We study theoretically the effect of inter-habitat migration on the distrib
ution of population sizes between two habitats, and compare this distributi
on with the expected ideal free distribution (IFD). Whenever emigration fro
m the two habitats is asymmetric, or when there is a survival cost during m
igration, the resulting equilibrium distribution of population sizes deviat
es from the IFD. This result holds irrespective of emigration rule, even th
ough a density-dependent fraction of emigrants generally produces a distrib
ution closer to the IFD than a constant fraction of emigrants. Environmenta
l stochasticity causes a linear relation between population sizes in the tw
o habitats, with slope and intercept only identical to the IFD when net int
er-habitat exchange is zero. The type and asymmetry of inter-habitat migrat
ion will influence how we should interpret data on population distribution
in different habitats. The resulting resource matching is also critically c
ontingent on the relative time-scales of population renewal and dispersal,
and when population size is measured in relation to reproduction and disper
sal. Therefore, data on population sizes cannot be used uncritically to ass
ess habitat quality.