In this project, sponsored by the Van Riemsdijk Stiftung, an automatic spelling analyzer for historical texts is developed that classifies spelling variation within lemma's and between lemma's (cliticization) in function of the spelling-external properties, such as historical form, linguistic context, and metadata, such as location and period of the texts in which the words occurs. The variation is rendered in function of these parameters using visualization tools. Random variation, i.e. variation that does not correlate with these preset parameters, is optionally removed and a new xml layer is provided as an intermediate form between transcribed layer and glossed form layer.