The Maiden Castle, Midlothian

This account is an extended version of the text on the display board at the site.

Location and Site                                       Grid reference NT 286644

Maiden Castle is located in the valley of the North Esk, in Midlothian, upstream from the small village of Polton, which is between the small towns of Bonnyrigg and Loanhead; the village of Roslin lies to the south-west. The site can be approached by a walk of no more than 1km from Polton, with more slopes, or a walk of nearly 2km from Roslin, with fewer slopes.

Maiden Castle occupies part of the meander core of the North Esk where the river has cut down, or become incised, to create steep slopes in several places.

Maiden Castle was probably a late Bronze Age or Iron Age site and, if so, would have been occupied for a time between about 1000BC and 200AD. It would have been one of many sites throughout Midlothian and East Lothian. The only visible remains today are two low mounds with an intervening ditch, which can be seen as you approach the site from a north-westerly direction. The outer bank is very short, but the inner one can be seen curving round for several metres. It is easy to walk over these low mounds without realising that they are the ramparts, as they are inconspicuous compared with the slopes which have been negotiated in order to reach the site.


The only records of Maiden Castle are four reports by inspectors from the Royal Commision on the Ancient and Historical Monuments of Scotland (rcahms) who visited the site in 1927, 1954, 1956 and 1983. These reports are very short, only one extends to two paragraphs. Each of these reports stresses the defensive role of the site. The following vocabulary is used, with the year of the report indicated.:

motte and bailey 1927

defensive works 1927, 1954

castle and keep 1954

fortress 1954 (quoted from NSA of Scotland 1845)

ramparts and ditch 1956

earthwork 1956

ditch and outer bank 1983.

Some of this vocabulary is open to question. Despite its name, Maiden Castle is not a castle, or even a keep, in the way we usually understand these words for medieval structure like Craigmillar Castle (284709) or the keep at Crichton Castle (380611). It is possible that the inspector in 1954 mistook the old quarry, which lies to the east of the ramparts, for Maiden Castle. The use of the word fortress also conjures up visions of a much grander structure than existed here. The 1954 report appears to have been influenced by the NSA of 1845 which it quotes  as follows, “Nothing now remains of this fortress but parts of the foundations which are still visible in some parts”, and then goes on to conclude that this supports “the theory that such a building existed here”. No mention is made of the quarry in this, or the preceding report. The small cliff face can give the impression of being an earth retaining wall of a an old building, but it is natural, or in situ, stone. The sandstone blocks which are lying around were the last to be quarried but were never moved. For identification of the quarry we must await the 1956 report. There are no remains of a building at this site.

It is difficult to understand why the claim was made that there was a motte and bailey here. The motte and Bailey is a simple defensive structure which was probably introduced into Britain after the Norman conquest. King David I (reigned 1124-53) had strong associations with the Norman court as his sister married the Conqueror’s son, who became Henry I. David endeavoured to extend royal control by inviting local and Norman lords to take up estates. They were expected to build a motte and bailey for defence and residential purposes. A motte is a large mound of earth, which would be round or oval in shape and flat on top. It would have required a great deal of labour to build, but servile labour was usually available. It would have had a palisade at the top or bottom, or both. The bailey, or area of ground around the motte, would also have been surrounded by a palisade. The motte and bailey were a statement of Norman authority and superiority.

It is impossible to find any structure at Maiden Castle which could be interpreted as the remains of a motte and bailey. Whilst motte and bailey strucrures can be found in various parts of Scotland, none are recorded in RCAHMS’ catalogue for the Lothians.The reason suggested for this is that is that the Lothians were the most developed part of the realm and that servile labour had already been commuted to wage labour. This made the construction of a motte and bailey too expensive. The local Lord may as well employ  stonemasons and build a proper castle. A good example of a motte can be found on the golf course at Carnwath; this apparently had no bailey. Other baileys are likely to have been destroyed by now and the land built on or cultivated.

Another feature at the site, near to the quarry, is a man made trench. Its purpose is open to conjecture.

Hillforts and Homesteads

The preceding comments on the reports by inspectors from RCAHMS, as well as the actual site, raise the question to what extent Maiden Castle was built for defence, if at all.

The term “hillfort” has entered the popular imagination from the early days of antiquarian studies with reference to all Iron Age sites. This is partly a consequence of its valid use for large forts like Maiden Castle in Dorset and on Eildon Hill North, near Melrose (NT 554328). This attitude is partly a response to the interpretation of history,which saw the Romans as part of the classical, literate and civilised world. Consequently the indigenous people who lived here prior to the arrival of the Romans were regarded as uncivilised. They were barbarians who were always fighting each other and who therefore had to live behind defences. This attitude led to practically every Iron Age site being called a hillfort. Work in recent decades has seriously questioned this view.

In his book “Iron Age Britain”, Prof. Barry Cunliffe points out that between the North York Moors and the Scottish Highlands only about 14 of approximately 1,500 sites are large enough to be called hillforts. The vast majority of sites cover less than one hectare. Maiden Castle obviously falls into this group, which is designated “homesteads”. The classification of Iron Age sites used by RCAHMS defines a homestead as having a maximum of three houses. As so little is known about Maiden Castle RCAHMS’ lower classification of “enclosure” may even be more appropriate, with its definition indicating that occupation is absent or unknown. It is conceivable that the site was not inhabited at all, but was used as a gathering centre or meeting place for people who lived around and was used for social, political, ceremonial or religious purposes.


The extent to which the site could be defended is problematic. It obviously has some defensive advantages. The incised meander and rock formations mean that it is almost surrounded by steep slopes, yet no attempt appears to have been made to enhance these advantages by defending the shortest line across the neck of the meander. Only part of the meander core was occupied.

The earth ramparts, now heavily degraded, can still be seen on the west side of the site. They would originally have been much higher, probably rising to as much as two metres above the intervening ditch, with a wooden palisade on top. Their construction would have required a great deal of labour. Both the purpose of these defences and the method of construction of the palisade are open to discussion.

Some people are of the opinion that these palisades were constructed of vertical timbers made from tree trunks. Others are of the opinion that well spaced vertical timbers would have had the intervening gaps filled with smaller timbers, like branches, laid horizontally and interwoven.

Two basic objections have been raised against the first idea of only using vertical posts. By this time much of the original forest had been cleared, consequently there may not have been sufficient timber available.  Even if trees were available questions can be asked about whether the enormous amount of labour required for felling, lopping, transporting and erecting theses timbers was available. The second type of construction therefore seems more likely, that is, with branches laid horizontally between vertical posts.

The terminology of “hillfort” and its association with “barbarians” has already been discussed. Questioning the validity of “hillfort” also includes questioning the purpose of these defences and asking if they really were built for strictly military purposes, or whether the people had  other aims. Cattle could have been brought inside at night to stop them from roaming away and also as a precaution against cattle thieving. Thieves themselves could have been deterred from entering. Defence against wild animals, like boar, may also have been necessary.

The site may have belongd to a local chief who wanted to impress his authority upon others. If this was the case he could probably have commanded more labour from people who lived nearby in “open”, or undefended houses, some of whom may have been held in a kind of client or subservient status. The Iron Age site of Chesters (380611), north of Haddington, is regarded as a good example of a site where social status was more important tham military defence. Its multiple ramparts are impressive, but it is overlooked by a hillside from which missiles could have been thrown into the site. It is neither on a hill nor is it a fort.

A further purpose of these defences would be privacy. In parts of rural tropical Africa today, compounds are still surrounded by a tall fence of woven  grass matting fixed between vertical posts. Inside, the whole area is referred to as the “house” and the individual huts as “rooms”. The fence does not offer a defence against anything more than roaming dogs or cattle, but does provide privacy. Visitors are not expected to just walk in through the narrow entrance, which can be closed at night, but to announce their arrival.


There are several Maiden Castles in Britain, as well as physical features which use the word Maiden, but the  origin of the word remains obscure. There is no guarentee that the origin of the word is the same in every locality, or that it originated at about the same time in history in every locality. Furthermore, the spelling of the word may have changed through history. Needless to say, the subject is ripe with conjecture.

A strong contender for Maiden is that it comes from the Celtic”maithdun”, meaning a large fort, and the Maiden Castle on the coast near Arbroath (NO670420) is often used to support this idea. Unfortunately our Maiden Castle cannot be described as large compared with some others in eastern Scotland. Local people, however, may not have known of these other sites and, even if they did, thought themselves justified in a degree of exaggeration!

Another claimant for Maiden is that it is derived from the gaelic “mai dun”, which means a great hill. This may seem reasonable in some places, but it can hardly be claimed for the Maiden Castle on the North Esk without some stretch of the imagination.

A claim which contradicts the the previous two is that it is derived from “moe din”, meaning a grassy plain, which would not appear to be a good place to build a castle in the first place. The fact that these words are also supposed to mean a “virgin fort” does not inspire confidence.

There are many claimants for this idea that the fort was never taken in battle and also for the idea that maidens gathered here. Unfortunately both of these ideas require a later definition of the word maiden. Although the word appears to have already been in use by early medieval times, this is still a long time after the Iron Age and therefore does not prove that the name dates back to the origin of any such sites.

A further idea can be found in “Scottish Place Names”, by W.C.Mackenzie (1931), p.4-5.

“The tradition about Edinburgh Castle is that the maidens of high birth were shut up there; presumably when their menfolk went to war. And it is a well authenticated fact that in ancient Scandinavia that was a customary thing to do.”

No dates are are mentioned, but presumably early medieval times are implied, unless “ancient Scandinavia” dates further back.


Apart from the degraded ramparts which can be seen today, nothing is known for certain about Maiden Castle. The approximate dating of the site depends upon comparing the ramparts with other sites with those features. Within that extensive period of twelve centuries we do not know exactly when the site was occupied or for how long, or whether it was occupied intermittently. In fact we cannot be certain that anyone lived there at all, as it may only have been used for ceremonial purposes. Even the origin of the name “Maiden Castle” remains elusive.

Howard Turner  April 2010

Back to Stories