Trial and error with a mis-mapped Scottish castle: Glancar Castle/Rough Island


This is the first in a series of blogs by Stirling doctoral students explaining aspects of their research.

We think of maps as reliable guides to place, and of castles as clear landmarks. But maps often take their bearings from the imperfect efforts of earlier cartographers, and incorporate all manner of uncertainty and conjecture in their representation of the physical world. Even the striking and monumental figure of a castle can slip through the resulting cracks of hearsay and doubt.

I was exploring the south-west of Scotland’s many stunning castles when I came across a seventeenth-century map which depicted a site labelled ‘Glancar’, in the middle of a loch. I knew this was a castle site because the logo used to illustrate it was identical to that of the nearby and better-known Threave Castle, ‘Freif C.’ on the map below, notable for being on an island in a river. I also knew roughly which loch this was; the ‘Ouer flu.’ is the Urr Water, so this must be Loch Urr.


Figure 1: Jenner & Hollar’s 1644 (reprinted 1752) map ‘The north-part of England and the south-part of Scotland. Quartermaster’s map.’ NLS shelfmark Nha.Map.b.21 (C) NLS:

I turned to Loch Urr, where there are three possible castle sites in and around the area. The two less likely candidates are the earthworks called White Isle, of uncertain date, and Loch Knowe, a promontory fort. The most obvious is the third candidate, Rough Island, which in the first edition of the Ordinance Survey’s 6-inch map is subtitled “Castle (Remains of)”.


Figure 2: Map of Loch Urr with Loch Knowe, Rough Island, White Isle (C) OS, RCAHMS

So far, so good; I could set aside the mystery of Glancar and carry on with my main research, which focuses on Scotland’s earliest stone castles, both as monuments and as part of wider social, economic and cultural phenomena. Except that I was curious as to why the castle was not called Glancar Castle on the most recent maps of the area. So I resolved to look at other maps from the early modern period to seek clarification. Here, things got a little complicated. One of the earliest maps, dated to 1595 AD, seems very familiar – and it’s likely that this inspired the map which I first looked at above.



Figure 3: Mercator, ‘[Scotiae Regnum south sheet]/ per Gerardum Mercatorem per privilegio.’ NLS shelfmark EMS.s.4a (C) NLS:

Here, ‘Glancar’ is very clearly indicated as a place on an island in a loch from which the ‘Ouer flu.’ flows. I could begin to say, therefore, that the site on the island in Loch Urr was a castle called ‘Glancar’, and was noted for being occupied, or a ruin, from the late sixteenth century onwards. Unfortunately, the story becomes confused, as the wider series of maps for the area begins to show:


See table below for image credits.


The maps above give a view of how the area around Loch Urr was perceived and changed over more than 200 years. Though the location and positioning of ‘Glancar’ changes, clearly there was a recognition of the loch and the name being associated. The last two maps, by Ainslie (1797) and Thomson (1821), are significantly more detailed, but here the name ‘Glancar’ is dropped entirely, although two distinct island sites in the loch are depicted.

Critically, however, it is a later map of Blaeu dated 1654 (oriented with west at top) which provides the key to unlocking the mystery of ‘Glancar’ castle.



Figure 4: Blaeu, ‘Praefectura Kircubriensis quae Gallovidiae maxime orientalis pars est. The Stuartirie of Kircubright, the most easterlie part of Galloway.’ NLS shelfmark EMW.X.015 (formerly WD.3B). (C) NLS:


Here we have Loch Urr with an eponymous associated settlement. And no ‘Glancar’. After Blaeu’s 1654 map, no further castles called ‘Glancar’ are asserted in Loch Urr. Only Coronelli’s 1696 map has the label ‘Glancar’ near the loch; it is worth noting that this last map is very similar to, indeed very likely a modified copy of, Speed’s 1610/1627 map, hence prior to Blaeu’s correction.

Three questions therefore remain; what or where is ‘Glancar’; was there ever a castle on Rough Island; and why was the error, corrected by Blaeu, perpetuated for so long prior to 1654?


‘Glancar’ was indeed a castle, though by a different name; Fergus and William of Glencairn appear as witnesses to charters in the early thirteenth century. Glencairn Castle is now lost to us, partly demolished and partly modified to make way for the (relatively) more modern Maxwelton House, a change which took place in the seventeenth century (according to RCAHMS’ database Canmore).

Rough Island?

A report submitted to the Dumfriesshire and Galloway Natural History and Antiquarian Society, published in their proceedings for 1900-5, reports on an excavation of Rough Island. Concluding the island to be likely artificial, the excavator was content to label it a crannog, a type of artificial island familiar across Scotland and built from prehistory to the seventeenth century. In the closing remarks of the brief report, however, he noted from correspondence with fellow antiquarians that “the presence of the stone buildings and the earthwork [White Isle] gave the crannog a character different from any which had yet been described.” Unfortunately the finds from the excavation were limited to undated pottery; consensus at the time suggested Rough Island was a post-Roman settlement, though since then pottery from around the thirteenth century AD was said to have been found there.

Timothy Pont

A further complexity is added when we consider that Blaeu’s map was in fact a published edition of a much earlier rough survey drawn by Timothy Pont between the late 1580s-90s. Here, however, we can see the kernel of resolving the confusion over ‘Glancar’ and Rough Island.

14 with arrows

Figure 5: Pont, [Nithsdale]. NLS Shelfmark Pont 35 (1) Nithsdale (C) NLS:

On the far left is the very vague outline of Loch Urr, the faint label below reading ‘L. Orr’ (left arrow); to the right, past streams, forests and farms, in the top-right corner is ‘Glenkairn Cast.’ (right arrow). It would seem that as early as the late sixteenth century, Pont had already created a detailed map which would have spared contemporary and future cartographers from the confusion of ‘Glancar’ and Rough Island.

Why the confusion?

It seems there are several points which led to confusion amongst most cartographers in the maps of Loch Urr and the presence, attested or otherwise, of a castle in the loch. The first is sheer proximity, and the translation of that proximity to early maps. The distance between Glencairn/Maxwelton and Rough Island is 7.8km (4.8 miles), which equates to very little when considering many of the earlier maps depicted Scotland as a whole. It is not clear why Glencairn was posited to be a castle on an island rather than by an unrelated river, but the suggestion stuck with cartographers from Mercator to Blaeu for over 50 years.

More subtly, it is possible that the persistence of ‘Glancar’ in Loch Urr was recognition of there having been a lived structure on Rough Island in the sixteenth and seventeenth centuries (as one of many phases of occupation), which may very well have been associated with Glencairn/Maxwelton Castle, something which the unverified thirteenth-century pottery piece would strengthen. Equally, Matthew Shelley’s PhD thesis on crannog sites has emphasized the links between crannog sites and their environs well into the late medieval period, which may suggest Rough Island as a castle site in its own right rather than a retreat for Glencairn/Maxwelton Castle. It may not be entirely coincidental that the change in cartographic policy, instigated by Blaeu in 1654, for labelling Rough Island and Loch Urr as ‘Glancar’, came about in the same century that Glencairn Castle was remodelled and became Maxwelton, thereby shedding its medieval name.

The small riddle of ‘Glancar’ and Rough Island has been resolved, after much trial and error. It serves as a useful lesson in the use of historic maps for research, and perhaps as a cautionary tale on distraction during research!


Corrie, J., ‘The Loch Urr Crannog’, TDGNHAS Vol. 17 (1900-5), pp 242-6:

Shelley, M.J.H., Freshwater Scottish loch settlements of the late medieval and early modern periods; with particular reference to northern Stirlingshire, central and northern Perthshire, northern Angus, Loch Awe and Loch Lomond (PhD thesis, University of Edinburgh, 2009):

Image credits, 9-map mosaic:



William’s research is funded by the AHRC, in partnership with RCAHMS. Edited by Victoria Hodgson.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s