Modelling the build-up of haloes is important for linking the formation of
galaxies with cosmological models. A simple model of halo growth is provide
d by Press-Schechter (PS) theory, where the initial field of density fluctu
ations is smoothed using spherically symmetric filters centred on a given p
osition to obtain information about the likelihood of later collapse on var
ying scales. In this paper the predicted halo mass growth is compared for t
hree filter shapes: Gaussian, top-hat and sharp k-space. Preliminary work i
s also presented analysing the build-up of haloes within numerical simulati
ons using a friends-of-friends group finder. The best-fit to the simulation
mass function was obtained using PS theory with a top-hat filter. By compa
ring both the backwards conditional mass function, which gives the distribu
tion of halo progenitors, and the distribution of halo mergers in time, the
build-up of haloes in the simulations is shown to be better fitted by PS t
heory with a sharp k-space filter. This strengthens previous work, which al
so found the build-up of haloes in simulations to be well matched to PS the
ory with a sharp k-space filter by providing a direct comparison of differe
nt filters and by extending the statistical tools used to analyse halo mass
growth. The usefulness of this work is illustrated by showing that the cos
mological evolution in the proportion of haloes that have undergone recent
merger is predicted to be independent of mass and power spectrum and to onl
y depend upon cosmology. Recent results from observations of field galaxies
are shown to match the evolution expected, but are not sufficiently accura
te to distinguish usefully between cosmological parameters.