Summary. The chemical reactivity of apatite phosphate rock (PR) from Sechura (Bayovar), North Carolina, Egypt (Hamrawein), Morocco (Khouribja) and Queensland (Duchess), was estimated using the following methods (indices): percentage of the phosphorus (P) present in the PR that is soluble in one extraction with 2% citric acid and 2% formic acid, and variation in crystal chemistry determined using X-ray diffraction to measure the length of the unit cell a dimension. The indices were related to the agronomic effectiveness of the PRs for pasture using percentage of the maximum (relative) yields for PR and superphosphate. At most sites there was a statistically significant trend for the agronomic effectiveness of the PRs to increase with increasing reactivity of the PRs as assessed by the 3 indices. However, this was not the case for sites where: (i) leaching of water-soluble P from freshly-applied superphosphate occurs in wet years; and (ii) the agronomic effectiveness using SV 50 values of all the PRs was uniformly low so that no differences between the PRs could be detected. The relationship between the agronomic effectiveness (SV50) of PR and chemical reactivity of the PR were generally different each year and in most cases this variability was related to the variation in total annual rainfall. From this study it is concluded that PRs in which greater than 65–70% of their total P content is soluble in 2% formic acid are likely to be highly reactive and therefore agronomically effective P fertilisers for direct application to appropriate soils in suitable environments.