Lost Castles

Lost Castles

THE CLAN ARCHIVE · LOST CASTLES

Lost Castles

Lorg an Chaisleáin
“Conor Mac Hale counted twenty. We have located and documented ten — these are the sites that still elude us.”

This page is a running record of everything we have not yet confirmed on the ground — the castles Mac Hale named but we cannot yet stand at.

I.  The count that won’t settle

Every history of the O’Dubhda begins with a number. The Wikipedia list, drawn from Conor Mac Hale’s research, names twenty castles. Tony Dowd’s family archives reference twenty-four castles and fifty-two towns. Local historians in Tireragh quietly suspect there are more still.

We have identified, documented, and published individual pages for ten of those sites — the castles you can find in the Castle Tour. This page is for the rest: the sites we know existed, but which we have not yet matched to standing masonry, to a verified location, or to a single identifiable place on a map.

A living ledger. Every entry below is a working stub. When a site graduates to a confirmed find — with a townland, a set of walls, and a defensible claim on the historical record — it moves off this page and into the main Castles index. The archive here grows shorter, not longer, as the research proceeds.

II.  What counts as “lost”

We keep the definition narrow. A castle is listed here only if no identifiable fortification remains are standing and no confirmed site on the ground has been matched to the name. The moment we confirm a set of walls, a mapped townland, or a surviving earthwork, the site is no longer lost — it may still be a ruin, but it is findable, and it belongs with the rest of the Castles pages.

Three categories of uncertainty appear below:

The Vanished. Known to have existed from historical records, but no physical trace survives and no map or tradition fixes them to a single place.

Mapped, but unverified. A townland is named in older records; walking the ground has not yet turned up confirmed remains. Either what survives is too eroded to read, or the recorded name may attach to a site we have already documented under a different name.

Later buildings on vanished castle sites. A later house or estate stands on or near a vanished O’Dubhda fortification. The modern structure is not the O’Dubhda castle, and in some cases is not even on the same spot. The original fortification is effectively gone.

III.  How entries get added

A site lands on this page when it meets at least one of the following:

  • It appears in Conor Mac Hale’s list (1990) or in Tony Dowd’s family archives, and is not currently matched to one of the ten documented castle pages.
  • It is named as an O’Dubhda fortification in the Annals, in the Down Survey of 1656–58 or its later redactions, or in John O’Donovan’s 1836 Ordnance Survey letters, and the location cannot be securely reconciled with a site currently mapped.
  • It appears in early English cartography — Baxter c.1600, Speed 1610, Boazio 1606, Down Survey barony maps — under a placename that no modern gazetteer or OS sheet can pin down. The Baxter map alone adds a cluster of unidentified towers along the Tireragh coast, three of which are listed below.
  • A credible local informant (historian, landowner, parish archivist) has reported walls, foundations, or a tradition that we have not yet been able to verify in the field.

If you have information that belongs here — or that would move a stub off this page and into the main Castles index — please get in touch.

The Count
Mac Hale’s list
20 castles
Tony Dowd archives
24 castles, 52 towns
Identified & mapped
On this page
8 stubs, and counting
New from the Baxter map
Help Us Find Them

We are actively searching for every castle, fortification, tower house, bawn and ruin connected to the O’Dubhda clan — and we know there are more out there.

If you have knowledge of old castle sites in the Mayo–Sligo area, references in documents or family stories, or simply a field with suspicious old walls in it — please share it with us.

Every lead matters. Every piece of local knowledge could unlock another chapter. Get in touch.

A Professional Survey
We are exploring a full archaeological survey of the O’Dubhda castle network, including LiDAR and historical-map cross-referencing.
If you’d like to sponsor this work, or join as a researcher or volunteer, contact us.
The Ledger

Sites We Have Not Yet Found

Each entry is a working stub. When a site is confirmed on the ground, it moves off this page and into the Castle Tour.

The Vanished
Doonecoy
Location unknown · Co. Sligo or Mayo
Appears in Mac Hale’s list of O’Dubhda castles and in older clan records. Not identified on any surviving map, Down Survey sheet or Ordnance Survey record. No tradition has fixed it to a specific townland. A true mystery — possibly a scribal corruption of a name we already know under another spelling, possibly a genuinely lost site.
Source: Mac Hale (1990)
The Vanished
Grangbeg
Tireragh coast · near Grangemore, Co. Sligo
Drawn as a distinct tower on the Baxter c.1600 map, sitting immediately north of Grangemore along the Tireragh coast. The name (“little Grange”) suggests a satellite ecclesiastical grange rather than an O’Dubhda fortification proper — Mac Hale notes that Grangemore itself “did not belong to the O’Dubhda family” but stood within their territory. Grangbeg is not on Mac Hale’s 1617 list and has not been matched to any surviving remains or modern townland. It may be a pre-dissolution grange site, now ploughed out.
Source: Baxter c.1600 (RMG P/49(7))
The Vanished
Dunbuoy
Tireragh coast · south of Aughris, Co. Sligo
Named as a castle tower on the Baxter c.1600 map, drawn on the Tireragh coast south of Aughris, within the large area labelled in red “O Dowdes Countrie.” The spelling renders Dún Buí, “the yellow fort” — a colour-name of a type well attested in Irish fortification toponymy. Dunbuoy does not appear in Mac Hale’s 1617 list and has not been correlated with any standing remains or surviving townland name. It is possible, but not established, that it is a Baxter-era rendering of a name recorded elsewhere under a different spelling; it is equally possible that it records a coastal tower now fully vanished.
Source: Baxter c.1600 (RMG P/49(7))
The Vanished
Duncarragh
Tireragh coast · inland of Aughris, Co. Sligo
Drawn as an inland tower on the Baxter c.1600 map, labelled between “Aghfure” (Aughris) and Grangemore, a little set back from the coast. The name resolves as Dún Cárrach, “the rough or rocky fort” — a standard Irish topographical form. Duncarragh is not on Mac Hale’s 1617 list, is not recorded on the Down Survey barony sheets of Tireragh, and has not been matched to a modern townland or set of remains. Because Baxter drew it on direct observation rather than copying an earlier map, a site matching the name should exist, or once have existed, in the hinterland of Aughris.
Source: Baxter c.1600 (RMG P/49(7))
The Vanished
Rath Maoilchatha (Rathmulcaha)
The royal rath at Castleconor · Baill Uí Dhubhda
In 1640 a herdsman accidentally rediscovered the entrance to an old fort at Castleconor that had been sealed for centuries. Duald MacFirbis went down into it himself and described nine cellars of smooth, well-cut stone — proof, he argued, that the Irish were building in dressed masonry before the Normans arrived. He named the rath as one of the ancient royal raths of Ireland and called the surrounding ground Baill Uí Dhubhda, the town of Dubhda. The openings have since been filled in by farmers.
Source: O’Reilly (1971), citing MacFirbis
Mapped, Unverified
Carahduff
Lackan townland · Co. Sligo
Named in older Tireragh records as a castle in the Lackan townland. The same townland contains the Lackan / Lecan Castle fragment associated with the MacFirbis scholars. Whether Carahduff is the same site under an older name, or a separate structure within the townland, has not been resolved. No confirmed independent remains have been identified.
Source: Mac Hale (1990), older Tireragh surveys
Mapped, Unverified
Castletown Castle, Easkey
Castletown (Baile an Chaisleáin) townland · Easkey, Co. Sligo
Stands — in tradition, at least — on the left bank of the Easkey River at its mouth, in the townland of Castletown (Baile an Chaisleáin, “the town of the castle”). Recorded as built by the Muldoons or Murphys in 1157 AD and subsequently held by the O’Dowds, with local tradition claiming that the O’Dowds had five castles in this single townland. A site near 54.29077, -8.95996 sits roughly 300m downriver from Rosslee Castle. The close proximity raises a real question: is “Castletown Castle” a separate structure at the river mouth, or is it a duplicate identification of Rosslee under the townland name? Field verification needed.
Source: Local tradition; relationship to Rosslee under review
Later Building on Vanished Site
Longford House (“Lomford”)
Near Ballysadare · Co. Sligo
Listed in older records under the spelling “Lomford” as part of the O’Dubhda castle ring. The current Longford House is a later estate residence, not the O’Dubhda structure, and no identified remains of the original castle survive on the site. Recently sold as a private residence, limiting access for field verification.
Source: Mac Hale (1990)
Later Building on Vanished Site
Bonniconlon House (“Beaufield”)
Bonniconlon · Co. Mayo
Listed in older records under the spelling “Beaufield.” Within the O’Dubhda territory of north Mayo, but details of the original castle are sparse and the current house bears no identified medieval fortification remains. Further archival work needed to pin down the original site.
Source: Mac Hale (1990)
Working list — more to come

The stubs above are the first entries, not the last. We are actively working through Mac Hale’s list, the surviving Down Survey terriers for Tireragh, O’Donovan’s 1836 letters and the newly-rediscovered Baxter c.1600 manuscript map. Additional sites will be added as they are verified — and, we hope, removed as they are found.

A Note from the Clan

These pages are volunteer-authored. The entries above are deliberately unfinished — we would rather publish a stub we can source than invent a history we cannot. If a site here has been found, misnamed or correctly identified under another entry, we want to know.

If you know of an O’Dubhda-associated castle that isn’t on this page or on the main Castles index, please get in touch.

Copyright © 2026. All rights reserved.