Across the animal kingdom, Hox genes are organized in clusters whose genomic organization reflects their central roles in patterning along the anterior/posterior (A/P) axis . While a cluster of Hox genes was present in the bilaterian common ancestor, the origins of this system remain unclear (cf. ). With new data for two representatives of the closest extant phylum to the Bilateria, the sea anemone Nematostella and the hydromedusa Eleutheria, we argue here that the Cnidaria predate the evolution of the Hox system. Although Hox-like genes are present in a range of cnidarians, many of these are paralogs and in neither Nematostella nor Eleutheria is an equivalent of the Hox cluster present. With the exception of independently duplicated genes, the cnidarian genes are unlinked and in several cases are flanked by non-Hox genes. Furthermore, the cnidarian genes are expressed in patterns that are inconsistent with the Hox paradigm. We conclude that the Cnidaria/Bilateria split occurred before a definitive Hox system developed. The spectacular variety in morphological and developmental characteristics shown by extant cnidarians demonstrates that there is no obligate link between the Hox system and morphological diversity in the animal kingdom and that a canonical Hox system is not mandatory for axial patterning.