Canga is a moderately hard iron-rich duricrust primarily composed of goethite as a result of the weathering of banded iron formations. Canga duricrusts lack a well-developed soil profile and consequently form an innate association with rupestrian plants that may become ferruginised, contributing to canga possessing macroscopic biological features. Examination of polished canga using a field emission scanning electron microscope (FE-SEM) revealed the biological textures associated with canga extended to the sub-millimetre scale in petrographic sections and polished blocks. Laminae that formed by abiotic processes and regions where goethite cements were formed in association with microorganisms were observed in canga. Biological cycling of iron within canga has resulted in two distinct forms of microbial fossilisation: permineralisation of multispecies biofilms and mineralisation of cell envelopes. Goethite permineralised biofilms frequently formed around goethite-rich kaolinite grains in close proximity to goethite bands and were composed of micrometre-scale rod-shaped, cocci and filamentous microfossils. In contrast, the cell envelopes immobilised by authigenic iron oxides were primarily of rod-shaped microorganisms, were not permineralised and occurred in pore spaces within canga. Complete mineralisation of intact rod-shaped casts and the absence of permineralisation suggested mineralised cell envelopes may represent fossilised iron-oxidising bacteria in the canga ecosystem. Replication of these iron-oxidising bacteria appeared to infill the porous regions within canga. Synchrotron-based Fourier transform infrared (FTIR) microspectroscopy demonstrated that organic biomarkers were poorly preserved with only weak bands indicative of aliphatic methylene (CH 2 ) associated with permineralised microbial biofilms. High resolution imaging of microbial fossils in canga that had been etched with oxalic acid supported the poor preservation of organic biomarkers within canga, indicating mineralogical replacement of organic biomarkers.