The merging of knowledge bases is a fundamental part of the collaboration in continuous knowledge construction. This paper introduces an operator for merging similar taxonomies, i.e. taxonomies that share the major part of their contents. Taxonomies have been chosen for the low time and space complexity of the classical inferences defined on them. A limit of this language is that it does not incorporate negations, thus the union of taxonomies is never inconsistent, though it is meaningful to consider that their merging does not coincide with their union. Thus, a way to extend the taxonomies' language is presented to allow the definition of a merging operator. This operator is algorithmically simple for the part of their contents on which the taxonomies agree, confining complexity to the part on which they do not. So it allows a low time and space complexity merging on similar taxonomies.