This broch sits on a small headland jutting out into the lower end of the Loch of Tankerness.
There are two ways to reach the site; Neither is straightforward and both requiring boots. One is to follow the mill stream and step on and over the wall where the stream comes to the loch shore, then follow the shoreline to your right. This way is very marshy and somewhat treacherous. Alternatively go up the hill from the Tankerness Mill and just round the road from Northwood House (the sound of doves comes from Glebe). Go to the field that is directly behind this, then downfield to the shore. It tends to be rather suckily muddy in many places. I would suggest going down the left-hand side, as the wall does a few strange things. To get to the headland requires going through a barbed wire fence , which fortunately has plenty of give in it or I'd never have managed it.
It is one of the least prepossessing brochs I have seen, though at the seaward end it stands rather higher than the likes of Oxtro. At first sight I thought that there was nothing. Going towards it you can certainly feel plenty of apparently isolated stones hidden beneath the grass but no sense of stucture that I could tell. Coming to the highpoint and taking the line around you certainly get a strong feel for where the inner and outer walls were, going by the turf.
At the top and found a roughly rectangular depression. Looking on the inner side I made out several slabs in a non-radial line. So I did some rooting about and tearing off of the overhanging grass. Definitely a straight wall, remaining about six foot long and a foot or so high. The other side is less well defined, a few loose stones as far as I could tell and not so obviously slabs. If you go poking about then gloves are a necessity - the place is covered in almost invisibly short nettles, I can fully feel the nasty little buggers several hours later! Maybe this is how such sites can sound so uninteresting in official reports , that only amateurs are daft enough to brush brush aside for a proper view.
The 1798 and 1842 Statistical Accounts describe this as semi-circular and 140 yards across, having walls 12' high and 9' thick. If there were internal compartments these were filled with rubbish at that time. Unfortunately the site was heavily 'quarried' for stone in order to make the glebe wall i.e. that around the minister's land. They believed the site might have once been used in part for burials. There were further walls on the outside of the "great wall". In the various compartments were found "peat ashes" along with animal bones (including deer horn), shells and small bone rings.
In 1931 a cist and some prehistoric fragments were excavated here (notes accompanying photo of ongoing excavations, no follow-up article found). Probably the cist is the 'box' I found.
In "The Archaeological Sites and Monuments of Scotland , 27..." by Raymond Lamb it is said this site is akin to Bretta Ness in Rousay, which upon excavation was found to be a crannog. As water levels have risen by a meter and it still isn't cut off from the land I somehow doubt this. But perhaps like Wasdale this is a possible causewayed island dun?
From RCAHMS I would add that there are several rectangular banks under a foot high to the S and W and that one of these connects to one of two probable defensive walls at 27m and 42m from the centre.