Authigenic terra rossa clay partly replaces the fossiliferous Salem limestone across a moving, 9-cm-wide metasomatic reaction front several feet below the earth's surface at Bloomington, Indiana. The front consists of a replacement zone, a bleaching zone ahead of it, and a thin opaque belt in the bleaching zone. The terra rossa consists of ferric-iron-bearing kaolinite, illite, and smectite, with small amounts of goethite and maghemite pisolites. Conclusive petrographic evidence of in situ replacement includes partly replaced fossils with conservation of shapes and volume, fossils in all stages of replacement from incipient to complete, and partly replaced calcite cement crystals whose unreplaced portions are in optical continuity. In many cases the clay partly replacing calcite forms sharp, intracrystalline microstylolites. This tiny hacksaw texture proves Maliva & Siever's fundamental new idea (Geology 1988) that a replacement happens because the new mineral grows and pressure-dissolves the host (not because the host chemically dissolves first and somehow “pulls” behind itself the growth of the guest mineral – the traditional but erroneous consensus). The petrographic and outcrop evidence not only indicates replacement of calcite by clay but simultaneously also negates both previous theories of terra rossa formation – the residual theory and the sedimentary (alluvial or ash-fall) theory. Terra rossas in the Bahamas, the Antilles, Yucatán, Florida, Texas, Kentucky, southern Europe, Israel, southern Australia, and elsewhere probably have the same origin as the Bloomington terra rossa, but a replacement origin, which can be detected only with an optical polarizing microscope, has never been suspected or sought. The iron, aluminum, and silicon needed for clay growth are provided probably by dust; this can be established by matching of strontium and neodymium isotope ratios. Saharan dust is certainly abundant in southern Europe and the Caribbean, both of which abundant in karst and terra rossa too. After settling, the dust dissolves at the surface, the solutes leak in, reach the reaction front at several feet of depth, and drive the precipitation of the red clay crystals that replace the limestone. Terra rossa is thus a unique laterite – one none of whose major chemical elements comes from its parent limestone. Many terra rossas support vineyards and wineries, such as at Cariñena, Spain; Chianti, Italy; Istria, Croatia; and Coonawarra, Southern Australia.

The isovolumetric replacement of limestone by clay produces acid. This is the acid that both bleaches an additional slice of limestone and dissolves voids in it, just ahead of the replacement subzone. The new voids locally increase the permeability, which propels advection of solutes to the reaction zone. This accelerates further replacement, which in turn produces more acid, which increases the porosity further. Through this unavoidable feedback – named the reactive-infiltration instability, first modeled mathematically in the 80s – an initially planar dissolution front, as it advances, must become a set of regularly spaced permeability fingers, which theoretically should convert to a set of funnels of greater wavelength. But fingers and funnels and sinkholes are precisely what karst is all about. We are thus led to realize that this is how karst forms. The terra rossa is thus not an alluvial mud that gets trapped in preexisting karst funnels or sinks, as traditionally held. Terra rossa, as it replaces limestone, carves the very karst sinks that contain it. That is why the two are associated.