A study conducted in the summers of 1997-1998 and 1998-1999 in Canterbury, New Zealand, revealed that just under one-third of the Listronotus bonariensis larval population was present in the ryegrass pasture thatch. Hitherto it had been assumed that a much smaller proportion of larvae existed in the thatch, typically in transit from tiller to tiller. Overall, neonate (non-feeding) first-instar larvae comprised 45% of the total larvae extracted from the tillers, whereas they made up only 8% of all the larvae extracted from the thatch. Conversely, 18% of the total fourth-instar larvae were extracted from the tillers compared with 27% from the thatch. The frequencies of occurrence of the other instars in the tillers and thatch were intermediate between the neonate first-instar larvae extremes. The lack of fit of the larval data to a simple ordinal categorical model indicated larval movement into the thatch after the first instar and net migration out of the thatch as late instars. A technique is presented whereby tiller and thatch larval populations can be extracted simultaneously. It is suggested that on the basis of this study, earlier L. bonariensis larval population data sets could be corrected to account for the thatch populations of larvae.