Assessment of infection is an essential part of many studies involving VA mycorrhiza. A summary is given of the range of techniques that have been used. We calculated the standard error of four methods of assessment based on observations of stained root samples either randomly arranged in a petri dish or mounted on microscope slides. The methods are based on presence or absence of infection at root/grid intersect points, on a visual estimate of percentage cortex occupied by fungus or on estimates of length, or presence or absence of infection in root pieces mounted on slides. The number of replicate observations required for a given standard error % infection can be read from the curves provided. The advantages of the different methods of assessment are discussed and reasons given why they all probably overestimate the true values.