Routine use of QFL, QpLsmLvm, and LmLsLv plate-tectonic provenance diagrams has led to the recognition of error populations. For these sandstones, the interpretation of tectonic setting based on data exclusive of petrography does not coincide with the interpretation of tectonic setting based on their location on compositional diagrams. The anomalous sandstones can be separated into at least four categories: i) sandstones deposited during the transition between tectonic regimes may be derived, in part, from relict source rocks; ii) the relative enrichment of sandstones in detrital quartz by weathering and/or depositional reworking may lead to an inaccurate interpretation of tectonic setting from compositional data; iii) sandstones deposited in tectonic settings as yet unrepresented on the provenance diagrams may plot between the provenance fields or overlap existing fields; iv) the role of detrital carbonate rock fragments directly affects the location of data points on provenance diagrams, and is an unresolved problem. The first two categories represent permanent problems in the use of sandstone composition as an indicator of plate-tectonic setting, whereas the last two may only be temporary problems that can be solved by more data and a consensus on determining detrital modes. Recognition of anomalous sandstones in the rock record will reduce the probability of erroneous interpretations of tectonic setting and enhance the use of sandstone petrology as an indicator of plate-tectonic setting.