Or so the cartographers of the mid-seventeenth century believed. The precise location has never been established, which gives this entry an unusual quality: it is a place defined almost entirely by its own absence.

The evidence comes from the Down Survey, a remarkable mapping project carried out between 1655 and 1656 under the direction of William Petty, Surveyor General to the Cromwellian administration in Ireland. The survey was undertaken to document land ownership across the country following the upheavals of the 1641 rebellion and the subsequent Cromwellian conquest, with confiscated lands needing to be measured and recorded before redistribution. The resulting maps are among the earliest systematic cartographic records of the Irish landscape, and they show a castle at the townland of Carmanhall, a name now largely superseded by the more familiar Leopardstown. Compiled in the records consulted by Geraldine Stout and Padraig Clancy, the entry carries a candid note: not precisely located. The castle appears on the historical record and then, in spatial terms, dissolves.

For anyone curious enough to go looking, the townland of Carmanhall lies in the southern suburbs of Dublin, roughly between Sandyford and the Leopardstown Road. The area has been heavily developed since the late twentieth century, with office campuses and retail estates replacing much of the older agricultural landscape. There is no marker, no ruin, and no clear candidate structure remaining above ground. What a visitor is really engaging with here is the cartographic record itself rather than any physical remnant, and the Down Survey maps are accessible online through the Trinity College Dublin digital collections, where the relevant sheets can be examined directly. If there is something to be found on the ground, it would require archaeological investigation to confirm it, and none appears to have been published to date.