This paper describes the development of a systematic method for creating domain ontologies. We have chosen to explicitly recognise the differing needs of the human domain expert and the machine in our representation of ontologies in two forms: a conceptual and a logical ontology. The conceptual ontology is intended for human understanding and the logical ontology, expressed in description logics, is derived from the conceptual ontology and intended for machine processing. The main contribution of our work is the division of these two stages of ontology development, with emphasis placed on domain experts themselves creating the conceptual ontology, rather than relying on a software engineer to elicit knowledge about the domain. In particular, this paper concentrates on the creation of conceptual ontologies and analyses the success of our methodology when tested by domain experts.