Quarrying by wedge

Reading the Rock Before Striking It

Granite is not isotropic. It was laid down under pressure, and that pressure left a preferred plane of cleavage — quarrymen call it the rift — along which the stone will part more cleanly than against it. Get this wrong and the fracture wanders; get it right and a block of several tonnes separates from the face with a crack that runs true for metres.

The method itself is old and remains largely the same. A line of holes is bored along the intended split, spaced at intervals that depend on the stone's texture and the size of the block required — tighter spacing for fine-grained granite, wider for coarse. Into each hole go a pair of thin iron feathers, curved plates that line the hole walls, and between them a tapered iron plug, or in the older practice a dry wooden wedge. Where iron was not available, dried hardwood was hammered in and then wetted. The wood swells, the feathers transmit the load to the hole wall, and the tensile stress accumulates along the line connecting each hole. When it exceeds the stone's tensile strength at that plane, the split runs.

The critical reading happens before the first hole is bored. The quarrier looks for existing natural joints and any surface expression of the rift — slight colour variation, a line where lichen grows differently, a faint lineation in the crystal structure. A test tap with a hammer and listening to the return ring tells an experienced hand something about what lies beneath. The holes are then laid out parallel to the rift, not across it. A split that must fight the grain requires more wedges, more force, and still tends to give a rougher, more variable face.

Depth matters too. A shallow hole produces a split that opens at the surface but closes at depth, leaving the block still attached below. The holes must be deep enough that the propagating fracture reaches the full intended bed height of the block. For a column drum or a major lintel — the kind of stone discussed in the monolithic shaft — this could mean holes driven to half a metre or more.

The result, when the grain has been read correctly, is a face that requires relatively little further dressing: the natural fracture surface is already broadly planar. What it leaves on the bed — the bearing surface between one course and the next — and what labour then goes into truing that surface is the separate question of preparation before laying.