As discussed in Chapter 3, the methodology of the excavation was conducted in order to record the pebbles in a very detailed way. As a result, pebbles from a great number of excavated squares were measured, weighed and described in terms of colour and unique features, which provided a comprehensive account of the raw material used to build Tor Cairn (TC), Little Tor Cairn (LTC) and Twin Cairn A (TCA). The methodology also allowed for a comparison between the pebbles selected to build the cairns and the pebbles found in natural test samples.

General statistics on the cairn pebbles

Tor Cairn

All the pebbles used for the construction of the excavated SE quadrant of Tor Cairn were counted. They numbered 7,705. Pebble frequency in the individual layers of the 1 m excavation units varied between c. 100 and 200 pebbles. We estimate on the basis of these figures that the total number of pebbles used to construct the cairn was around 31,000. This number of pebbles could easily have been collected and used to build the cairn by a small kin group in a matter of two to three weeks. The amount of labour time expended would have been relatively small if they were collected from the ground surface in the vicinity of the cairn up to 100 m or so away or quarried from shallow scoops made in the valley slope to its west nearest to the cairn. This would have been far simpler than digging vertical pits or trenches and was the modern quarrying method for obtaining pebbles used in the eighteenth and nineteenth centuries in the area (see Chapter 13). However, this account assumes that the building of the cairn was a single and continuous event, which it was not, as discussed in Chapter 3.

Little Tor Cairn

In total 4,560 pebbles were recorded from the entire cairn. In the western half excavated in 2009 there were four distinct pebble layers; on the eastern side, excavated in 2010, five; and in the centre an additional two layers covering an area of approximately 0.5 m in diameter with a total of 452 pebbles (10 per cent of the total). These made up the inner core of the cairn. The total weight of these pebbles is 723 kg. If all the pebbles were laid in a row it would stretch for 85 m. But this is a very modest structure that could have been constructed by a small group of people in a matter of days.

Twin Cairn A

Since the pebbles at TC and LTC had been studied in great detail prior to the Twin Cairn excavations and no evidence was found of deliberate spatial patterning of different pebble shapes or colours in the various pebble layers or different areas of these cairns, a decision was made to document in detail only the pebbles in the NE quadrant of TCA. In all there were 11 distinct levels of pebbles in the SW quadrant and 12 in the NE quadrant. In addition there were three pebble layers in the central depression (Tables 4.3–4.8). The total number of pebbles in the NE quadrant excluding those from the depression was 947, so we can estimate that the cairn itself was made up of around 4,000 pebbles. This is roughly the same number as those counted for the entirety of LTC: 4,560.

Statistical pebble analysis: length, weight, shape, colours and specials

One characteristic of the geology of the Pebblebeds is that the pebbles are ungraded or only crudely sorted by natural processes, contrasting with beach pebbles that are graded by the tides. The pebble strata in the cliffs, which in places reach a maximum thickness of 30 m, are characteristically interleaved with silty sand and grit lenses formed by the great Triassic river. Pebbles of markedly different sizes and shapes, 3 cm or less in size to a maximum diameter of 45 cm, can occur together where they can be observed in the vertical cliffs along the coast at Budleigh Salterton or inland in quarry exposures (Edwards and Scrivener 1999: 91 and see the discussion in Chapter 1).

1. Pebble length

After sorting out the broken pebbles, the unbroken ones had their longest length measured and recorded. Statistics for each cairn and a comparison between them and the natural test samples are presented below.

According to a geological definition repeated in some excavation manuals the definition of a pebble is that it is a water-rounded stone up to 6.4 cm maximum dimension. The term cobble is used for water-rounded stones larger than that. We do not find this arbitrary distinction useful and indeed reject it as having any descriptive or analytical value in discussing the cultural use and value of pebbles.

Tor Cairn

We recorded pebble size of unbroken pebbles in terms of their maximum length. The vast majority of them in Tor Cairn, c. 90–100 per cent (the precise figure varying a little between the different levels of the cairn and from one 1-m-square excavation unit to another), measure between 5 cm and 15 cm, or about the same size as a baking potato. Very few pebbles are more than 15 cm long and between 5 and 10 per cent are less than 5 cm in length (or diameter in the case of round pebbles). Larger or smaller pebbles are not clustered in different areas of the cairn. However, smaller pebbles less than 5 cm long are more frequent in the three pebble layers of the primary cairn (up to 9 per cent) and none here are over 15 cm long. So the secondary cairn differs from the primary cairn in that there is a higher proportion of larger pebbles and only a few very large pebbles are present: a slight change in the selection criteria for the appropriate size of pebble (for the structure of the cairn see Chapter 3). As a matter of comparison, the mean length of pebbles in the excavated test sample trenches near to the cairn was only 5 cm, while in the 11 more distant test samples the mean length varied between 7 cm and 8 cm.

Little Tor Cairn

There is a consistent patterning of the pebbles in terms of size in Little Tor Cairn. Disregarding the inner core consisting of small pebbles 5 cm or less in size, the largest pebbles are found in the top and the bottom levels of the cairn on the western side: levels 1 and 4. About 92 per cent of the pebbles in the top and basal levels measure between 5.1 cm and 20 cm. In the middle levels this figure drops to around 75 per cent. The presence of small pebbles (up to 5 cm long) in the middle levels is around 25 per cent. Thus the cairn has the structure of a sandwich, with smaller pebbles being more frequent in the middle layers.

Twin Cairn A

Most of the pebbles used to construct the Twin Cairn A were between 5 cm and 15 cm long with only minor variations from level to level. The largest pebbles were those used to construct the kerb (Table 4.3 and Table 4.4).

Figure 4.3 compares the length of the recorded pebbles from the five units of analysis and demonstrates that in all analysed squares the majority of pebbles measured between 5.1 cm and 15 cm. Although some similarities can be noticed between the pebbles from the cairns and the test sample squares, chi-square test results support the idea that there was a positive selection of pebbles in the construction of the three cairns, since the frequency pattern found in the cairns differs substantially from that observed in the natural and beach test squares at the highest significance level.

2. Pebble weight

All unbroken pebbles from natural samples and the cairns were weighed. The results of the analysis of the weight of the cairn pebbles are discussed below.

Tor Cairn

The weights of the individual pebbles in Tor Cairn may be up to 1 kg or more but most weigh between 100 g and 300 g and it is clear that these smallerand lighter pebbles were chosen for preference. The heaviest pebble in TC, weighing 5.5 kg, was found in the basal layer at the centre of the cairn. One way to understand the choice of smaller and lighter pebbles for the construction of the primary cairn in the approximate centre of the cairn (context 019 – see Chapter 3) may be that they were picked up in the immediate vicinity. The secondary cairn seems to have been built with pebbles from further afield, as the excavated test trenches near TC are composed only of lighter pebbles weighing around 100 g. The average weight of the pebbles from the 11 more distant test samples was between 250 g and 450 g.

Little Tor Cairn

The proportion of pebbles weighing over 1 kg remains constant throughout levels 1–4 of quadrant N1 (10–11 per cent). The highest frequency of small pebbles weighing less than 250 g occurs in levels 2 and 4, replicating the sandwich effect of smaller and larger pebbles in relation to pebble size noted above. In S2 the picture is markedly different, with the proportion of pebbles gradually decreasing in weight down the cairn from only 29 per cent weighing less than 250 g in layer 1 to 60 per cent in layer 4. In N4 between 10 and 15 per cent of the pebbles weigh over 1 kg in layers 1–3. This falls to 5 per cent in layer 4 with only around 0.5 per cent in the basal layers L5 and L6. In L1 42 per cent of pebbles weigh less than 250 g, declining to between 27 per cent and 33 per cent in levels 2 and 3 and then successively rising from 60 per cent in layer 4 to 96 per cent in layer 6. In S3 the frequency of heavy pebbles weighing over 1 kg is much lower in all levels and overall there is a significant decrease in pebble weight with depth from 55 per cent weighing less than 250 g in L1 to 77 to 82 per cent in basal levels L6 and L7. In all levels the majority of the pebbles, 62–97 per cent, weigh less than 500 g.

Twin Cairn A

As is the case for LTC, the vast majority of the pebbles in all the layers of Twin Cairn weighed between c. 100 g and 500 g, with comparatively little variation from level to level. In each level there were just a few heavy pebbles weighing over 1 kg and the number of pebbles weighing more than 750 g is small in all cases. The cairn is thus fairly uniform in construction in terms of pebble weight.

Figure 4.4 provides a comparison of all excavated units in relation to the frequency of pebble weight. Again the statistics do not point to any substantial differences in the choice of pebbles used to build the three cairns. However, chi-square tests strongly support the idea that pebbles were positively selected to build the three cairns because their weight differs from all pebbles collected from the natural test squares and the beach samples (see section on chi-square tests below).

3. Pebble shape

The unbroken recorded pebbles had their shape classified according to three categories: oval, round and irregular.

Tor Cairn

Most pebbles in Tor Cairn were irregular or oval in form with very few (5 per cent or less) being round. Frequencies of these varied between the pebble layers in the cairn and according to excavation unit but there was no systematic patterning, that is, positive choices for irregular as opposed to oval pebbles in different areas or levels of the cairn.

Little Tor Cairn

Round pebbles account for around 6–8 per cent of the pebbles in all layers and quadrants in Little Tor Cairn. Irregular pebbles make up between 20 and 52 per cent of the total with the lowest frequency occurring in the pebbles in quadrant N1 (20–38 per cent). Oval pebbles make up between 17 and 23 per cent of the pebbles in various levels.

Twin Cairn A

The majority of the pebbles in all the layers of Twin Cairn A, as in TC and LTC, are irregular in shape with few round pebbles and variable frequencies of oval pebbles (normally within the range of 15–20 per cent). Again this varies little between the pebble layers in TCA, except in a few cases where the total number of pebbles is anyway low. Figure 4.5 shows the results of all excavated units. It is clear that the pebbles from the beach differ substantially from both the pebbles found in the natural soil across the heathlands and in the built cairns in terms of shape and frequencies of broken and unbroken pebbles.

Broken and unbroken pebbles

Soil acidity has acted to deteriorate the constitution of the heathland pebbles and make them more fragile and prone to cracking, a condition that does not occur on the beach, where most pebbles are unbroken. At LTC the frequency of broken pebbles varies between 35 per cent and 58 per cent in the different layers (around 45 per cent for the whole cairn). Similarly, the frequency of broken pebbles at TCA varies between 31 per cent and 52 per cent in the different layers (around 45 per cent for the whole cairn). No consistent pattern was identified in LTC or TC. The relative frequencies of broken pebbles in TCA varies between 26 per cent and 49 per cent (37 per cent for the whole cairn).

4. Pebble colours

The colours of the pebbles were recorded in terms of seven categories: white/quartz, red, brown, grey, yellow, black and Carter Blue Stone (CBS), a category first identified by George Carter in his work, discussed in Chapter 2.

Tor Cairn

Despite exhaustive detailed recording of the colour of the individual pebbles that make up the cairn, we could find no patterning whatsoever in terms of preferentially using, for example, white pebbles in one area or pebble layer of the cairn as opposed to, for example, yellow or brown ones.

Little Tor Cairn

The relative frequencies of yellow, red, brown and grey pebbles in Little Tor Cairn vary from level to level without any consistent patterning. Overall there is no evidence for choosing pebbles of different colours in the different layers or areas of the cairn. Like the natural pebbles in the Pebblebeds, the chief characteristic of the cairn is that it is a multicoloured structure. The only exception to this is the arrangement of some of the blue stones. Blue stones are found in almost every level throughout Little Tor Cairn in low absolute and relative frequencies. In all 71 Carter Blue Stones were recorded. These, like the special pebbles, were most frequent on the NE side of the cairn in squares S3 and N4.

Twin Cairn A

Frequencies of quartz/white pebbles are fairly consistent from level to level throughout Twin Cairn A except in the basal levels and in the central area, where relative frequencies are twice as high as in the upper levels. Red pebbles are surprisingly infrequent in levels 7 and 8 or the middle levels of TCA. Brown, grey and yellow pebbles vary significantly from level to level. Only eight blue stones were recorded, seven of these from the central basal level of the cairn (Table 4.6). In the SW section through the cairn another interesting feature was documented – a column of blue stones going through the whole structure of the cairn. The column was located near the E–W profile in the middle between the centre and the western perimeter of the cairn. Similar features were documented in plotting the blue stones in TC (see Chapter 3).

As regards the colours of the pebbles the test samples showed a similar range of colour variations. In other words the colours of the pebbles in the cairns simply mimic those found across the Pebblebed heathlands. There appears to be no preferential choice for more pebbles of one colour (e.g. yellow or red) to include in the cairn construction. The cairns were meant to be multicoloured like the pebbles of the Pebblebeds themselves. So the cultural product – the cairn – mimics the landscape of pebbles on which it stands. The exception to this is the blue stones. In all the test samples there were either no blue stones present or only one or two compared with 72 recorded and individually plotted at LTC, 8 in the NE quadrant at TCA and 152 at TC. They must have been deliberately curated. To find so many in the cairns, given their general rarity, provides positive evidence for their differential selection and curation. At TC and TCA some blue stones were arranged in vertical columns through the levels of the cairn structure, a fugitive patterning within the cairn itself and not visible from the outside. In LTC there were clusters of these stones deposited in the pebble layers on the NE side but no columns. Figure 4.6 shows the frequency of colours as they appear in the three cairns and the test samples.

5. Special pebbles

Excavation is always a theatrical performance, bringing together a team of people with different backgrounds, interests, gender and competencies to conduct the work (Tilley 1989; Bender et al. 2007). Many of our discussions inevitably involved the form and character of the pebbles themselves and their potential significance. At the initial stage the project brief that we had set ourselves was simply to record only two visual aspects of the pebbles: their shape and colour. However, having uncovered the entire upper surface of Tor Cairn we realized that this did not do sufficient justice to the materiality of the pebbles themselves. Some were multicoloured in a quite extraordinary way: how were we to record that? Others had remarkable quartz veins and inclusions: should they be ignored as if they did not exist? Might not these pebbles have a particular significance themselves in the process of the construction and colouring of the cairn? Following the suggestion of Wayne Bennett, we started to record another category of pebble that we termed ‘specials’. These were recorded separately following the excavation of each individual excavation unit and layer, bagged up and taken away for later analysis (see Chapters 5 and 6). We also recorded the frequency of specials from our test samples and took these away for comparative purposes.

Tor Cairn

In all 968 specials were recorded from the two completely excavated quadrants of the cairn. This may be compared with only 83 special pebbles from all the test samples. Special pebbles were significantly more frequent in almost all of the cairn squares and levels than those recorded from the natural samples. The only exceptions were the two beach samples, in which special pebbles were far more common than in the 13 other test samples. In these special pebbles made up between 19 per cent and 24 per cent of the total number of pebbles recorded. The fact that there were far more special pebbles recorded from the beach samples than from either the cairns or the other natural samples is easily explained. The beach pebbles are clean and completely devoid of any soil staining. Since they are continuously being scoured by salt water, it is easy to recognize special and unusually patterned pebbles with their multiple colours and quartz veins and inclusions. By comparison, the pebbles recorded from the cairns are stained and discoloured by the soil matrix in which they are embedded, as is the case for the other natural samples. As a consequence, the frequency of special pebbles that can be recognized, even after washing and scrubbing the pebbles, is inevitably considerably lower. From the beach samples we can conclude that about 20 per cent of the pebbles in the Budleigh Salterton Pebblebeds as a whole are ‘special’.

The more valuable comparison to be made is between the frequencies of special pebbles in the cairn and those found in the other samples from the natural. To take one example, for Square 9 of the SE quadrant of Tor Cairn the special pebbles varied between 4 per cent and 10 per cent of the total number of pebbles in nine recorded levels, whereas none were recorded from the four excavated test samples in its immediate vicinity. Overall the frequency of special pebbles in the other test samples from the natural varied from 1 per cent to 5 per cent. By contrast the frequency of special pebbles in the excavated metre squares and pebble layers of the cairn varied between 3 per cent and 20 per cent, with the highest frequency being the pebbles recorded from the pit under the cairn (22 per cent). It should be noted that the frequencies of special pebbles in some of the cairn excavation units and levels match those of the natural beach samples despite the fact that their frequency without being scoured by salt water is inevitably considerably depressed. The mean frequency of special pebbles from the cairns was c. 12 per cent, while that from the natural samples, excluding the two beach samples, was only 2 per cent. Furthermore the majority of the specials in the beach samples were ‘mottled’ pebbles with fine gradations in colours. These colour gradations do not show up well where pebbles are buried and earth-stained The vast majority of the specials from Tor Cairn were those with striking veins and quartz inclusions that are much easier to recognize and show through the discoloration of the pebble surface caused through burial under sandy and peaty soil. Considering this together with the striking contrast between the high frequencies of the specials at the cairns when compared with all the other samples from the natural leads us to conclude that special pebbles were deliberately selected during the construction of the cairn and that they occur in considerably greater numbers than we would expect from a random pebble selection. Their frequency varies considerably from metre square to metre square and between the various pebble levels from the top to the bottom of the cairn. There is no evidence for it increasing from top to bottom of the cairn or vice versa or from one part of the cairn to another. Special pebbles are found throughout and everywhere in Tor Cairn. The entire cairn is thus composed of both multicoloured and special pebbles.

Little Tor Cairn

Special pebbles range in relative frequency from 3 per cent to 26 per cent in Little Tor Cairn. They occur in consistently higher frequencies in N4 and S3 on the northeast side of the cairn than in N1 and S2 on the southwest side (between 3 per cent and 10 per cent on the SW side and 10–26 per cent on the NE side). They are thus between two and three times more numerous in the NE part of the cairn. In N1 the frequency of special pebbles is highest in level 2 and thereafter declines with depth, from 10 per cent to 4 per cent. In S2 the lowest frequencies of special pebbles occur (only 3–5 per cent). In N4 frequencies of special pebbles are twice or three times greater (9–12 per cent) in all levels except level 5 (19 per cent). In S3 most occur (11–26 per cent) with again the highest frequency in level 5, a basal level of the cairn.

Twin Cairn A

At TCA we recorded 101 specials in the NE quadrant and the centre of the cairn (10 per cent of the total). Special pebbles vary in relative frequency from 4 per cent to 42 per cent in the individual layers of TCA. They are particularly frequent in levels 6 and 7 (the central layers of the cairn) and in the basal level of the bowl shaped depression at the centre of the bottom of the cairn (Table 4.6).

Overall the frequency of special pebbles in the test samples from the natural across the heathlands varied from 1 per cent to 5 per cent. This strongly suggests positive selection of these pebbles for inclusion in the cairn construction. Some may have been collected in the vicinity, others possibly from much further away.

Choice and randomness: similarities and differences amongst cairn structures and pebble test samples

In order to formally verify whether there was a pattern in the construction of the cairns and to check similarities and differences between their structure and the natural disposition of pebbles on the landscape, a number of chi-square tests were performed. In other words, these tests are intended to ascertain whether there was positive selection in the process of collecting pebbles from nature to build a cairn or whether they were randomly picked up. Rejecting the null hypothesis of no difference between the cairn pebbles and the samples from the natural would lead to the conclusion that observed differences did not occur by chance, an indication that there was positive selection. In other words, if the test rejects the null hypothesis, it confirms that there was positive selection. If not, the patterns observed in cairns are not statistically different from the natural samples. The three cairns were tested against the natural test samples, and the results are shown in Table 4.7.

For all tests, the null hypothesis was strongly rejected (all p-values are far below 5 per cent), indicating that there are statistically significant differences for all categories across every pair of squares and, therefore, there was positive selection of pebbles in the process of building Tor Cairn, Little Tor Cairn and Twin Cairn A. These results can also be represented in the form of graphs, as in the three examples shown in Figure 4.7, that show the difference between the observed and expected frequencies for shape in a comparison between LTC and the natural, for weight in a comparison between TC and the natural, and for length between TCA and the natural (Figure 4.7) All cairns andcategories of analysis were tested in the same way against the natural samples.

Chi-square tests were also applied to verify similarities or differences in the choice of pebbles used to construct the cairns. To do this, the cairns were tested in pairs against one another (TC vs. LTC, TC vs. TCA and LTC vs. TCA) according to each analysed category (shape, weight, length and colour). The null hypothesis affirms that there is no statistically significant difference between the frequencies of pebbles across cairns, for a given category. The chi-square statistics (χ²) can be found in Table 4.8 alongside their respective p-values. The null hypothesis was strongly rejected (all p-values are far below the 5 per cent confidence level that this might occur by chance) in all tests, indicating that there are statistically significant differences for all categories across every pair of cairns and that there was no strict pattern in the composition of pebbles in these monuments. In other words, the tests indicate that each cairn was built in a different way, using a selection of pebbles that was not random.

All tests confirmed that the observed frequency of pebbles in all four categories of analysis for every square did not replicate the expected frequency and therefore rejected the null hypothesis. Figure 4.8 below illustrates this conclusion for all tested pairs of cairns.

The statistical analysis presented in this chapter allowed for two strong conclusions: first, that the construction of cairns involved intentional or positive selection and curation of pebbles. Second, that the building of an individual cairn was a singular event, that is, it was in part improvised rather than following strict patterns for the selection and the deposition of pebbles. The action of cairn building differed from one case to another even though many similarities can be observed in terms of the overall statistics for pebble length, weight, shape and colour.