The anterior-to-posterior extension of the vertebrate body axis is driven by bipotent neuromesodermal progenitors (NMps) located in the tailbud. NMps express both neural (sox2) and mesodermal (T/Brachyury) markers, and differentiate to either lineage depending on Wnt signalling (low Wnt leads to neural fates, high Wnt to mesodermal). Hox13 genes were proposed to halt axis extension by downregulating Wnt signalling, but studies have relied heavily on overexpression in amniotes. Now, Zhi Ye and David Kimelman re-examine this hypothesis by generating loss-of-function mutants for abundantly expressed zebrafish Hox13 paralogs. hoxa13b mutants show posterior defects in the presence of a cold-sensitive mutation in tbxta (the zebrafish T/Brachury), which has reduced Tbxta activity at a lower temperature; these phenotypes – which do not involve loss of notochord - are enhanced when hoxd13a is also deleted. In double hoxa13b and tbxta cold-sensitive mutants, expression of mesodermal markers, as well as Wnt genes, is substantially reduced, while sox2 expression increased, demonstrating that the Hox13 genes and tbxta work synergistically to produce mesoderm from the NMps. Finally, overexpression of Hoxa13b leads to a severe posterior truncation, with expression analysis showing that, for many genes, overexpression does not lead to an opposite of loss of function, rather it has the same effect; overexpression of Hox genes has the potential to be misleading. Thus, the authors conclude that Hox13 genes promote axis extension by promoting the formation of mesoderm from the NMp population.