We revisited the brachiopod fold hypothesis and investigated metamorphosis in the craniiform brachiopod Novocrania anomala. Larval development is lecithotrophic and the dorsal (brachial) valve is secreted by dorsal epithelia. We found that the juvenile ventral valve, which consists only of a thin layer that was previously described as periostracal, is not a valve and is not secreted by the same epithelia as the dorsal valve. It is secreted by the attachment area of the larva at the posterior-most tip of the posterior larval lobe. The same attachment area is used by larvae of rhynchonelliform brachiopods during metamorphosis to cement their pedicle to the substrate. N. anomala is therefore not initially attached by a valve but by material corresponding to pedicle cuticle. This is different to previous descriptions, which had led to speculations about a folding event in the evolution of Brachiopoda. We show that the “brachiopod fold hypothesis,” which argues that brachiopods are transversely “folded” across the ontogenetic anterior–posterior axis, should be rejected at least with respect to the craniiforms. The data now suggest that the Craniiformea may be a derived group within the Rhynchonelliformea. This interpretation suggests that the last common ancestor of the Craniiformea has lost the pedicle and the ventral valve in early juvenile development. Characters that have previously been considered to be shared between the Craniiformea and the Linguliformea (clade Inarticulata), such as a through-gut and missing hinge articulation, may thus be secondarily derived characters of the Craniiformea within the Rhynchonelliformea.