Bracket 14 at the Cwm Hesgin slate workings is a flat plate 190mm by 95mm with two bolts at 140mm centres, and it has outlasted the shed it was screwed to. The shed is a rectangle of nail-stained turf now. The bracket is still there at chest height on the north face of the incline house, holding nothing, canted about eight degrees off level because the mortar under the lower bolt gave up before the upper one did.
I record four things for each one: overall dimensions, bolt pattern, mounting substrate, and corrosion state. That is enough to sort them and not so much that I stop bothering on a wet afternoon. The tape is a folding boxwood rule I keep because it does not skid on slate the way a steel tape does.
Mounting styles fall into three groups in the workings I have covered so far. Plate-and-bolt straight into dressed stone is the commonest and the most durable. Then there is the socket type, a cast collar mortared into a drilled pocket, which fails when water gets behind the collar and jacks it out over about forty winters. Last is the strap type, a wrought band round a timber post, which almost never survives because the post goes first and the band drops into the spoil and gets buried.
Slate holds water against the metal and keeps it there. Brackets on slate go to a dense black-brown scale that flakes in sheets roughly 1mm thick, and the bolt shanks thin out at the exact line where the plate meets the rock. Limestone is drier and more alkaline at the contact and the rust is a looser orange that comes off as powder. On limestone I have twice found brackets where the plate is sound but the wall behind has spalled away, leaving the whole assembly standing proud on its own bolts like a shelf bracket in a demolished house.
Bolt pattern is my rough dating tool, and I want to be honest that it is rough. Two bolts on a vertical centreline, square-headed, generally reads as pre-1880 to me. Two bolts side by side with hexagonal heads sits later. Four bolts in a rectangle usually means the bracket carried a heavier acetylene fitting and belongs after 1905. The gap between those groups is wide enough that a bracket reused from an older working will lie to you completely.
Bracket 22, in the Bryn Ithel limestone quarry, is the one that keeps bothering me. Square heads, vertical pair, 165mm centres, but the plate has a machined boss on the front that no early gas fitting would need. Reused stock, or a local smith working to an old habit. I have left it unclassified since March and I expect it stays that way.
Next trip is the eastern galleries at Cwm Hesgin, where the roof water drips steadily and I expect nothing to have survived at all. Worth checking. Absence is a measurement too, provided I write down where I looked.