The accuracy of a systematic sample cannot be calculated within the framework of classical sampling theory. A simple example relating to populations biology (insects counting) is used to show that the transitive theory of regionalized variables allows this calculation. Results are compared with those obtained from sub-sampling simulation within data.