One of our big challenges on the Building Stones project is directly tracing a stone in a building to a quarry. Detailed fieldwork can be really effective for working out the range of rock types used and to give some idea of the areas these may have come from but, in general, for our project, has fallen short of providing a strong link to any one quarry. Much of this is down to the inherent variability of the typical rocks in this area. Both Old Red Sandstone and Silurian limestones – which together constitute almost all of the building stone in Herefordshire – can be highly variable within a single quarry. An additional problem in buildings is the lack of stratigraphic context once blocks are removed from a quarry and jumbled up in a building. This is particularly true of the Silurian rocks of the area. While a very detailed stratigraphy exists, this is largely based on pattern recognition and sparse fossil evidence; both of which are largely lacking in buildings. Looking purely at the lithology of a rock, it is often not clear whether it is from one formation or another. Diagnostic range fossils are rare, the main exception being Kirkidium knightii, a brachipod unique to the Aymestry Limestone which has been useful in confirming the use of this formation in several buildings.

Because of this I have been interested in exploring ways which may allow us to “fingerprint” stone from a given quarry or formation more uniquely. One such technique is analysis of thin sections of rock. By studying these under a polarised light microscope we can get useful information such as the porosity, cement and sediment mineralogy. Quite extensive use of this has been employed in Bromyard on samples kindly donated by homeowners, mostly derived from building work. This has given us some indication that there are useful variations in mineralogy which, in some cases, seem to be unique to a single quarry. However, there are significant drawbacks. Firstly we risk inaccuracies by characterising a quarry or building on the basis of only a few samples. Secondly any differences present are subtle and so, to properly understand, them requires “point counting”; an extremely laborious process whereby the mineral present at up to 500 points on a grid is identified and tallied to give a quantitative breakdown of the makeup of the rock. This takes up to 2 days per section. Lastly, inherent in this process is the removal of stone from a building. This limits the sites we can look at, by and large, to those which have current or recent building works.
As a result I have been exploring the possibility of using portable X-Ray Fluoroscopy to characterise stone buildings and quarries. In contrast to the mineralogical information yielded by thin section analysis, this gives us information about the chemistry of the rock which is ultimately related to the mineralogy as well as the diagenetic and weathering history of the rock. The machine is a handheld device with something of a Star Trek prop about it. Although surprisingly portable, it is rather heavy to hold at arm’s length for extended periods (as required if you don’t want to give yourself an imprudent dose of radiation). It works by firing X-rays at the sample. These excite electrons within the constituent atoms which, as they return to their natural energy state, release X-rays of their own or, in other words, fluoresce. The wavelengths of these X-rays are particular to each element and, by measuring the intensity of emitted rays at each wavelength, the instrument can determine the concentration of each in the sample.

With the instrument hired for a week I set about two main case studies: the first being Bromyard, where we could compare the results of thin section analysis and XRF for the same samples. The second was the Ludlow Breadwalk and Mortimer Forest sections which give a stratigraphically well-constrained section through the Silurian.
The results thus far are promising, with a few provisos. Both Bromyard and Ludlow seem to show some strong distinguishing features; in Bromyard, between samples from buildings in the town and buildings and quarries on the Bromyard Downs. This is interesting as it backs up the hypothesis on the basis of fieldwork that much of the stone in Bromyard was quarried within the town from several (now lost) quarries and possibly from the excavation of bedrock cellars too.

Around Ludlow, some of my attempts to get field data were stymied by poor weather. A big drawback with the instrument is that it is strongly affected by air and surface moisture. Heavy rain and fog on several field days along the Mortimer Forest trail rendered the data essentially useless which was very disappointing. Nonetheless, we had a good clear day to sample the Breadwalk Section which runs along the west bank of the Teme from Dinham Bridge to Ludford Corner. In addition, a large number of measurements were made on the curtain wall of Ludlow Castle in good conditions. This provides strong evidence that the castle is indeed built from the Lower Whitcliffe Beds which form the bluff on which it stands, rather than other formations which outcrop in various quarries across the river.

I am planning to do more work on this technique, particularly to see if it is possible to come up with a calibration to mitigate against the effects of water. Although as in all things with this project there is rarely a magic bullet, this technique does show a lot of promise as a cheap, quick and non-destructive way of comparing and linking stones. With sensible reservations, and an understanding that differences between formations will likely only ever be locally applicable, this could be a powerful addition to the Building Stones toolbox.