Knowing the injectivity of a gelant, i.e. the pressure gradient required to achieve a given flow rate, enables the designer to program the rate/volume schedule of a water shut-off treatment. Bulk gelation times determined from bottle testing have often been used to estimate available pumping time, but such tests do not measure the evolution of gelant rheology as crosslinking proceeds, nor do they measure the tendency of the gelant to undergo filtration within a porous medium. These factors can affect injectivity significantly, resulting in pumping times much less than the bulk gelation time. In this paper we describe techniques for discriminating between vis-cosification and filtration by means of experiments in porous media, making use of multiple pore volume injection, step changes in flow rates, and in-line capillaries. Tests with phenol-formaldehyde crosslinked polymer gelants, which possess a number of characteristics favorable to high temperature applications, show that filtration of pre-gel aggregates in sandpacks and cores can play the dominant role in reducing the mobility of the gelant. The filtration behaviour is not universal, however; one gelant formulation showed no loss of mobility at all by this mechanism. When filtration does occur, it is strongly correlated to the age of the gelant (i.e. time the gelant has spent at temperature) but is relatively insensitive to the permeability of the porous medium or the presence of residual oil. The viscosity of the gelants tested evolves in similar ways in bulk and during flow in porous media.