as background variability, additional components of complexity have been revealed here by the more disparate contributions from the separate axes in the pillow region (relative to the baseline), indicating a substantial distortion in that area, and the existence of a modulated pattern measured primarily in the crucial area. Therefore, it would seem that more indexes of magnetic complexity have been shown for these crucial regions reported here than in many other field studies. Because the laboratory studies highlight complexity as a crucial component, support for these dimensions spontaneously occurring here would seem important. The magnetic signatures we have measured clearly index an environment that is distinguishable from baseline areas surveyed at varying levels of proximity inside the TR, other rooms in the same location, and from other separate locations. It is therefore perhaps no coincidence that a relatively specific given point in space (the pillow area) has produced some remarkable instances of haunt-type reports.

Finally, the present findings are part of a long-term investigation that has been in place at this location since 1992. Future studies are planned to investigate the magnetic environment further as part of this longitudinal approach detailing a promising spontaneous case of a classic English castle haunting. If the laboratory studies are correct, then the predictions from their findings must be pursued and applied in the natural setting, particularly at promising locations that have been, and continue to be, associated with haunt-type experiences. A major aim of the future research at this location is to provide a detailed analysis of the actual frequency components available. Typically, geological and artificial fields occupy different regions in the frequency spectrum and should be easily identifiable (i.e., standard man-made sources = 50 Hz in the UK; geologically defined fields would be much lower). By carrying out a detailed frequency analysis we can ascertain how this amplitude variability is being shared across frequency components and better evaluate the contributions and their potential sources. Furthermore, these return studies will include comprehensive spatiotemporal surveys and investigations of the fields measured, both of which will add new dimensions to the scientific investigation of a reputed haunting.

REFERENCES

AKHURST, M.C., CHADWICK, R.A., HOLLIDAY, D.W., McCORMAC, M., McMILLAN, A.A., MILLWARD, D., & YOUNG, B. (1997). The geology of the West Cumbria district. Memoir of the British Geological Survey, Sheets 28, 37 and 47 (England and Wales).

BELL, G.B., MARINO, A.A., & CHESSON, A.L. (1992). Alterations in brain electrical activity caused by magnetic fields: Detecting the detection process. Electroencephalography and Clinical Neurophysiology, 83, 389-397.

BELL, G.B., MARINO, A.A., & CHESSON, A.L. (1994). Frequency-specific responses in the human brain caused by electromagnetic-fields. Journal of Neuro Sciences, 123, 26-32.

BRAITHWAITE, J.J. (2004). Magnetic variances associated with 'haunt-type' experiences: A comparison using time-synchronised baseline measurements. European Journal of Parapsychology, 19, 3-28.

BRAITHWAITE, J. J. (submitted). Using digital magnetometry to quantify anomalous magnetic fields associated with spontaneous strange experiences: The Magnetic Anomaly Detection System (M.A.D.S). Journal of Parapsychology.

BRITISH GEOLOGICAL SURVEY (1999). Geological Map 1:50000 Sheet 37 Gosforth, Solid and Drift.

COOK, C.M., & PERSINGER, M.A. (2001). Geophysical variables and behavior: XCII. Experimental elicitation of the experience of a sentient being by right hemispheric, weak magnetic fields: interaction with temporal lobe sensitivity. Perceptual and Motor Skills, 2, 447-448.

FULLER, M., DOBSON, J, WIESER, H.G., & MOSER, S. (1995). On the sensitivity of the human brain to magnetic fields: Evocation of epileptiform activity. Brain Research Bulletin, 36, 155-159.

GEARHART, L., & PERSINGER, M.A. (1986). Geophysical variables and behavior: XXXIII. Onset of historical poltergeist episodes with sudden increases in geomagnetic activity. Perceptual & Motor Skills, 62, 463-466.

HOURAN, J. (2000). Toward a psychology of "entity encounter experiences." Journal of the Society for Psychical Research, 64, 141-158.

HOURAN, J., & BRUGGER, P. (2000). The need for independent control sites: A methodological suggestion with special reference to haunting and poltergeist field research. European Journal of Parapsychology, 15, 30-45.

LANGE, R., & HOURAN, J. (1997). Context-induced paranormal experiences: Support for Houran and Lange's model of haunting phenomena. Perceptual and Motor Skills, 84, 1455-1458.

LANGE, R., & HOURAN, J. (2001). Ambiguous stimuli brought to life: The psychological dynamics of hauntings and poltergeists. In J. Houran and R. Lange (Eds.), Hauntings and poltergeists: Multidisciplinary perspectives (pp. 280-306). Jefferson, NC: McFarland.

MAHER, M.C. (2000). Quantitative investigation of the General Wayne Inn. Journal of Parapsychology, 63, 47-80.

MAILER, M.C., & HANSEN, G.P. (1997). Quantitative investigation of a legally disputed "haunted house." Proceedings of Presented Papers: The Parapsychological Association 40th Annual Convention, 184-201.

MAKAREC, K., & PERSINGER, M.A. (1987). Electroencephalographic correlates of temporal lobe signs and imagings. Perceptual and Motor Skills, 64, 1124-1126.

McCUE, P.A. (2002). Theories of haunting: A critical overview. Journal of the Society for Psychical Research, 66 (866), 1-21.

MILLWARD, D., JOHNSON, E.W., BEDDOE-STEPHENS, B., YOUNG, B., KNEELER, B.C., LEE, M.K., & FORTEY, N.J. (2000). Geology of the Ambleside District. Memoir of the British Geological Survey, Sheet 38 (England and Wales).

NICHOLS, A., & ROLL, W.G. (1998). The Jacksonville water poltergeist: Electromagnetic and neuropsychological aspects. Proceedings of Presented Papers: The Parapsychological Association 41st Annual Convention, 97-107.

NICHOLS, A., & ROLL, W.G. (1999). Discovery of electromagnetic anomalies at two reputedly haunted castles in Scandinavia, Proceedings of Presented Papers: The Parapsychological Association 42nd Annual Convention, 219-236.

PERSINGER, M.A. (1974a). The paranormal: Part I: Patterns. MSS, New York.

PERSINGER, M.A. (1974b). The paranormal: Part II: Mechanisms and models. MSS, New York.

PERSINGER, M.A. (1983). Religious and mystical experiences as artifacts or temporal lobe function: A general hypothesis. Perceptual and Motor Skills, 57, 1255-1262.

PERSINGER, M.A. (1984). Propensity to report paranormal experiences is correlated with temporal lobe signs. Perceptual and Motor Skills, 59, 583-586.

PERSINGER, M.A. (1985). Geophysical variables and behaviour: XXII. The tectonic strain continuum of unusual events. Perceptual and Motor Skills, 60, 59-65.

PERSINGER, M.A. (1988). Increased geomagnetic activity and the occurrence of bereavement hallucinations: Evidence for a melatonin mediated microseizuring in the temporal lobe? Neuroscience Letters, 88, 271-274.

PERSINGER, M. A. (1993a). Average diurnal changes in melatonin levels are associated with hourly incidence of bereavement apparitions: Support for the hypothesis of temporal (limbic) lobe microseizuring. Perceptual & Motor Skills, 76, 444-446.

PERSINGER, M.A. (1993b). Transcendental meditation and general meditation are associated with enhanced complex partial eplileptic-like signs: Evidence for "cognitive" kindling? Perceptual and Motor Skills, 76, 80-82.

PERSINGER, M.A. (1999a). Increased emergence of alpha activity over the left but not the right temporal lobe within a dark acoustic chamber: Differential response to the left but not to the right hemisphere to transcerebral magnetic fields. International Journal of Psychophysiology, 34, 163-169.

PERSINGER, M.A. (1999b). Near-death experiences and ecstasy: A product of the organization of the human brain? In S. Della Sala (Ed.), Mind myths: Exploring popular assumptions about the mind and brain (pp. 85-99). New York: Wiley.

PERSINGER, M.A., & CAMERON, R.A. (1986). Are earth faults at fault in some poltergeist-like episodes? Journal of the American Society for Psychical Research, 61, 49-73.

PERSINGER, M.A., & KOREN, S.A. (2001). Predicting the characteristics of haunt phenomena from geomagnetic factors and brain sensitivity: Evidence from field and experimental studies. In J. Houran & R. Lange (Eds.), Hauntings and poltergeists: Multidisciplinary perspectives (pp.. 179-194). Jefferson, NC: McFarland

PERSINGER, M.A., KOREN, S.A., & O'CONNOR, R.P. (2001). Geophysical variables and behaviour: CIV. power-frequency magnetic field transients (5 microtesla) and reports of haunt experiences within an electronically dense house. Perceptual and Motor Skills, 3, 673-674.

PERSINGER, M. A., LUDWIG, H. W., & OSSENKOPP, K. P. (1973). Psychophysiological effects of extremely low frequency electromagnetic fields: A review. Perceptual and Motor Skills, 36, 1131-1159.

PERSINGER, M.A., & MAKAREC, K. (1993). Complex partial epileptic-like signs as a continuum from normals to epileptics. Normative data and clinical populations. Journal of Clinical Psychology, 49, 33-45.

PERSINGER, M.A., & RICHARDS, P.M. (1994). Quantitative electroencephalographic validation of the left and right temporal lobe indicators in normal people. Perceptual and Motor Skills, 79, 1571-1578.

PERSINGER, M. A., RICHARDS, P. M., & KOREN, S. A. (1997). Differential entrainment of electroencephalographic activity by weak complex electromagnetic fields. Perceptual and Motor Skills, 84, 527-536.

PERSINGER, M. A., & ROLL, W. G. (Eds.). (1985). The temporal lobe factor in psi phenomena. Metuchen, N.J. & London: Scarecrow Press.

PERSINGER, M. A., TILLER, S. G., & KOREN, S. A. (2000). Experimental stimulation of a haunt experience and elicitation of paroxysmal electroencephalographic activity by transcerebral complex magnetic fields: Induction of a synthetic "ghost"? Perceptual and Motor Skills, 90, 659-674.

RICHARDS, P. M., PERSINGER, M. A., & KOREN, S. A. (1993). Modification of activation and elevation properties of narratives by weak complex magnetic field patterns that stimulate limbic burst. International Journal of Neuroscience, 71, 71-85.

ROLL, W. G., MAHER, M. C., & BROWN, B. (1992). An investigation of reported haunting occurrences in a Japanese restaurant in Georgia. Proceedings of Presented Papers: The Parapsychological Association 35th Annual Convention, 151-168.

ROLL, W. G., & NICHOLS, A. (1999). A haunting at an army post. Proceedings of Presented Papers: The Parapsychological Association 39th Annual Convention, 289-307.

ROLL, W. G., & PERSINGER, M. A. (2001). Investigations of poltergeists and haunts: A review and interpretation. In J. Houran & R. Lange (Eds.), Hauntings and poltergeists: Multidisciplinary perspectives (pp. 123-163). Jefferson, NC: McFarland.

TANDY, V. (2000). Something in the cellar. Journal of the Society for Psychical Research, 64, 129-140.

TANDY, V., & LAWRENCE, T. R. (1998). The ghost in the machine. Journal of the Society for Psychical Research, 62, 360-364.

WISEMAN, R., WATT, G., GREENING, E., STEVENS, P., & O'KEEFFE, C. (2002). An investigation into the alleged haunting of Hampton Court Palace: Psychological variables and magnetic fields. Journal of Parapsychology, 66, 387-408.

WISEMAN, R., WATT, C., STEVENS, P., GREENING, E., & O'KEEFFE, C. (2003). An investigation into alleged "hauntings." British Journal of Psychology, 94, 195-211.

Behavioural Brain Sciences Centre

School of Psychology

University of Birmingham

Edgbaston, Birmingham, B15 2TT, UK

j.j.braithwaite@bham.ac.uk

ACKNOWLEDGEMENTS

We would like to thank the Pennington family and all the staff at Muncaster Castle for their continued support with our research over the years. This research was supported, in part, with a research grant from the Society for Psychical Research (SPR) awarded to the first author. We would also like to thank John Reid and all at ASSAP (the Association for the Scientific Study of Anomalous Phenomena: an educational charity) for financially supporting the MADS.

NOTES

(1.) Braithwaite (2004) introduced the generic term EIFs (Experience Inducing Fields) to encompass the variety of magnetic fields/electromagnetic fields that could have stimulatory properties. It is a helpful term that functionally describes the outcome of exposure to such complex environments (be they AC or DC). Of course the general term of EIFs could also be equally effectively applied to encompass other variables as well (i.e., infrasound).

(2.) Note: we take the artificially created EIFs from laboratory studies as a guide only to the crucial components, both necessary and sufficient, for spontaneous EIFs. However, this does not imply that naturally occurring fields with experience-inducing properties will be a literal incarnation of them. The laboratory studies have identified one method and the most potent components for stimulation to occur. This does not mean that it is the only method, and as such, other types of fields could have experiential properties (though this remains to be seen). Nevertheless, it is more than reasonable to apply the fundamental properties identified from these studies to the natural situation in the first instance.

(3.) The use of parametric statistics on these data might, at first, appear problematic. This is mainly because magnetic-field data are known to be inherently non-stationary and thus seriously violate important assumptions about using such procedures. There are a number of ways around this problem. One way is to use nonparametric procedures that do not make such normality assumptions. The other way is to correct for the nonstationarity by either a subtracting process or averaging the raw measurements into discrete mean averages. These mean values are normally distributed around a central mean value (known as the sampling distribution of the mean). Here, we adopted the second option and corrected the data by calculating a series of means. The ANOVAs were carried out on these corrected values.

(4.) We note that there is a slight increase in variability after the pulse event occurs. This is consistent with some device/settings switching on and placing a higher demand on the power supply.

(5.) Note that such patterns would be completely missed by most commercially available magnetometers employing much slower sampling periods than those used here.

TABLE 1
DESCRIPTIVE STATISTICS FOR BOTH SENSOR A (PILLOw AREA) AND SENSOR B
(BASELINE AREA) AVERAGED ACROSS ALL SESSIONS, WITH ALL VALUES GIVEN
IN NANOTESLA (NT)

           Sensor A (pillow area)  Sensor B (baseline)

                    Std                     Std
           Mean     Dev    Range   Mean     Dev    Range

Mag X      00076   00065   00417   00033   00034   00213
Mag Y      00052   00031   00206   00017   00009   00060
Mag Z      00022   00017   00103   00020   00009   00062
MagTotal   00116   00033   00195   00052   00021   00131

TABLE 2
DESCRIPTIVE STATISTICS FOR PULSE AMPLITUDE MEASURED
ACROSS BOTH SENSOR LOCATIONS, AMPLITUDE SHOWN IN
NANO TESLA (nT), VALUES SHOWN AS OVERALL. CHANGES
ABOVE THE BACKGROUND VARIABILITY (PEAK-TO-PEAK) AND AS
INCREASES AND DECREASES (+/-)

Pulse #   Session #   Amplitude (pillow area) Amplitude (baseline)
                      Overall       + / -     Overall     + / -

   1          1        216  nT     108 nT      122 nT     61 nT
   2          2        250  nT     125 nT      100 nT     50 nT
   3          2        222  nT     111 nT      92  nT     46 nT
   4          3        228  nT     114 nT      108 nT     54 nT
   5          4        234  nT     117 nT      88 nT      44 nT
   6          5        244  nT     122 nT      156 nT     78 nT
   7          6        259  nT     129 nT      110 nT     55 nT
Average       --       236  nT     118 nT      111 nT     55 nT

Note. The intensity of the pulses was greater (more than double)
in the pillow area than in the baseline area.

TABLE 3
DESCRIPTIVE STATISTICS FOR PULSE DURATION, MEASURED
ACROSS BOTH SENSOR LOCATIONS, DURATION GIVEN IN MILLISECONDS (MS)
ROUNDED-UP TO THE NEAREST 100 MS

Pulse      Session    Duration (pillow    Duration (baseline)
Number     Number     area)

   1          1            500 ms               300 ms
   2          2            400 ms               400 ms
   3          2            400 ms               400 ms
   4          3            400 ms               200 ms
   5          4            500 ms               300 ms
   6          5            500 ms               700 ms
   7          6            300 ms               500 ms
Average      --            500 ms               400 ms

COPYRIGHT 2004 Parapsychology Press
COPYRIGHT 2008 Gale, Cengage Learning