We discuss the different definitions of the mass of a halo in common use and how one may convert between them. Using N-body simulations we show that mass estimates based on spherical averages are much more tightly correlated with each other than with masses based on the number of particles in a halo. The mass functions pertaining to some different mass definitions are estimated and compared to the "universal form"of Jenkins et al. ([CITE]). Using a different simulation pipeline and a different cosmological model we show that the mass function is well fit by the Jenkins et al. ([CITE]) fitting function, strengthening the claim to universality made by those authors. We show that care must be taken to match the definitions of mass when using large N-body simulations to bootstrap scaling relations from smaller hydrodynamical runs to avoid observationally significant bias in the predictions for abundances of objects.