The structure stood about eighteen metres long, quadrangular in plan, a few yards south of the medieval tower house that still marks the site. What drew attention to it was the gap between what people said it was and what it most likely had been.

When the historian and clergyman William Carrigan visited in the early twentieth century and published his account in 1905, he found the building already broken down. He noted the local tradition that it had served as a chapel, but was sceptical, suggesting it looked more like a hall, mansion, or dwelling house attached to the castle, perhaps pressed into use as a place of worship only briefly and secondarily. A more mundane explanation has since been proposed: the long range, oriented northeast to southwest, may have been a stable block associated with Bonnetstown House, which was built in 1737. That would make it a piece of eighteenth-century estate infrastructure rather than anything older or more devotional. The building appears on the first edition Ordnance Survey six-inch map, surveyed around 1839 to 1840, still standing at that point roughly five metres south of the tower house. By the time the map was revised in 1945 to 1946, it had gone entirely.

What remains, then, is a gap on the modern ground where something once stood, and a small puzzle about what that something actually was: a stable, a hall, a chapel, or some combination of all three at different moments in its life.