Duntulm Castle

HES dates the courtyard castle to the fourteenth and fifteenth centuries, with later additions. The older RCAHMS description places the principal structure in the fifteenth century. Both accounts make a sixteenth-century starting date too late for the complex as a whole. They do not supply a secure year or a confidently identified individual builder. The possibility of an earlier fort called Dun David belongs to tradition; an Ordnance Survey visit found no trace of that supposed predecessor.

A description by Donald Monro in 1549 names the castle and associates it with Donald Gorm. This provides a documentary landmark distinct from the broader dating of the masonry. The later history unfolded within the contested lordship of Skye, including MacDonald and MacLeod rivalry, but a general clan conflict should not be turned into an invented siege at Duntulm. The evidence needs to identify the event as well as the families involved.

In 1617 the Privy Council required Sir Donald MacDonald of Sleat to make Duntulm his seat, and to build or rebuild if the existing accommodation was unsuitable. That instruction is more specific than a vague claim that the family merely occupied an old ruin. The castle’s later domestic additions belong to a continuing history of adaptation. Changes to accommodation should be distinguished from the earlier courtyard defences rather than compressed into one building campaign.

The surviving layout also records that complexity. HES describes alterations to the curtain walls connected with sixteenth-century defensive developments and later buildings within the enclosure. A main residential tower once dominated the site. A smaller adjoining tower and a seventeenth-century house added further accommodation. The sea gate and the landward ditch show different relationships with the surrounding terrain, while the vaulted lower structure preserves a trace of the household spaces beneath the lost upper rooms.

Around 1732 Duntulm was given up in favour of Monkstadt. The Highland historic environment record notes that its stone was reused there, while the record for Monkstadt distinguishes the start of work in 1732 from completion in 1741. This was a process of changing residence and reusing materials, not proof that every wall collapsed on one date. Later weathering and structural failure continued to alter the site; HES records a tower collapse in 1990.

Duntulm is a scheduled monument, and the surviving masonry has archaeological value even where a visitor sees only broken surfaces. A 2026 watching brief recorded ash deposits during work at a new access gate. The record recommends further examination; it does not yet provide a radiocarbon date or prove an early fort. For a visit, read the castle from safe ground and respect barriers rather than approaching unstable edges. A broad view brings sea, ditch and ruins into one composition. The most compelling story here is the changing residence that the evidence reveals, with its gaps left honestly visible.