Opaline silica (opal-A) has formed in marine, lacustrine and geothermal environments throughout geological time. During diagenesis opal-A normally changes to opal-CT, then opal-C, and finally to quartz. Such changes commonly destroy the original fabrics and any fossils that opal-A contained. The physical changes that accompany the opal-A to opal-CT transition, however, are known poorly. X-ray diffraction analyses, electron microprobe analyses and high-resolution, high-magnification scanning electron microscope imagery of siliceous sinters from the Geysir geothermal area in Iceland show that opal-A is formed of heterometric arrays of randomly packed microspheres (up to 5 μm diameter) with neighbouring spheres commonly being joined by small connection pads. In contrast, enlarged spheres, lepispheres, inverse opal (two types) and spindle frameworks with hexagonal motifs characterize opal-CT. The textures in opal-CT, which vary on a microscale, reflect the complex interplay between dissolution (e.g. inverse opal) and precipitation (e.g. enlarged spheres, spindle frameworks) that probably was mediated by groundwater in a near-surface environment. The processes deciphered from these young rocks should, however, be applicable to sedimentary opal-A and opal-CT of all ages, irrespective of their origin.