Adult sex ratio is a basic component of breeding systems. Estimates of sex ratios of moa (Aves: Dinornithiformes) have ranged from near balanced to significantly female-biased. However, ratios have usually been estimated by simple ratios of females to males identified by some level of sexual size dimorphism or, at most, tested against a balanced ratio by χ2 test. Application of binomial tests confirmed a great heterogeneity, and high levels of uncertainty in estimates of moa sex ratios from different areas and from different kinds of fossil deposits. Large samples gave more constrained estimates than small, but even for some of the larger, binomial analysis often revealed a range of possible ratios, including one with a bias to males. Some causes of extreme values for swamp and lake bed deposits, including sexual differences in territorial behaviour, have been suggested before. However, a new issue – significant and sometimes abrupt changes in female and perhaps male body size through time – was identified here from series of genetically identified and radiocarbon dated moa from North Canterbury, New Zealand. The size changes compromise allocation of individuals to sex by morphometrics of limb bones, especially in undated samples. Intensive radiocarbon dating of series of genetically sexed moa of different taxa from a range of areas will be required to identify potential regional and temporal differences in their sex ratios before any interpretation of the evolution of size dimorphism and breeding systems based on moa sex ratios will be possible.