Abstract
The age of digitally accessible datasets has transformed palaeontology, enabling previously impossible macroevolutionary insights. However, a substantial reservoir of generally inaccessible 'dark data' resides within museum collections, which may alter our understanding of ancient groups and their ecological and evolutionary history. We demonstrate how the addition of data held exclusively in museums impacts our macroevolutionary understanding of an entire taxonomic group, using a dataset of Palaeozoic echinoids containing the majority of museum occurrences for the clade. We find that museum 'dark data' shows clear differences in composition compared to data available in the published literature and strongly impacts biogeographic patterns, increasing the average geographic range size of taxa by 35%. Global model results assessing drivers of diversity are also significantly affected by the addition of museum-only data. Conversely, 'dark data' have a more limited impact on the temporal ranges of taxa or estimates of overall diversity and are impacted by similar socio-geographic biases as the published record. These findings show that unpublished museum data are necessary to obtain a complete understanding of macroevolutionary patterns in deep-time, illustrating the importance of the collection, curation, digitization and continued care of 'dark data' in the age of 'Big Data' in palaeobiology.