In this article an account is given of the collection of data for Ømålsordbogen (Dictionary of Danish Insular Dialects), including the organisation of work and the kinds of methods used. Special attention is given to the relations between the dialectologists and the informants, examined on the basis of letters to and from the informants and a card index of informants. Having descriptions of the informants in some detail the card index gives an idea of the qualities that were highly valued in the informants. It also gives examples of problems connected with the collection of data, for example that an informant now and then recorded words of the written language like “rod shaped bacterium”.