Because of their causal role in establishing the anterior-posterior axis of bilaterians, Hox genes are useful tools to address not only macroevolutionary changes within bilaterian body plans, but the origin of these body plans as well. The origin of one body plan in particular, that of the phylum Chordata, has received much attention over the past 130 years. Probably the most elegant, and hence most enduring, scenario put forth to explain chordate origins is the auricularia hypothesis of Garstang. Garstang hypothesized that in the early evolutionary history of chordates the lateral halves of the feeding ciliated band of the ancestral dipleurula (auricularia or tornaria) larva and its associated nerve tract came to the dorsal mid-line and formed the neural tube. Because in the chordate Hox genes are expressed in the neural tube, but not in the non-neural ectoderm, Garstang’s hypothesis predicts that Hox genes should be expressed in the aboral ectoderm of the dipleurula larva. To test this prediction we isolated eight Hox genes from the "enteropneust" hemichordate Ptychodera flava. In the juvenile, spatial colinearity was found with presumed 3’ genes expressed anteriorly and presumed 5’ genes expressed posteriorly. Most genes were detected in mesodermal populations, and none were detected in neuronal structures including the collar cord. In the tornaria larva, quantitative PCR against total RNA revealed that only four of these Hox genes are transcriptionally active, and in situ hybridization detected transcripts in various larval structures, but none were detected in the aboral ectoderm nor in the feeding ciliated band. Because Garstang’s primary prediction is not met, it is argued the auricularia hypothesis is no longer tenable as an explanation for the origin of the phylum Chordata.