Rocks in Roslin Glen: A record of a swampy past

Print Friendly, PDF & Email

Mark Wilkinson and Claire Jellema (UK)

Midlothian is an area of central Scotland that lies to the west of Edinburgh and is an area with strong geological connections due to a history of mining for both coal and oil shale. As a part of the annual Midlothian Science Festival (, the School of GeoSciences at the University of Edinburgh offered a walk to look at some local geology and a talk about climate change research on the Greenland icecap. In addition, a ‘Dino and Rocks Day’ was attended by 380 people, proof (as if it were needed) that dinosaurs continue to fascinate the general public. The Edinburgh Geological Society also contributed with a session about Midlothian Fossils and a local historian talked about the history of coal mining in the area.

The geology walk visited local exposures, in this case Carboniferous sediments including what may be the best exposed fluvial sediments in the area. The walk was advertised as “Rocks in Roslin Glen: a Record of a Swampy Past” and all 25 spaces were quickly booked. The location was Roslin Glen, which may sound familiar if you’ve seen the film, The Da Vinci Code, based on the novel by Dan Brown. We have not misspelled the name of the glen incidentally. For some reason, Rosslyn Chapel lies on the edge of Roslin Glen and the country park of the same spelling. The glen itself is a steep-sided valley of around 20m in depth, which carries the River North Esk roughly west to east towards the North Sea. The glen is wooded and is a popular location for short walks.

The chosen Saturday in October was a cool but dry autumn day and good for the walk. We met in a car park inside the country park, by the river North Esk. Following a short introduction, we crossed the river by a footbridge, arriving almost immediately at the first stop – an exposure of sandstone in a small hollow that looks like it was quarried out at some point in the past (Fig. 1).

Fig. 1. The first stop – a small sandstone quarry. The top half of the ‘exposure’ is actually a stone wall. (Photo credit: Lee Live

This seemed like a good place to talk about what the area would have been like when the sandstone was laid down as river sand, in the Early Carboniferous. As the land, which eventually became Scotland, drifted across the equator during this time, it would have been a good deal warmer than a present day October; and it would have looked rather different too. Although there were trees around, they were tree ferns and not the modern angiosperms we have today, which did not evolve until the early Cretaceous.

Modern tree ferns are quite modest in size, a few metres tall judging by the modern ones growing in Madeira (a great place to visit for a geology, incidentally). However, fossil tree trunks found during quarrying of Carboniferous sandstones in the Edinburgh area show that some of the tree-ferns where huge. You may have seen one example from Edinburgh outside the Natural History Museum in London, although 150 years of London air haven’t done much for its finer details. This specimen was found in 1854 and then lost for around 15 years, according to the book, Building Stones of Edinburgh by McMillan et al. (1999).

How you lose a rock that is around 12m long we are not sure. Other examples of the tree trunks can be found in the Botanic Gardens in Edinburgh. The smaller plants in our modern woods are also largely angiosperms too, so rather different from their Carboniferous equivalents. However, the ferns and occasional horsetails growing below the modern trees look pretty similar to their Carboniferous ancestors.

The Carboniferous landscape would have sounded different too – there were no birds, which didn’t appear until the Jurassic, so no birdsong. Presumably, the huge dragonflies that could reach one metre in wingspan didn’t make much noise, but some of the amphibians might have had calls, as do modern frogs for example. There were also alarmingly large creepy-crawlies, with millipedes a metre long. There are no traces of these to be seen locally, but there is a great example of a millipede track on the Isle of Arran, a few tens of kilometres to the west. We briefly talked about the sandstone and how it was brought here as sand by a river. We didn’t spend time discussing the stratigraphy of the Scottish Carboniferous, although the sediments here are part of the Passage Formation, which is late Namurian in age and marks the transition from relatively limestone rich beds below to the overlying coal measures.

Moving up a set of steps, we arrived at the entrance to Roslyn Castle (Fig. 2). This is set dramatically on a headland surrounded by cliffs on three sides, with a bridge spanning a deep defile on the fourth side. Of the castle itself, only a few ruined walls survive, although the stonework is beautifully clean and shows great cross bedding. We had brought along a few hand lenses, so these were handed around and instructions given on how to use them. we are always surprised by how many of our geology students want to hold the hand lens too far from their eyes.

When looking at rocks at ground level, this is understandable, as lying flat on the ground isn’t always pleasant, especially if it’s been raining. However, here, the vertical walls provide a great opportunity to get up close and see the individual sand grains in a degree of comfort and to imagine the spaces between them. These spaces, or porosity, are important as oil and gas are found in these in places such as the North Sea, from where we obtain our liquid fossil fuel supplies. Much of the UK’s drinking water is also derived from sandstone aquifers.

Fig. 2. Rosslyn Castle, complete with clean sandstone for hand-lens practise. (Photo credit: Lee Live at:

We then retraced our steps back to the car park and along the road westwards. After negotiating a road crossing on a steep hairpin bend, we left the road to walk along a pleasant and traffic-free wooded lane. Incidentally, if you ever lead a group in the field, then the thing you worry about most isn’t getting lost or forgetting what to talk about – it’s getting someone injured doing something that we all do every day and crossing the road is probably the most dangerous thing most field parties are likely to attempt. Getting lost, on the other hand, is only embarrassing, provided of course you can find your way back.

The lane runs close to the river North Esk and the river valley quickly becomes narrower. On the right-hand side of the track, there are alcoves carved into the valley side, around ten metres square, the purpose of which soon becomes apparent. The track enters a clearing, which was the site of a gunpowder works for around 150 years, closing in 1954. The works used charcoal made from local wood, sulphur from Italy and saltpetre from India. Ground and mixed together, these make gunpowder.

Needless to say, the whole activity is rather hazardous and the alcoves we’d passed earlier were the site of the mixing operations. The buildings were alcoves, so that if one exploded then the blast went towards the river, rather than along to the other buildings. Apparently, explosions were not especially uncommon. Looking at the peaceful woodland scene today, it is hard to imagine that not much more than 50 years ago, this was an industrial complex. We briefly discussed why the works were here, given that the two geological materials needed for the manufacture of gunpowder were imported. There were three other factors: solitude (essential in the light of the safety record); a supply of wood for charcoal; and the river for power before it was replaced by fossil fuels. There is a dam upstream, and the remains of a building that housed a water wheel at the next stop.

Continuing along the track brought us to the highlight of the trip – a 20m-high exposure of sediments known as Hare Craig, exposing more Namurian sediments, now in the slightly younger Upper Limestone Formation (Fig. 3). The stratigraphic name of the rocks is rather misleading, as there is no limestone exposed here. Partly for this reason, putting rock types into stratigraphic names has fallen out of fashion. The modern practise is to name stratigraphic units after a ‘type locality’, usually where the unit is well exposed. The Upper Limestone Formation is part of the Clackmannan Group, named after Clackmannanshire – a county near the town of Stirling, some 50km to the northwest. Ironically, the Clackmannan Group does not appear to be well exposed here, although it may have been previously exposed in quarries or mines that are now infilled.

Fig. 3. Hare Crag, with the channel sandstone at the top of the cliff (the overhanging smooth surface area), cutting into seat earth and bedded overbank sands and silts.

Back in Roslin Glen, the cliff is on the other side of the river from the path but is easily visible, especially in winter, when the leaves have fallen from the trees (Fig. 3). The top half of the crag is a sandstone, deposited as a river channel. The base of the channel clearly cuts down into the underlying sediment, so has eroded this sediment. The sandstone is cut by vertical joints that are picked out by weathering at a spacing of a metre or more, suggesting that, when the river undercuts the base of the cliff sufficiently, the sandstone must fall as substantial blocks. We were asked the interesting question of why there are no such blocks lying in the stream – certainly the present stream doesn’t look like it is going to move metre-scale blocks of rock, although presumably it must flood occasionally. One good suggestion was that, as the sandstone would have to fall 10 or 20m before hitting the stream bed, and the sandstone isn’t particularly well consolidated, the blocks might disintegrate on impact – whatever the answer, we wouldn’t like to be there to find out the hard way.

Immediately below the sandstone lies around three or four metres of a pale grey rock, which, from a distance, looks almost artificial. In fact, it is a seat earth, developed below the Carboniferous forest floor. The plants of the forest removed so many nutrients from the underlying soil, that they bleached it, leaving the distinctively pale grey sand and siltstone. The plants themselves might have formed a coal layer, but were eroded by the overlying river channel – a common fate for what would have been peat at the time, which is soft and easily removed. Below the seat earth are bedded sands and silts, presumably deposited by the river on the floodplain to the side of the fluvial channel. Each bed probably represents a single flood. While it isn’t possible to inspect the rocks in the cliff closely, they are exposed more accessibly a little further upstream in the river bank. However, as this requires a scramble over a fence and some slippery rocks, we gave it a miss.

A point worth considering is the formation of Roslin Glen itself. There are a number of relatively deeply incised rivers in this part of Scotland, which presumably have a common origin. The area has been rapidly uplifted since the end of the last glaciation, which might have forced the rivers to concentrate all their erosive power on cutting downwards to keep pace. The problem with this idea is that, if the glen were entirely the result of post-glacial erosion, then at around 20m deep, it represents the sort of erosion rates that are normally associated with active mountain ranges such as the Himalayas. So perhaps the glen was at least partly cut by meltwaters from the waxing glaciers themselves. Or maybe a shallower version of the glen was present before deglaciation – perhaps as a subglacial channel.

That concluded the walk, and people wandered back to the car park. They seemed to have enjoyed themselves and we’d answered any number of questions, which is always a good sign. The younger members of the party specially enjoyed the ruined buildings at the last stop, where there are small tunnels and bridges to run in and around on.

One rock type we didn’t get to see, which is important for understanding what the area would have been like in the Carboniferous, is coal. Exposures of coal seams are hard to find in the UK. This is partly because coal is a rather soft rock, so doesn’t easily form good natural exposures. More importantly, any coal seam that was exposed has been dug away over the last thousand years or so, with recorded digging of coal in the Lothian area in the twelfth century ( However, a coal seam within the Westphalian Coal Measures has been recently exposed in the area and can be visited (Fig. 4).

Downstream from Roslin Glen, the River North Esk joins with the South Esk to form, with a notable lack of imagination, the River Esk. This flows through Dalkeith Country Park ( The park itself offers free parking, a shop and cafe, as well as trails for walking and cycling and is understandably popular. Oddly, the coal seam isn’t listed as one of the parks attractions. However, at UK grid reference NT 343 693 on the southeast bank, where the river flows under a bridge for a new section of trunk road, a 20 to 30cm seam of coal sits a metre or two above a track, which is not shown on even the most recent Ordnance Survey map.

Fig. 4. The coal in Dalkeith Country Park. The seam is 20cm to 30cm thick, with a pale yellow river sandstone above and a light grey seat earth below.

There is a pale grey siltstone below the seam, which is a seat earth or fossil soil in which the plants grew that formed the coal. There are no obvious roots in the seat earth and certainly no large tree-type roots that are common fossils elsewhere. So, the coal probably formed from small plants and would have accumulated as a peat in a bog or swamp, probably on the floodplain of a river.

The coal has millimetre-scale laminations, which would have been more like one centimetre thick when the peat formed. Because peat is largely water and very low density once dried out, as anyone who has burned peat on a fire will know, peat compacts to around only one tenth of its original thickness by the time it has turned into coal. These laminations might represent annual growth, each layer being a single growing season, although, as tropical climates are not very seasonal, perhaps not.

Above the coal is a cross-bedded sandstone, which represents a river deposit. Often, the channels in rivers cut down into the underlying sediments as they migrate sideways across the landscape. This often removes any peat that has formed on the surface, as we saw earlier in Roslin Glen. Here however, the sands do not cut into the coal, suggesting they might represent flood waters spilling from a river channel, rather than the channel itself. The sands are well exposed in a separate exposure a few metres to the southeast and can be seen for several hundred metres in the opposite bank of the river.

Summarising what we know about the area back in the early Carboniferous, this was an area of rivers a good few metres deep, running across flood plains with extensive vegetation in a tropical climate. In other places, there were lakes and overall the area might have been not much different to the Everglades today. Much of the Carboniferous of Scotland is fairly similar. We do not find the classic Carboniferous Limestone – Millstone Grit – Coal Measures sequence that is found further south in northern England, probably because the Scottish Highlands were an upland area that generated a steady supply of sediment and prevented the formation of thick limestones. There are thin limestones and these have been quarried extensively for agricultural lime, but they are for another day.

The Roslin Glen walk can be followed by anyone and more information can be found in the Lothian Geology guidebook by McAdam and Clarkson, which is online:

About the authors

Mark Wilkinson is Senior Lecturer in Geological Carbon Storage at the University of Edinburgh, UK, where he runs a Masters (MSc) course in GeoEnergy, as well as teaching geology to students. Claire Jellema (MSc Sustainable Cities from Kings College London) is the Midlothian Science Festival Manager, coordinating a two week-long science festival every October. Now in its sixth year, the 2017 festival attracted near 8,000 people.

Further reading

Scottish Fossils, by Nigel H Trewin, Brocken Dunedin Academic Press Ltd, Edinburgh (2013), 118 pages (hardback), ISBN: 978-1-780460-019-2

Leave a Reply