This chapter details the experimental burning of a quartz assemblage and the subsequent thin sections taken on burnt and unburnt samples. Section 8.2 refers again to previous research outlined in Section 5.8. Section 8.3 describes the selection of material, measurements, and the fire setting, while Section 8.4 details the results of the experiment and the thin sectioning, detailing the effects on the quartz appearance, fragmentation, and spatial distribution in the hearth and beyond. Section 8.5 concludes the chapter with a discussion of the results, with the full report on the thin sections in Appendix A- 49.
As discussed in Section 5.8, Gonick (2003) and Ballin (2008) have undertaken experiments with burning and boiling, and burning quartz respectively. Ballin (2008) noted that burnt quartz did not appear to be recognised in the Scottish lithic reports that he examined, and he surmised that this was probably due to a lack of identification, rather than a lack of burning. Unfortunately, Ballin has not published his quartz burning experiments beyond the mention in the 2008 publication. During the literature review of Irish quartz finds, few references were found for burnt quartz.
Ballin (2008, 51) commented that his burning experiments showed that “the quartz was generally characterized (sic) by (i) pitting and ‘peeled-off ’ surfaces, (ii) a dull and opaque appearance (where fresh quartz tends to be clear and vitreous), (iii) various degrees of ‘granulation’ and disintegration, and (iv) occasional areas with either a reddish or a pink hue”. Ballin (2008, 51-2) was unable to reproduce a certain yellow-brown colour in his experimental burning:
"[a]s the yellow-brown pieces of quartz seems to be mainly associated with later prehistoric sites from the Northern and Western Isles, and not the Mesolithic sites of the western mainland and the Southern Hebrides, the author assumed that the burning of peat, particularly characteristic in Scottish later prehistory, might have caused the differences in appearance. It has not been possible to reproduce experimentally the yellow-brown colour of burnt quartz from the Northern or Western Isles, but the author believes this discolouration to be the result of either the accidental burning of quartz in peat fires, or the deposition of the burnt pieces in a peaty environment (e.g. in peat ash deposited in domestic middens). As the experimental burning of quartz in a peat fire did not produce the anticipated colours, the author expects the discolouration to probably be the combined result of (1) weakly developed ‘granulation’ due to the exposure of heat/fire, making the quartz slightly more porous, (2) deposition in iron rich peat or peat ash and (3) time".
From his interpretation of the visible attributes of burnt quartz, Ballin (2008, 51) has recognised extremely high proportions (over 50% for some assemblage) of burnt quartz in some Scottish assemblages; he comments that these high ratios of burnt quartz remain unexplained. Ballin notes that he observed varying degrees of disintegration of his material, but does not quantify it.Top of Page
Following from the results of Sergant et al. (2006) discussed in Section 5.8, this burning experiment focused on placing artefacts directly into a fire. Three effects were investigated; the visible signs of burning, the fragmentation rate of the quartz, and the spatial distribution of artefacts.
The four quartz sources used for the experimental burning were the same as used for the first experimental assemblage – B.Q., M.Q., P.Q., and R.Q. 10 artefacts from each source were used for each of the seven size grades (Table 8-1). This gave a total assemblage of 280 artefacts, totalling 1.7kg. The instruments for measurements, weights, and sieving were as outlined in Chapter 5. Two hues were noted for the materials; the primary hue and the secondary hue, with the latter defined as the hue cause by iron and so forth in the quartz matrix.
The fuel source used was pine (Pinus); the dead wood was collected from a forest floor and consisted of small twigs and branches up to 5cm in diameter. Pine was a predominant species in the North Mayo forest during the date range of the Belderrig quartz scatter (O'Connell and Molloy 2001). A series of plywood sheets, covering 36m², were placed surrounding the hearth in order to determine the extent of artefact dispersal from the fire. In order to prevent fire damage to the sheets, a layer of peat was placed surrounding the hearth, laying over the edges of the plywood.
|Size grades||B.Q. (g)||R.Q. (g)||M.Q. (g)||P.Q. (g)||Total (g)|
Table 8-1 Pre-burning weight for 10 artefacts per size grade
The fire was set and fuelled for one hour; the assemblage was then placed in the centre of the hearth, in a manner to allow all of the assemblage an equal chance of burning. While the temperature of the fire was not taken, it can be estimated at over 500ºC (Urbas and Parker 1993). After a further two hours the fire had cooled enough to begin removing the assemblage from the hearth and begin the sieving, using the 5mm and 1mm sieves. The peat surrounding the hearth was also sieved and the hearth was trowelled back until unburnt soil was exposed.Top of Page
Once placed in the fire, the immediate visible sign of burning was what appeared as a bleaching of the assemblage in the fire, which occurred at the same time as a significant amount of crackling was heard. Most of the artefacts turned a pure snow white, losing their generally metallic grey hue. Even when in direct contact with flames and red hot coals almost none of the quartz exhibited blackening at all. Another concurrent immediate effect was the development of a salmon pink and burnt orange hue in places on many of the artefacts. Both these colour alterations were noted immediately. Figure 8-1 shows the fire once cooled, three hours after the assemblage had been deposited into the hearth.
Once the artefacts were removed from the fire it was clear that the four sources reacted differently, albeit subtly, depending on the mineral inclusions present within the matrix of the vein quartz. Table 8-2 gives the pre- and post-burn characteristics for the four sources. In terms of the primary hue (Figure 8-2), all the materials began as metallic grey and all but the R.Q. turned white; the R.Q. only partially turned white.
The secondary hue – caused by iron and inclusions of the psammite matrix – of the materials was not apparent on all the pre-burnt artefacts, and when present it was usually confined to small patches, apart from the B.Q. when it was more extensive (Figure 8-2). For the M.Q. and B.Q., the secondary hue was formed by iron causing a pale yellow hue, while the P.Q.’s grey hue was formed by an intermixing of the parent psammite into the vein quartz matrix as well as iron. The M.Q.’s secondary hue changed to a burnt orange, and the P.Q.’s and B.Q.’s changed to a salmon pink/burnt orange.
|Material source||Stage||Primary Hue||Secondary Hue||Grain||Opacity|
|Metadolerite||Pre-burn||Metallic grey||Pale yellow||Smooth||Semi-opaque|
|Rose Cottage||Pre-burn||Metallic grey||-||Sugar/smooth||Translucent|
|Post-burn||White||Salmon Pink/Burnt Orange||Sugar||Semi-opaque|
|Beach||Pre-burn||Metallic grey||Pale yellow||Sugar/smooth||Semi-opaque|
|Post-burn||White||Salmon Pink/Burnt Orange||Sugar/smooth||Semi-opaque|
Table 8-2 Pre- and post-burn characteristics
Figure 8-1 Quartz in hearth three hours after initial deposition
Figure 8-2 Colour change. Top left: B.Q.. Top right: M.Q.. Bottom left: P.Q.. Bottom right: R.Q.
Figure 8-3 Backlit artefacts, highlighting opacity change on similarly thick artefacts. Left: B.Q.. Right: M.Q.
In terms of the texture, or granularity, of the materials, the unburnt materials were described as either sugar, sugar/smooth, or smooth. The ‘sugar’ texture describes the quartz crystal grains that appear like granulated sugar that has been moistened and subsequently dried, allowing some of the grains to meld. The smooth grain texture is where the crystal grains are not readily apparent, even though the thin sections discussed in Chapter 6 have shown that the individual crystals are nevertheless large (1-5mm). Sugar/smooth lies between the two textures. The B.Q. and R.Q. are sugar/smooth. Over all, there was no visible change to the texture. Analysis of the thin sections noted little to no effect of burning on the texture or crystal structure:
"undulose extinction and most sub-grain boundaries are little affected, if at all, by heating. However the lamellar structures are absent in the burnt samples. The reason for the loss of the lamellar structures is unknown but might reflect the release of strain in the crystals during heating” (Menuge 2009b).
While no change was noted to the texture, the changes in the materials’ hue, especially the secondary hue, accentuated some of the grain patterning by creating a greater degree of contrast in the material. This didn’t apply to the M.Q., which retained its smooth appearance, or to R.Q. which had less of a colour change and no secondary hue.
The unburnt materials’ opacity was divided into translucent and semi-opaque. The M.Q.’s opacity changed from semi-opaque to opaque, and the change was the most uniform of all the materials. The B.Q. began and ended as semi-opaque, but the level of opacity increased slightly. Both the P.Q. and R.Q. changed from translucent to semi-opaque with the latter’s opacity increasing to a lesser extent. Figure 8-3 uses a back light on pre- and post-burnt B.Q. and M.Q. artefacts with similar thicknesses to highlight the change in opacity. The thin sections suggest that two factors may create the reduced transparency in the samples.
1. Decrepitation of fluid inclusions. Fluid inclusions are formed from the hydrothermal fluid in which the quartz crystals grew; on cooling, bubbles of fluid are trapped, within which is a bubble of vapour. During the burning experiment the heat from the fire, as well as the consequent fractures, led to the escape of the fluid, called decrepitation. (The pre- and post-burn fluid inclusion can be seen in Appendix A- 49.) This “loss of liquid increases the refraction of light passing between quartz crystal and the inclusion and so decrepitated inclusions block more light than those which are mainly liquid filled” (Menuge 2009b, 4).
2. Formation of microfractures. There was a distinct increase in the quantity of microfractures in the burnt samples. “Microfractures can also reduce the transmission of light in a similar way to inclusions, by refraction of light when it crosses a fracture. Whilst light can also travel along fractures so that they appear bright in thin section...the overall effect of many microfractures in many orientations in a hand sample should be to reduce light transmission.
However, Menuge (2009b) is uncertain whether these two factors are the main cause of the increasing opacity.
All the quartzes had an initial vitreous lustre; the P.Q. has a lesser vitreous lustre because of the contamination of the psammite parent into the vein quartz matrix giving a general duller appearance in places, but where the quartz was less contaminated, it had a vitreous lustre. After burning the artefacts were generally dulled giving a less vitreous lustre, especially the M.Q. which became the uniformly white and duller. But the material all remained vitreous, just less vitreous than the unburnt material.Top of Page
The initial 280 artefacts resulted in an assemblage of 1534 ≥1mm artefacts giving a greater than fivefold increase in artefacts. When removing the quartz from the fire, the quartz had increased in brittleness and a number of the larger of the artefacts snapped and/or crumbled – on returning to the artefacts two days later, the artefacts had become less brittle and did not crumble as easily as when initially taken out of the hearth. Figure 8-4 highlights the changed proportions for the size ranges of the artefacts – they began with equal proportions, with no <10mm artefacts – while Figure 8-5 graphs the change in weight for the size grades. The largest artefacts (≥40<50mm), were reduced from 40 artefacts to 31, and the ≥35<40mm artefacts had a similar reduction in quantity. While the pre-burnt assemblage contained no ≥5<10mm artefacts, the post-burnt assemblage consisted of 12.8% of that size range; <15mm artefacts initially accounted for 14.3% of the assemblage, and after burning they accounted for 80.5% of the assemblage, with the <5mm artefacts accounting for the majority of that proportion. The thin sections showed an increase in both macrofractures and microfractures in the burnt samples (Menuge 2009b). The macrofractures in the burnt samples developed along quartz grain boundaries and/or within crystals.
Overall, 93% of the ≥5<10mm artefacts exhibited signs of burning (Figure 8-6). The ≥1<5mm fragments proved difficult to identify as burnt so this category was not quantified. Similarly, the ≥5<10mm group had the lowest rate of visible burning – this is not necessarily because these fragments were not ‘burnt’, but that these smaller fragments are more difficult to identify as burnt. The R.Q. was the most difficult to identify as burnt, possibly because this quartz contained less fluid to turn white, thus appearing unburnt; conversely, some of the artefacts may in fact have remained unburnt, and coincidentally have been only the R.Q.. This, however, does not seem likely. For the burnt ≥25mm artefacts, it was reasonably easy to categorise the source materials of burnt artefacts (Figure 8-7). For the <25mm artefacts, however, it was harder to categorise them by source material due to the general similarities of the sources, so they were labelled as “?” source.
Figure 8-4 Post-burning size grades. Pre-burn assemblage had similar proportions (14.3%) of each of the seven initial size grades and no <10mm debitage.
Figure 8-5 Pre- and post-burning weight
Figure 8-6 Percentage burnt
Figure 8-7 Post-burning quartz count by source and size grade. Pre-burn assemblage had 40 artefacts per size gradeTop of Page
As discussed above, a 36m² area was covered with plywood sheets in order to catch any artefacts that were expelled from the hearth (Figure 8-8). A minor proportion (0.85%; n=13) of the artefacts were expelled outside of the hearth – the hearth zone is defined here as any area of burning, and the final hearth was approximately 1.4m². The layer of peat surrounding the hearth was low, and therefore would not have blocked pieces being expelled. The expelling of the small fragments started within minutes of the deposition of the assemblage into the fire and continued sporadically for at least 1½ hours. None of the artefacts were expelled in the prevailing breeze’s direction. The furthest expelled artefact noted was 2.8m from the hearth’s centre. This may imply that some pieces could have been expelled further than the area monitored. Most of the expelled artefacts fell within 1m radius of the hearth centre. The greater bulk of the artefacts’ movement was limited to within the hearth itself. Figure 8-8 shows the initial and final position of the quartz, the extent of the hearth, and the position of the 13 expelled artefacts; within the hearth, the final area covered by the quartz was double the initial area.
Figure 8-8 Spatial distribution of hearth extent and artefacts expelled from hearthTop of Page
The burnt quartz experiment sought to investigate three aspects of burnt quartz in order to develop a framework for identifying and analysing burnt quartz in the archaeological record: visible characteristics, fragmentation rate, and spatial distribution. The effect of burning on the visual characteristics of the quartz was apparent immediately on placing the artefacts in the fire, with the assemblage turning a snow white. However, differences were apparent with the materials, with the R.Q. component changing hue to a lesser degree than the other materials. The quartz with more abundant impurities changed as well, from grey and yellow to burnt orange and pink. The change in opacity was apparent for all materials but varied; for the M.Q., the least granular of the quartzes, the change was the most uniform, while for the B.Q. the change was slighter. The burning dulled all the materials, giving them a less vitreous appearance; this was especially the case for the M.Q.. The lustre of the quartz is difficult to be precise about, and difficult to assess possible heat related changes without a similar sample to compare with. All the materials were still in effect vitreous, just less so. There was no difference in the granularity of any of the materials. The only effect noted was that the hue change created a contrast, accentuating the grain texture – this accentuation of grain texture might be due to some grain boundaries becoming the sites of macrofractures (Menuge 2009a pers. comm.): however, this effect was slight. The burning also increased the amount of visible macrofractures on the artefacts. While Ballin (2008, 51) noted pitted and peeled off surfaces as characteristic of burning, neither of these were visible on the Belderrig burnt quartz.
In terms of macroscopic visual characteristics of burnt quartz in archaeological contexts, therefore, a key consideration is what an unburnt sample of the raw material looks like. While a given artefact may appear as dull, opaque, with numerous fractures, the source of that material should be examined to determine whether the dullness and opacity is related to its source or burning; multiple fractures are not a useful characteristic for identification as these are common anyhow. The impurities, such as iron, in the quartz changed their hue the most, therefore these can be a useful indicator of burning. In terms of the microscopic visual characteristics, the thin sections have shown that the decrepitation of fluid inclusions is a signature of burning at high temperatures.
The effect of burning on the fragmentation of the quartz artefacts was dramatic, with a greater than fivefold increase in the assemblage’s quantity. The 80 ≥35mm artefacts were reduced to 61, while the ≥15<25mm size grades almost doubled in quantity. The most dramatic difference was the smaller artefacts – the ≥1<10mm began with no artefacts and ended with 1128. This has serious implications for any analysis of a burnt quartz assemblage, or assemblages that contain burnt quartz. Problematically, it was noted that these smaller fragments are the most difficult to identify as being burnt. However, assemblages with significantly high proportions of smaller fragments may suggest evidence of burning, other depositional and post-depositional factors notwithstanding.
The spatial analysis of the hearth has shown that, while the quartz fractures dramatically, there is less of an explosive expelling of artefacts from the hearth than with flint (see Sergant et al. 2006). Very few artefacts were expelled outside of the hearth, with most movement occurring over centimetres, resulting in a reasonably contained final hearth spread. The lack of explosive movement compared to the flint may relate to the lesser proportion of fluid in quartz compared with flint, resulting in less pressure during heating. In terms of a comparison to the fragment rate with the flint, this is not possible, as Sergant et al. (2006) do not provide a breakdown of the pre- and post-burn size grades.