Somite number and vertebrate evolution

Abstract

Variation in segment number is an important but neglected feature of vertebrate evolution. Some vertebrates have as few as six trunk vertebrae, while others have hundreds. We examine this phenomenon in relation to recent models of evolution and development. Surprisingly, differences in vertebral number are foreshadowed by different somite counts at the tailbud stage, thought to be a highly conserved (phylotypic) stage. Somite number therefore violates the 'developmental hourglass' model. We argue that this is because somitogenesis shows uncoupling or dissociation from the conserved positional field encoded by genes of the zootype. Several other systems show this kind of dissociation, including limbs and feathers. Bmp-7 expression patterns demonstrate dissociation in the chick pharyngeal arches. This makes it difficult to recognise a common stage of pharyngeal development or 'pharyngula' in all species. Rhombomere number is more stable during evolution than somite number, possibly because segmentation and positional specification in the hindbrain are relatively interdependent. Although developmental mechanisms are strongly conserved, dissociation allows at least some major evolutionary changes to be generated in phylotypic stages.