Longer answer: the paper in PLoS describes a procedure for generating homeotic mutations in cnidarians by manipulating the expression of Cnox genes in hydrozoans. Knock outs of various cnidarian Hox-like genes and the medusae develop extra manubriae, or the tentacled part at the oral end of the animal, which the authors colloquially call “heads” (and they do usually put the word in quotes). These structures aren’t homologous to the things we consider heads—they are associated with mouth structures and express genes that we usually consider markers of the anterior end, so “head” is just easy shorthand.
I wrote a bit about Cnox genes in cnidarians before. These are Hox-like genes, alright, and they resemble anterior Antp class genes in triploblasts, but their expression is weird and species specific in diploblasts. The suspicion is that the last common ancestor of cnidarians and triploblasts had Hox-like genes, but they weren’t yet committed to a specific role of defining anterior-posterior patterning. The bilaterian ancestor coopted the Hox genes and committed them to specific axial roles, roles that are retained to this day. The descendants of the earliest cnidarian ancestor coopted them in different and diverse ways. All we can say is that all of them use these Cnox transcription factors to define spatial positions in the developing animal, but they do it in different ways.
The important result of the paper, though, is that they’ve got a protocol. For a long time, evo-devo work in these organisms has been rather descriptive: they’ve got genes X, Y, and Z, and the genes are expressed in spatial domains A, B, and C. We can learn a lot from that, but what we really need to see are manipulations of the genes that allow us to work out the interactions between the genes—it’s the regulatory network that is key, not just the list of components. Jakob and Schierwater have documented some tools that will allow epistatic interactions in these genes to be characterized. The multiple manubriae (“heads”) is just one cool phenotype.
Jakob W, Schierwater B (2007) Changing Hydrozoan Bauplans by Silencing Hox-Like Genes. PLoS ONE 2(8): e694. doi:10.1371/journal.pone.0000694