SEQUENTIAL SAMPLING OF GRASS GRUB, COSTELYTRA-ZEALANDICA

Sequential sampling schemes for testing a paddock for infestation by grass grub (Costelytra zealandica (White) (Coleoptera: Scarabaeidae)) larvae are developed. Instead of taking a fixed number of samples per paddock, decisions are made during sampling as to whether sampling needs to be continued. The hypothesis that the larvae have a negative binomial distribution is tested and accepted. The performance of the schemes is assessed theoretically for larvae populations with varying degrees of clumping. Their performance in practice is compared with that of schemes which use a fixed number of samples, and it is found that the sequential schemes significantly reduce the number of samples required with little loss of accuracy.