We address the question of the predictability of skarn textures and their role in understanding the evolution of a skarn system. Recent models of skarn formation show that skarns are ideal for application of self-organisation theory, with self-patterning the rule in fluid-rock interaction systems rather than the exception. Zonation in skarn deposits, a consequence of infiltration-driven metasomatism, can also be treated in terms of self-organisation. Other less commonly described features, such as scalloping, fingering and mineral banding, can be understood by application of reactive infiltration and hydrodynamics at the skarn front. Devolatilisation may trigger formation of back-flow fluxes that overprint previously formed skarn. The range of textures formed from such events can be used to discriminate between prograde and retrograde stages. Refractory minerals, such as garnet, magnetite and pyrite, readily retain overprinting events. Skarns are also composed largely of minerals from solid solution series (garnet, pyroxene, pyroxenoids, etc.) and therefore skarn mineralogy helps to establish trends of zonation and evolution. The same minerals can act as ‘chemical oscillators’ and record metasomatic trends. The Ocna de Fier-Dognecea deposit was formed in a ∼10 km deep skarn system. Zonation and evolution trends therefore represent only the result of interaction between magmatically derived fluids emerging at the source and limestone. From the same reason, the transition from prograde to retrograde regime is not influenced by interaction with external fluids. Thirdly, the mineralisation comprises Fe, Cu and Zn-Pb ores, thus facilitating comparison with skarn deposits that commonly are formed in shallower magmatic-hydrothermal environment. Copper-iron ores (magnetite+Cu-Fe sulphides), hosted by magnesian (forsterite+diopside) skarn, occur in the deepest and central part of the orefield, at Simon Iuda. Their petrological character allows interpretation as the core of the skarn system formed from a unique source of fluids emerging from the subjacent granodiorite. It formed first as a consequence of the local setting, where a limestone indented in the granodiorite permitted strong reaction at ∼650 °C and focussed the up-streaming, buoyant fluids. The first sharp front of reaction is seen at the boundary between the Cu-Fe core and Fe ores hosted by calcic skarn (Di_70-90-And_70-90), where Cu-Fe sulphides disappear, and forsterite gives way to garnet in the presence of diopside (Di₉₀). Following formation of forsterite, devolatilisation and transient plume collapse is interpreted from a range of piercing clusters and trails. We presume lateral flow to have been initiated at the source, as the emerging fluids are in excess to the fluids driven into reaction by the plume. Formation of the other orebodies, up to 5 km laterally downstream in both directions, is interpreted as skarn fingering at the limestone side. The metasomatic front is perpendicular to the flow along the channel of schists placed between the limestone base and the granodiorite. A metal zonation centred onto the source is defined, based on metal distribution: Cu-Fe/Fe/Zn-Pb. The second front of reaction, at the boundary between the Fe and Zn-Pb zone, has a sulphidation/oxidation character, with diopside giving way to a Fe-Mn-rich pyroxene, (HedJoh)_>60+pyroxmangite±bustamite; garnet is minor. Johannsenite-rich pyroxene (Di_20-40Hed_20-40Joh₄₀) is found in proximal skarn at the upper part of Simon Iuda, stable with Zn_0.95Fe_0.05S, at an inferred 570 °C. In distal skarn from Dognecea and Paulus, Mn-hedenbergite (Di_<10Hed₇₀Joh_20-30) formed at ∼400 °C is stable with Zn_0.84Fe_0.16S. Extensive …