Given the lack of success we had with the parasite samples (both samples came back negative), pollen analysis has only been done on the waterlogged deposits found below the bodies of the Scottish soldiers. This is because waterlogged soils generally preserve botanical remains really well. There is ongoing debate about the origin of this waterlogged soil because we were unable to explore the full extent of it due to the restricted size of the excavation area, in the same way that we were unable to fully excavate all of the human remains. A hazel nutshell within the waterlogged soil has been radiocarbon dated and dates to the 12th / 13th century, so quite a long time before the Scottish soldiers were buried above it. It may be that it formed in a defensive ditch extending around the castle’s wall, but, whatever its origin, the fact that it is waterlogged means it provides excellent preservation for botanical remains.

Bulk soil samples were taken from the two waterlogged deposits (context 520 and 508). These samples were processed as part of the initial post-excavation stage. Soil processing involves adding water to the soil and gently mixing it, which causes any charred or waterlogged botanical remains to float to the surface. The soil processing produced plant remains typical of wet or damp ground habitats, including sedges, marsh-marigold, spike-rushes, bogbean and lesser spearwort. The absence of aquatic plant remains suggests that in the 12th / 13th century this was a marshy environment rather than an area of permanently standing water. Ruderal weeds, including hogweed, henbane, nipplewort and common chickweed were also present in the samples and would have grown on areas of nearby rough ground. Some of the arable weeds may have formed an additional component of this flora, with others having been transported to the site with crops. Hazel nutshells and oak wood fragments were also found in the samples.

These plant remains help us to build a picture of the immediate environment in the 12th / 13th century. Pollen analysis offers a method of providing additional information about the local environment and a snapshot of the broader landscape at that time. So, while this sample won’t tell us anything about the Scottish soldiers themselves, it will be able to tell us something about the history of the site before the soldiers were buried there.

Extracting pollen from the soil

The pollen extraction process is quite lengthy – it took a whole morning (8am – 12.45pm) for the material to be processed. One millilitre of soil is placed in a test tube. Distilled water and acid are added to the soil sample along with a tablet containing a known number (approximately 18583) of Lycopodium spores. Lycopodium (clubmosses) grow in mountain heaths and are unlikely to occur in most archaeological soil samples. Their distinctive spores act as ‘exotic’ markers, which, when added to a known volume of sample, allow the concentration of pollen to be calculated. Once the Lycopodium tablet has dissolved, sodium hydroxide is added to the sample which is then placed in a heated water bath in order to break down the soil sample and release the pollen grains. The sample is then poured through a coarse sieve to remove large particles such as grit and sand, and over a very fine sieve to remove silts and clays. The contents of the fine sieve undergo an acid digestion of any remaining humic material, before sodium polytungstate is added. This changes the density of the liquid trapping any heavy particles below it and causing the pollen grains to float on top if it where they can be pippetted off. The pollen is then washed several times and suspended in silicone oil. This allows it to be spread on a slide and viewed under a microscope for identification.

red selecting the soil
Taking the soil sample
soil acid water lycopodium tablet
The two soil samples in test tubes with the lycopodium table dissolving
red adding the sodium polytungstate
Adding the sodium polytungstate
red coming out the centrifuge
Dark layer of pollen grains floating on the sodium polytungstate in the test tube.

What the pollen tells us

The pollen grains that we found were frequently degraded, but the quantity of pollen recovered was in the average range for typical sediments of this date. Herbaceous plants, particularly cereal pollen and grasses, dominated the pollen assemblage. Arable and ruderal weeds were also present and included members of the daisy, plantain and buttercup families. The few tree/shrub pollen grains were from heather, hazel and willows.

Plantain 1
Plantain pollen grains

The presence of Centropyxis, a testate amoeba of wet freshwater habitats, reflects the damp waterlogged conditions that botanical remains had already indicated. However the degraded condition of some of the pollen grains may indicate that there had been periods of drying. Interestingly, although there were macrofossil remains of sedges, no sedge pollen was found in this analysis. This may be because sedge pollen didn’t survive, or it could be that sedges were not growing locally, but were brought to the site from late summer/autumn for use as hay or packing/roofing material.

Cereal pollen grains were abundant. Although cereal pollen is difficult to separate, the grains lacked the characteristic oblong outline of rye, and the large size of the majority suggested either oats or wheat. Of these, oats is more likely, as weeds associated with spring-sown crops such as oats and barley dominated the weed flora recorded in the plant macrofossil assemblage.

Cereal type pollen C520 image 1 cropped
Cereal type pollen grain

Oats, barley and wheat are notorious for poor pollen dispersal as the pollen can become trapped in the hulls and therefore usually represent just a few percent of a pollen assemblage. The samples that we took contained 58% cereal pollen! Substantial proportions of cereal pollen are often indicative of nearby or very widespread cultivation but on Durham peninsula it seems more likely that other processes are the cause. For example it might be that the high number and diversity of agricultural weeds (represented both as seeds and pollen) may indicate the dumping of crop processing waste in the marshy area. Experiments have shown that threshing considerably increases pollen dispersal, so the high cereal pollen grain count could be explained by the cereal crop being brought to the peninsula in sheaf form and being processed near the site.

The almost total absence of tree pollen reveals a lack of woodland cover on the peninsula in the 12th / 13th century. The single records of hazel, willow and heather may indicate occasional shrubs or small trees growing locally (a few hazel nutshells were recorded in the macrofossil assemblage), or their presence in the regional pollen catchment zone. The absence of oak pollen indicates that the large oak wood fragments recovered from the waterlogged layers derive from trees that were felled elsewhere and brought onto the peninsula, probably for construction purposes.

Charlotte O'Brien

Author: Charlotte O’Brien

Charlotte manages the Palaeoenvironmental Archaeology service of Archaeological Services Durham University. She has a doctorate in palaeoecology and specialises in the use of plant remains for the reconstruction of past environments. She investigates changes in diet, crop husbandry techniques and woodland management through the analysis of plant macrofossil, pollen, wood and charcoal assemblages.

Pollen Analysis